Lesson Plan Model Using the lesson plan model effectively to plan your instruction.
Ontario’s Integrated Power System Plan · 2020. 2. 7. · long terms. Without a working...
Transcript of Ontario’s Integrated Power System Plan · 2020. 2. 7. · long terms. Without a working...
Discussion Paper 2:Load Forecast
Ontario’s Integrated Power System Plan
September 7, 2006
i
120 Adelaide Street West
Suite 1600
Toronto, Ontario M5H 1T1
T 416-967-7474
F 416-967-1947
www.powerauthority.on.ca
September 7, 2006 To Ontario’s Electricity Consumers and Stakeholders: Today, I am pleased to deliver for your consideration “Discussion Paper #2: Load Forecast”, the Ontario Power Authority’s (OPA’s) second of eight envisaged papers on the Integrated Power System Plan (IPSP). Building on OPA’s “Discussion Paper #1: Scope and Overview”, released in June, this series of papers is intended to focus on specific aspects of planning. Together, the papers will provide insights and building blocks for the IPSP, and the feedback they generate will be important guidance for the preparation of the plan. Please see the table on the next page outlining the envisaged list of IPSP discussion papers. The load forecast paper is one of the starting points for the whole plan. While a forecast of long-term demand is inherently uncertain, the strength of having a forecast is the understanding it provides of the drivers of demand and the range of possibilities and uncertainties. In addition to various analytical models used for forecasting, equally important elements of planning for the IPSP include the adaptive nature of the plan we develop and the ongoing opportunities for adjustments. OPA very much welcomes your input and participation in the web and tele-conference to discuss the load forecast paper, which is planned for September 14th. For details on stakeholder input and participation opportunities (and other IPSP matters), please see www.powerauthority.on.ca/IPSP/, the OPA’s dedicated IPSP web page. In the months ahead, I look forward to receiving your advice, thoughts and comments through the IPSP consultation process and to sharing with you the additional planning documents as they are developed. In addition to the comprehensive report we are releasing today, OPA will be releasing more data and assumptions in the coming two weeks in support of certain aspects of the forecast. I strongly believe that developing a shared understanding of the planning challenges and the concrete steps needed to address them will focus the discussions, improve the dialogue, and ultimately result in a better plan for the benefit of all Ontarians. Yours sincerely,
Amir Shalaby Vice-President, Power System Planning
ii
OPA’s IPSP Discussion Papers
# Discussion Paper Title Release 1 Scope and Overview June 29 2 Load Forecast Sept. 7 3 Conservation and Demand Management Sept. 20 4 Supply Resources Oct. 20 5 Transmission Oct. 20 6 Safety, Environmental Protection, Environmental
Sustainability Oct. 30
7 Resources and Transmission Integration Oct. 30 8 Options for Procurement TBD
NB: For details on stakeholder input and participation opportunities (and other IPSP matters), please see www.powerauthority.on.ca/IPSP/, the OPA’s dedicated IPSP web page.
iii
TABLE OF CONTENTS
Table of Contents............................................................................................................... iii List of Tables.....................................................................................................................iv List of Figures....................................................................................................................iv Summary of Key Load Forecast Findings .............................................................................. 1 Introduction ...................................................................................................................... 2 1. Scope and Purpose...................................................................................................... 3 2. Background................................................................................................................. 4 2.1 Electricity Use is Driven by Many Variables............................................................. 4 2.2 Historical Use of Electric Energy in Ontario............................................................. 4
3. The Basis of Load Forecasting Analysis.......................................................................... 5 3.1 Adoption of Demand Side Working Group National Study ........................................ 7 3.2 Peak Forecast Methodology .................................................................................. 9 3.3 Key Variables and Assumptions............................................................................. 9 3.4 Sensitivity Analysis..............................................................................................12
4. Current Results ..........................................................................................................12 4.1 Electricity Use by Sector for 2025.........................................................................13 4.2 Components of Residential Sector Demand ...........................................................14 4.3 Components of Commercial Sector Demand..........................................................15 4.4 Components of Industrial Sector Demand .............................................................16 4.5 Comparison to Supply Mix Results ........................................................................17 4.6 Sensitivity Results ...............................................................................................18
5. Next Steps .................................................................................................................18 Appendices.......................................................................................................................19
iv
LIST OF TABLES
Table 4.1 – Average Annual change in electricity intensity over the reference case by industry sub-sector ........................................................................................................................17
LIST OF FIGURES
Figure 2.1 – Historical Electricity Use in Ontario – Energy Demand (TWh)............................... 5 Figure 2.2 – Historical Electricity Use in Ontario – Peak Demand (MW) ................................... 5 Figure 3.1 – Schematic of Information Flows Leading to IPSP End-use Model Reference Forecast........................................................................................................................................ 8 Figure 4.1 – End-use Model Reference Forecast, Energy – 2005-2025 ...................................12 Figure 4.2 – End-use Model Reference Forecast, Peak – 2005-2025 ......................................13 Figure 4.3 – Sectoral Contribution to Reference Forecast in 2025: Energy..............................13 Figure 4.4 – Growth in Residential Electricity Use, Energy, 2005 - 2025 .................................14 Figure 4.5 – Electricity End-Use in 2005, 2010 and 2025, Residential Sector: Energy ..............14 Figure 4.6 – Growth in Commercial Electricity Use, Energy, 2005 - 2025................................15 Figure 4.7 – Electricity End-Use in 2005, 2010 and 2025, Commercial Sector: Energy .............15 Figure 4.8 – Growth in Industrial Electricity Use, Energy, 2005 - 2025 ...................................16 Figure 4.9 – Electricity Demand by Sub-sector, Industrial Sector, 2005 and 2025: Energy .......16 Figure 4.10 – Electricity End-Use in 2000, 2010 and 2025, Industrial Sector: Energy...............17 Figure 4.11 – End-Use Model Reference Forecast Compared to Supply Mix Forecast, Energy – 2005-2025........................................................................................................................18 Figure 4.12 – End-Use Model Reference Forecast Compared to Supply Mix Forecast, Peak – 2005-2025........................................................................................................................18
IPSP Discussion Paper Load Forecast
1 September 7, 2006
Summary of Key Load Forecast Findings
• Load forecast is a first step for planning conservation demand management, supply and
transmission for Ontario.
• The plan that is developed for Ontario will adapt, change and adjust as the future unfolds.
This brings to focus two important features:
o to develop a robust and flexible plan
o to recognize the adaptive nature of planning. It is this adaptive nature of plans,
and the recognition of risks and uncertainties inherent in all its elements that
make the forecast a good basis to start.
• The results indicate that growth in the period from 2005-2015 is at about 1%/year. The
growth beyond that period is more uncertain as it depends on the nature and content of
manufacturing industries in Ontario, relative energy prices, and other drivers to demand
such as population and employment.
• This load forecast provides a reasonable basis for seeking approvals for projects and actions
within the first five years of the plan. For this reason, the 2005-2015 period is more crucial
than the last 10 years of the forecast.
• OPA is building on ongoing forecasting efforts to gain insights into electricity use and
future demand in Ontario.
• A number of data sources and tools were brought together to produce this analysis
• The data required for comprehensive forecasting is not consistently available for Ontario.
This causes uncertainties around some of the specific results, particularly on detailed
assessment of peak demand drivers.
• Significant improvements in data quality and consistency can result from ongoing data
check, better sources from market research and validation. This work will take place over
many months or years.
• The analysis provides support for and elaborates on generally known trends and patterns,
for example, the increasing proportion of electricity used for space cooling and its
significant input on summer peak.
• The framework developed is a good basis for power system planning. It includes
uncertainty bands around the forecast and provides a framework for assessing CDM
potential on a consistent basis with forecasting.
• More details than contained in this discussion paper will be prepared and posted on OPA’s
website over the coming short period. The extensive volume of the data that is currently
internal to models requires efforts to make it more accessible to stakeholders.
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 2
Introduction
Objectives
The mandate of the Ontario Power Authority, as established by the Government of Ontario, is
to ensure an adequate, long-term supply of electricity for the province. An integral part of
fulfilling that mandate is preparation of an Integrated Power System Plan (IPSP) for Ontario
that will take into account supply planning and procurement, transmission planning, energy
conservation and demand management opportunities, and many other factors.
Clearly, however, we cannot undertake practical and effective power system planning without
a working estimation of the province's future electricity needs. This discussion paper therefore
projects future electricity use in Ontario – the expected overall rate of energy-use growth
between 2005 and 2025 as well as growth in peak-demand-period use, and the expected usage
contributions of the residential, commercial and industrial sectors.
The methodologies used in preparing this load forecast are discussed in some detail in the body
of the discussion paper, along with the key assumptions, drivers and sensitivities we have taken
into account. Essentially, we have incorporated and built on earlier work in this area, including
proven macroeconomic and other models and further analysis of existing data.
That said, forecasting energy demand is a complex task. It involves understanding current
consumption levels and estimating possible future development in consumption patterns,
taking into account variables such as population and economic growth, income growth,
industrial structure, energy prices and the pace at which new energy-use or energy-saving
technologies might be introduced. Even these variables themselves are difficult to determine
and project with any certainty. As we continue to prepare the IPSP for Ontario, we will continue
to focus on understanding the variables and their potential impact on electricity load growth in
the province.
Seeking Stakeholder Input
A fundamental principle of all OPA activities is to consult with all individual and organizations
who have an interest in or concern with Ontario's electricity supply today and in the future. For
this reason, the load forecast is being issued as a discussion paper, in order to attract comment,
input and insight from stakeholders. This may include comments on the forecasting
methodology, data or assumptions, as well as the interpretation of uncertainties. In particular,
we would welcome feedback on the following questions:
Do the results, based on the methodology, key drivers and assumptions, provide a sound base
for planning? If not, please elaborate on your concerns.
Are there any other considerations that could be addressed in this analysis? If so, please
elaborate.
Please submit your comments on this discussion paper by October 2, 2006, using the mechanism
provided on the OPA website.
IPSP Discussion Paper Load Forecast
3 IPSP Stakeholder Engagement
1. Scope and Purpose
This load forecast, a projection of future electricity use in Ontario, is a foundational step
necessary for planning to meet the province's electricity supply needs over the medium and
long terms. Without a working estimation of future electricity use, we cannot plan effectively
for supply, transmission, conservation and demand management or other vital elements.
This discussion paper is the second in a series issued by the OPA as part of our process for
preparing an Integrated Power System Plan (IPSP) for Ontario. The first, an IPSP “scope and
overview” discussion paper, was released on June 25, 2006 and can be found on the OPA
website. This Load Forecast Discussion Paper builds on issues with respect to load forecasting
that were raised in the earlier paper, and provides an overview of additional analysis that we
have been working on in this area. This includes methodology, key drivers and assumptions,
currently available results, and next steps. All of these issues will also be discussed as part of
the stakeholder engagement process.
Also, while energy conservation and demand management (CDM) is closely related to the load
forecast, this topic will be addressed separately in a future IPSP discussion paper.
When the IPSP is submitted to the Ontario Energy Board, the OPA will be seeking approval for
projects and actions that will be required within the first five years of the plan. For this reason,
the 2005-2015 forecasting period is more crucial than the later years, which are also subject to
greater uncertainty. However, because the IPSP will be revisited every three years, we will be
able to consider new or additional information to either validate or update the current forecast
results. We are also carrying out sensitivity analysis on the load forecast (for instance, we are
analyzing the impact of much higher or lower than expected economic or population growth).
This is an area where we have significant flexibility to incorporate stakeholder feedback. We
believe that sensitivity analysis is an essential component of effective planning, with an effective
plan being both prudent and flexible while allowing for continuous adaptation to new
information.
In this discussion paper we have taken a province-wide approach to load forecasting. How the
province-wide load forecast is disaggregated among various regions of the province is also a
subject of on-going work. Because the regional transmission area plans to be completed under
the IPSP will consider all options (CDM, distribution and transmission, and supply options) on
a more localized basis, we will develop region-specific load forecasts in parallel with these area
plans. This should enhance consistency between the area load forecasts and the provincial load
forecast.
In order to maintain transparency of IPSP related information, work is in progress at OPA to
convert data files, which form the basis of the analysis, into a more accessible information but
tabulating and describing the data. When completed, they will be posted on the OPA’s web site.
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 4
2. Background
2.1 Electricity Use is Driven by Many Variables
A key aspect of load forecasting is to gain a better understanding of the variables that affect
overall electricity demand and/or peak period demand, their interdependence and their impact.
These variables include population, number of households, income growth, general economic
performance, industrial structure, relative energy prices (especially between electricity and
natural gas), and the pace at which new energy-using and energy-saving products are
introduced.
There are several models available to map these dynamics and interactions, each with unique
features that simulate energy consumption. A key dynamic relates to how some variables (e.g.
gross domestic product or population) contribute to uncertainty. We gain insight into the
impact these variables have on the load forecast through sensitivity analysis. Through this type
of analysis, we can assess the sensitivity of the forecast to variation in key parameters..
The OPA’s current load forecasting effort represents the beginning of our capability to collect
and build end-use data. As we continue to prepare for the IPSP, we will maintain our focus on
understanding the variables that contribute to uncertainty and their impact on projected load
growth in Ontario.
2.2 Historical Use of Electric Energy in Ontario
As shown in Figure 2.1, use of electric energy in Ontario has more than doubled since 1970 from
approximately 63 TWh to 158 TWh in 2005. This has occurred as Ontario’s population has
increased, its economy has grown and further industrialized, and its homes and offices have
acquired more electrical appliances and devices.
IPSP Discussion Paper Load Forecast
5 IPSP Stakeholder Engagement
In terms of peak requirements (as shown in
Figure 2.2), demand has also
approximately doubled since 1970,
growing from 11,000 MW to 26,000 MW in
2005.
The changes in growth rates are driven by
complex interactions of the demand
drivers. While the work on this subject
continues, our work to date includes the
report prepared for the Chief Conservation
Officer entitled “Electricity Demand in
Ontario”1 and the Supply Mix Advice
report’s examination of historical patterns
of electricity use in Ontario.2 Both reports
are available through the OPA website.
3. The Basis of Load Forecasting Analysis
The OPA's Supply Mix Advice projected provincial electricity demand to 2025 based on an
extrapolation of the Independent Electricity System Operator’s (IESO) July 2005 10-Year
1 Ralph Torrie, ICF Consulting. “Electricity Demand in Ontario – A Retrospective Analysis”. Prepared for Chief
Conservation Officer, Ontario Power Authority. Revised November 2005. Available from OPA on request. 2 Ontario Power Authority, Supply Mix Advice Report, Volume 3, Part 4 “Conservation and Demand Management
Issues”, December 9, 2005. Available from:
http://www.powerauthority.on.ca/Storage/18/1360_Part_3-4_Conservation_and_Demand_Management_Issues.pdf.
Figure 2.1 – Historical Electricity Use in Ontario – Energy Demand (TWh)
Energy
0
20
40
60
80
100
120
140
160
180
1970 1975 1980 1985 1990 1995 2000 2005
Ene
rgy
(TW
h)
Source: IESO Note: Data is at generator level and represents actuals measured.
Figure 2.2 – Historical Electricity Use in Ontario – Peak Demand (MW)
Peak
0
5000
10000
15000
20000
25000
30000
1970 1975 1980 1985 1990 1995 2000 2005
Pea
k D
eman
d (M
W)
Source: IESO Note: Data represents actuals measured.
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 6
Outlook load forecast.3 Beginning with a 2005 estimate of energy demand of 155 TWh, the
supply mix projection grew to 185 TWh by 2025.
The supply mix load forecast was based on an econometric analysis -- that is, it was based solely
on macroeconomic variables such as gross domestic product (GDP) and population. Analysis of
the final end-uses of electricity (i.e. use of electricity by refrigerators, office equipment, and
industrial machinery) was derived from the econometric forecast.
Understanding the projected end-uses of electricity is important for conservation and demand
management (CDM) planning, as it lends insight into the potential for CDM in the province and
accounts considerably for efficiency of use. For this reason, it is important that both load
forecasting and CDM analysis are carried out using the same data and the same methodological
framework. In the Supply Mix Advice, we noted that additional work would be needed.4 This
additional work would require more sophisticated end-use analysis techniques, and for load
forecasting and CDM analysis to be carried out on the same basis.
Building on the work completed for the Supply Mix advice, the OPA will capture the strength
of two forecasting methodologies in preparing the IPSP. The first methodology is based on the
Canadian Integrated Modelling System (CIMS) and the second is based on the IESO’s
econometric framework.
CIMS is an integrated end-use model, in that it determines the overall electricity forecast from
the individual end-uses, but does so using techniques that include econometric (and other)
algorithms. By using this framework we can ensure that the IPSP load forecast and CDM
potential results will be carried out on a consistent basis in terms of variables, data, assumptions
and methodology, and using the same tools.
The IESO’s econometric framework is useful because it increases understanding around the key
sensitivities of extreme weather and higher or lower than expected growth in GDP. Specifically,
we will use the relationship of weather and economic drivers to load as determined
econometrically by the IESO.5
By using both an integrated end-use and an econometric framework, we can address some of
the uncertainty caused by variables such as GDP and population growth. This aspect of the load
forecasting analysis is discussed in greater detail in the Sensitivity Analysis, Section 3.4.
3 For a summary of the Supply Mix Advice load forecast, please see, Volume 1, Part 2, “Advice Report and
Recommendations”, pp.12-15. For a discussion of the load forecasting methodology used in the Supply Mix Advice,
please see Part 6, “What Methodology and Assumptions Have OPA Adopted?”, December 9, 2005. Available from:
http://www.powerauthority.on.ca/Page.asp?PageID=924&SiteNodeID=157. 4 Supply Mix Report, Volume 1, p. 16. 5 A discussion of the econometric forecasting methodology can be found in the sensitivity analysis section below and
on the IESO website at http://www.ieso.ca/imoweb/pubs/marketReports/Methodology_RTAA_2006jun.pdf.
IPSP Discussion Paper Load Forecast
7 IPSP Stakeholder Engagement
3.1 Adoption of Demand Side Working Group National Study
As mentioned above, using a common data set and methodological framework for load
forecasting and CDM analysis adds to the robustness, and consistency of planning analysis. To
this end, we have adopted the Ontario results of the July 2006 national study of the Demand
Side Management Working Group. 6 Adoption of this study is also consistent with one of the
key planning principles, which is to build on established foundations and studies.
The national study, which was commissioned by the Council of Energy Minister’s working
group on CDM, was undertaken to provide a consistent and updated understanding of
Canadian conservation and demand management potential. It involved the participation of
representatives from federal and provincial governments, the energy utility industry, major
energy users, and non-government organizations. The study was undertaken by M.K. Jaccard
and Associates (“MKJA”) and Marbek Resource Consultants Ltd. (“Marbek”) and is based on
data and analytical techniques that were discussed and documented with the participants.
3.1.1 Extensions and Updates to National Study
The OPA was drawn to the national study because:
• it employs integrated end-use methodology
• it maintains consistency between the load forecast and CDM estimates
• it models Canadian energy consumption over a 25 year period from 2000–2025, and does so
with provincial/territorial and sectoral break downs and
• it is accepted by the study participants.
However, since the study was focused at the national level we decided that it was advisable to
commission additional analysis in developing the IPSP for Ontario. In particular, we asked that
the Ontario-specific results be updated to include provincial data for the 2001-2005 period, (an
appendix detailing this extension/update will be posted on the OPA website by
September 22, 2006). The resulting IPSP Reference Model End-Use Forecast was defined as
projected energy use, taking into account energy efficiency and conservation measures that are
driven by normal market forces (commonly referred to as “naturally occurring” CDM).
For a visual schematic illustrating how the national study was updated and extended, please
see Figure 3.1.
6Marbek Resource Consultants Ltd. And M.K. Jaccard and Associates, Inc. Demand Side Management Potential in
Canada: Energy Efficiency Study. Submitted to Canadian Gas Association, May 2006. Available from:
http://www.canelect.ca/en/News2006/EE-DSM_Final%20Report.pdf
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 8
Figure 3.1 – Schematic of Information Flows Leading to IPSP End-use Model Reference Forecast
Source: OPA
3.1.2 The Model Used for End-use Analysis
The IPSP Reference Model End-Use Forecast (please see Figure 3.1 for an illustration) is based
on CIMS, which was developed and is maintained by the Energy Research Group at the School
of Resource and Environmental Management at Simon Fraser University. CIMS is an integrated
set of economic and energy models capable of a range of combined energy and economic
analysis. The model tracks energy flows through the Canadian economic system, from
production processes through to eventual end-use.
As noted above, CIMS is a hybrid simulation model that incorporates the benefits of both
end-use and econometric analysis. End-use analysis is based on an extensive database of
energy-using stock7 (such as appliances and lighting) across all sectors of the economy,8 as well
as the technology options available for new, replacement or retrofit stock. In turn, decisions
about the purchase of these technologies are shaped by consumer choices and macroeconomic
factors.
7 The definition of stock depends on the sector (residential, commercial, industrial) under discussion. Generically,
stock refers to consumer durables and capital formation, for instance, refrigerators, photocopiers, or industrial
machines. It can also refer to buildings. 8 With the exception of agriculture, non-metal mining and transportation sectors which were not analyzed in the
national study.
IPSP Discussion Paper Load Forecast
9 IPSP Stakeholder Engagement
Being an energy model, CIMS does not produce results in terms of power needed to estimate
peak demand. The need for conversion to a forecast of peak demand is discussed in section 3.2.
3.2 Peak Forecast Methodology
Because the power system needs to meet both energy and peak period needs, capacity planning
requires a forecast of peak demand. The energy results that are the outcome of the modelling
effort described above have been translated into hourly demand load profiles, based on end-use
and system load profiles.9 This builds on the approach taken in the Supply Mix Advice and
incorporates additional load profiles that more fully account for end-uses. As a result, the IPSP
will include a more comprehensive analysis of end-uses than was possible for the Supply Mix
Advice.
The peak forecast methodology is based on 2003 load profiles from Michigan and New York
State. While the most recent energy data is for the year 2005, the most recent complete and
consistent load profile data is for the year 2003. No comprehensive Ontario-specific end-use
profiles were available as collection of this data has not been undertaken during the last
15 years in Ontario. Given the similarity of Michigan and New York State to Ontario weather
and time-of-use patterns, we expect that these load profiles provide a good starting point for
load analysis.
Given the current data and tools available for the Ontario context, these profiles are the best that
are presently obtainable. OPA will continue to develop this analysis and intends to share with
you the results of ongoing efforts.
3.3 Key Variables and Assumptions
Generally speaking, overall provincial electricity demand is driven by forecast demand for
energy services, including air conditioning, lighting and physical industrial production (e.g.,
tonnes of pulp and paper or molten steel). Energy service demands grow at different rates.
Some uses, such as refrigeration, may only change at the rate of stock replacement or
population growth, while other end-uses, like air conditioning, may also grow with increased
market penetration, when, for example, more air conditioning is added to houses and
workplaces. In addition, on-going change in product mix and manufacturing processes affect
electricity use in the industrial sector.
The IPSP Reference Model End-Use Forecast (as illustrated in Figure 3.1), in analyzing these
subjects, is based on a number of assumptions regarding key variables. These relate particularly
9 This work has been conducted by Navigant Consulting Ltd. using the same analytic approach employed in the
supply mix advice, but with the new annual energy end-use numbers generated by Jaccard.
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 10
to fuel price forecasts, economic growth projections and the expected efficiency of future
technologies that are anticipated but are not yet in widespread use. Assumptions regarding
"naturally occurring" CDM and the potential role of future structural change were also
important to the overall forecast.
In order to maintain consistency, in adopting the national study analysis we have been careful
only to update more recent consumption data and a new set of output drivers produced by the
federal government. For example, updated drivers include projections of household numbers,
commercial floor space, and industrial output.
An illustration of the data and analysis flows leading to the development of the IPSP end-use
model reference forecast is presented in Figure 3.1.
3.3.1 Fuel Price Forecasts
The IPSP Reference Model End-Use Forecast and the national study from which it was derived
are consistent with assumptions reflected in a National Energy Board (NEB) scenario
constructed to examine different energy futures. The electricity prices used in the national study
and the IPSP Reference Model are based on the “techno-vert” price scenario in the NEB’s report
entitled Canada’s Energy Future.10 This scenario represents a world in which technology
advances rapidly and Canadians take broad action with respect to the environment and
demonstrate a preference for environmentally friendly products and cleaner-burning fuels.
Under the “techno-vert” scenario, Ontario electricity prices rise from an average of 10.3 ¢/kWh
in 2005 to 11.7 ¢/kWh in 2010, falling off again to 9.9 ¢/kWh in 2025 (all prices in $2005). These
prices provide a signal on which consumers make energy efficiency choices, select technologies
and select fuels (for example, between electricity and gas).
3.3.2 Economic Growth Projections
The CIMS forecast is based on projected demand for energy services. These, in turn, are based
on physical drivers such as growth in households, commercial floor space and physical
industrial production.11 Over the period 2005-2025, the number of households in the province is
projected to grow at an average annual rate of 1.2%, while commercial floor space grows at an
average annual rate of 1.9%.
These physical drivers were derived from economic growth rates assumed in Natural Resources
Canada's new energy emissions outlook (expected to be released in Fall/Winter 2006-2007). The
national study was based on NRCan’s previous Outlook (2000), and the Ontario results were
updated using the new outlook rates. It has been assumed that physical production growth will
10 (see the “Supply Push” and “Techno-Vert” scenarios), National Energy Board, Canada’s Energy Future: Scenario’s for
Supply and Demand to 2025.. http://www.neb-one.gc.ca/energy/SupplyDemand/2003/index_e.htm 11 The exception to this is the “other manufacturing” sub-sector, which is based on gross output.
IPSP Discussion Paper Load Forecast
11 IPSP Stakeholder Engagement
be equivalent to economic (i.e., dollar value) production growth, in order to translate the
financial figures available into physical terms.12
3.3.3 Expected Efficiency of Future Technologies
The load forecast is also affected by assumptions regarding the expected efficiency of future
technologies. It is assumed that energy efficiency will improve steadily over time as the stock of
existing appliances and equipment continues to be replaced by more energy efficient
technologies. Efficiency improvements are assumed to be based on currently available
technologies, or those that are expected to be available in the near future. The potential for
further improvement is projected to decline once all of today’s equipment has been replaced by
equipment with a higher level of efficiency. Consequently, the dampening effect of technology
improvement on demand growth is less in the latter years of the forecast.
3.3.4 "Naturally Occurring" CDM
Forecast growth in energy demand is mitigated by the on-going contribution of conservation
and demand management measures taken before 2000 and by ‘natural conservation’ that occurs
due to incremental technology improvements, changes in the economy that reduce energy
intensity, and old energy-consuming assets being replaced with new and more efficient
technologies. These efficiency improvements are not brought about through market
intervention, but result from competition among technologies on the basis of life-cycle cost and
non-financial factors that are known to affect purchasing decisions. In contrast, policy-driven
CDM measures and programs induce people and organizations to change behaviour beyond
what they would have done "naturally".
3.3.5 Economic Structural Change
Future economic and social structural changes could have an impact on the role electricity plays
in Ontario’s economy. The current forecast recognizes the emergence of other manufacturing
industries replacing the historical role of resource industries in Ontario. In addition to these
changes, there will be more future changes that would be captured in future forecasts.
12 For additional detailed information on these assumptions, please see Jaccard, “Modelling and Scenario
Documentation,” Draft Report, September 6, 2006, accompanying this report.
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 12
3.4 Sensitivity Analysis
While we recognize that all forecasts are subject to significant uncertainties, sensitivity analysis
allows us to study the impacts of key factors that are the sources of the most significant
uncertainty. By adjusting a key variable or assumption that drives the model, the resulting
change in the forecast can be evaluated, patterns identified, and conclusions explored.
We have used the econometric analysis underlying the IESO forecast to investigate the impact
of two key variables on the forecast: namely, GDP and extreme weather.
The IESO’s approach to determining a high and low GDP growth forecast, as well as an extreme
weather forecast, is documented in the IESO’s 2005 10-Year Outlook.13 As with the Ontario
specific data modifying CIMS, we will post an appendix detailing the sensitivity analysis on the
OPA website by September 22, 2006.
4. Current Results
As mentioned in Section 3.2, these results are based on efforts to date and incorporate the tools
and data that are presently available in the Ontario context. We are continuously improving our
load forecasting capability and working towards the development of a complete and consistent
Ontario end-use data set.
Over the 2005–2025 period, as shown in
Figure 4.1, OPA forecasts that energy
demand will grow from 155 TWh in 2005
to 196 TWh in 2025. This corresponds to an
average annual growth rate of 1.2%.
During the first fifteen years of the
forecast, the growth rate is almost 1%.
Over the five years from 2000 to 2005, as
show in Figure 2.1, demand in Ontario
grows at an average annual rate of 1.0% on
a weather corrected basis but at 1.3% in
actual terms.
Demand during peak-load periods is expected to grow at an average annual rate of 1.2%. It rises
from 25,823 MW in 2005 to 32,531 MW in 2025, whereas the 2005 starting point in the supply
mix analysis was 23,672 MW. The different 2005 peak values result from the IESO’s change to a
13 www.ieso.ca/imoweb/pubs/marketReports/10YearOutlook_2005jul.pdf.
Figure 4.1 – End-use Model Reference Forecast, Energy – 2005-2025
Energy
0
20
40
60
80
100
120
140
160
180
200
2005 2010 2015 2020 2025
Ene
rgy
(TW
h)
Source: Jaccard, Navigant, OPA
IPSP Discussion Paper Load Forecast
13 IPSP Stakeholder Engagement
seasonal peak methodology from the prior weekly peak methodology.14
In terms of sectoral breakdown, residential
grows from 52 TWh in 2005 to 61 TWh in
2025, corresponding to an average annual
energy growth rate of 0.8%. Growth in
minor appliances accounts for 80% of the
growth in energy demand in this sector.
The commercial sector grows from 55 TWh
in 2005 to 70 TWh in 2025, corresponding
to an average annual energy growth rate of
1.2%. Lighting, office equipment (auxiliary
equipment), and ventilation are expected
to continue to be main contributors to
commercial energy demand.
The industrial sector grows from 48 TWh in 2005 to 66 TWh in 2025, corresponding to an
average annual energy growth rate of 1.6%. The growth in energy demand by the “other
manufacturing” industries will outpace that of the “big four” industries (pulp and paper, iron
and steel, industrial chemicals and mining) with “other manufacturing” accounting for 80% of
increased industrial energy demand.
4.1 Electricity Use by Sector for 2025
We expect that by 2025, residential use will
account for 31% of total provincial electricity
demand, the commercial sector for 36% and
the industrial sector for 33%. This is
illustrated in Figure 4.3.
14 IESO, The Ontario Reliability Outlook, June 2006, v1 i2. In particular, please see p. 5 “Changes in Planning
Assumptions”. Available from: http://www.ieso.ca/imoweb/pubs/marketReports/ORO_Report-2006-1-2.pdf.
Figure 4.2 – End-use Model Reference Forecast, Peak – 2005-2025
Peak
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
2005 2010 2015 2020 2025
Pea
k D
eman
d (M
W)
Source: Jaccard, Navigant, OPA
Figure 4.3 – Sectoral Contribution to Reference Forecast in 2025: Energy
Energy
Residential30.9%
Commercial35.6%
Industrial 33.4%
Source: Jaccard, OPA, Navigant
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 14
4.2 Components of Residential Sector Demand
As shown in Figure 4.4, residential electric energy use grows from 52 TWh in 2005 to 61 TWh in
2025, reflecting an average annual growth rate of 0.8%. While energy use grows over the
forecast period, electricity intensity (measured as kWh used per household per year) declines at
an average rate of 0.2% per annum.
Figure 4.5 illustrates the major sources of
growth in residential energy demand. An
increasing number of households account
for part of this growth, while increasing
saturation of appliances and demand for
new appliances is also a driver.
The largest contributor to energy growth is
minor appliances –the collection of
electricity consuming devices such as
televisions, answering machines,
computers, battery chargers and DVDs.
Lighting and air conditioning also grow in
the forecast. According to the National Energy Use Database, the penetration of air conditioners
by household in Ontario is growing significantly, particularly for central systems, as discussed
in “Modelling and Scenario Documentation”.12
Beyond 2005, the forecast assumes that most
standard appliances – those other than air
conditioning – will be at their saturation
point, with no increased penetration
included in the forecast. Dishwashers,
clothes washers and clothes dryers
increased their penetration over the past
five years.
Lighting grows with household growth,
with a slight increase in the penetration of
compact fluorescent lighting over the
forecast period. The decline in electricity use
for space and water heating reflects both
increased efficiency and fuel-switching
towards natural gas.
Figure 4.4 – Growth in Residential Electricity Use, Energy, 2005 - 2025
Energy
0
10
20
30
40
50
60
70
2005 2010 2015 2020 2025
Ele
ctric
ity D
eman
d (T
Wh)
Source: Jaccard
Figure 4.5 – Electricity End-Use in 2005, 2010 and 2025, Residential Sector: Energy
Energy
0
10
20
30
40
50
60
70
80
2005 2010 2025
Ele
ctric
ity C
onsu
mpt
ion
(TW
h/ye
ar)
Source: Jaccard, Navigant
IPSP Discussion Paper Load Forecast
15 IPSP Stakeholder Engagement
4.3 Components of Commercial Sector Demand
In the commercial sector, as shown in
Figure 4.6, electricity use is expected to
rise by a 1.2% average annual growth rate
from 55 TWh in 2005 to 70 TWh in 2025.
Growth in this sector results from
increasing commercial floor space and an
associated increase in end-use activity.
Commercial electricity intensity declines
at an average rate of 0.7% per annum
(measured in kWh/m2 of commercial floor
space). This occurs because new buildings
use less electricity than existing buildings.
Aging of the existing stock (in this case,
stock refers to commercial buildings)
results in significant renovation and demolition activity.
Growth in commercial use of electric energy is shown in Figure 4.7. Increasing electricity use for
commercial space heating, lighting and office equipment are the main contributors to overall
growth in the sector, followed closely by
ventilation. Lighting, ventilation and office
equipment continue to be the largest
end-uses in the commercial sector in terms
of energy consumed.
Most commercial end-uses grow at the rate
of stock replacement and new floor space
growth, with the exception of an increase
in the commercial floor space heated by
electricity. According to estimates by the
NRCan’s Office of Energy Efficiency, this
trend has been demonstrated over the past
five years and is assumed to continue
during the forecast period.
Figure 4.6 – Growth in Commercial Electricity Use, Energy, 2005 - 2025
Energy
0
10
20
30
40
50
60
70
2005 2010 2015 2020 2025
Ele
ctric
ity D
eman
d (T
Wh)
Source: Jaccard
Figure 4.7 – Electricity End-Use in 2005, 2010 and 2025, Commercial Sector: Energy
Energy
0
10
20
30
40
50
60
70
80
2005 2010 2025
Ele
ctric
ity C
onsu
mpt
ion
(TW
h/ye
ar)
Source: Jaccard, Navigant
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 16
4.4 Components of Industrial Sector Demand
In aggregate, the industrial sector, as seen
in Figure 4.8, is projected to grow at an
average annual rate of 1.6%, from 48 TWh
in 2005 to 66 TWh in 2025.
While the “big four” industrial sub-sectors
are expected to continue recent trends of
slight to negative growth in electricity
demand, strong growth is expected in
other manufacturing industries, as shown
in Figure 4.9. The “other manufacturing”
sector, as noted above, includes industries
such as automotive
manufacturing. This is consistent with
trends experienced over the last several
years.
Electricity intensity, which is noted in
Table 4.1, is forecast to decline at varying
rates by industry, with the exception of
iron and steel, which shows an increase
due to increasing production of secondary
steel. Of the growth in electricity use that
occurs in the industrial sector, almost 80%
is accounted for by growth in “other
manufacturing” (not in the “big four”
industries). Jaccard provides a detailed
discussion of these and other modeling
assumptions in his paper on “Modelling
and Scenario Documentation”, which is appended to this paper.12
Figure 4.8 – Growth in Industrial Electricity Use, Energy, 2005 - 2025
Energy
0
10
20
30
40
50
60
70
2005 2010 2015 2020 2025
Ele
ctric
ity D
eman
d (T
Wh)
Source: Jaccard
Figure 4.9 – Electricity Demand by Sub-sector, Industrial Sector, 2005 and 2025: Energy
Energy
0
10
20
30
40
50
60
70
80
2005 2010 2025
Ele
ctric
ity C
onsu
mpt
ion
(TW
h/ye
ar)
Source: Jaccard, Navigant
IPSP Discussion Paper Load Forecast
17 IPSP Stakeholder Engagement
In terms of end-use trends, “machine drives” – the motor systems used for pumping, air
displacement, compression, conveyance and direct drive – dominate electricity used in
industry. Little anticipated improvement
in motor efficiency is expected, but motor
efficiency is already relatively high. There
are improvements, however, in the
auxiliary systems that use motors. Direct
heat, the next largest end-use, grows
fairly significantly. This end-use applies
mainly to iron and steel production and
“other manufacturing”, which have
greater than average production growth
during the forecast (particularly for other
manufacturing).
4.5 Comparison to Supply Mix Results
The load forecast that informed the Supply Mix Advice grew from 155 TWh to 185 TWh in 2025,
as described in Section 3.This energy forecast corresponded to a forecast of peak demand of
23,672 MW in 2005, growing to 30,441 MW in 2025. The Supply Mix load forecast is compared
to the current end-use model results for the IPSP in Figure 4.11 and Figure 4.12. The Supply Mix
Advice and end-use results for the IPSP Reference End-Use Model Forecast track quite closely
for the period 2005-2020, but then deviate for the last five years of the forecast.
Table 4.1 – Average Annual change in electricity intensity over the reference case by industry sub-sector
Sector Change in Electricity Intensity
Pulp and Paper -0.8%
Chemical Manufacturing -0.2%
Iron & Steel 0.3%
Metal, Smelting & Refining -1.1%
Metals and Mineral Mining -0.9%
Other Manufacturing -0.4%
Industrial Minerals -0.5%
Petroleum Refining -1.7% Source: Jaccard
Figure 4.10 – Electricity End-Use in 2000, 2010 and 2025, Industrial Sector: Energy
Energy
0
10
20
30
40
50
60
70
80
2005 2010 2025
Ele
ctric
ity C
onsu
mpt
ion
(TW
h/ye
ar)
Source: Jaccard, Navigant
Load Forecast IPSP Discussion Paper
IPSP Stakeholder Engagement 18
4.6 Sensitivity Results
We are currently developing and analyzing the sensitivity results. These results and a
description of the methodology will be posted on the OPA website by September 22, 2006.
5. Next Steps
The OPA would appreciate receiving stakeholder insights, advice and experience related to the
load forecast. This may include comments on the forecasting methodology, data or
assumptions, as well as the interpretation of uncertainties. In particular, we are seeking
feedback on the following question:
• Do the results, based on the methodology, key drivers and assumptions, provide a
sufficiently robust basis for planning? If not, please elaborate your concerns.
• Are there any other considerations that could be addressed in this analysis? If so, please
elaborate your concerns.
We would also welcome feedback on other load forecasting topics that are relevant to long term
planning. As one of the first step for the first integrated power system plan for Ontario in over a
decade, the end-use forecast results represent a starting point. Every three years the IPSP will be
revisited, allowing for an iterative approach that builds and improves on existing analysis. As
such, any comments and evidence presented at the current time may also help shape the next
round of IPSP development.
Figure 4.11 – End-Use Model Reference Forecast Compared to Supply Mix Forecast, Energy – 2005-2025
Energy
0
20
40
60
80
100
120
140
160
180
200
2005 2010 2015 2020 2025
Ene
rgy
(TW
h)
Source: Jaccard, Navigant, OPA, ICF
Figure 4.12 – End-Use Model Reference Forecast Compared to Supply Mix Forecast, Peak – 2005-2025
Peak
0
5000
10000
15000
20000
25000
30000
35000
2005 2010 2015 2020 2025
Pea
k D
eman
d (M
W)
Source: Jaccard, Navigant, OPA, ICF
IPSP Discussion Paper Load Forecast
19 IPSP Stakeholder Engagement
Important dates related to customer and stakeholder discussion of load forecasting include:
• IPSP Load Forecast Stakeholder Paper released September 6, 2006
• Load Forecast Stakeholder Webcast September 14, 2006
• OPA intends to post additional September 22, 2006
appendices (on extension/update of the national
study and on the sensitivity analysis methodology)
• Stakeholder comments on Load Forecasting October 2, 2006
Discussion Paper due
• Load Forecast Stakeholdering Summary to be posted October 15, 2006
Appendices
1. M.K. Jaccard and Associates, Inc. “Modelling and Scenario Documentation Draft Report.”
September 6, 2006 DRAFT. Prepared for the Ontario Power Authority.
2. Ralph Torrie, ICF Consulting. “Electricity Demand in Ontario – A Retrospective Analysis”.
Prepared for Chief Conservation Officer, Ontario Power Authority. Revised November
2005. Available from OPA on request.
3. Historical Trends
a. Historical Energy Demand: Generator Level – Actual, 1970-present
b. Historical Peak Demand: Generator Level – Actual, 1970-present
c. Historical Energy Demand: Generator Level – Weather Corrected, 1970-present
d. Historical Peak Demand: Generator Level – Weather Corrected, 1970-present
e. Historical Energy Demand: Customer Level, 1958-present
4. Marbek Resource Consultants Ltd. And M.K. Jaccard and Associates, Inc. Demand Side
Management Potential in Canada: Energy Efficiency Study. Submitted to Canadian Gas
Association, May 2006.
R E P O R T
Modelling and Scenario Documentation
Instalment #1
Reference Case Documentation
Draft Report
September 6, 2006
Prepared for: Ontario Power Authority Power System Planning
Prepared by:
M.K. Jaccard and Associates, Inc.
#414 675 West Hastings Street Vancouver, BC V6B 1N2
Bryn Sadownik Chris Bataille
Jotham Peters Jacqueline Sharp
Mark Jaccard
Modelling and Scenario Documentation
i
Table of Contents
Table of Contents .............................................................................................................i
Table of Tables...............................................................................................................iii
Table of Figures ............................................................................................................. iv
1 Introduction .............................................................................................................5
2 General Methodology ..............................................................................................6
2.1 Study Scope.....................................................................................................7
3 Analytical Approach: The CIMS Model...................................................................8
3.1 Structure and Simulation Process .....................................................................8
3.2 Algorithms.....................................................................................................11
3.2.1 Technology Competition Algorithm .......................................................11
3.3 Fuel Switching in CIMS.................................................................................12
3.3.1 Describing fuel switching in the results ..................................................13
3.3.2 Controlling for fuel-switching in the simulation......................................13
4 Reference Case ......................................................................................................14
4.1.1 Energy End Use Profiles.........................................................................14
4.1.2 Energy Prices .........................................................................................15
4.1.3 Economic growth projections .................................................................16
4.1.4 DSM Programs in the Baseline...............................................................18
4.2 Reference Case Results ..................................................................................19
4.2.1 Base year and 2005 ................................................................................19
4.2.2 Forecast Simulation to 2025 ...................................................................20
APPENDIX A - Sector Detail........................................................................................27
Industry .....................................................................................................................27
Auxiliary Systems..................................................................................................29
Process Specific Systems .......................................................................................30
Residential.................................................................................................................46
Space Heating........................................................................................................46
Apartments ............................................................................................................47
Attached Housing (Row) .......................................................................................47
Modelling and Scenario Documentation
ii
Single Family Dwellings........................................................................................47
Non-Appliance Hot Water Use ..............................................................................48
Lighting and Major Appliances..............................................................................48
Minor Appliances ..................................................................................................48
Renewable Energy Sources and other Distributed Energy ......................................48
Water Heating........................................................................................................48
Commercial ...............................................................................................................51
Lighting.................................................................................................................53
Shell/HVAC ..........................................................................................................53
Refrigeration, Cooking, Hot Water, Plug Load.......................................................54
Plug Load ..............................................................................................................54
Domestic Hot Water ..............................................................................................55
Refrigeration..........................................................................................................55
Cooking.................................................................................................................55
Cogeneration .........................................................................................................55
APPENDIX B – Decomposition Analysis, Reference Case............................................56
General Method.........................................................................................................56
Decomposition Results ..........................................................................................58
Modelling and Scenario Documentation
iii
Table of Tables
Table 1: Discount Rates used in the Residential, Commercial and Industrial Sub-models......................................................................................................................................12
Table 2: Ontario Energy Prices Used in the Model ........................................................16
Table 3: Coal prices ($1995/GJ) used in CIMS..............................................................16
Table 4: Average Annual Ontario Growth Rates............................................................18
Table 5: Annual Industrial Production /Household and Floor space Demand .................18
Table 6: Base Year Electricity Demand Calibration Results...........................................19
Table 7 Comparison of Electricity Demand in CIMS to IESO Demand, 2000 and 2005 (TWh/year)....................................................................................................................19
Table 8: Reference Case Electricity Demand to 2025 (TWh/year) .................................20
Table 9: Average annual change in electricity intensity over the reference case by industry sub-sector ........................................................................................................25
Modelling and Scenario Documentation
iv
Table of Figures
Figure 1: CIMS Logistic Curve .....................................................................................11
Figure 2: Reference case forecast by sector....................................................................21
Figure 3: Electricity Consumption (TWh/year) by End-use in 2000 and 2025, Residential Sector............................................................................................................................22
Figure 4: Electricity Consumption (TWh/year) by End-use in 2000 and 2025, Commercial Sector ........................................................................................................24
Figure 5: Electricity Consumption (TWh/year) by Sub-sector, Industrial Sector, 2000 and 2025 ..............................................................................................................................25
Figure 6: Electricity Consumption (TWh/year) by end-use in 2000 and 2025, Industrial Sector............................................................................................................................26
Figure 7: Energy Flow Model of the Steel Products Industry .........................................27
Figure 8: Auxiliary Flow Model Diagram......................................................................28
Figure 9: Process Services Flow Model, Chemical Product Manufacturing....................31
Figure 10: Process Services Flow Model, Industrial Minerals ........................................33
Figure 11: Process Services Flow Model, Iron and Steel................................................35
Figure 12: Process Services Flow Model, Metals Mining...............................................37
Figure 13: Process Services Flow Model, Metal Smelting and Refining (Non-ferrous) ..39
Figure 14: Process Services Flow Model, Petroleum Refining.......................................41
Figure 15: Process Services Flow Model, Pulp and Paper ..............................................43
Figure 16: Process Services Flow Model, Other Manufacturing.....................................45
Figure 17.......................................................................................................................50
Figure 18 Commercial Flow Model ...............................................................................52
Figure 19: Decomposition applied to the Reference Case (in 2025), all sectors..............58
Figure 20: Decomposition applied to the Reference Case (in 2025), residential sector ...59
Figure 21:Decomposition applied to the Reference Case (in 2025), commercial/institutional sector ......................................................................................60
Figure 22: Decomposition applied to the Reference Case (in 2025), industrial sector.....60
Modelling and Scenario Documentation
5
1 Introduction
In 2005 the Council of Energy Ministers demand-side management (DSM) Working Group commissioned a study to estimate an achievable demand-side management (DSM) potential in Canada. The DSM Working Group comprises representatives from the federal government (Natural Resources Canada), provincial governments and the utility industry (Canadian Gas Association, Canadian Electricity Association).1 Marbek Resources Consultants Ltd. and M.K. Jaccard and Associates Inc completed the study jointly.
While this study (‘National DSM study’) was conceived as a high-level policy exercise to improve the discussion of the role of DSM in Canada, it involved the developed of regional and energy-specific end-use demand forecasts to estimate DSM potential. The Ontario Power Authority (OPA) felt that leveraging the national study work would be advantageous in developing load forecast and conservation demand management (CDM) potential for use in for the development of the 20 year Integrated Power System Plan.
CDM as defined in this project includes energy efficiency, fuel substitution, cogeneration and distributed generation. Three CDM scenarios were developed for the National DSM study: an Economic Potential and two Achievable Potential scenarios (Scenarios 1 and 2). Scenario 1 looks at the policy response from subsidies that target the uptake of energy efficient technologies, while Scenario 2 analyzes a broader policy package that includes building and appliance standards, marginal cost pricing, carbon pricing, land-use measures, and subsidies to both energy efficient technologies and end-use renewables
This report documents the modelling and scenario assumptions used in the National DSM study, including specific adjustments to the modelling (including changes to the reference case) to meet the needs of the OPA. Elements of the reference case and CDM scenario results are also discussed including key factors that drive the results as well as caveats to using the data provided.
This report accompanies the provision of detailed spreadsheets to the OPA that describe:
• Reference forecast electricity demand: total, by sector, by sub-sector (industry, commercial service), by end-use (in 5-year intervals).
• CDM scenario results: total, by sector, by sub-sector, by end-use. In 5-year intervals, for the 3 scenarios developed for the National DSM Study.
• Costs associated with conservation measures, including capital costs and program costs. (Caveats relating to the costs presented are outlined in Appendix D)
• Energy efficiency measures database including the sorts of technologies assumed and their penetration.
1 Marbek Resource Consultants Ltd. and M.K. Jaccard and Associates. Demand Side Management Potential in Canada: Energy Efficiency Study. Summary Report Draft. May 2006. Submitted to the Canadian Gas Association.
Modelling and Scenario Documentation
6
• The ‘decomposition’ of the reference case and CDM scenario into the key underlying factors (energy efficiency, cogeneration, fuel-switching). Note: Fuel-switching in this decomposition exercise has a more global meaning. For instance, we are describing whether a sector uses more or less electricity relative to other fuels overall, including changes in the relative importance of certain end-uses.
This report is structured as follows. We first provide a general overview of the National DSM study and methodology (Section 2), followed by a description of the CIMS model (Section 3). The reference case, economic potential and achievable potentials are then described in more detail including an overview of the key results of these scenarios (Sections 4-6). Appendices are included which describe the CIMS sub-models in more detail (Appendix A), the decomposition methodology and results (Appendices B and C) and the costing method and caveats (Appendix D).
2 General Methodology
The analysis for the National DSM study was conducted using the CIMS model, supported by Marbek DSM tools and databases.2 CIMS is an integrated energy-economy model that simulates technology acquisition in the economy over time. Using this model it is possible to alter model inputs to simulate alternative forecasts and policy scenarios. The model is described in more detail in section 3 (and in Appendix A).
CIMS was applied to develop four scenarios: a reference case, an economic potential, and two achievable potential scenarios. The scenarios are defined as follows:
Reference Case: A projection of energy demand to 2025, in the absence of any new and incremental institutional market interventions. It is the baseline against which the scenarios of energy savings are calculated. The reference case forecast includes “natural conservation”, i.e., changes in end-use efficiency over the study period that are projected to occur in the absence of new and incremental market interventions.
Economic Potential (or Techno-Economic Potential)3: An estimate of the energy demand that would occur if all equipment and building envelope energy management actions that pass a ‘Total Resource Cost’ test (not including administration costs) were implemented in the target markets. These actions are applied at either natural stock turnover or retrofit rates.
Achievable Potential: An estimate of the energy demand that would occur as a result of market intervention to influence the take up of energy management actions. The
2 The Energy and Materials Research Group and Simon Fraser University developed the CIMS model. 3 We suggest that ‘Techno-economic potential’ would be a better term for this potential given that it does not necessarily represent what is economic (from a societal perspective) nor an upper bound to which CDM measures can be targeted. The scenario reflects the potential for CDM actions that are net benefits when evaluated at a social discount rate, exclusive of consumer heterogenity, risk, and non-financial preferences, which can be real economic costs. Full inclusion of externalities would also alter what would be economic from a societal perspective. For consistency with the National DSM study we continue to refer to this potential as the Economic Potential.
Modelling and Scenario Documentation
7
potential is estimated in two policy scenarios. The first focuses on the response from subsidies to specifically target the uptake of actions identified in the (Techno-) Economic potential. The second scenario includes the energy demand response to broader based policy instruments (marginal cost pricing and carbon pricing), land-use measures, building and equipment standards and renewables subsidies. Both scenarios include an energy savings multiplier to represent the possible impact of information programs.
Independent potentials were not developed for electricity, natural gas, oil etc. Rather, the model simulated the response by all energy types to changes in technology or energy price-related parameters.
2.1 Study Scope
The modelling scope of the source study scope is defined as follows:
• Sector Coverage: The National DSM study addresses three sectors: residential, commercial/institutional (referred to as commercial) and industrial. Energy supply sectors (electricity, upstream oil and gas and coal) are not included in the study.
• Geographical Coverage: The National DSM study results are presented for seven provinces and regions, including British Columbia and the territories, Alberta, Saskatchewan, Manitoba, Ontario, Quebec, and the Atlantic region.
• Energy Types: All energy types are covered including natural gas, electricity, refined petroleum products and other fuels such as biomass.
• CDM Coverage: Includes energy efficiency, fuel substitution, cogeneration and distributed generation.4 The analysis considers all technologies that are expected to be commercially viable through to 2025.
• Jurisdictions: CDM measures are contemplated for utilities and for all levels of government in Canada (including municipal, provincial and federal).
• Study Period: This study covers a 25-year period. The base year is 2000, with milestone periods at 5-year increments: 2005, 2010 2015, 2020, and 2025. The assessment of CDM potentials occurs from 2005 through 2025. The modelling makes predominant use of the base year 2000 for source information/data used in the model, but we also relate simulated energy demand in 2005 to actual Independent Electricity System Operator demand data.
• How the DSM Impact is Reported: The CDM scenarios analysed in the study describe changes in end-use energy demand that must be met by utilities and other energy suppliers. Therefore, the reduction in energy demand in the study’s scenarios
4 Cogeneration (or combined heat and power) produces both electricity and useful thermal energy simultaneously from the same fuel (or fuels). This allows cogeneration to generate electricity and thermal energy using less input fuel than the stand-alone alternatives, such as boilers and centralized utility generation of electricity.
Modelling and Scenario Documentation
8
affects the amount of purchased energy supply due to energy efficiency and distributed electricity generation (‘self-generation’) within the sectors covered in this analysis. Any electricity generated within these sectors serves to offset the amount of electricity that needs to be supplied. No attempt was made in this report to relate the savings to load curves, or to consider savings relative to base and peak energy demand.
3 Analytical Approach: The CIMS Model
3.1 Structure and Simulation Process
CIMS is an integrated energy-economy model that simulates the technological evolution of fixed capital stocks (mostly equipment and buildings) and the resulting effect on costs and energy use. Technologies are represented in unique sub-models that meet energy service demands in the residential, commercial, transportation, electricity supply, and industry sectors. It is therefore possible to specifically represent the evolution of a technology, or group of technologies, in a reference case forecast and to alter model inputs to simulate alternative forecasts and policy scenarios.
CIMS has been in continuous development since 1986 by the Energy and Materials Research Group in the School of Resource and Environmental Management at Simon Fraser University.5 The model emphasizes the micro-economic level of analysis in that it simulates in considerable detail the equipment and building decisions of firms and households in response to changes in financial costs, consumer preferences, market conditions and availability of alternatives.6 However, it can also incorporate indirect feedbacks that are normally associated with macro-economic models, namely the interaction between sectors that use energy and those that produce or transform it, and shifts in the demand for final and intermediate products as their costs of production change. This capability was not employed in the National DSM study.
As noted, energy use in the Canadian economy is represented by unique sector sub-models in CIMS that are described uniquely for different regions. These sub-models represent stocks of technologies that produce and or consume energy in a particular sector, in terms of the annual quantity of intermediate and final products or services they provide (i.e., tonnes of newsprint, cubic metres of refined petroleum products). Product and energy service demands are linked in sector flow models that describe the sequence of key energy consuming activities required to generate that product or service (see Appendix A for a description of the flow models by sector). The sum of energy represented by these technologies in a sector, or where possible sub-sector, is calibrated to official statistical data of energy demand by sector from the Office
5 The energy demand component of the model, previously called ISTUM, was first developed in the early 1980s by the U.S. Department of Energy as an energy use model of the industrial sector. 6 In this respect CIMS resembles models developed and applied by the electric utility industry in the 1980s for estimating the effects of policies intended to influence technology choices for energy efficiency and fuel switching objectives. CIMS has been used by electric and gas utilities in Canada for this purpose.
Modelling and Scenario Documentation
9
of Energy Efficiency, Natural Resources Canada for the year 2000. CIMS was calibrated to the amount of fuel use, for each sector to within 5%. For the OPA, base year and 2005 electricity demand was also calibrated against Independent Electricity System Operator (IESO demand). This is discussed in more detail (Section 4.2).
In this National DSM study, the following sector sub-models are used for Ontario:
• Residential
• Commercial
• Chemical Products
• Coal Mining
• Industrial Materials
• Iron and Steel
• Metal Smelting
• Mining (metals, uranium, potash)
• Other Manufacturing
• Petroleum Refining
• Pulp and Paper
Forecasts of service demands drive the sub-model simulations in five-year increments, thus allowing for detailed assumptions about industrial output (by product), growth in commercial floor space and the number of households. The rate of technological change is influenced by a retirement function that captures the normal, technical lifespan of energy-using equipment, as well as the technology acquisition algorithm that determines the new stocks required to meet additional growth.
The CIMS simulation in this study involves four basic steps:
1. Assessment of Demand: Technologies are represented in the model in terms of the quantity of service they provide. This could be, for example, vehicle kilometres travelled, tonnes of paper, or m2 of floor space heated and cooled. A forecast is then provided of growth in energy service demand.7 This forecast drives the model simulation, usually in five year increments (i.e., 2000, 2005, 2010, 2015, etc.).
2. Retirement: In each future period, a portion of the initial-year's stock of technologies is retired. Retirement depends only on age.8 The residual technology stocks in each
7 The growth in energy service demand (e.g., tonnes of steel) is often derived from a forecast provided in economic terms (e.g., dollar value of output from the steel sector). 8 There is considerable evidence that the pace of technology replacement depends on the economic cycle, but over a longer term, as simulated by CIMS, age is the most important and predictable factor.
Modelling and Scenario Documentation
10
period are subtracted from the forecast energy service demand and this difference determines the amount of new technology stocks in which to invest.
3. New Technology Competition / Retrofit Competition: Prospective technologies compete for this new investment. The objective of the model is to simulate this competition so that the outcome approximates what would happen in the real world. Hence, while the engine for the competition is the minimization of annualized life cycle costs, these costs are substantially adjusted to reflect market research of past and prospective firm and household behaviour.9 Thus, technology costs depend not only on recognised financial costs, but also on identified differences in non-financial preferences (differences in the quality of lighting from different light bulbs) and failure risks (one technology is seen as more likely to fail than another). Even the determination of financial costs is not straightforward, as time preferences (discount rates) can differ depending on the decision maker (household vs. firm) and the type of decision (non-discretionary vs. discretionary). The model also allocates market shares among technolgoies probabilistically.10 More detail regarding the technology competition algorithm is provided in the Algorithms section.
Retrofitting: In each time period, a similar competition occurs with residual technology stocks to simulate retrofitting (if desirable and likely from the firm or household's perspective).11 The same financial and non-financial information is required, except that the capital costs of residual technology stocks are excluded, having been spent earlier when the residual technology stock was originally acquired.
4. Output: Since each technology has net energy use and costs associated with it, the simulation ends with a summing up of these. The difference between a reference case simulation and a policy simulation (economic and achievable potentials) provides an estimate of the likely achievement and cost of the scenario.
Additional steps are introduced if the model if the energy market (supply and demand) integration and macro-economic functions in the model are used.
9 With existing technologies there may be data on consumer behaviour. However, with emerging technologies (especially the heterogeneous technologies in industry) firms and households need to be surveyed (formally or informally) on their likely preferences. 10 In contrast, the optimizing models will tend to produce outcomes in which a single technology gains 100% market share of the new stocks.
11 Where warranted, retrofit can be simulated as equivalent to complete replacement of residual technology stocks with new technology stocks.
Modelling and Scenario Documentation
11
3.2 Algorithms
3.2.1 Technology Competition Algorithm
New market shares of competing technologies in CIMS are simulated at each end-use (competition node) based on their life cycle cost, according to the following formula:12
Error! Objects cannot be created from editing field codes. (1)
Where MSj = market share of technology j, CC = capital cost, MC = maintenance and operation cost, EC = energy cost, i = intangible cost, r = private discount rate, and v = measure of market heterogeneity.
The main part of the formula (the part inside the square brackets) is simply the standard levelized life cycle (LCC) cost of each technology. In this formulation, the inverse power function acts to distribute the penetration of that particular technology ‘j’ relative to all other technologies ‘k’. A high value of ‘v’ means that the technology with the lowest LCC captures almost the entire new market share. A low value for ‘v’ means that the market shares of new equipment are distributed fairly evenly, even if their LCCs differ significantly. Figure 1 is a graphical representation of the simple case where two technologies (A and B) with different life cycle costs are competing for new market share with different values of ‘v’.
Figure 1: CIMS Logistic Curve
0%
10%
20%30%
40%50%60%
70%
80%
90%100%
0 1 2
Ratio of LCC A and B
Pro
babi
lity
of c
hoos
ing
alte
rnat
ive
A
1
3
5
10
15
The ‘v’, ‘i’ and ‘r’ preference parameters in CIMS are estimated from empirical studies of consumer and business decision-making, in some cases based on past consumption patterns and in some cases (especially with new technologies) based on surveyed preferences for specific technology attributes. The default value for ‘v’ in CIMS is 10, meaning that where a technology has an LCC advantage of at least 15% over its competitor(s) it would capture at
12 CIMS can employ a number of hard controls to limit the penetration of technologies to certain levels (e.g., a maximum of one washing machine per household) as well as a declining capital cost function to simulate learning-by-doing and economies of scale exhibited particularly for new technologies.
v
Modelling and Scenario Documentation
12
least 80% of new stock. The default value for ‘i’ is zero. However, there are some cases in which research suggests a specific value for ‘i’. Also, ‘i’ can be used as a calibration parameter when the values for ‘v’ and ‘r’ are inadequate for simulating the historical penetration rate of certain technologies.
Table 1: Discount Rates used in the Residential, Commercial and Industrial Sub-models Sector Technology Lit. Range
% Source Rate Used
Commercial Building HVAC 25-50 30%* 40%
Other End-uses 25-50
Lohani and Azini 1992
40%
Residential Space heat / Shell 26 - 79 Hartman and Doane 1986
35%
Refrigeration 61 - 108 Cole and Fuller 1980 65%
Other Appliances 30 - 70 Reported in Train 1985 35%
Industrial Process 20 - 50 Hassett and Metcalf 1993
35%
Auxiliary >50 DeCanio 1993 50%.
*A lower discount rate is used for publicly owned buildings (schools and hospitals)
Table sources:
Cole, H., and R. Fuller. Residential Energy Decision-making: An Overview with Emphasis on Individual Discount Rates and Responsiveness to Household Income and Prices. Columbia, MD: Hittman Associates Report, 1980.
DeCanio, S. J. Barriers within firms to energy-efficient investments. Energy Policy 21, no. 9 (1993): 906-914.
Hartman, R. S., and M. J. Doane. Household discount rates revisited. The Energy Journal 7, no. 1 (1986): 139-148.
Hassett, K. A., and G. E. Metcalf. Energy conservation investment: Do consumers discount the future correctly? Energy Policy 21, no. 6 (1993): 710-716.
Lohani, B.N., and A.M. Azini. Barriers to energy end-use efficiency. Energy Policy 20, no. 6, 1992.
Train, K. Discount rates in consumers’ energy-related decisions: A review of the literature. Energy 10, no. 12 (1985): 1243-1254.
3.3 Fuel Switching in CIMS
The electricity forecast produced by CIMS is a function of the technology choice outcomes in the model rather than being directly built from assumptions about changes in ‘energy
Modelling and Scenario Documentation
13
efficiency’, ‘fuel-switching’, and even structure.13 Thus fuel switching is not an assumption in the modelling but an outcome of the input parameters that inform lifecycle cost.
3.3.1 Describing fuel switching in the results
Decomposition method
We have attempted to provide a picture of fuel switching in the reference case and CDM scenarios using a decomposition approach (described in detail in Appendices B and C). This method is applied to sector results to show how much of the change in electricity demand can be ascribed to activity variable (e.g., change in floor-space heated, production outputs etc. over the study period), efficiency (e.g., how much energy was required to heat the floor-space, or to produce a unit of industrial output) and fuel-mix (e.g., was electricity substituted for another fuel, or vice versa?). The fuel mix value represents the change in electricity use due to fuel switching in the sector, as well as differences in efficiency between the fuels. We apply this method after first separating out the influence of cogeneration of electricity with heat.
Note: Structural change is reflected in the fuel switching term. For instance in the reference case, the residential sector shows an overall ‘fuel-switching’ in the forecast towards electricity in the decomposition. This occurs mainly due to growth in minor appliance and air conditioning end-uses cause residential electricity demand to grow more quickly than for natural gas and oil. In the decompositions, fuel switching does not simply reflect change in fuel choice at each end-use, but also the change in the relative importance of end-uses.
In general, fuel switching calculated this way should be interpreted as a change to the overall mix between electricity and other fuels to produce a unit of product (in terms of industry) or to meet the energy services required by households or commercial floor space (in terms of the residential and commercial sectors). It doesn’t exclusively represent a switch at a particular end-use. So for example, a situation in which both electricity and gas use decline, but gas declines more significantly, will show up as a positive fuel switching term in the decomposition (overall the mix of energy types such that the relative share of fuels is less than electricity – relative to the reference case).
Other description
We are able to comment on observed patterns of fuel switching at certain end-uses. However, no effort was done to quantify fuel switching by end-use.
3.3.2 Controlling for fuel-switching in the simulation
In the Economic Potential we chose to limit the technology competitions so that energy efficient technology options would be assessed within each fuel share type. This was to make the results more comparable to other conservation potential studies. This is described more in the Economic Potential Methodology (section 5.1).
13 Many important end-uses are nested – their service demand depends on the simulation outcome of other end-uses.
Modelling and Scenario Documentation
14
4 Reference Case
The Reference Case is a projection of future energy consumption to 2025, in the absence of any new and incremental utility and other institutional market interventions after the base year. It is the baseline against which the energy savings from the economic and achievable potential scenarios are compared. The Reference Case takes into account that, even in the absence of incremental market interventions, there will be an independent and price induced “natural conservation” effect that will drive a change in the average intensity of energy end-uses over time.
To develop the Reference Case three main sub-tasks were carried out with the aim of verifying and updating a wide range of numerical values in the CIMS model, including:
1. Energy end use profiles,
2. Fuel prices, and
3. Economic growth projections.
4.1.1 Energy End Use Profiles
The model contains data on the initial market shares of equipment stocks, by sector and region. Individual types of equipment are characterized in terms of capacity, capital cost, unit energy consumption (and output for energy conversion equipment), non-energy operating cost, expected lifespan and first year of market availability for new technologies. These data are collected from multiple sources including public statistical agencies, energy utilities, literature reviews, industry associations and surveys of sector experts.14 The characterizations of existing equipment stocks are derived from diverse surveys and thus have some degree of inaccuracy, especially in terms of the current operating characteristics of older equipment. To deal with this challenge, data on existing stocks are tracked with disaggregated, sector-specific (and where possible end-use specific), energy consumption data.15
While some review of the technology and end-use profiles took place as part of the National DSM study, this was an internal review whose scope included all sectors and energy forms for all regions in Canada, all of which occurred within a fairly tight budget and time frame.16 The data and profiles would likely be improved with external-expert consultation that is typical of detailed conservation potential reviews. Nevertheless the technology database and profiles available in CIMS model likely represents the most detailed characterization of disaggregate energy consumption in Ontario currently available. 14 The Energy and Materials Research Group, which developed CIMS, operate one of the energy data centres funded primarily by the Canadian government, along with contributions from utilities and industry associations. 15 End-use energy consumption estimates are available regionally for the commercial and residential sector in the National Energy Use Database <<oee.nrcan.gc.ca/corporate/statistics/neud/dpa/data_e/databases.cfm?attr=0>>; Industrial energy data is available on this site as well, though for different types of manufacturing, though typically not for individual end-uses. 16 This was important in that Marbek Resource Consultants Ltd, who have developed significant in-house databases and tools, had the opportunity to review the residential and commercial data in CIMS.
Modelling and Scenario Documentation
15
The technology options and interrelationships in CIMS are not as detailed as many engineering and accounting models. CIMS attempts to bridge the top-down econometric models (with no technological detail) with bottom-up engineering and accounting models (with significant technology detail but no representation of consumer behaviour and macro-economic feedbacks). The model reflects decisions to portray key energy relationships and technology options in the national energy system (as detailed regionally).
CIMS describes technology acquisition and retrofit options for equipment, process and buildings. Behavioural actions are not included. These are energy-efficiency actions that involve no equipment expense, but which require time, thought and action by the end-user (for instance, switching lights off, shutting or opening windows). This focus may lower the potentials derived in the National DSM study, though the limitation to technology adoption is limited the reduction to changes that result in savings that will more likely be permanent.
4.1.2 Energy Prices
The energy prices used in the model are from the National Energy Board’s (NEB) Canada’s Energy Future.17 Two scenarios are presented by the NEB that bound possible outcomes in the future:
• The Supply Push scenario represents a world in which technology advances gradually and Canadians take limited action with respect to the environment. The main theme of this scenario is security of continental energy supply and the push to develop known conventional sources of energy.
• The Techno-Vert scenario represents a world in which technology advances rapidly and Canadians take broad action with respect to the environment and the accompanying preference for environmentally friendly products and cleaner-burning fuels.
The Techno-Vert, or high fuel price scenario, was used in the National DSM study because it was thought to provide the most conservative approach (in that it will trigger a low potential estimate since it decreases the estimate of what can be achieved assuming more energy efficiency in the baseline), since energy prices are higher to reflect the emissions price embodied in fossil fuels. As well, continued growth of gas demand in the power generation sector and in particular the possible replacement in Ontario of coal-fired generation with natural gas lends itself to a scenario with higher energy prices. The Techno-Vert scenario also provides the greatest allowance for technology to maximize conventional natural gas extraction from the Western Canadian Basin, which is one of the most salient differences between the scenarios. In contrast, the Supply Push scenario assumes that conventional natural gas from this source runs out somewhere around 2010 and is assumed replaced with coal bed methane and shale gas. Despite this, the price of gas falls steadily and smoothly over this period. The Techno-Vert scenario assumes that more technology (side drilling, better seismics, etc) will allow Canada to maintain the energy supply status quo until approximately
17 National Energy Board, “Canada’s Energy Future” Scenario’s for Supply and Demand to 2025. (Supply Push and Techno-Vert scenarios). http://www.neb-one.gc.ca/energy/SupplyDemand/2003/index_e.htm
Modelling and Scenario Documentation
16
2025. This fits much better with the smooth fuel price forecasts that are provided in the Techno-Vert scenario.
Table 2 shows the energy prices used for Ontario in the model. Prices were converted to $1995 from $1986 based on consumer price index information from Statistics Canada.18
Table 2: Ontario Energy Prices Used in the Model Ontario 2000 2005 2010 2015 2020 2025
Residential ($1995/GJ)
Electricity 25.35 26.62 31.53 30.32 28.89 26.88
Natural Gas 7.60 9.69 10.11 9.86 9.60 9.30
Light Fuel Oil 12.92 10.88 11.35 11.29 11.20 11.07
Commercial ($1995/GJ)
Electricity 20.28 23.61 26.23 25.13 23.88 22.19
Natural Gas 5.46 7.96 8.38 8.14 7.88 7.57
Light Fuel Oil 12.64 10.59 10.53 10.30 10.06 9.78
Heavy Fuel Oil 6.36 4.62 4.56 4.37 4.16 3.92
Industrial ($1995/GJ)
Electricity 18.80 22.07 24.41 23.22 21.96 20.36
Natural Gas 4.81 6.69 7.11 6.87 6.61 6.31
Heavy Fuel Oil 5.26 5.05 4.99 4.78 4.55 4.29
One exception to the use of the NEB price data is coal, which we believe is exceedingly low in the NEB scenario. After a review of coal prices, we decided to retain the original coal prices used in CIMS, which were from NRCan’s CEOU99 update 2000 and are shown in Table 3.
Table 3: Coal prices ($1995/GJ) used in CIMS Ontario 2005 2010 2015 2020 2025
Industry ($1995/GJ) 2.36 2.30 2.25 2.25 2.25
Electricity Generation ($1995/GJ) 2.06 2.01 1.96 1.96 1.96
4.1.3 Economic growth projections
Almost all sector sub-models in CIMS describe a base year physical demand and use annual growth rates to calculate forecasts over time. The exception to this is ‘other manufacturing’ which is based on gross output.
Growth rates for all sectors represent NRCan growth rates assumed in its new energy and emissions outlook (anticipated release Spring/Summer 2006). The CIMS sub-models were updated with these rates specifically for the OPA project, while the source project (National DSM project) growth rates are based on those used in its older Outlook.
18 Consumer Price Index, Historical Summary. CANSIM, table 326-0002 and Catalogue nos. 62-001-XPB and 62-010-XIB. <www40.statcan.ca/l01/cst01/econ46.htm>
Modelling and Scenario Documentation
17
While most industry sub-models in CIMS are driven by physical production, only economic growth rates are only available from NRCan (in terms of gross output by sector). We assume that economic production growth will be equivalent to physical production growth.
Table 3 provides the provincial annual growth estimates for electricity and the industrial, commercial, and residential sub-sectors. Table 5 shows physical demand assumed in each period.
Modelling and Scenario Documentation
18
Table 4: Average Annual Ontario Growth Rates Sector 2001-2005 2006-2010 2011-2015 2016-2020 2021-2025
Residential 1.44% 1.38% 1.23% 1.11% 1.11%
Commercial 1.58% 1.89% 1.82% 1.98% 1.98%
Industry
Pulp and paper -1.23% 0.04% 0.59% 0.59% 0.59%
Chemical man. -0.20% 1.96% 1.29% 0.64% 0.64%
Iron & steel -0.67% 0.34% 1.09% 1.05% 1.05%
Metal smelting & refining 2.36% 1.32% -0.58% 0.54% 0.54%
Metals and mineral mining 2.26% 2.16% 1.70% 0.38% 0.38%
Other manufacturing 1.41% 2.63% 2.23% 3.58% 3.58%
Industrial minerals 2.82% 1.35% 2.86% 2.08% 2.08%
Petroleum refining 0.00% 5.78% 1.52% 1.53% 1.53% Note: Petroleum refining growth in the first period directly based on Statistics Canada refining output data.
Table 5: Annual Industrial Production /Household and Floor space Demand Sector Unit 2000 2005 2010 2015 2020 2025
Residential million households 4.62 4.96 5.31 5.64 5.96 6.30
Commercial million m2 floor space 219.60 237.47 260.76 285.44 314.80 347.19
Industry
Pulp and paper Mt of pulp and paper 5.17 4.86 4.87 5.02 5.17 5.32
Chemical man. Mt chemicals* 5.90 5.84 6.44 6.86 7.08 7.31
Iron & steel Mt molten steel 12.09 11.69 11.89 12.55 13.23 13.94
Metal smelting & refining
Mt refined product 0.43 0.49 0.52 0.51 0.52 0.53
Metals and mineral mining
Mt mineral ores 26.84 30.01 33.40 36.34 37.04 37.75
Other manufacturing GDP ($1997 billions) 73.34 78.67 89.58 100.00 119.23 142.15
Industrial minerals Mt of clinker 6.50 7.47 7.99 9.19 10.19 11.30
Petroleum refining billion m^3 of refined products
22.65 22.65 30.01 32.35 34.91 37.66
* Only energy-intensive chemicals explicitly described in CIMS
4.1.4 DSM Programs in the Baseline
Initial technology stocks and their energy intensities reflect past purchasing choices that are autonomous, price induced, and CDM influenced. Calibrating CIMS to known energy intensities19 by end-use ensures that the initial stock in the Reference Case base year, 2000, reflects past energy efficiency uptake. CDM programs implemented after the base year 2000 do not further influence the Reference Case.
19 National Energy Use Database. Office of Energy Efficiency, Natural Resources Canada
Modelling and Scenario Documentation
19
4.2 Reference Case Results
4.2.1 Base year and 2005
The base year for the simulation is 2000. Electricity demand in the sub models is calibrated to within 2% of electricity sales described in the NRCan Office of Energy Efficiency databases (Table 6).20
Table 6: Base Year Electricity Demand Calibration Results Electricity Demand TWh Residential Commercial Residential Total
CIMS 41.9 43.7 48.0 133.5 Source data 42.7 44.2 48.6 135.5 % difference -1.9% -1.2% -1.3% -1.4%
We also relate base year electricity demand to that of the IESO, both in the 2000 and in 2005 (Table 7). The second period represents the outcome of the first simulation period in CIMS. As applied in this project, CIMS represents 97% of electricity demand in Ontario, but does not include electricity demanded by transportation, agriculture, and non-metal mining, including upstream oil and gas extraction. We add in an estimate of energy demand in these sectors to make the comparison with IESO demand.
CIMS explicitly represents base year cogeneration in the demand sectors according to installations described in the CIEEDAC cogeneration database.21 Electricity produced by cogeneration offsets industrial, commercial and residential sector demand in the sub models. However, because cogenerated electricity also represents supply contacts by the IESO, we adjusted the share of electricity that offsets demand in the CIMS sub-models to be consistent with the IESO demand. In consultation with the OPA, we assume that 30% share of electricity generated by cogeneration offsets demand.
Table 7 Comparison of Electricity Demand in CIMS to IESO Demand, 2000 and 2005 (TWh/year) 2000 2005
Sectors represented in CIMS 133.5 139.0 Transportation 0.9 0.9 Agriculture 2.3 2.3 Non-metal mining 0.96 1.0 Sum 137.7 143.2 IESO demand 147.4 154.7 Implied transmission and distribution losses 6.6% 7.4%
20 Calibration ensures that the disaggregate picture of energy demand by end-uses described in the sub-models is consistent with statistical data. All energy and fuels described in the model are calibrated. 21 Cogeneration (combined heat and power) is the combined production of both heat and electricity.
Modelling and Scenario Documentation
20
4.2.2 Forecast Simulation to 2025
Table 8 and Figure 2 show the results of the reference case simulation. The reference case scenario describes end-use energy demand that must be met by utilities and other energy suppliers net of distributed electricity generation, which offsets demand. The following are key determinants of the forecast trend:
• Improvement in end-use efficiency – this occurs as a function of the cost-effectiveness of the options available to replace existing technology stocks.
• Fuel choice – this is based on the relative difference in electricity prices and other fuel prices, the relative installed costs of equipment, and other consideration such as the ease of switching.
• Activity – this represents expected growth in households, floor space and industrial production.
• Structural change - Energy service demands grow at different rates. For instance, some energy services may only be growing at the rate of stock replacement and new population or demand growth, while other end-uses (such as air conditioning, appliances) also grow in terms of household saturation In terms of industry, structural change represents change in product mix and process choice.
We discuss these factors in relation to how they relate to end-use trends in the forecasts below. As part of the OPA study deliverables, we have provided a quantitative picture of these underlying factors. This task is challenging because the electricity forecast produced by CIMS is a function of the technology choice outcomes in the model rather than being directly built from assumptions about changes in ‘energy efficiency’, ‘fuel-switching’, and even structure.22 We apply a decomposition analysis to the forecast. Appendix B describes the methodology and the results of the decomposition.
Table 8: Reference Case Electricity Demand to 2025 (TWh/year) Electricity Demand (TWh/year) 2000 2005 2010 2015 2020 2025
Industry 43.65 42.78 44.68 46.86 51.77 58.80Commercial 47.98 50.84 52.54 55.16 59.16 64.43
Residential 41.90 45.36 47.66 48.62 50.74 53.83Total 133.54 138.97 144.88 150.64 161.67 177.06
Note: The total represents 97% of consumer demand.
22 Many important end-uses are nested – their service demand depends on the simulation outcome of other end-uses.
Modelling and Scenario Documentation
21
Figure 2: Reference case forecast by sector
-
10
20
30
40
50
60
70
2000 2005 2010 2015 2020 2025
Ele
ctri
city
De
ma
nd
(T
Wh
)
industry
commercial
residential
A detailed breakdown of this forecast by end-use is provided to the OPA. We discuss some key underlying trends by sector.
Residential Sector
Electricity demand grows by 28% over the forecast period. Electricity intensity declines (GJ/hsld) at an average rate of 0.2% per annum. To put this in context, electricity intensity in the past has generally declined (according National Energy Use Database (NEUD), NRCan), but has grown recently. Between 1993 and 2003, intensity declined by 0.5% annually, while between 1998 and 2003 intensity increased by 1.6 % annually).23 Figure 3 shows the contribution of end-uses to electricity consumption in 2000 and 2025.
23 Past intensities calculated based on data retrieved from the National Energy Use Database, Office of Energy Efficiency, Natural Resources Canada.
Modelling and Scenario Documentation
22
Figure 3: Electricity Consumption (TWh/year) by End-use in 2000 and 2025, Residential Sector
-
10.00
20.00
30.00
40.00
50.00
60.00
TW
h/year
2000 2025
Minor Appliances
Water Heating
Large Appliances
Lighting
Air Conditioning
Space heating
The largest contributor to growth is minor appliances – these are the collection of electric-consuming devices such as televisions, answering machines, garage door openers, battery chargers, DVDs etc. According to the National Energy Use Database, electricity demand to power these devices in the residential sector increased by 63 percent from 1990 to 2003. In the reference case forecast we assume continued growth. A similar pace of increase is maintained in the short term (5% per annum), gradually slowing to 2% per annum by 2025.
Air conditioning also grows in the forecast. According to NEUD data, the penetration of air conditioners by household in Ontario is growing significantly, particularly for central systems. This forecast projects recent growth trends of 4.1% per annum and 0.1% for room air-conditioners, to a saturation point in 2010, whereby 70% of household in Ontario are assumed to have central systems and 20% have room systems. It was felt unlikely penetration overall for air conditioning would increase beyond 90%, or that central systems would supplant room units. For instance, in the southern United States, 93% of households have an air conditioning (with room air accounting for 25% of households).
Beyond 2005, we assume that most standard appliances are assumed to be at their saturation point, with no increased penetration assumed in the forecast. This is based on a review of NEUD growth rates for these appliances. The exceptions to this are for dishwashers and for clothes washers / clothes dryers, which have shown to have increased their penetration over the past 5 years. In the forecast we assume that,
• The penetration of clothes washers and dryers is assumed to reach 90% (reached in 2005 for washers and 2010 for dryers).
• The penetration of dishwashers is assumed to reach 73% (reached in 2025).
Modelling and Scenario Documentation
23
The penetrations are guided by a breakdown of appliance penetration by income level reported by the US EIA.24 Top income levels report clothes washer/dryer penetration is 92% for top income bracket, and 83% for dishwasher. Electricity intensity typically declines per appliance – at a level in keeping with past performance improvements. The exception to this is dishwashing where models are moving towards those that heat water directly using electricity relative to using water heated by the household water furnace.
Lighting grows with household growth (staying at a penetration of 29 light bulbs per household). There is slight increase in the penetration of compact fluorescent lighting over the forecast (from meeting 2% of lighting demand in 2005 to 10% in 2025).
Electricity for space heating and water heating reflect both increase efficiencies and fuel switching towards natural gas
Commercial Sector
Electricity demand grows by 34% over the forecast period. Electricity intensity declines at an average rate of 0.7% per annum. To put this in context, electricity intensity in the past has shown a general increase. Between 1993 and 2003, intensity increased by 1.3% annually, while between 1990 and 2003 intensity increased by 0.3 % annually).25 According to the National Energy Use Database for Ontario, this trend is occurring from fuel switching to electricity in space heating and growing lighting intensity. While the CIMS forecast does include some fuel switching to electricity for space heating, it occurs at a more modest rate than recent NRCan trends suggest. CIMS shows a decline in electricity intensity over the forecast due to efficiency gains. This is in keeping with Marbek in-house data, which shows lower lighting intensities in new commercial buildings space.
However it is likely that energy efficiency will be accelerating, particularly after 2015, due to:
Aging of the existing stock which will force significant renovation and demolition activity
The green construction revolution will significantly accelerate and likely transform the market after 2015
Research efforts to improve the efficiency in some end-uses such as lighting will bear fruit after 2015 with available products.
Continued pressure on energy prices will bring about more aggressive energy efficiency
Figure 4 shows the contribution of end-uses to electricity consumption in 2000 and 2025. Commercial energy intensity (kWh/ m2) is declining because new building intensities are lower than existing building. Aging of the existing stock will force significant renovation and demolition activity. For instance, according to Marbek’s files, heating intensities in new building are 30% lower; cooling 18% lower, relative to existing buildings. 24 U.S. Energy Information Administration. 2004. The Effect of Income on Appliances in U.S. Households. http://www.eia.doe.gov/emeu/recs/appliances/appliances.html 25 Past intensities calculated based on data retrieved from the National Energy Use Database, Office of Energy Efficiency, Natural Resources Canada.
Modelling and Scenario Documentation
24
Figure 4: Electricity Consumption (TWh/year) by End-use in 2000 and 2025, Commercial Sector
-
10.00
20.00
30.00
40.00
50.00
60.00
70.00
TW
h/year
2000 2025
Auxiliary equipment
Water Heating
Lighting
Ventilation
Space Cooling
Space Heating
Building segments within the commercial sector grow at differential rates. Retail, warehouse and miscellaneous segments grow relatively less. This is based on assumptions in the NRCan Commercial Energy Use model. While different segment in CIMS assume different electricity intensities, the relative mix of the change doesn’t impact electricity consumption during the forecast.
The demand for most end-uses are growing at a rate of stock replacement and new floor space growth. The main exception to this an increase in the commercial floor space heated by electricity. This has occurred over the past 5 years, according to OEE estimates, and is assumed to continue during the forecast.
Industrial Sector
Electricity demand grows by 35% over the forecast period. Electricity intensity declines at varying rates by industry (Table 9), with the exception of iron and steel, which shows an increase due to increasing production of secondary steel.
Modelling and Scenario Documentation
25
Table 9: Average annual change in electricity intensity over the reference case by industry sub-sector Sector
Pulp and Paper -0.8%Chemical Manufacturing -0.2%Iron & Steel 0.3%Metal Smelting & Refining -1.1%Metals and Mineral Mining -0.9%Other Manufacturing -0.4%Industrial Minerals -0.5%Petroleum Refining -1.7%
Figure 5 shows the contribution of end-uses to electricity consumption in 2000 and 2025, and by sub-sector in Figure 6.
Figure 5: Electricity Consumption (TWh/year) by Sub-sector, Industrial Sector, 2000 and 2025
-
10.00
20.00
30.00
40.00
50.00
60.00
2000 2025
petroleum refining
industrial minerals
other manuracturing
metals and mineral mining
metal smelting & ref ining
iron & steel
chemical man.
pulp and paper
Modelling and Scenario Documentation
26
Figure 6: Electricity Consumption (TWh/year) by end-use in 2000 and 2025, Industrial Sector
-
10.00
20.00
30.00
40.00
50.00
60.00
70.00
TW
h/year
2000 2025
Lighting
HVAC
Heat Production
Electrochemical
Machine Drive
Many sub-sectors show minimal growth in electricity demand due to both low production growth assumed in the forecast and continued efficiency improvements. The exception to this is other manufacturing and iron and steel. Other manufacturing’s share increases due to sizeable production growth assumed in the forecast (2 - 3.5% per annum). Electricity demand grows for iron and steel production because most new growth is accommodated in secondary steel rather than in primary, integrated mills. Secondary steel is considerably more electricity-intense.
Electricity demand declines in pulp and paper due to negligible production increases. We also assume a relative decline in newsprint relative to other papers, and relatively higher growth in chemical pulping relative to mechanical pulping. Both of these trends also lower electricity consumption.
In terms of end-use trends, machine drive dominates electricity used in industry. Machine drive represents motor systems used for pumping, air displacement, compression, conveyance and direct drive. There is little improvement in motor efficiency during the forecast (which is already quite high). However, there are improvements to the auxiliary systems that use motors.
Direct heat, the next largest end-use, grows fairly significantly. This end-use applies mainly to iron & steel production and ‘other manufacturing’, which have greater than average production growth during the forecast (particularly for other manufacturing).
Modelling and Scenario Documentation
27
APPENDIX A - Sector Detail
Industry
Each industrial sub-model in CIMS has its own driving variable, usually the total amount of final product produced or the amount of raw input processed (e.g., tonnes of steel, tonnes of mineral ore throughput, m3 refined petroleum products). The driving variables are determined exogenously in the reference case (but can be can respond to production cost changes due to policy simulation if the macro-economic capability of the model is used).
In CIMS, the product and energy service demands in a sub-sector are linked in a flow model that describes the sequence of activities required to generate that product. The energy flow model for the iron and steel industrial branch is shown in Figure 20 below as an example (detailed flow models for all sectors follow). A CIMS flow model is geared towards representing technology evolution and energy consumption rather than economic criteria (as in an econometric model where units are typically in monetary terms) or actual mechanical processes (as in the blueprints or process flow diagrams used by engineers). Because the emphasis is on energy consumption and not material flow, the nodes in the flow model represent process stages in which energy consumption can be distinctly estimated. The flow model describes hierarchical nodes, linked by engineering ratios. Technology competitions take place at the lowest level nodes in the hierarchy.
Figure 7: Energy Flow Model of the Steel Products Industry
Nodes shaded in green represent unique industry processes, while nodes shaded in other colours represent auxiliary systems. These are generic energy services that are supplied to the major process technologies and are shared by all CIMS industry sub-models. The auxiliary systems fall into four general categories: steam generation (boilers and cogenerators), lighting, heating, ventilation and air conditioning (HVAC), and electric motor systems including pumps, fans, compressors and conveyors. Figure 21 shows the energy flow diagrams for auxiliary systems not included in Figure 20. In some cases, the energy service
Modelling and Scenario Documentation
28
meets the direct need for steam, pumping or compression, while in other cases, it serves only to provide suitable conditions for production to continue, as in the case of lighting and HVAC systems.
Figure 8: Auxiliary Flow Model Diagram
The sections that follow describe how key auxiliary systems and the CIMS industrial model captures process specific components. The section on auxiliary systems covers steam generation and electric auxiliary services. Lighting and HVAC are not described in detail because the energy demanded by these services is small. The section on process specific systems covers each of the CIMS industrial sub-models.
Modelling and Scenario Documentation
29
Auxiliary Systems
Steam Generation
A number of industry process technologies generate steam as a by-product. CIMS accounts for steam generated by these technologies in the overall steam demand in each industry. Boilers and cogenerators must meet any shortage of steam.
Boiler efficiency can vary greatly depending on boiler design, age, and fuel used. For modern oil and gas boilers, thermal (first law) efficiencies may be 85% or higher. Boiler efficiencies can be increased by introducing non-condensing and condensing heat recovery systems and by installing regenerative burners with computerized fuel/air mixtures to maximize fuel efficiency.
Cogeneration is the sequential production of electricity and useful thermal energy, usually as steam or shaft drive power. In cogeneration a boiler is used in conjunction with a turbine system to generate electricity, with the "waste" steam going to meet process steam requirements. In comparison to conventional electricity generation, cogeneration facilities are capable of conversion efficiencies of up to 85% (including both the steam and electricity produced) as compared to utility condensing turbine systems, which have efficiency in the 30% to 40% range.
Electric Auxiliary Systems
The vast majority of electricity consumed by industry is used by motor systems. A motor is the core component of a much broader system of electrical and mechanical equipment that provides services, including hydraulic power, compressed air, motive power and air flow. Opportunities for efficiency improvement exist in both the motor itself, and in the latter systems – pumping, air displacement, compression, conveyance as well as other types of machine drive that are unique to a given production process (direct drive).
Figure 21 above is a generic energy flow model representing motor and steam auxiliary systems as they appear in the various CIMS industrial sub-models. As the figure indicates, CIMS simulates six auxiliary groups related to process drive (in red). The first five demand-shaft or mechanical drive provided by the sixth group, motors (machine drive). Each group simulates technology evolution of a specific system.
Increased disaggregation in some of the groups reflects differing end-uses and levels of efficiency. For example, compressor systems vary in electrical efficiency by size. Therefore, competition between different technologies must be separated by size. The energy flow model reflects this by displaying two sizes of compressors, which are available for use by the model.
Pumping, air displacement, compressing and conveyance systems have been represented as packages within CIMS. The components of each of these systems are associated with the key end-use. For example, a pipe, throttle and speed control device is attached to a pump, and the entire package is called a pump system. In the model, there are at least two types of systems:
1.) The least efficient system comprised of the least efficient components.
Modelling and Scenario Documentation
30
2.) The most efficient system comprised of the most efficient components.
The system efficiency is then a function of the efficiency of all the components within that system. In some cases intermediate efficiencies have also been listed. New technologies may be added to reflect the intermediate range.
Electronic variable speed drives (VSDs) provide speed control by varying the rotating speed of the induction motor to match variable speed requirements of driven equipment. Pumps and fans, for example, are often required to deliver varying qualities of fluid flow. When this control procedure is used the energy consumption of the fan or pump is dramatically reduced during part load operation, as compared to the conventional practice of using a valve or damper to reduce flow.
Process Specific Systems
Process specific flow models are described for each industry sector sub-models in CIMS. These represent direct energy consumption as well as processes that draw on auxiliary services.
Chemicals Manufacturing
Unlike some of the other industrial sectors modelled in CIMS, the chemical products sector is not dominated by a single energy intensive product to which all others can be associated. As a result, several products and processes must be modelled to accurately reflect the energy flow within the industry. The dominant chemical products, from an energy consumption perspective, are: chlorine, caustic soda (sodium hydroxide), sodium chlorate, hydrogen peroxide, ammonia, methanol, ethylene, propylene and polymers. Sub-nodes in the model represent different process stages in which energy consumption can be distinctly represented.
Ethylene, methanol, toluene and xylene are the primary petrochemicals produced. Crude oil and natural gas feedstocks provide the raw material from which these are produced. Ammonia, chlorine (often produced in complement with caustic soda) and sulphuric acid are the major inorganic chemicals produced. Like petrochemicals, natural gas provides the principal raw material used in the manufacture of ammonia. Ammonia, in turn, forms the primary feedstock for urea, ammonium phosphate, ammonium nitrate, ammonium sulfate and nitrogen solutions commonly sold as fertilizers.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
31
Fig
ure
9: P
roce
ss S
ervi
ces
Flo
w M
odel
, Che
mic
al P
rodu
ct M
anuf
actu
ring
Modelling and Scenario Documentation
32
Industrial Minerals
The industrial minerals sub-model of CIMS generally includes all industry involved in the production of all non-metallic mineral products. It encompasses the cement, lime, glass and brick industries, with cement and lime as the major energy consumers. These key products both require significant quantities of thermal energy per unit production during the calcining process. Other energy-demanding processes involved in cement or lime production typically include raw materials preparation (this excludes extraction of the raw materials from their in situ locations, even though these may be within the plant gate), and finishing or finished materials preparation. Calcining remains the most energy intensive part of the processes, consuming large quantities of combustible fuel (85% or more of the total energy consumed). Initial and final preparation of the materials requires electricity to drive motors that are attached to various auxiliary and grinding devices.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
33
Fig
ure
10:
Pro
cess
Ser
vice
s F
low
Mod
el, I
ndus
tria
l Min
eral
s
Modelling and Scenario Documentation
34
Iron and Steel
The CIMS iron and steel model represents this industrial branch as a process whole, including the production of coke from coal, the incorporation of iron ore (including any agglomeration that occurs on site) and ending with the set of end products. End products include the three basic forms (slabs, blooms and billets) and, in many cases, their sub-forms. The iron and steel industrial branch has a high energy-intensity per tonne of product.
Presently, two different processes generate Canadian steel. In the first process, coke, a coal derivative, reduces iron oxides in ore to pig iron in a blast furnace. Basic oxygen furnaces (BOFs) then purify this liquid iron along with some scrap by injecting high purity oxygen, which is itself an energy-intense product. In the second process, electric arc furnaces (EAFs) recycle 100% scrap metal. In both cases, the molten steel (with carbon content less than 2%) may be cast, inspected, reheated and finished. Each of a wide variety of finished products requires varying inputs of heat and mechanical energy to its manufacture.
The steel industry is rapidly modernizing. Most of these technologies concentrate on energy-saving measures and reuse of what had historically been waste heat. Direct reduced iron and direct smelted iron processes, presently in pilot scale plants, may eventually make the coke-dominated process redundant. More advanced finishing processes such as direct rolling and thin slab and thin strip casting will eliminate reheating of steel (typically done after initial inspection).
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
35
Fig
ure
11:
Pro
cess
Ser
vice
s F
low
Mod
el, I
ron
and
Stee
l
Modelling and Scenario Documentation
36
Mining
The CIMS mining model represents metal mines as well as potash mining. Other non-metal mines are not modelled in CIMS, as they currently consume a relatively small percentage of total mining energy. Activities related to coal mining and oil and gas extraction are covered elsewhere in CIMS.
Metal mines are typically categorized as either open-pit or underground. While the general production processes that occur in both categories are about the same, specific aspects of the mining technologies can differ significantly. For example, underground mining operations must address air quality issues in the mineshaft, which may require: cooling, heating and ventilation.
The basic processes for metal mining include most of the following generic steps (although not always in this order): initial breaking, transport (hoisting, trucking, conveying), crushing, grinding, concentration (mineral separation) and waste disposal (rock and tailings). Energy consumption in the final smelting and refining stage is not included in this sub-model but has been included in the CIMS metal smelting model.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
37
Fig
ure
12:
Pro
cess
Ser
vice
s F
low
Mod
el, M
etal
s M
inin
g
Modelling and Scenario Documentation
38
Metal Smelting and Refining
The metal smelting and refining model of CIMS includes establishments that are primarily engaged in manufacturing finished metal products excluding iron and steel. These processes consume very high amounts of energy per unit of product. The CIMS sub-model explicitly includes processes related to aluminium, nickel, copper, zinc, lead, magnesium and titanium. Though economically important, minor elements such as gold, silver, platinum and cadmium are not represented explicitly because they are often processed in conjunction with the metals listed or are processed in too small a quantity to require direct representation.
In metal smelting, metal is typically extracted from the concentrate by leaching (hydrometallurgical recovery) or through heat (pyrometallurgical recovery). In Ontario, for example, all zinc smelters use hydrometallurgical techniques while copper and nickel smelters use pyrometallurgy. Smelting processes generate products that contain between 95% and 99% pure metal. Refining, if it occurs, depends on the type of smelting technique used. Products prepared through hydrometallurgy tend to use electrowinning for refining while products prepared through pyrometallurgy tend to use electrolysis or fire refining. Refined products are at least 99.99% pure metal.
Purification of the aluminum is accomplished utilizing the Hall-Heroult process, which requires large quantities of electrical energy. Although Canada has no aluminum ore body of sufficient concentration to warrant extraction, readily available, inexpensive electrical energy attracts companies who wish to refine the metal from alumina (concentrated ore). Recycled aluminium requires only 5% of the energy of original production. Increasing demand for conservation and recycling may have a significant effect on the production of virgin metal.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
39
Fig
ure
13:
Pro
cess
Ser
vice
s F
low
Mod
el, M
etal
Sm
elti
ng a
nd R
efin
ing
(Non
-fer
rous
)
Modelling and Scenario Documentation
40
Petroleum Refining
This industrial sector includes establishments primarily engaged in manufacturing petroleum products. Activities such as the manufacture of lubricating oil and grease or asphalt and coal/coke products are included in CIMS where they are carried out in these plants. Establishments that are primarily engaged in these activities are not included; however, because of their comparatively small size.
The basic process involves refining crude oil into a number of products including: ethane, propane, butane, motor gasoline (typically 40% of total product), naphthas, jet fuel, petrochemical feedstocks, distillate fuel oil, diesel fuel, residual fuel oils, lubricants, coke and asphalt. All crude includes these products in varying concentrations. However, it is possible to increase output of more desirable products per unit crude input by using cracking and reforming processes.
The type and quality of crude and the processing requirements to generate the end products determine the refinery’s complexity and have significant impact on energy consumption in the plant. In spite of this, however, the basic processing procedures are common to all refineries. The major process steps include: atmospheric distillation, vacuum distillation, cracking, desulphurization, alkylation, isomerization, reforming, sulphur recovery and hydrogen production. The last two processes should not be considered part of the actual refining process; they act as utility functions to clean out sulphurous compounds or provide hydrogen for specific in-plant operations.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
41
Fig
ure
14:
Pro
cess
Ser
vice
s F
low
Mod
el, P
etro
leum
Ref
inin
g
Modelling and Scenario Documentation
42
Pulp and Paper
The pulp and paper industry includes establishments primarily engaged in manufacturing pulp, paper, paperboard and building and insulation board. These establishments are included in the CIMS pulp and paper sub-model because they are large in size and their products are highly energy intensive when compared to other industry products. Paper products, which do not form part of the sub-model include: asphalt roofing, paper box and bag, and other converted paper product industries. These are included in the other manufacturing model of CIMS. The pulp and paper industrial branch has a high energy-intensity per tonne of product.
Pulping processes can be broken down into three basic types: chemical, mechanical, and recycled. Mechanical pulping, which tends to be very electricity intensive, uses wood fibre much more efficiently than does chemical pulping. Recycled pulp consumes considerably less energy than either of the other two major processes but input supply is limited and the collection and transportation of this supply can significantly increase the energy intensity of the product. The products generated from each of these processes are distinct enough that they generally do not compete on the open market.
Each pulp product serves as the stock for many types of paper products. Newsprint, the largest category in terms of production (> 60%), can use feedstock from each of the pulping processes. Other paper products incorporated into this sub-model includes tissue paper, uncoated and coated paper (or wood-free paper), and linerboard, each of which may have varying degrees of stock inputs.
The following steps in the production of pulp and paper are common to both chemical and mechanical mills where raw logs serve as feedstock: wood debarking, wood chipping, pulp making, pulp washing, pulp bleaching, pulp drying, paper stock preparation, paper sheet formation, paper pressing and paper drying. Mills that use chips or recycled paper as feedstock may exclude some of these steps.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
43
Fig
ure
15:
Pro
cess
Ser
vice
s F
low
Mod
el, P
ulp
and
Pap
er
Modelling and Scenario Documentation
44
Other Manufacturing
The CIMS industrial sub-model for other manufacturing industries has been created to capture those industries within a region, which, on their own, do not consume enough energy to merit the development of a separate model. When considered as a group, however, they may consume a significant (if not the dominant) portion of the energy within any one region. The industries considered part of the other manufacturing sector are: food, beverage, tobacco, rubber products, plastic products, leather and allied, primary textiles, textile products, clothing, wood, furniture and fixture, printing and publishing, fabricated metal products, machinery, transportation equipment, electrical and electronic, and other manufacturing.
The other manufacturing group of industries typically includes a large variety of technologies and processes. Usually, a simplified flow model of generic energy services represents adequately the energy consumption of these industries as a whole. Steam boilers, process heat, space heat, and electricity for lights and electric motors and their attached auxiliary devices (pumps, conveyors, fans and the like) constitute the bulk of energy-using services.
Unlike other “major” industries the other manufacturing industry as a group does not have any single product which dominates the production processes and to which the production of all other products may be linked. As a proxy for the physical output measures used in the sub-models major industries (i.e., tonnes of steel or cubic metres of gasoline) the output of the other manufacturing industry is represented in monetary terms such as the RDP (real domestic product) of the industry.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
45
Fig
ure
16:
Pro
cess
Ser
vice
s F
low
Mod
el, O
ther
Man
ufac
turi
ng
Modelling and Scenario Documentation
46
Residential
The CIMS residential sector sub-model attempts to capture all energy demand by residential buildings. This includes single family dwellings, duplexes, row housing, walk-ups and large apartments. Large apartment buildings are often considered commercial buildings by utilities, but, since their energy consumption patterns better fit within the residential sector, they have been included as part of this sub-model.
Figure 30 presents CIMS’ residential sector flow sub-model. The number of residences required in any year drives the model. The flow model demonstrates the linkages between the driving variable and the energy end-uses in the residential sector. The number of residences are determined exogenously, based on future growth estimates of population and economic activity. The end-uses are linked to the driving variable through engineering ratios based on historic relationships between a residence and each end-use. This variable directly determines the demands for all services, except for water heating and furnace fan demand. The technologies for services that require hot water and furnace fans establish those demands.
The following sections describes energy end-use services incorporated into the model, the variables that affect the amount of energy consumed in providing the services, the relationship between the end-use and the driving variable at the primary node (number of households) and any other interrelationships that permit a more accurate simulation of the end-uses.
Space Heating
A variety of factors uniquely determine the energy consumption required to heat a household. These factors include the physical building envelope, the type of heating system, maintenance of both the envelope and heating system, climate and exterior landscape features, and thermostat settings. Modelling this service requires many assumptions and generalizations. Due to the multitude of housing types, the sector must be condensed into groups consisting of thousands of households, which are represented, in the model by one typical house archetype. All heating systems cannot be included, due to the large number of possibilities; thus the models contain only the most common systems and those considered likely to gain significant market shares in the future. Case studies and surveys provided estimates for geographical and behavioural assumptions.
Each space heating technology in the residential CIMS sub-model combines a shell archetype and a heating system. Each archetype, which is a set of physical measurements for components such as floor space, window area, insulation levels, and rate of air exchange, represents a group of households. Modellers use archetype descriptions to calculate the heat load and cost of the typical building as well as the cost and energy savings of retrofitting. For each shell archetype, CIMS includes a choice of four or five different heating systems. These systems include oil furnaces, natural gas furnaces (typically three efficiency levels with 78% to 90% efficiency), high efficiency space heat/hot water systems, wood stoves, electric baseboard heating, and electric air source heat pumps.
In contrast to the other energy services in the residential sector, the amount of space heating demanded by a residential unit can vary considerably depending on the building type. As a
Modelling and Scenario Documentation
47
result, the space heating service node of the model is divided into three basic housing types: apartment, attached (row), and single family detached dwellings. The apartment category includes both high and low rise apartments while the attached category consists of duplexes, triplexes, and townhouses. As described above each of these housing types have been divided up into different archetypes to describe the differences in building heat loads and the fuels and efficiencies of the heating systems.
Apartments
In the apartment node, two basic technology archetypes are available, standard and ‘improved shell’. The standard archetype generally represents the existing or a baseline shell design that demands a relatively high heat load while the improved apartment shell design represents better insulation, windows, etc., demanding a lower heat load. Each of these archetypes is divided into several different heating system types based on the type of fuel and efficiency of the heating system, e.g., oil heating, several efficiency levels of natural gas heat (62%, 78%, 90%) and electric heat. During each model run new apartments can be purchased either from the options available from existing stock or from a selection of new buildings with an improved shell design and better heating systems. There are no retrofit options currently available for existing apartment stock.
Attached Housing (Row)
The attached housing category is modelled in the same manner as the apartments.
Single Family Dwellings
The greatest level of technological variety in the residential sector exists in the types of systems for heating single family dwellings (SFDs). The possibility exists to model the existing homes in two separate climates (described as Primary and Secondary), however, this distinction is only used currently in CIMS for British Columbia, where significant climate variation occurs between coastal and interior regions of the province.
In the basic CIMS residential sub model, SFDs are divided up into two categories, existing and new housing. Two archetypes are used to model all existing houses. Archetype A houses have very little insulation and, therefore, higher heat loads than Archetype B houses. Where wood consumption is present, wood stoves meet part of heat load demand. CIMS allows houses in either of these archetypes to retrofit to improve insulation levels. The six possible heating systems for the current housing stock are oil furnace, electric baseboard heating, air-source heat pump, standard natural gas furnace (78% efficient), high-efficiency natural gas furnace (92% efficient), and high efficiency space heat/hot water systems.
In each region, three types of houses compete for new stock: standard, improved and R-2000 archetypes. R-2000 houses meet the current heating budget specified by R-2000 standards for the region. The improved archetype is based on the ratio between standard and improved archetypes. Houses can be upgraded to better insulated categories through retrofits. Heating systems are the same as those modelled for existing homes. In some regions wood stoves may be chosen to heat standard and improved archetypes.
Modelling and Scenario Documentation
48
Non-Appliance Hot Water Use
This service meets the hot water required for baths, showers, hand washing, and cleaning. The competing technologies, shower heads and faucets with varying flow (litres/minute) levels, consume hot water rather than energy. The water demand determined by this node contributes to the hot water demand that drives the hot water heating service.
Lighting and Major Appliances
Each of the following appliances is modelled as a separate service:
Refrigerators Freezers Dryers Ranges Dish Washers Clothes Washers Air Conditioners Lighting
The stock is measured in number of appliance units (i.e., number of refrigerators) and the ratio between the service and the driving variable (number of households) is the market penetration of the appliance. Generally, there is a range of efficiency levels for new stock in each major appliance service.
Clothes washers demand both electricity and hot water. Dish washers are disaggregated into two categories: machine, which demand electricity and hot water, and non-machine, which only consume hot water.
Minor Appliances
This service contains many diverse technologies including televisions, car block heaters, lawn mowers, and computers. Although, these appliances collectively consume a significant portion of household energy, individually either their consumption or their market penetration is low. As a group, this end-use is growing considerably. Efficiency improvements are outweighed by increase in number of goods.
Renewable Energy Sources and other Distributed Energy
A number of renewable technologies (in addition to wood stoves) can be applied at the household level to provide both passive and active energy. These include photovoltaic panels, geothermal, solar hot water heating and passive solar design. Options are also increasingly becoming available to generate electricity using combined heat and power generations at the household level, for instance through fuel cells and through micro-CHP (for instance the stirling engine). Some of these options are modelled at the end-use in which they meet thermal demand, rather than as a unique ‘supply’ node in the flow model.
Water Heating
The water heating demand depends on the technologies used to supply services of non-appliance hot water, dish washing, and clothes washing. For most purposes, hot water must be at 60oC and, unheated water generally enters the household at 15oC. Both storage tank and non-storage tank water heaters are modelled. Cogeneration is modelled at this node to meet water heat loads
Modelling and Scenario Documentation
49
(while generating electricity for household use). Most types of cogeneration systems are modelled only for apartments. The exception to this are fuel cells and stirling engines are modelled for both apartments and non-apartments at this end-use
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
50
Fig
ure
17
Ene
rgy
Flo
w M
odel
for
the
Res
iden
tial
Modelling and Scenario Documentation
51
Commercial
In CIMS, the definition of the commercial sector excludes light industrial facilities, which are included in the other manufacturing sub-model, and high-rise residential apartments, which are included in the residential sub-model.
For more information about the development of the commercial sector sub-model please see ISTUM-C: Canadian commercial Building Sector Energy End-Use Model (Strickland, 1996). The model was adjusted during a 1999 project, which represented greenhouse gas emissions abatement actions from the Buildings Issue Table in the National Climate Change Process. These were reviewed during the source study for this report.
Figure 31 shows CIMS’ commercial sector energy flow model. CIMS considers commercial floor space as being the driving variable (primary node) of the commercial sector. The flow model demonstrates the linkages between the driving variable (commercial floor space) and the energy end-uses in the commercial building sector. The amount of commercial floor space is determined exogenously, based on future growth estimates of population and economic activity. The end-uses are linked to the driving variable through engineering ratios based on historic relationships between commercial floor space and each end-use.
Mo
delli
ng a
nd S
cena
rio D
ocu
me
ntat
ion
52
Fig
ure
18 C
omm
erci
al F
low
Mod
el
Modelling and Scenario Documentation
53
Lighting
In the commercial building sector, lighting consumes more electricity than any other end-use. Lighting is measured in terms of the number of square meters of commercial floor space that it serves. Therefore, the lighting service has a one to one ratio with commercial floor space.
Lighting is not modelled separately by building segment. CIMS simulates lighting at different energy efficiency levels. The base technology represents the average efficiency of stock in 2000. The ‘new’ technology represents the marginal energy efficiency of new stock built in 2000. Other options represent actions to adopt more efficient lighting equipment (super T8, advanced HID, compact fluorescent lighting) and improving lighting design.
The lighting end-use category is modelled as three sub-categories: general area lighting, high bay lighting and service lighting. General area lighting is the dominant category of lighting and refers to the lighting found throughout general office areas and uses predominately fluorescent tube lights. Service area lighting refers to lighting systems used in corridors, lobbies, vestibules, washrooms, and any non-critical areas adjacent to areas illuminated by general area lighting. Lastly, service lighting refers to the lighting systems used in high bay areas of a building - large atria and lobbies, gymnasiums etc. The energy intensity of any sub-category depends on the efficiency of the lighting equipment and the use rate.
CIMS does not currently account for the energy consumed for exterior lighting such as street lighting, architectural lighting or parking lot lighting.
Shell/HVAC
The Shell/HVAC category is disaggregated into nine representative building segments (sub-categories). These segments, listed below, combine two or three different building types to group buildings with similar HVAC system types and usage. For example, the schools category includes elementary schools, colleges and universities.
A. Warehouses and Wholesale Outlets
B. Hotels and Motels
C. Schools and Universities
D. Small Office Buildings
E. Large Office Buildings
F. Small Retail
G. Large Retail
H. Hospitals and Nursing Homes
I. Miscellaneous buildings (not including residential or light industrial).
Modelling and Scenario Documentation
54
The mix of buildings in the segments is derived from Natural Resources Canada, Office of Energy Efficiency NEUD data for the base year. The eight segments each have different occupancy and use rates, system types, and shell types. The relative shares of the building types within each segment varies over the study period.
Each building segment encompasses a variety of shell and building HVAC technology options. Each is represented by a ‘shell’ node, which competes alternative shell options. This represents a demand for heat load, which is met by a corresponding HVAC service node for each building segment (see flow model). The link between the shell and HVAC competitions attempts to capture the relationship between decisions about building shell characteristics and the amount of energy demanded of the heating and air conditioning services. HVAC systems are modelled as ‘single’ technologies to represent the interdependence of choices about heating, cooling and ventilation systems.
In the flow model, separate node are described for HVAC1 and HVAC2. We represent options for fuel heating on one node and for electrical heating on another. However, all of these options do compete together in the simulation.
Refrigeration, Cooking, Hot Water, Plug Load
The end-uses in the commercial sub-model are linked to the driving variable based on historic ratios of use per unit floor space and the weighted average of all the various commercial sub-sectors (e.g., a hospital will use more hot water than a warehouse per unit area). The weighted ratios can change over time, depending on changes in the proportion of the different kinds of floor space assumed over that period.
As shown on the flow model, refrigeration, cooking, hot water and plug load are each represented by a competition node. At these nodes, technologies compete based on life-cycle cost to provide the increased energy end-use services demanded due to change or growth in floor space. The number of technologies available for each of these end-uses varies.
The model treats the six end-use categories (HVAC, lighting, refrigeration, cooking, hot water and plug load) independently, meaning that the demand for energy in one category does not depend on the demand in another. While there are interaction between lighting (and to a lesser extent other end-uses) and HVAC, these interactive effects are very difficult to quantify in CIMS. Models exist that calculate these interactions for a specific building type but this type of analysis is beyond the scope of this current CIMS sub-model. The focus in the model is on capturing the main interactive effects between shell choice and HVAC demand.
Plug Load
The plug load is related to the primary node by the installed outlet wattage per square meter of commercial building floor space. This category represents the energy use of electric office equipment such as photocopiers, computers and microwave ovens. It does not include refrigerators, freezers or hot water heaters, which are represented separately in the model.
Modelling and Scenario Documentation
55
Domestic Hot Water
Domestic hot water (DHW) end-use represents the hot water used for all building services except space heating and is measured in cubic meters of hot water demand for the whole commercial building sector. It is related to the primary node by the ratio of cubic meters demand to commercial building floor space. Hot water is used for showering, washing, cleaning and food preparation.
Refrigeration
The refrigeration end-use includes all commercial refrigeration except for cooling required for air conditioning and is measured in gigajoules of refrigeration demand per square meter of commercial building floor space. This end-use is represented by two types of technologies: stand-alone plug-in refrigerators and large built-up refrigeration systems. These two types have different efficiency improvement opportunities. The refrigeration category in the model combines these two technologies and its energy intensity is the weighted average of the two system types. The ratio of the two system types is assumed to remain constant over time in a typical simulation.
Cooking
The cooking end-use is served by either electric or gas ovens, stoves and fryers and is measured in gigajoules of cooking demand per square meter of commercial floor space. This end-use also consists of different system types and the energy intensity of this end-use is the average for all types of commercial cooking equipment. As in the other end-use services, the shares of the various equipment types are assumed to remain constant over the simulation period.
Cogeneration
Cogeneration systems are sized to provide 100% of the facility’s heating load. In certain instances this results in a surplus of electricity beyond that required by the shell/HVAC system. This surplus is assumed to be sold back to the electric utility for the same price that the facility is charged for electricity.
Modelling and Scenario Documentation
56
APPENDIX B – Decomposition Analysis, Reference
Case
General Method
The OPA project Terms of Reference asks for a picture of the underlying factors which are driving the reference case forecast. This task is challenging because the electricity forecast produced by CIMS is a function of the technology choice outcomes in the model rather than being directly built from assumptions about changes in ‘energy efficiency’, ‘fuel-switching’, and even structure.26
A common approach to isolating determinants of historical energy demand into explanatory factors is an approach called decomposition analysis. We apply this method to the CIMS forecast. We do this to isolate changes due to activity (e.g. change in floor-space heated over the study period), efficiency (e.g. how much energy was required to heat the floor-space to a given temperature at the beginning and end of the study period) and fuel-mix (e.g. was electricity substituted for another fuel, or vice versa?). We first separate out the influence of cogeneration of electricity with heat.
Fuel-switching in this decomposition exercise has a more global meaning. For instance, we are describing whether a sector uses more or less electricity relative to other fuels overall, including changes in the relative importance of certain end-uses.
‘Sub-sector structure’, which we define as the relative changes in the importance of energy end-uses in a sector, is not a term in the decomposition equation. Inclusion of sub-sector structure in the equation is not simple and would require a considerably more complex and data-intensive calculation. Because this is not included the impacts of structure are ‘buried’ in the other terms. This explains why the results may seem counter-intuitive. Structural change is reflected most strongly in the fuel switching term. For instance in the reference case, the residential sector shows an overall ‘fuel-switching’ in the forecast towards electricity in the decomposition. This occurs mainly due to growth in minor appliance and air conditioning end-uses cause residential electricity demand to grow more quickly than for natural gas and oil.
Detailed Method
Decomposition analysis starts with an identity that accounts for the factors that influence energy consumption. We use the identity described in Equation 1.
∑
⋅⋅≡
j
j
E
E
A
EAE
(1)
Where:
26 Many important end-uses are nested – their service demand depends on the simulation outcome of other end-uses.
Modelling and Scenario Documentation
57
A = Sector activity
E = Energy consumption
j indexes fuel types
To compare energy consumption in two time periods, T (2025) and 0 (the present), we apply equation (1) as follows:
OT EEE −=∆ (2)
O
j
j
j
T
j
E
E
A
EA
E
E
A
EAE ∑∑
⋅⋅−
⋅⋅=∆
(3)
FuelMixEfficiencyAct EEEE ∆+∆+∆=∆ (4)
Where:
∆EAct = Difference due to activity
∆EEfficiency = Difference due to energy intensity
∆EFuelSwitiching = Difference due to fuel mix
Each of the subcomponents of ∆E is calculated using a parametric approximation to an integral of equation (4). Using the LMD I approach, the components of equation (4) are calculated as follows:27
( ) ( )∑
−=∆
jO
T
Oj
Tj
Oj
Tj
Activity A
A
EE
EEE ln
ln
(5)
( ) ( )∑
−=∆
jT
T
Oj
Tj
Oj
Tj
Efficiency A
E
A
E
EE
EEE
0
0
lnln
(6)
( ) ( )∑
−=∆
jO
Oj
T
Tj
Oj
Tj
Oj
Tj
FuelMix E
E
E
E
EE
EEE ln
ln
(7)
Using this method, the efficiency value represents the change in electricity use due to the overall change in energy efficiency in the sector. The fuel mix value represents the change in electricity use due to fuel switching in the sector, as well as differences in efficiency between the fuels.
27 Ang, B. 2005. “The LMD I Approach to Decomposition Analysis: A Practical Guide” Energy Policy, 33, pp. 867-871.
Modelling and Scenario Documentation
58
Decomposition Results
Detailed results are provided in spreadsheets provided to the OPA. We summarize the decompositions in the Reference Case in Figures 32 to 35. The purple bar in each figure shows the overall change in electricity demand between 2000 and 2025. The blue bars breaks this overall change down and shows the change in electricity demand due to activity, fuel-switching (change in fuel-mix), efficiency and cogeneration.
In the reference case, the fuel switching towards electricity at this aggregate level represents a structural shift towards a greater share of energy-services dominated by electricity in the commercial and residential sectors (air conditioning and plug load).
Figure 19: Decomposition applied to the Reference Case (in 2025), all sectors
All Sectors - Changes in demand due to impact of activity, efficiency, fuel switching and cogeneration (2000-2025)
-30
-20
-10
-
10
20
30
40
50
60
70
Change indemand
activity efficiency fuelswitching
cogeneration
TWh
Modelling and Scenario Documentation
59
Figure 20: Decomposition applied to the Reference Case (in 2025), residential sector
Residential - Changes in demand due to impact of activity, efficiency, fuel switching and cogeneration (2000-2025)
-10
-5
-
5
10
15
20
Change indemand
activity efficiency fuelswitching
cogeneration
TWh
Modelling and Scenario Documentation
60
Figure 21:Decomposition applied to the Reference Case (in 2025), commercial/institutional sector
Commercial/Institutional - Changes in demand due to impact of activity, efficiency, fuel switching and cogeneration (2000-2025)
-15
-10
-5
-
5
10
15
20
25
30
Ch
an
ge in
de
ma
nd
activity
efficie
ncy
fue
lsw
itchin
g
coge
ne
ratio
n
TWh
Figure 22: Decomposition applied to the Reference Case (in 2025), industrial sector
Industry - Changes in demand due to impact of activity, efficiency, fuel switching and cogeneration (2000-2025)
-10
-5
-
5
10
15
20
25C
ha
ng
e in
de
ma
nd
activity
efficie
ncy
fuel
switchin
g
cog
ene
ratio
n
TWh
Electricity Demand in Ontario – A
Retrospective Analysis
Prepared for: Chief Conservation Officer
Ontario Power Authority
Revised November 2005
Prepared by: ICF Consulting
277 Wellington St. West, Suite 808 Toronto, ON M5V 3E4
TABLE OF CONTENTS
1. INTRODUCTION & OVERVIEW OF ELECTRICITY DEMAND............................................................1 2. ELECTRICITY DEMAND IN ONTARIO, 1993-2004 ..................................................................................5
THE BIG PICTURE ......................................................................................................................................................5 3. SECTOR ANALYSIS ........................................................................................................................................8
METHOD....................................................................................................................................................................8 RESIDENTIAL SECTOR................................................................................................................................................9 COMMERCIAL & INSTITUTIONAL SECTOR................................................................................................................15 INDUSTRIAL SECTOR (INCLUDING TRANSPORTATION AND AGRICULTURE) .............................................................22
4. DEMAND IMPLICATIONS...........................................................................................................................26 5. SUMMARY ......................................................................................................................................................32
LIST OF TABLES
Table 1 Electricity by New Major Appliances, Efficiency Improvements, 1990 and 2001......... 13 Table 2 Appliance Saturations and Replacement Rates ............................................................... 13 Table 3 Changes in Ontario Residential Appliance Efficiency .................................................... 14 Table 4 Growth in Electricity Use vs. Floor Area by Sub-sector ................................................. 17 Table 5 Electricity Intensity by Subsector .................................................................................... 18 Table 6 Reported Lighting Conservation Features for Ontario – 2000 Survey............................ 19 Table 7 Lighting Intensity by Subsector....................................................................................... 20 Table 8 Comparison of Growth in Power Consumption and GDP by Sub-Sector....................... 25 Table 9 Comparison of Energy and Demand Contributions by Sector (1999 data) ..................... 27 Table 10 Ontario Electrical Demand Contributions to Peak Demand (as % of peak hour) ......... 31 Table 11 Ontario Electricity Use by End Use............................................................................... 33
List of Figures Figure 1 Electricity Final Demand in Ontario, 1958-2004............................................................. 1 Figure 2 Electricity Generated in Ontario by Source, 1958-2004 ................................................. 2 Figure 3 Final Demand for Fuels and Electricity in Ontario, 1958-2004....................................... 2 Figure 4 Relative Growth in Electricity and Electricity per Capita and per Dollar of GPP, 1958-2003................................................................................................................................................. 3 Figure 5 Electricity Productivity in Ontario, 1958-2004 ................................................................ 4 Figure 6 Ontario Electricity Supply, 1958-2004............................................................................. 6 Figure 7 Ontario Electricity Price Index......................................................................................... 7 Figure 8 Relative Growth of Residential Electricity Use and Key Drivers.................................. 10 Figure 9 Ontario Residential Electricity End Use Shares, 2003................................................... 11 Figure 10 Residential Electricity Use in Ontario.......................................................................... 11 Figure 11 Residential Electricity Use by End Use........................................................................ 12 Figure 12 Changes in Residential Electricity Use – 1990 to 2003 ............................................... 15 Figure 13 Relative Growth of Commercial Sector Electricity Use and Key Drivers................... 16 Figure 14 2003 Service Sector Electricity Use by Sub-Sector .................................................... 16 Figure 15 Energy Intensity for Services Sector ........................................................................... 17 Figure 16 2003 Service Sector Electricity Use by End Use ........................................................ 18 Figure 17 Service Sector Electricity Use -- Structural vs. Efficiency Impact .............................. 20 Figure 18 Contributions to Service Sector Electricity Growth by End Use – 1990 to 2003 ....... 21 Figure 19 Contributions to Service Sector Electricity Growth by Sub-Sector – 1990 to 2003 ... 22 Figure 20 Relative Growth of Industrial Sector Electricity Use and Key Drivers ....................... 23 Figure 21 Electricity per Dollar of Output for Selected Industries............................................... 24 Figure 22. 2003 Industrial Electricity Use.................................................................................... 24 Figure 23 Contributions to Change in Industrial Electricity Use - 1990-2003............................. 25 Figure 24 Industrial Electricity Use -- Structural Vs. Intensity Effects........................................ 26 Figure 25 Residential Group - Typical Weekday Load Pattern in July ...................................... 28 Figure 26 Commercial Group - Typical Weekday Load Pattern in July ...................................... 28 Figure 27 Small Industrial Group – Typical Weekday Load Pattern in July............................... 29 Figure 28 Direct Industrial Load Shapes for Average July Weekday ......................................... 29 Figure 29 Ontario Peak Demand (MW) for Typical July Weekday............................................ 30 Figure 30 Ontario Peak Demand (MW) for Typical January Weekday ...................................... 30
Revised November 2005 Page 1 of 33
1. Introduction & Overview of Electricity Demand Since 1958, the demand for electricity has increased fivefold, from 28,000 GW.hours to over 140,000 GW.hours in 2004. Over this long term, growth averaged 3.8 percent per year but it has been slowing down in recent years. There has also been a shift over the years in the share of total demand from each of the main consuming sectors. In 1958, the industrial sector accounted for 55 percent of Ontario’s electricity use and commercial and institutional buildings only 16 percent; today the industrial share of electricity demand has dropped to 29 percent and commercial and institutional buildings account for fully 38 percent of total demand.
Figure 1
Electricity Final Demand in Ontario, 1958-2004
-
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
GW.hours
Industrial Commercial Residential
As shown in Figure 2, there has also been a transformation in the supply of electricity over this period. In 1958, most of Ontario’s electricity was supplied by hydroelectric power from the large power dams on the St. Lawrence River, at Niagara Falls, and various other sites around the province. There has been little growth in the supply of hydropower since then, however, and the current system relies on a mix of coal, nuclear, hydro, gas and other power plants.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 2 of 33
Figure 2
Electricity Generated in Ontario by Source, 1958-2004
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002
GW
.hou
rs
Hydro Coal Nuclear Gas Other
Figure 3
Final Demand for Fuels and Electricity in Ontario, 1958-2004
Electricity
Fuel
-
500
1,000
1,500
2,000
2,500
1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002
PJ
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 3 of 33
Figure 3 illustrates the contribution of electricity to Ontario’s total demand for energy. About 12-14 percent of our energy use consists of applications that can only be provided with electricity: lighting, small motors and appliances, electronics, etc. In addition to this “captive market”, electricity is also used to provide space and water heat, boosting its overall share of Ontario’s energy use to 21 percent. The overall pattern of electricity demand is illustrated in Figure 4 over the 1958-2004 period. After 30 years of steady growth, the demand for electricity started to decelerate in the 1990’s Per capita electricity demand peaked in 1988 and was down by 15 percent by 2004. The electricity intensity of the economy, expressed in electricity demand per dollar of GPP leveled off in the 1970’s and has been declining since the early 1990’s. In fact, the electricity intensity of the Ontario economy in 2004 was lower than it has been in fifty years.
Figure 4
Relative Growth in Electricity and Electricity per Captia and per Dollar of GPP, 1958-2003
0.00
1.00
2.00
3.00
4.00
5.00
6.00
1958 1963 1968 1973 1978 1983 1988 1993 1998 2003
1958=1
Electricity Per Capita Electricity per Dollar GPP Electricity Demand
A closer look at the long term trends reveals that there have been three distinct phases in the evolution of the demand for electricity in Ontario since 1958. These phases are best illustrated by tracking the trend in the electricity productivity of the Ontario economy – the dollars of economic output generated per kilowatt-hour of electricity used. This indicator – the inverse of the electricity intensity illustrated in Figure 4 – is illustrated for the 1958-2004 period in Figure 5.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 4 of 33
Figure 5
Electricity Productivity In Ontario, 1958-2004
2.00
2.20
2.40
2.60
2.80
3.00
3.20
3.40
1958 1963 1968 1973 1978 1983 1988 1993 1998 2003
1997 Dollars of GPP per kW.hour
In the first phase, from 1958 to 1974, electricity demand grew at nearly 7 percent per year while the economy grew at 5 percent and the population at 2.2 percent. The economy was booming and the “Live Better Electrically” advertising campaign encouraged the use of electricity for space and water heating. Large coal-fired power plants were built to meet the growing demand. From 1974 to 1993, electricity demand growth rates slowed down but not as much as for other fuels. While the demand for oil and gas fuels “decoupled” from GPP growth, electricity demand continued to grow at the same rate as the economy, about 2.5 percent per year, prompting Ontario Hydro to put forward a capital expansion program consisting of dozens of coal and nuclear power plants. The third phase of electricity productivity in Ontario began in the early 1990’s and is ongoing. During this period, Ontario’s economic growth and electricity growth curves diverged sharply. While the economy grew at an average rate of 4 percent (and population growth continued at 1.4 percent), electricity demand growth has averaged 0.8 percent per year. This period corresponds to the restructuring of Ontario Hydro, the shutting down of utility demand side management programs, and a cessation of capital investment in large new power plants. The output of the nuclear program peaked during this period and began to decline. It is this third phase, from 1993 to 2004, that is the subject of the rest of this analysis.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 5 of 33
2. Electricity Demand in Ontario, 1993-2004
The Big Picture
Between 1989 and 1993 the Ontario economy went into recession, and the demand growth for electricity stalled with it. Given the strong correlation between economic and electricity growth rates throughout the 1970’s and 1980’s, the slowdown in electricity growth was initially attributed to the economic downturn. But when Ontario’s economic output returned to pre-recession levels in 1994 and began to grow again, electricity growth did not rebound with it. In retrospect it is now apparent that the economy that emerged from the recession was less electricity-intensive than the pre-recession economy and that economic growth and electricity demand growth had “decoupled” in the manner that petroleum demand and the economy had decoupled twenty years earlier. The electricity productivity of the economy – the GPP produced per kilowatt-hour of electricity consumed – is an aggregate indicator and reflects both the structure of the economy and the technological efficiency of electricity use. A relative shift in output from electricity intensive activities such as smelting and paper making toward general manufacturing or services will cause electricity productivity to increase. The same effect is caused by shifts within the electricity intensive industries toward higher value-added output. The electricity productivity of the economy will also improve if there is a shift away from the use of electricity for heating applications, or if the efficiency of electricity using technologies, from lights to computer screens, improves. All of these factors have been at work in Ontario over the past fifteen years, although there is very little research and understanding of their relative importance. What is clear from the aggregate indicator illustrated in Figure 5 is that the electricity productivity in Ontario is now higher than it has been in fifty years and that it is continuing to improve on a steep curve. One way to appreciate the significance of this trend is to think of the electricity productivity as a new “source” of power, and to consider how much additional generating capacity Ontario would have been required in 2004 if not for the productivity improvement that has occurred synced 1993. Viewed this way, as illustrated in Figure 6, electricity productivity emerges as the biggest story in Ontario’s electricity economy; by 2004 the productivity improvement was displacing the need for over 60,000 GW.hours of generation, equivalent to 150 percent of the output of all the hydro dams, or twice as much as the output of all the coal plants, or about equal to the power output of ten large CANDUs operating with 80 percent capacity factors. Considering the size and the economic value of the productivity “resource”, we have a relatively poor understanding of its makeup and internal dynamics, for example how much is due to fuel switching and technology efficiency improvements and how much is due to structural change in the economy. We can see from the aggregate statistics that the services sector has been growing faster than the manufacturing sector in Ontario and this will generally result in an increase in
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 6 of 33
electricity productivity, but in general there is a dearth of data and quantitative analysis for understanding both the historical trends in and the future potential for electricity productivity improvement in Ontario.
Figure 6
Ontario Electricity Supply, 1958-2004(Electricity productivity relative to 1993)
-
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
180,000
200,000
220,000
1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002
GW.hours
Hydro Coal Nuclear Gas Other Productivity
It is interesting to note that the electricity productivity improvements took place during a period in which electricity prices were generally either stable or declining and during which utility demand side programs all but disappeared from the Ontario electric market. Electricity prices did increase in the period leading up to 1994, but remained then declined slowly until the rates were partially unbundled in 2001 in preparation for Market Opening. Error! Reference source not found. provides a price index showing how electricity prices changed over the period for industrial customers with greater than and less than 5,000 kW demand per month. The pattern of price changes for residential and commercial/institutional customers would have been similar up to the point of market opening. While electricity prices increased by approximately 20 percent in the early 1990s, public perception of this increase may have differed from the actual price change. Announcements by the Minister of Energy at that time led consumers to expect that there would be a 45 percent price within a few years. While that price increase didn’t occur, consumers may have responded based on their expectations of future prices.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 7 of 33
In May 2002, the Ontario electricity market was opened to competition. At that point, electricity charges were “unbundled” to separate out elements that were open to competition from those that were deemed to be ‘non-competitive’. After May 2002, individual customers may have paid commodity costs based on market prices or a contract price received from a Retailer. At the opening of the market, approximately 15 to 20 percent of residential customers had contracted with a Retailer for commodity. Generally these prices were above what eventually became the fixed or capped price for commodity. For customers who did not choose to contract with a Retailer, commodity prices were based on the Wholesale Spot Price from May 2002 to December 2002. In December 2002, the government capped the commodity price for approximately 50 percent of the market or about 98 percent of participants. This meant that small volume consumers and most institutional customers paid a fixed price for commodity (4.3¢ per kWh) while larger accounts (generally those with demands over 50 kW per month) paid a variable, hourly price based on the spot market. Essentially, most customers saw very little change in electricity prices in the period
under review. There was however, considerable
uncertainty with regards to both prices and how electricity would be supplied in the last 2-3 years of the period. Consumers also experienced considerable price volatility for the commodity portion of the bill during 2002. The subsequent ‘refunding’ of some of these costs may also have affected consumers’ perceptions of costs.
___________________________
The remainder of this report takes a closer look at electricity demand trends at the sector level in Ontario, focusing on the period from 1990 through 2003.
40
60
80
100
120
140
160
180
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Year
Inde
x (1
997=
1)
Over 5 MW Under 5 MW
Figure 7. Ontario Electricity Price Index
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 8 of 33
3. Sector Analysis
Method
The analysis of Ontario’s electricity use carried out in this report is primarily based on data made available by the Office of Energy Efficiency (OEE) at Natural Resources Canada. The data is published on a national and provincial basis in the “Comprehensive Energy Use Database” (website: www.nrcan.oee.gc.ca). Information for Ontario and more specifically for electricity use in the province was extracted from this database for analysis. Information on energy pricing was obtained from Statistics Canada and from comparisons published by Quebec Hydro. The OEE reports all energy use in terms of joules (Petajoules PJ, Megajoules MJ, etc.). In order to make the discussion more relevant to electricity, all values were converted to electrical terms (i.e. kilowatt hours kWh, Megawatt hours MWh or Gigawatt hours GWh). As part of the analysis, the impacts of changes in structural and electricity intensity were calculated for the Service and Industrial sectors. In these cases, a model was constructed to calculate how electricity use would have changed had one variable (i.e. electricity use per square metre of floor area) been held constant. The difference between actual historic use and the value calculated represents the impact of the change in structure or intensity. In the discussion of electricity demand patterns, very little data for Ontario was available in the public domain. Load shape data published by the Independent Electricity System Operator (IESO) on its website (www.iemo.com) was used to construct an illustrative demand model. This load data is from 1998, and is assumed to have been collected by Ontario Hydro. The IESO was contacted to obtain information on the data, but while staff were generous in their efforts to assist, they had little information on how the original data had been gathered. The sector descriptions used for the load shape data did not match exactly to those used in the OEE energy data. Where the descriptions differed, judgment was used in selecting the most appropriate load shapes available to match the sub-sectors and end use definitions used by OEE. The load shape data is published for typical weekdays and weekend days for each month. In most instances, the data had been normalized for a 1 MW load. In order to construct a model for the Ontario system as a whole, the annual energy associated with each load shape was calculated. The energy use represented by the load shape was then compared to total Ontario energy use for the selected load in order to calculate the percentage of total Ontario load represented by the load shape. Each load shape was then weighted by its proportion of total Ontario electricity use to calculate a load shape for each sub-sector or end use for a typical weekday in January or July. Once a peak day load shape had been created, the peak hour for that load shape was determined. The percentage contributions of each load shape (i.e. sub-sector or end use) to the peak demand were then calculated. To simplify presentation, any sub-sectors which contributed less than 5 percent of the total peak were aggregated into Other Industry or Other Services.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 9 of 33
This model is intended to provide only a general illustration of demand contributions by economic sector and end use. Two caveats should be borne in mind:
• There is limited information regarding the underlying load shape data used in the model. It would be useful to have information that is more current and to know that the sample size was sufficient to ensure that it reasonably represents Ontario loads.
• The data is for a “typical” weekday. Obviously the critical loads which drive the summer peak, and to a lesser extent the winter peak, behave differently on very hot or cold days than on typical days. For this reason, the model is assumed to under-represent the contribution of Residential and Service sector cooling loads.
It should be noted that there is a great deal of hourly load information available in Ontario. Most Local Distribution Companies have had hourly interval meters installed on large customers for well over 10 years and many have extended the use of interval metering down to customers of 2-300 kW per month in recent years. Utilities are also in the process of collecting load shape and consumption data on different customer classes and loads as part of the process of reviewing
Residential Sector
The residential sector accounted for 33.8 percent of electricity used in 1990, and fell slightly to 32.9 percent by 2003; the relative growth rates of some of the key drivers are shown in Figure 8. While GDP would not be expected to be strongly correlated with growth in residential electricity use, it is shown as a point of reference for the observed decoupling of total electricity use from total economic activity. More importantly, while population and households were up 19 percent and 25 percent, respectively, and total residential floor area was up 32 percent (a trend toward bigger houses continued throughout the period), residential electricity actually declined from 1990-1998, and by 2003 had only just regained its 1990 level of about 47,000 GW.hours. The average electricity use per household declined fully 16 percent during this period, as compared with a decline in per household energy use (including all fuels) of 6.1 percent. Figure 9 shows the distribution of residential electricity use by end use in 2003. The figure is an illustration of the share of total residential electricity use, but the average electricity use per household and the corresponding breakdown by end use will vary depending on whether electricity is used for space heating and/or water heating. The sensitivity of residential electricity use to both the size of the major end use categories and the market share held by electricity is further illustrated in Figure 10. While the average household in Ontario consumes a little over 10,000 kW.hours per year, not too many households consume the average. Households that use electricity only for lighting, appliances and air conditioning use about 8,000 kW.hour per year on average. Electric water heating adds another 5,000 kW.hour to this total, but only about 33% of Ontario households use electricity for water heating. Electric space heating will add nearly 17,000 kW.hours to household electricity use but only about 15% of Ontario households heat
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 10 of 33
with electricity. Relatively small changes in electricity’s share of the space and water heating market have relatively large impacts on the total residential electricity use in the province.
Figure 8
Relative Growth of Residential Electricity Use and Key Drivers
0.80
0.90
1.00
1.10
1.20
1.30
1.40
1.50
1990 1992 1994 1996 1998 2000 2002
1990=1
GDP PopulationHouseholds Actual Electricity UseElectricity (1990 weather)
Over the 1990-2003 period, the relative shares of end uses in the makeup of residential electricity changed as illustrated in Figure 11. Energy use for water heating, refrigerators and freezers declined, however, increases in air conditioning, home electronics (other appliances) and lighting, more than offset these gains. Efficiency gains in major appliances (refrigerator use declined 33 percent, freezer use declined 44 percent, dishwasher use decreased 7 percent and clothes washers by 4 percent) amounted to approximately 2.9 GW.h over the period. By contrast, consumption by Other Appliances increased by about 2.7 GWh as shown in Figure 11. Electricity use for space heating grew by 4.9 percent over the period, although its share of the heating market declined from 12.8 percent to 11.3 percent. Part of this decline in market share reflects the increased use of natural gas heating in new construction. The actual number of homes using electric heat, including heat pumps and dual-fuel heating systems, increased by 13 percent over the period. Most of this increase was for heat pump systems, often in conjunction with gas heating. The stock of homes with baseboard heating declined by 5 percent over the period, leaving approximately 666,000 homes with that type of system by 2003.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 11 of 33
Figure 9. Ontario Residential Electricity End Use Shares, 2003
Figure 10
Residential Electricity Use in Ontario
- 5,000 10,000 15,000 20,000 25,000 30,000
Lights and Appliances(100% saturation)
Water Heat (33%saturation)
Space Heat (15%saturation)
Air conditioning (50%saturation)
Total
kW.hours per year
Average use per household, including all householdsAverage use for those households using electricity for this end use
Other Appliances15.5%
Lighting 15.4%
Refrigerator8.2%
Water Heating15.2%
Cooling7.6%
Heating23.3%
Freezer2.7%Dishwasher
0.3%
Clothes Washer0.5%
Clothes Dryer5.5%
Range6.0%
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 12 of 33
Electric water heating energy use decreased by 16.4 percent from 1990 to 2003. As a result, the share of residential electricity used for water heating declined from 19 percent to 15 percent over the period. Electricity’s share of the domestic water heating market fell from 33 percent to 22 percent; the number of homes with electric water heating fell from a high of 1.7 million homes in 1991 to a low of 1.48 million in 1991, but then started to increase again. By 2003, there were 1.61 million households with electric water heating in the province.
Figure 11 Residential Electricity Use by End Use
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1990 2003
Other AppliancesRangeClothes DryerClothes WasherDishwasherFreezerRefrigeratorCoolingWater HeatingLighting Heating
Cooling energy use increased by over 100 percent from 1990 to 2003, reflecting the increased prevalence of central air conditioning. Over the period, energy use by central air systems increased by 121 percent while use by room air conditioners increased by under 20 percent. Part of this increase in consumption reflects differences in the weather. Cooling degree days were approximately 20 percent higher in 2003 than in 1990. If the summer of 2003 had been comparable to 1990, the increase in air conditioning electricity use would have been closer to 70 percent. While air conditioning energy use represents only 7.6 percent of annual electricity use in the residential sector, it has a disproportionate impact in terms of demand. Among appliance end uses, new major appliances became substantially more efficient. As the stock of appliances was replaced, this was reflected in decreased average electricity use (i.e. refrigerators on average used 16.4 percent less over the period). There was essentially no change in the average number of refrigerators or freezers per household over the period. Table 1 shows how the average electricity use of new appliances changed between 1990 and 2001, and combined with the saturation rates shown in Table 2, the change in actual electricity use per household and per appliance over the 1990-2003 period is shown in Table 3.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 13 of 33
Table 1.
Electricity Use by New Major Appliances, Efficiency Improvements, 1990 and 2001
1990 2001
Refrigerators 950 559
Freezers 714 384
Dishwashers 1025 634
Electric Ranges 772 762
Clothes Washers 1218 810
Clothes Dryers 1102 916
Source: Energy Consumption of Major Household Appliances Shipped in Canada; Trends for 1990-2001;
Natural Resources Canada, December 2003
Table 2.
Appliance Saturations and Replacement Rates
Appliance Type
1993 Stock ('000)
% of all
Households
1994/95 Purchases
('000)
% of all
HouseholdsRefrigerators 10,313 99.6% 1,156 10.2% Range 10,359 1.0% 950 8.4% Dishwasher 4,566 44.1% 804 7.1% Freezer 6,226 60.1% 505 4.4% Washing Machine 8,020 77.4% 1,057 9.3% Clothes Dryer 7,645 73.8% 882 7.8% Central Air Conditioning 1,579 15.2% 210 1.8% Window Air Conditioner 1,134 15.2% 166 1.5% Source: The Household Equipment of Canadians, Features of the 1993 Stock and the 1994 & 1995 Purchases, Analysis Report, Natural Resources Canada, March 1997
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 14 of 33
Table 3
Changes in Ontario Residential Appliance Efficiency
Electricity use by “Other” appliances, such as home electronics, increased by 57 percent over the period 1990 to 2003. Lighting use increased by almost 20 percent. There is limited information available on the appliances included in the “other appliance” category within the OEE database. Given the increasing load associated with these appliances, further investigation of these loads, their efficiency and saturations would appear warranted. The total housing stock in Ontario increased by 25.2 percent or 958,000 units between 1990 and 2003 growing at 1.7 percent per year. The most rapid growth occurred in Single Attached homes (47 percent) followed by Single Detached (27.9) and Apartments (12.5 percent). The number of mobile homes declined over the period by almost 15 percent. By the end of the period, the housing stock was comprised of 57 percent single detached homes, 18 percent apartments, and 14 percent single attached homes. The more rapid growth in houses vs. apartments may account for the increase in average floor area per home. The data available from the OEE provides limited information on energy use by heating fuel type (i.e. electricity vs. gas or oil) by building type or vintage. As a result, it was not possible to examine trends in energy use between building types.
Absolute Total % % Change1990 2003 Change Change per Year
Average Use per Household (all households)Refrigerator 1,603 852 (751) -88.1% -4.7%Freezer 630 278 (351) -126.2% -6.1%Dishwasher 42 31 (11) -35.2% -2.3%Clothes Washer 64 49 (15) -31.2% -2.1%Clothes Dryer 780 654 (126) -19.2% -1.3%Range 723 715 (8) -1.1% -0.1%Other Appliances 1,285 1,614 328 20.4% 1.8%
Use per Appliance (for actual stock)Refrigerator 1,301 676 (625) -92.4% -4.9%Freezer 1,102 519 (583) -112.5% -5.6%Dishwasher 107 61 (46) -76.2% -4.3%Clothes Washer 91 64 (26) -41.2% -2.6%Clothes Dryer 1,132 836 (295) -35.3% -2.3%Range 731 717 (14) -2.0% -0.2%Other Appliances 188 177 (12) -6.6% -0.5%
Notes: Some homes have multiple major appliances. Use per Household shows averageuse for all households including those with no appliance or those with multiple appUse per Appliance shows average use based on the actual stock of appliances.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 15 of 33
Usage per household (intensity) fell from 12,474 kWh annually in 1990 to 10,445 kWh per year in 2003 (870 kWh per month). Figure 12 shows how changes by end use contributed to the overall change in electricity use for the sector between 1990 and 2003.
Figure 12 Changes in Residential Electricity Use – 1990 to 2003
Commercial & Institutional Sector
The Commercial and Institutional (Service) sector accounted for 36.6 percent of total electricity used in 2003. This represents an increase in share from 1990 when it represented 30.4 percent of the total. Electricity use in the Services sector grew more rapidly than for any other sector. From 1990 to 2003, electricity use increased 30.1 percent or 2.0 percent per year. Total floor area for the sector grew 24.9 percent reflecting a modest increase in electricity intensity. The relative growth rates of commercial sector electricity and some of the key drivers are shown in Figure 13. Services sector GDP now dominates Ontario’s total GDP and tracked it closely over the 1990-2003 period. Building floor area tends to grow much slower than output, and commercial building electricity use tracks building floor area fairly closely.
Figure 14 shows the distribution of commercial sector electricity use by sub-sector for 2003. Over half of the electricity used by the sector went to Offices (54 percent) with Accommodation and Food (16 percent) and the Retail sector (11 percent) representing the next largest users. Electricity use grew more rapidly than floor area in several sub-sectors between 1990 and 2003, representing an increase in electricity intensity for those building types.
-2000000
-1500000
-1000000
-500000
0
500000
1000000
1500000
2000000
2500000
3000000
1
MW
h
HeatingCoolingWater HeatingRefrigeratorFreezerDishwasherClothes WasherClothes DryerRangeOther AppliancesLighting
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 16 of 33
For the Services sector as a whole, energy intensity increased slightly, from 1.9 to 2.2 GJ per m2, between 1990 and 2003. Electricity intensity also increased slightly, from 250 to 260 kWh per square metre over the period. Figure 15 compares the electricity intensity, in kWh per square metre, for each sub-sector for 1990 and 2003. While the Accommodation and Food industry represents less than 5 percent of total floor area, its higher energy intensity results in it representing over 15 percent of total electricity use.
Table 4
Figure 13
Figure 14 2003 Service Sector Electricity Use by Sub-Sector .
Educational Services 4%
Health Care & Social Assistance
6%
Wholesale Trade 4%
Other Services 1%
Accommodation & Food Services
9%
Arts, Entertainment and Recreation
2%
Retail Trade 12%
Information & Cultural Industries
1%Transportation &
Warehousing 2%
Offices59%
Relative Growth of Commercial Sector Electricity Use and Key Drivers
0.80
0.90
1.00
1.10
1.20
1.30
1.40
1.50
1.60
1990 1992 1994 1996 1998 2000 2002
1990=1
GDP Commercial Sector GDP Building Floor Area Actual Electricity Demand
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 17 of 33
For the Services sector as a whole, energy intensity increased slightly, from 1.9 to 2.2 GJ per m2, between 1990 and 2003. Electricity intensity also increased slightly, from 250 to 260 kWh per square metre over the period. Figure 15 compares the electricity intensity, in kWh per square metre, for each sub-sector for 1990 and 2003. While the Accommodation and Food industry represents less than 5 percent of total floor area, its higher energy intensity results in it representing over 15 percent of total electricity use.
Table 4
Growth in Electricity Use vs. Floor Area by Sub-Sector
Growth in Electricity Use
Growth in Floor Area
Wholesale Trade 1.2% 3.0% Retail Trade 8.8% 7.3% Transportation & Warehousing 1.5% -1.1% Information & Cultural Industries 0.0% 43.8% Offices 41.7% 34.1% Education 38.7% 29.8% Health Care & Social Assistance 26.3% 28.5% Arts, Entertainment & Recreation 49.2% 49.4% Accommodation & Food Services 21.7% 24.2% Other Services 10.0% 14.1%
Figure 15 Energy Intensity for Services Sector
0 100 200 300 400 500 600
Wholesale Trade
Retail Trade
Transportation &Warehousing
Information & CulturalIndustries
Offices
Educational Services
Health Care & SocialAssistance
Arts, Entertainment andRecreation
Accommodation & FoodServices
Other Services
kWh/sq.m.
20031990
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 18 of 33
Electricity intensity increased in the Offices (5.6 percent), Education (6.8 percent), Transportation & Warehousing (2.7 percent) and the Retail sector (1.4 percent) between 1990 and 2003; intensities fell in all the other sub-sectors.
Table 5
Electricity Intensity by Sub-Sector (kWh/sq. m.)
1990
2003
Wholesale Trade 222 218 Retail Trade 225 229 Transportation & Warehousing 144 148 Information & Cultural Industries 110 110 Offices 325 343 Education 64 68 Health Care & Social Assistance 229 225 Arts, Entertainment & Recreation 178 177 Accommodation & Food Services 501 490 Other Services 100 96
Figure 16 2003 Service Sector Electricity Use by End Use
Lighting represents the largest single end use within the Service sector, accounting for over 37 percent of electricity use. Auxiliary motors (24 percent), auxiliary equipment (21 percent) and cooling (14 percent) were the next largest end uses (see Figure 16).
Auxiliary Motors21.2%
Lighting37.3%
Space Cooling 16.1%
Auxiliary Equipment18.6%
Space Heating4.9%
Water Heating0.1%
Street Lighting1.8%
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 19 of 33
Electric space heating, which represented just 2.3 percent of electricity use in the sector in 1990, grew by 185 percent (over 8 percent per year) over the period. As a result, its share of total electricity use more than doubled to 4.9 percent. Water heating electricity use, by contrast, fell by 80 percent. By the end of the period, its share of total electricity use had fallen to just 0.1 percent. Electricity use for air conditioning rose by 52 percent reflecting increasing its share of total electricity use, as in the residential sector. Over the period, however, natural gas began to capture an increased share of the cooling market. In 1990, electricity supplied over 96 percent of cooling for the service sector. By 2003, approximately 15.5 percent of cooling was provided by natural gas. Lighting energy use increased by 33 percent over the period being reviewed, while total floor area for the sector increased by 24.9 percent. This implies a slight increase in lighting energy use per square meter. In 2000, the OEE carried out a Survey of Commercial & Institutional Building Energy Use. They reported that in Ontario, 43 percent of buildings had installed energy efficient ballasts and that 31 percent used energy efficient lamps (see Table 6). Given that approximately 70 percent of commercial/institutional lighting is fluorescent and that new T8 lamps are at least 25 percent more efficient than traditional T12 lamps and ballasts, the lack of improvement in lighting efficiency reflected in the OEE database is therefore somewhat surprising.
Table 6
Reported Lighting Conservation Features for Ontario – 2000 Survey
Conservation Measure
% of Buildings Reporting Measure
Reflectors 23.7% Energy Efficient Ballasts 42.6% Daylight Controls 13.9% Occupancy Sensors 9.0% Time Clocks 26.3% Manual Dimmer Switches 23.0% Energy Efficient Lamps 31.1% Other 10.0%
Source: Survey 2000, Commercial & Institutional Building Energy Use, Detailed Statistical Report,
Natural Resources Canada, December 2002. Table 7 shows lighting intensities in kWh per square metre for the Commercial & Institutional sub-sectors. As the table shows, lighting intensities increased in all of the sub-sectors, apparently in spite of improved efficiency lighting systems.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 20 of 33
Table 7 Lighting Intensity by Sub-Sector (kWh/sq.m.)
1990 2003 Wholesale Trade 71 75 Retail Trade 71 73 Transportation & Warehousing 75 77 Information & Cultural Industries 40 41 Offices 134 142 Educational Services 27 28 Health Care & Social Assistance 86 88 Arts, Entertainment and Recreation 64 65 Accommodation & Food Services 96 99 Other Services 36 37
It should be noted that changes in lighting intensity also impact space cooling loads. As a rule of thumb, every watt of lighting reduction is normally assumed to reduce cooling loads by 0.25 watts. Street lighting electricity use increased by 15.2 percent between 1990 and 2003. While no activity measure is readily available for street lighting, the number of households increased by 25 percent. Assuming that the number of streetlights should increase roughly in proportion to the number of households, this implies an increase in lighting efficiency for street lighting.
Figure 17
Service Sector Electricity Use – Structural vs. Efficiency Impact
30,000,000
35,000,000
40,000,000
45,000,000
50,000,000
55,000,000
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
MW
h
Electricity Use if Efficiency held as in 1990 Actual Use Electricity Use if Efficiency held as in 1990
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 21 of 33
In order to measure the relative contributions structural change versus changes in energy intensity, two analyses were carried out to calculate 2003 electricity use: 1) if the relative share of floor area in 1990 (i.e. structure) was held constant, and 2) if 1990 electricity intensities were held constant. The outcomes of these analyses are displayed in Figure 17. Actual electricity use by the sector increased by 30.1 percent between 1990 and 2003. Had the sub-sectoral shares of floor area remained fixed at 1990 levels, electricity use by the Service sector would have increased by 28.8 percent over the same period. Had electricity intensity remained unchanged at 1990 levels, electricity use would have grown by 26.2 percent by 2003. In energy terms, 2003 usage would have been 1,560 GWh lower had 1990 shares of floor area been maintained, or 2,589 GWh lower if electricity intensity had not increased. Figure 18 and Figure 19 show how the changes in end use and sub-sector growth contributed to total Service sector electricity use.
Figure 18 Contributions to Service Sector Electricity Growth by End Use – 1990 to 2003
-1000000
0
1000000
2000000
3000000
4000000
5000000
6000000
SpaceHeating
Water Heating AuxiliaryEquipment
Space Cooling AuxiliaryMotors
Lighting Street Lighting
MW
h
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 22 of 33
Figure 19 Contributions to Service Sector Electricity Growth by Sub-Sector – 1990 to 2003
Industrial Sector (including Transportation and Agriculture)
The industrial sector consumed 28.6 percent of all electricity used in Ontario in 2003. Industrial electricity use declined by 9.3 percent between 1990 and 2003. Agricultural electrical use, which accounted for 1.69 percent of total electricity grew by 12.4 percent over the period, while transportation use (0.3 percent of total) declined by 1.8 percent. Industrial electricity use and some of its key drivers are illustrated in Figure 20. Overall, industrial GDP has tracked total GDP for the province, but there is an important difference here between “heavy industry” and other industry.1 Total output from the electricity intensive primary industries declined throughout the period, while output from the other, much less electricity intensive industries grew slightly faster than the average economic growth rate.
1 Including the Heavy Industry grouping in thief analysis are Pulp and Paper, Iron and Steel, Industrial Chemicals, Petroleum Refining, Mining and Primary Metals Smelting and Cement.
0
1000000
2000000
3000000
4000000
5000000
6000000
7000000
8000000
9000000
10000000
Wholes
ale Trad
e
Retail T
rade
Transp
ortati
on & W
areho
using
Inform
ation
& Cult
ural In
dustr
ies
Offices
Educa
tiona
l Serv
ices
Health
Care
& Social
Assist
ance
Arts, E
nterta
inmen
t and
Rec
reatio
n
Accom
modati
on & Foo
d Serv
ices
Other S
ervice
s
MW
h
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 23 of 33
Figure 20
Relative Growth of Industrial Sector Electricity Use and Key Drivers
0.80
0.90
1.00
1.10
1.20
1.30
1.40
1.50
1.60
1990 1992 1994 1996 1998 2000 2002
1990=1
GDP ElectricityIndustrial GDP Heavy Industry GDPOther Industrial GDP Heavy Industry ElectricityOther Industry Electricity
The electricity use per dollar of in the primary industries is an order of magnitude greater than for the secondary manufacturing industries, so the divergent growth rates of these two parts of Ontario’s industrial economy has and continues to have a profound impact on industrial electricity demand. Further, within the secondary manufacturing industries there has been a marked increase in the value added produced per kilowatt-hour consumed. The net result of these factors has been stable or declining electricity consumption in Ontario’s industrial sector. While industrial electricity use declined, the economic output of the sector grew by 42 percent. As a result, electricity intensity, measured in MWh per 1997$ GDP, decreased 26 percent.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 24 of 33
Figure 22. 2003 Industrial Electricity Use
Figure 22 shows how electricity was used by sub-sector in 2003, while Figure 23 shows the relative contributions to the change in electricity use by industrial sub-sector.
Smelting and Refining 5%
Chemicals 10%
Pulp and Paper21%
Petroleum Refining 4%
Iron and Steel 14%
Mining 7%
Other Manufacturing 37%
4,101
4,460
4,386
2,874
2,659
1,896
181
0
1,024
3,295
702
4,835
2,892
3,921
1,685
356
0
1,391
0 1,000 2,000 3,000 4,000 5,000 6,000
Pulp and Paper
Smelting and Refining
Petroleum Refining
Cement
Chemicals
Iron and Steel
Other Manufacturing
Forestry
Mining
2003 1990
Figure 21. Electricity per Dollar of Output for Selected Industries
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 25 of 33
Figure 23
Contributions to Change in Industrial Electricity Use - 1990 to 2003
(3,000,000)
(2,500,000)
(2,000,000)
(1,500,000)
(1,000,000)
(500,000)
-
500,000
1,000,000
1,500,000
Pulp and Pap
er
Smelting an
d Refining
Petroleu
m Refining
Cemen
t
Chemicals
Iron an
d Steel
Other Man
ufacturin
g
Forestry
Mining
MW
h
Table 8. Comparison of Growth in Power Consumption and GDP by Sub-Sector
Growth in Electricity Use Growth in GDP 1997 $ Construction 0% 17% Pulp & Paper 8% -13% Smelting & Refining 98% -69% Petroleum Refining -23% -15% Cement 9% 9% Chemicals -35% -4% Iron & Steel 25% 11% Other Manufacturing -15% 67% Forestry 0% 11% Mining -38% -16%
In order to measure the relative contributions structural change versus changes in energy intensity, two analyses were carried out to calculate 2003 electricity use: 1) if 1990 structure was held constant, and 2) if 1990 electricity intensities were held constant. The results are summarized in Figure 24.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 26 of 33
Actual electricity use by the sector decreased by 9.3 percent between 1990 and 2003. Had the sub-sectoral shares of GDP remained fixed at 1990 levels, industrial electricity use would have increased by 31 percent over the same period. Had electricity intensity remained unchanged at 1990 levels, electricity use would have been 19.8 percent higher in 2003. In energy terms, 2003 usage would have been 18,418 GWh higher had industry structure not changed, or 13,308 GWh higher had electricity intensity not decreased.
4. Demand Implications Ontario’s electricity system has traditionally peaked during the winter months, either during the Christmas period as a result of seasonal outdoor lighting, or in the coldest month, due to electric heating requirements. In 1998, for the first time, the system peaked during the summer. Since that time, the annual peak has occurred during hot summer weather in every year except 2000. The IESO reports that while winter peak demand grew by 0.7 percent per annum between 1987 and 2004, the summer peak has grown by 1.3 percent. Table 9 shows a comparison of the contribution by sector to peak demand versus energy use.
Indsutrial Electricity Use -- Structural Vs. Intensity Effects
40,000
45,000
50,000
55,000
60,000
65,000
70,000
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
GW.hours
Actual Electricity Use With 1990 industry output sharesWith 1990 electricity/output ratios
Figure 24
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 27 of 33
Table 9. Comparison of Energy and Demand Contributions by Sector (1999 data)
Contribution in % to
Peak Demand Percentage of Total
Energy Use Industrial 33.6 33.1 Commercial 38.3 34.2 Residential 26.1 30.5 Agricultural 1.4 1.9 Transportation 0.6 0.3
Notes: Demand information from the IESO web site. Energy information based on OEE database. While this data relates to the period when the system had just become summer peaking, it does show that the Residential sector makes a disproportionate contribution to peak demand. The higher average load factors in the industrial and commercial sectors result in their contributions to the peak being relatively lower than their contributions to energy use. Space cooling energy use has increased rapidly in both the Residential and Services sector over the past 13 years. Air conditioning loads have a very low annual load factor and a very high coincidence with the system peak. As the saturation of residential air conditioning has increased, the system has become much more weather sensitive. This has resulted in the Ontario electricity system moving from a winter to a summer peak. In order to gain a clearer picture of the drivers of the system peak, a model of Ontario electrical demand was constructed. Information on representative load shapes for industrial and Commercial/Institutional sub-sectors, as well as residential end uses was obtained from the IESO (Figure 25 to Figure 28). These load shapes were then weighted based on 2003 electricity use derived from the OEE. The result provides an approximation of the contributions of the various sectors and end uses to a typical day load shape. It should be noted that the load shapes, particularly those for the residential end uses, represent an average day. The weather sensitive loads, such as air conditioning and heating, would be expected to be more pronounced on hotter or colder days when the peak demand typically occurs. There is also little information available on how the load patterns were gathered (i.e. how representative the sample may be). Despite these limitations, the resulting load shapes do provide a reasonable picture of the loads which drive the system peak. Figure 29 and Figure 30 show a typical day load curve for an average weekday in July and January. Error! Reference source not found. below, shows the calculated contributions to a peak summer and winter day.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 28 of 33
Figure 25
Residential Group - Typical Weekday Load Pattern in July
0
1
2
3
4
5
6
7
8
9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
HourNote - not stacked, no allowance for diversity
Load
(MW
)
Baseload Central Air Water Heating
Figure 26
Commercial Group - Typical Weekday Load Pattern in July
0.5
0.7
0.9
1.1
1.3
1.5
1.7
1.9
2.1
2.3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
HoursNote - not stacked, no allowance for diversity
Load
(MW
)
Food Stores Other Retail Stores ServicesWholesale Warehouses Offices Health FacilitiesEducation Hotels and Other Accommodation Recreational FacilitiesMulti-residential
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 29 of 33
Figure 27 Small Industrial Group – Typical Weekday Load Pattern in July
Figure 28 Direct Industrial Load Shapes for Average July Weekday
Direct Industrial Load Shapes for Average July Weekday
0
200
400
600
800
1000
1200
1400
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
HourNote - Graph shows stacked line, no allowance for load diversity
Mw
Pulp and Paper Smelting and Refining Petroleum Refining CementChemicals Iron and Steel Other Manufacturing Mining
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
HourNote - Loads are stacked (cumulative)
MW
Dem
and
Residential Water Heating
Residential Central Air
Residential Baseload
Mining
Forestry
Other Manufacturing
Iron and Steel
Chemicals
Cement
Petroleum Refining
Smelting and Refining
Pulp and Paper
Construction
Other Services
Accommodation and Food Services
Arts, Entertainment and Recreation
Health Care and Social Assistance
Educational Services
Offices
Information and Cultural Industries
Transportation and Warehousing
Retail Trade
Wholesale Trade
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 30 of 33
Figure 29
Ontario Peak Demand (MW) for Typical July Week Day
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
20,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
HourNote - Loads are stacked (cumulative)
MW Demand
Wholesale Trade Retail Trade Transportation and Warehousing
Information and Cultural Industries Offices Educational Services
Health Care and Social Assistance Arts, Entertainment and Recreation Accommodation and Food Services
Other Services Construction Pulp and Paper
Smelting and Refining Petroleum Refining Cement
Chemicals Iron and Steel Other Manufacturing
Forestry Mining Residential Baseload
Residential Central Air Residential Water Heating
Figure 30
Ontario Peak Demand (MW) for Typical January Weekday
0
5,000
10,000
15,000
20,000
25,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
HourNote - Loads are stacked (cumulative)
MW Demand
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 31 of 33
Table 10 2003 Ontario Electrical Demand Contributions to Peak Demand (as % of peak hour)
Typical Week Day for Sector/End Use July January Offices 22.1% 22% Retail Trade 5.3% 3.9% All Other Services 8.0% 8.8% Pulp and Paper 4.6% 4.4% Smelting and Refining 3.5% 4.9% Other Industries 5.9% 4.9% Iron and Steel 3.5% 4.9% Other Manufacturing 10.6% 2.5% Residential Base load 19.7% 23.4% Residential Central Air 12.3% 0% Residential Water Heating 4.4% 4.9% Residential Electric Heating 0% 15.4%
Total - 99.9% 99.9% While there is very limited current load shape information publicly available a great deal of information is potentially available in Ontario. Local Distribution Companies (LDCs) in Ontario have had interval meters installed on many of their commercial and industrial customers for years. Current efforts to gather information for Cost Allocation studies and the implementation of Smart Meters should result in data being available on most classes and sizes of electricity consumers in Ontario.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 32 of 33
5. Summary Overall electricity use in Ontario grew by less than 1 percent per year (0.6 percent) between 1990 and 2003 for a total increase of 7.9 percent. In addition to being far lower than the historic growth rates prior to 1980s, this represented a growth rate well below that for either the economy or household formation. In effect, the electricity intensity of the Ontario economy decreased over the period as less electricity was used per household or per dollar of GDP. Most of the growth in electricity use over the period under review took place in the Services sector. Residential electricity use increased by just over five percent at 0.4 percent per year. Industrial electricity use declined by 0.7 percent per year (9 percent in total) while output, as measured by real GDP increased by 2.7 percent per year (42 percent in total). In the Residential sector, the electrical efficiency of major appliances increased significantly reducing their use (i.e. refrigerator use declined 33 percent). At the same time, the growing use of home electronics, computers and other small appliances resulted in an increase in the “Other Appliance” category increasing use by 57 percent. Residential lighting energy use also increased by 20 percent during the period. The reduction in electricity use by major appliances (2.9 GWh) effectively offset the increase in “Other Appliances” (2.7 GWh). Residential electric heating use increased by 6 percent during the period while electric water heating declined by 16 percent. In the Services sector, the majority of growth in electricity use occurred in the Offices sector. Offices accounted for almost 40 percent of the floor area in the sector and over 53 percent of electricity use and saw the greatest rate of growth over the period. Electricity intensities and even lighting intensities (energy used per square metre) evidently increased over the period. Lighting represents the largest single end use within the Service sector, accounting for over 37 percent of electricity use, followed by auxiliary motors (24 percent), auxiliary equipment (21 percent) and cooling (14 percent). As in the Residential sector, space cooling grew rapidly over the period increasing its electricity use by 52 percent. The industrial sector decreased its electricity use over the period, in part due to a structural change to less electricity intensive industries and in part due to increases in efficiency. Overall electricity intensity decreased for the sector, with some exceptions. The Smelting & Refining and Pulp & Paper industries increased their electricity use per dollar of product. Table 11 shows a comparison of how electricity was used in Ontario, by end use, in 1990 versus 2003. The values shown for the Residential and Service sectors are based on data derived from the Office of Energy Efficiency database. The distribution of electricity by end use for the Industrial sector was calculated based on historic values and industry knowledge.
Factor Analysis of Recent Trends in Ontario’s Electricity Use
Revised November 2005 Page 33 of 33
Table 11. Ontario Electricity Use by End Use
1990 2003 (1) Industrial Motors (2) 31,973,558 29,002,795Industrial Lighting (2) 5,937,946 5,386,233Industrial Process & Electro-technologies (2) 7,765,007 7,043,536Service Sector Space and Water Heat 1,181,090 1,076,047Service Sector Cooling 5,627,263 11,801,178Service Sector Lighting 14,845,321 19,004,744Service Sector Motors & Equipment 18,309,796 20,197,956Street lighting 781,322 781,322Residential Space Heat 9,375,684 10,705,537Residential Water Heat 8,682,320 7,255,950Residential Cooling 1,902,127 4,530,707Residential Base load 24,242,077 25,625,070Notes: 1. For Commercial and Institutional Sector 2002 values were used as 2003 space heating value appears to be inconsistent. 2. No end use breakdown is available for the Industrial sector. These values assume 70% of energy is used for motors, 23 percent for lighting and 7 percent for electro-technologies (based on industry knowledge).
Source: all data from IESO except "Energy - Customer Level Actual, GWh" which is from ICF.
Year Energy - Generator Level Actual (MWh)
Energy - Generator Level Monthly-Weather-Corrected (MWh)
Peak - Generator Level Actual (MW)
Peak - Generator Level Monthly-Weather-Corrected (MW)
Energy- Customer
Level Actual (GWh)
1958 28,3301959 31,3341960 32,319,871 32,9431961 33,569,282 33,9761962 35,488,920 35,9371963 37,288,830 37,9511964 40,305,468 41,0121965 43,511,760 44,6041966 48,054,064 46,6261967 51,355,505 51,5611968 55,788,375 55,5181969 59,424,443 59,0921970 64,287,606 11,203 63,6151971 68,132,582 68,1381972 73,495,941 12,676 72,6601973 78,162,478 13,496 77,1831974 82,695,005 13,479 81,6211975 84,220,853 14,375 80,6041976 90,851,777 15,810 86,6951977 92,853,682 15,838 90,8821978 95,371,714 16,078 93,2031979 98,126,082 16,269 97,0581980 100,173,722 16,749 98,2111981 101,659,190 17,086 102,1771982 100,835,599 17,992 98,8111983 106,071,164 18,750 102,3901984 112,293,482 18,783 110,4451985 116,048,994 20,393 112,8031986 120,573,918 20,586 118,9571987 126,455,099 20,448 123,6571988 134,394,697 22,933 130,1161989 140,770,186 23,491 135,7251990 136,744,134 22,272 132,4711991 136,965,556 23,046 132,5381992 134,376,269 23,463 131,7791993 133,477,115 21,964 131,1541994 134,874,024 23,857 129,5311995 137,037,990 135,945,465 22,812 22,462 134,2191996 137,418,082 136,734,271 22,072 22,103 134,9831997 138,370,766 138,270,615 22,030 22,093 134,1401998 139,931,347 141,165,170 22,403 22,169 134,1841999 144,094,149 143,968,842 23,433 23,002 136,4582000 146,945,129 147,740,780 23,301 23,801 140,0932001 146,912,113 147,642,051 25,239 23,402 138,5142002 152,959,790 151,599,147 25,414 24,272 145,1042003 151,716,756 151,106,402 24,753 23,893 145,0802004 153,437,275 153,924,115 24,979 24,618 146,4122005 156,969,047 155,042,710 26,160 25,823 148,092
Notes:
- Weather-corrected data only available after 1995. Note these numbers were monthly-corrected rather than seasonal-conrrected.- Historical peak demand at the customer level Not available.
“These results do not correspond directly to the updated results for the IPSP. This document is included as an illustration of the CIMS/National Study
methodology, which was produced using an older Canada Emissions Outlook produced by Natural Resources Canada.”
Marbek Resource Consultants Ltd. M.K. Jaccard & Associates, Inc. 300222 Somerset Street West 414 675 West Hastings Street Ottawa, Ontario K2P 2G3 Vancouver, B.C. V6B 1N2 Tel: 613.523.0784 s Fax: 613.523.0717 Tel: 604.6831252 s Fax: 604.6831253 www.marbek.ca
DEMAND SIDE MANAGEMENT POTENTIAL IN CANADA: ENERGY EFFICIENCY STUDY
–Summary Report–
Submitted to:
Canadian Gas Association
Submitted by:
Marbek Resource Consultants Ltd.
and
M.K. Jaccard and Associates, Inc.
May 2006
Table of Contents
1. INTRODUCTION .........................................................................................................1
1.2. Study Scope .........................................................................................................2 1.3 Study Context ......................................................................................................4
2. METHOD EMPLOYED ...............................................................................................6
2.1 Modelling Platform ..............................................................................................6 2.2 The Study Scenarios.............................................................................................6
3. RESULTS ....................................................................................................................14
3.1 Reference Case Forecast ....................................................................................14 3.2 Economic Potential Results ................................................................................15 3.3 Achievable Potential Results ..............................................................................17
4. DISCUSSION ..............................................................................................................24
Appendices:
Appendix A Reference Case Report Appendix B Economic Potential Report Appendix C Achievable Potential Report
List of Exhibits
Exhibit 1.1: Explanations for Lower than Expected Energy Efficiency Investment ....................5 Exhibit 2.1: Technovert National Energy Prices .......................................................................8 Exhibit 2:2: Summary of Policies Instruments Applied in Each Scenario .................................13 Exhibit 3.1: Reference Case Energy Demand (PJ), Commercial Sector....................................15 Exhibit 3.2: Reference Case Energy Demand (PJ), Residential Sector ......................................15 Exhibit 3.3: Reference Case Energy Demand (PJ), Industrial Sector.........................................15 Exhibit 3.4: All Sectors National Economic Potential Energy Demand Reduction by Milestone
Year and Fuel (PJ)................................................................................................16 Exhibit 3.5: National Economic Potential by Sector Share of Energy Reduction in 2025 .........17 Exhibit 3.6: Total Enduse Energy Demand by Scenario, All Sectors.......................................18 Exhibit 3.7: Energy Demand, by Milestone Year: Achievable Potential Scenarios vs. Reference
Case and Economic Potential................................................................................18 Exhibit 3.8: Energy Savings by Milestone Year: Achievable Potential Scenarios vs Reference
Case and Economic Potential................................................................................19 Exhibit 3.9: Comparison of Achievable Potential Scenario 2 Savings and 19902003 Energy
Demand Growth...................................................................................................19 Exhibit 3.10: National Achievable Potential by Sector Share of Energy Reduction in 2025:
Scenarios 1 and 2 .................................................................................................20 Exhibit 3.11: All Sector Savings According to Fuel...................................................................22 Exhibit 3.12: Added Cogeneration Generation by Sector ..........................................................23 Exhibit 3.13: Achievable Potential Performance Range From Recent DSM Studies...................28
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 1
1. INTRODUCTION
This report summarizes the findings of a highlevel study to estimate the achievable demandside management (DSM) potential in Canada. The study was conducted for the Council of Energy Ministers demandside management (DSM) Working Group with the goal of bringing DSM to the forefront of the energy and economic policy discourse in the country. The DSM Working Group comprises representatives from the federal government (Natural Resources Canada), provincial governments, the utility industry, major energy users, and nongovernmental organizations.
The report culminates a comprehensive analysis of three key sectors of the economy: industrial, residential and commercial/institutional (hereafter, referred to as commercial). The study comprised three important scenarios, reference case (businessasusual), economic potential and achievable potential; each of those milestones are documented in separate reports which are presented in appendices as follows:
Reference Case ReportAppendix A Economic Potential ReportAppendix B Achievable Potential ReportAppendix C.
This report summarizes the findings of these three reports.
The study findings indicate that the total achievable reduction in energy demand in 2025 for the industrial, residential and commercial sectors could be reduced by between 3% and 10%. as a result of a diverse mix of policy instruments. 1 Moreover, this savings range means that achievable energy management can meet 16% to 56% of the projected energy demand growth to 2025. The estimated reduction in energy demand is due to a mix of energy efficiency, cogeneration and fuel substitution measures, driven by a range of policy instruments. This range of achievable potential savings, as determined from this study, represents a credible contribution to meeting Canada’s longterm energy supply needs.
The study was conceived as a high level, policy oriented exercise and, as such, the outputs should been seen as the foundation for future dialogue. This dialogue should further examine how to advance DSM to the forefront of energy policy circles and, hopefully, bring direction, certainty and action to the policy concepts presented herein, or to alternative policy mixes.
The study findings should not be taken as the platform for DSM program design. For some jurisdictions the study findings are based on aggregated regional data and do not necessarily reflect the actual situation of the individual jurisdictions within the region. Therefore, while the study provides a sound indication of DSM potential on a national basis, it is not intended to provide all details sufficient for the development of specific programs to meet the needs of individual jurisdictions.
This summary report is organized according to the following subsections:
1 The term “demand” used in the report refers to the demand for purchased energy to meet energy service needs.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 2
This introductory section, which includes the study context and scope. The method employed. The results which present the empirical outputs for the achievable and economic potential
scenarios. Discussion of the results.
1.2. STUDY SCOPE
The study scope is defined as follows:
Sector Coverage: The study addresses three sectors: residential, commercial/institutional (referred to as commercial) and industrial. Energy supply sectors (electricity, upstream oil and gas and coal) are not included in the study.
Geographical Coverage: The study results are presented for seven provinces and regions, including British Columbia and the territories, Alberta, Saskatchewan, Manitoba, Ontario, Quebec, and the Atlantic region.
Energy Types: All energy types are covered including natural gas, electricity, refined petroleum products and other fuels such as biomass.
DSM Coverage: For this study, DSM includes energy efficiency, fuel substitution, cogeneration and distributed generation. Cogeneration (or combined heat and power) produces both electricity and useful thermal energy simultaneously from the same fuel (or fuels). The analysis considers all technologies that are expected to be commercially viable through to 2025.
How the DSM Impact is Reported: The DSM scenarios analyzed in the study comprise energy efficiency, fuel substitution, cogeneration and distributed generation measures that affect changes in enduse energy demand among the three studied sectors. This has a resulting effect on the amount of purchased and nonpurchased energy supply required by these sectors. The study reports the total effect of the measures on energy demand, meaning that the outputs take into account both reduced secondary energy demand and changes in the mix of primary energy demand. No attempt was made in this study to relate the electricity savings to peak or average demand reduction.
Jurisdictions: DSM and energy efficiency measures are contemplated for utilities and for all levels of government in Canada (including municipal, provincial and federal).
Study Period: This study covers a 25year period. The base year is 2000, with milestone periods at 5year increments: 2005, 2010 2015, 2020, and 2025.
Metrics Used to Present Results: All of the national levels results are presented in metric energy units. 2
2 The factors used to convert to common units are: NG: 39.8MJ/m3, Fuel Oil (light): 38.68 GJ/m3, propane: 25.53 GJ/m3 (0.02553 GJ / litre), electricity: 0.0036 GJ / kWh
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 3
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 4
1.3 STUDY CONTEXT
During the past 25 years, governments at all levels, together with both natural gas and electric utilities, have delivered a wide array of market interventions in an effort to reduce overall demand for energy by residential, industrial, or commercial energy users. The energy efficiency of most equipment and buildings in Canada has steadily improved. Moreover, between 1990 and 2004, the energy intensity of industrial production declined. by 30%. 3
Notwithstanding these performance improvements, energy demand continues to climb for all sectors. Between 1990 and 2003, secondary energy use increased 22 percent, from 6,951 to 8,457 petajoules (PJ). 4 What is happening is that the effects of economic activity, namely, the growth of the housing and commercial building stock, larger homes, the market penetration of more energy using devices, and industrial production growth together offsets the effects of energy efficiency improvements. Hence, the energy demand curve continues to show an upwards trajectory. A difficult question for this study is how much and at what speed we can affect this trend and, consequently, bend the slope of the curve. Key challenges exist and two dimensions to this challenge are worth noting here.
At the risk of oversimplification, a good portion of the DSM “low hanging fruit” has already been attained in all three sectors, i.e., many of the lower cost, short payback measures have been implemented. This includes, for example, the penetration of higher efficiency appliances, motors and lighting. Unless economic circumstances change considerably, the potential that remains will be more difficult to capture for several reasons, including: i) the target submarkets become more challenging, e.g., small commercial, mid and highrise apartments, small and medium sized industry and ii) the solutions can become more complex, e.g., moving to process integration and balance of plant measures in industry; getting industry and commerce to effectively apply corporate energy management systems as the foundation for ongoing, sustainable and strategic management of energy.
Equally important is the degree to which policy can influence the adoption of greater energy efficiency in the economy by addressing fundamental market barriers. Experience with market intervention over the past two decades has shown that, while many energy efficiency opportunities can be shown to be costeffective, when the monetary value of energy savings is assessed against the initial capital cost outlays, consumers and firms forego apparently cost effective investments in energy efficiency. Energy users appear to discount future savings of energyefficiency investments at rates well in excess of market rates for borrowing or saving. This has often been referred to as the energyefficiency "gap". 5 Exhibit 1.1 lists some of the
3 Based on gross output. This is for ‘Total Industry’ (NAICS 100000). ‘Total Manufacturing Industry’ (NAICS 100001) shows a similar trend. Canadian Industrial Energy Enduse Data and Analysis Centre (CIEEDAC), Development of Greenhouse Gas Intensity Indicators for Canadian Industry, 1990 to 2004 , Burnaby: Simon Fraser University, 2005. 4 Office of Energy Efficiency, Energy Efficiency Trends in Canada, 1990 to 2003, Ottawa: Natural Resources Canada, 2005. 5 For example, see A. Jaffe and R. Stavins, “The EnergyEfficiency Gap: What Does it Mean?” Energy Policy 22, 10 (1994): 804810; J. Scheraga, “Energy and the Environment: Something New under the Sun?” Energy Policy 22, 10 (1994): 811818; R. Sutherland, “The Economics of Energy Conservation Policy,” Energy Policy 24, 4 (1996): 361370.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 5
crosscutting barriers, market behaviours and failures identified in the literature to explain why the take up of energyefficiency is lower than expected.
Exhibit 1.1: Explanations for Lower than Expected Energy Efficiency Investment
Category Explanation
Price Signals Energy pricing at levels that do not integrate externalities associated with the cradle to grave lifecycle (full cost accounting).
Energy pricing signals that do not reflect realtime costs. Consumer Awareness and Preferences
Awareness that energy efficiency opportunities & products exist Awareness of benefits – cost and cobenefits. Consumer technical ability to assess the options. Consumer offsetting preferences (e.g., large single detached homes). Lack of public perception/understanding of infrastructure needs/ resource
constraints/ the functionality, cost, drivers and challenges are unknown to the public.
Product and Service Availability
Local or national product availability. Existence of a viable infrastructure of trade allies. Vendor or trade ally awareness of the efficiency options and their
understanding of the technical issues. Technology and Innovation An energy efficient technology may not be a perfect substitute for another,
accepted technology for an enduse. An energy efficient technology may not be costeffective for all
consumers, even if it is costeffective for the average consumer. Lack of enabling tools and techniques to facilitate market adoption of
sustainable energy solutions. Financing Access to appropriate financing.
Uncertain future energy prices, combined with the irreversible nature of energy efficiency investments.
Size of required energy efficiency investment vs. asset base. Payback ratio – actual vs. required.
Transaction Costs Level of effort/hassle required to become informed, select products, choose contractor(s) and install.
Perceived Risk/Reward Level of perceived risk that the energy efficient product may not perform as promised.
Level of positive external/personal recognition for “doing the right thing” by installing the efficiency measure(s).
Split Incentive/Motivation Level to which the incentives of the agent charged with paying for the energy efficiency measure are aligned with those of the person(s) that would benefit.
Institutional and Regulatory Codes or standards that prohibit implementation of innovative energy efficient technologies.
Limited horizontal cooperation/coordination to integrate policies and implementation.
Municipal policies and land planning processes that supported, even encouraged, development of greenfield areas and subsidized the practice through low development fees.
Disconnect between longevity of infrastructure and shortterm horizons on crucial decisions, such as budget allocations for maintenance and rehabilitation and rate structures.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 6
2. METHOD EMPLOYED
2.1 MODELLING PLATFORM
The analysis was conducted using the CIMS model, supported by Marbek DSM tools and databases. 6 CIMS is an integrated energyeconomy model that simulates technology acquisition in the economy over time. Technologies are represented in unique submodels that meet energy service demands in the residential, commercial, transportation, electricity supply, and industry sectors. It is therefore possible to specifically represent the evolution of a technology, or group of technologies, in a forecast and to alter model inputs to simulate alternative forecasts and policy scenarios.
The takeup of DSM technologies in CIMS is driven by a model construct that tries to reflect the financial and nonfinancial considerations affecting energy user decisions and choices. CIMS is a platform for a competition among various DSM technologies. While the engine for this competition is the minimization of annualized life cycle technology costs, energy user decisions not only depend on recognised financial costs (capital, energy and other operating and maintenance costs), but also respond to:
Identified differences in nonfinancial preferences (e.g. differences in the quality of lighting from different light bulbs).
The preferences of firms and households with respect to the risk of newness and risk of irreversible investments. Thus the lifecycle cost is calculated with effective ‘private’ discount rates that are revealed from market data. 7
The nondeterministic nature of market behaviour. Market shares are allocated among technolgoies probabilistically according to a variance parameter. 8
The preference parameters in CIMS are set using a combination of literature review, original survey research, expert judgment, and model validation.
2.2 THE STUDY SCENARIOS
2.2.1 Scenario Definitions
In this project CIMS was applied to develop four scenarios: a reference case, an economic potential, and two achievable potential scenarios. Given that energy systems in Canada differ significantly by region, the national potential for energy demand reduction is derived from the analysis of regional potentials (rather than a single national potential). This is done according to the disaggregation currently available in the CIMS model. Unique submodels represent British Columbia, Alberta, Saskatchewan, Manitoba, Ontario, Quebec and a combined Atlantic region. The CIMS model is not currently set up to model the Atlantic region on a provincial basis and, therefore, the analysis of the
6 The CIMS model is developed by the Energy and Materials Research Group and Simon Fraser University. 7 Revealed discount rates cover both of these factors because the new technologies of interest to energyeconomy modellers are those that increase energy efficiency through irreversible, long payback investments. 8 In contrast, the optimizing models will tend to produce outcomes in which a single technology gains 100% market share of the new stocks.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 7
Atlantic region potential does not reflect the diversity of energy systems, fuel availability, prices and mix, and electricity prices in the Atlantic provinces. 9
The scenarios are defined as follows:
Reference Case: A projection of energy demand to 2025, in the absence of any new and incremental institutional market interventions after 2005. It is the baseline against which the scenarios of energy savings are calculated. The reference case includes "natural conservation", i.e., changes in end use efficiency due to stock replacement, energy prices and other factors over the study period that are projected to occur in the absence of new and incremental market interventions.
Economic Potential: An estimate of the energy demand that would occur if all equipment and building envelope energy management actions that pass a ‘Total Resource Cost’ test were implemented in the target markets. These actions are applied at either natural stock turnover or retrofit rates.
Achievable Potential: An estimate of the energy demand that would occur as a result of market intervention to influence the take up of energy management actions. The potential is estimated in two policy scenarios. The first focuses on the response from subsidies to specifically target the uptake of actions identified in the ‘Total Resource Cost’ test in the Economic potential. The second scenario includes the energy demand response to broader based policy instruments, landuse measures and ‘aggressive’ building and equipment standards and renewables subsidies.
2.2.2 Reference Case Elaboration
The reference case forecast is strongly influenced by three factors: energy prices, economic growth, and the saturation and mix of energy using equipment in the existing buildings and industrial stock. The CIMS base year in all regions is calibrated to within +/5% of the latest 2000 energy supply and demand data from Statistics Canada and, consequently, 2000 is the start year of the study analysis. The most critical challenge was to update the pricing assumptions to ensure a robust and credible modeling foundation.
Prior to this study, the energy prices in CIMS were based on Natural Resources Canada (NRCan)’s Canada’s Emissions Outlook: An Update 2000 which is, of course, outdated. At the time when the Reference Case was to be constructed, NRCan had not yet completed a new national energy use and price forecast. With the support of the DSM Working Group, the consulting team decided to completely update the energy price schedule in CIMS. After consultations with the DSM Working group, we adopted the price forecasts of one of the two scenarios embodied in the National Energy Board (NEB)’s Canada’s Energy Future, referred to as the “TechnoVert” scenario. 10
9 The Atlantic region accounts for 7% of the enduse energy (in 2000) for sectors represented in the study.
10 National Energy Board, Canada’s Energy Future” Scenario’s for Supply and Demand to 2025. (Supply Push and TechnoVert scenarios). http://www.nebone.gc.ca/energy/SupplyDemand/2003/index_e.htm
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 8
The NEB scenarios represented the only recent forecast available with provincial and sectoral coverage. The TechnoVert scenario was selected as the more realistic of the two options because: i) it projects higher energy prices and ii) due to the higher energy prices it embodies a higher rate of “natural conservation”. Exhibit 2.1 presents the national prices in the Technovert forecast. 11 Interestingly, even under the more aggressive price forecast, it shows declining or stable price trends over the study period.
Exhibit 2.1: Technovert National Energy Prices
Canada 2000 2005 2010 2015 2020 2025
Residential ($1995/GJ) Electricity $21.24 $21.08 $22.51 $21.95 $21.35 $20.52 Natural Gas $7.60 $8.90 $9.33 $9.10 $8.87 $8.62 Light Fuel Oil $12.83 $11.67 $12.77 $12.65 $12.51 $12.38 Commercial ($1995/GJ) Electricity $17.21 $18.83 $19.85 $19.35 $18.76 $17.92 Natural Gas $6.27 $7.80 $8.23 $8.02 $7.79 $7.52 Light Fuel Oil $12.99 $11.05 $11.50 $11.61 $11.81 $11.67 Heavy Fuel Oil $7.18 $5.24 $5.22 $4.99 $4.64 $4.02 Industrial ($1995/GJ) Electricity $12.39 $13.32 $14.02 $13.66 $13.22 $12.63 Natural Gas $4.19 $5.59 $6.02 $5.76 $5.51 $5.23 Heavy Fuel Oil $5.42 $5.11 $5.06 $4.83 $4.59 $4.31 Coal $2.36 $2.30 $2.25 $2.25 $2.25 $2.36
There was also considerable effort invested to review and update the DSM technologies in the CIMS submodels. The update addressed the following parameters: i) coverage of DSM technology candidates, ii) energy performance and iii) installed costs.
2.2.3 Economic Potential Scenario
The economic potential is defined as a future in which energy efficiency investments are adopted by all producers and consumers (at the rate of technology stock turnover and/or accelerated takeup through retrofit opportunities), if the life cycle cost (LCC) of the investment is lower than the longrun cost of energy supply. In the economic potential, three major parameters affect the life cycle cost competition and, therefore, drive the economic potential: i) the energy long run marginal cost (LRMC) used for screening the economics of the candidate technologies ii) the discount rate and iii) the variance parameter.
The LRMC valuation combines the costs of generation, production, transmission and distribution and is a two step exercise: i) separate valuation methods are employed to establish the LRMCs for electricity versus natural gas and Refined Petroleum Products and ii) a carbon liability value is added to all of the energy forms.
11 The regional price forecasts from this scenario were adopted in CIMS.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 9
In CIMS, the electricity LRMC value is derived to estimate the supply price of a new combined cycle gas turbine (CCGT) in each jurisdiction to which is added the costs of transmission and distribution, while taking into account line losses. Among other things, the supply price estimates are regionalized by setting the CCGT variable fuel cost in each 5year period at the regional market price for industrial natural gas. Again, as this is a high level, policy oriented study, it does not fully capture all of the regional and provincial realities and drivers affecting long run power generation baseload and peaking supply. We recognize that the CCGT option will not necessarily apply to all regions. A carbon price of $15 / t CO2e is also incorporated into the energy prices (based on the carbon content of the affected fuels) as a financial cost liability that is considered in a full calculation of the long run marginal cost. 12
As the economic potential is a societal perspective, the life cycle cost analysis uses a social discount rate of 10% real for all regions and technologies. The technology competitions which occur in the reference case and achievable potential projections use a schedule of private discount rates that are typically much higher than the social discount rate. Changing the discount rate from a private to a social perspective has two effects in the competition of technologies in CIMS. First, more energy efficiency measures are likely to pass the life cycle cost test generating a positive net present value. Second, among the larger number of measures that become candidates for competition, an increasing number of higher performing measures are selected as the least cost option.
CIMS contains a variance parameter (‘v’) that represents sensitivity of the technology adoption to relative life cycle costs. A high v value means that the technology with the lowest life cycle cost captures almost all of the market for new equipment stock, a “winner takes all result”. A low v value means that new equipment market shares are distributed more evenly among competing technologies, even if their lifecycle costs are different. The value of the v factor is set low for the economic potential scenario thereby enabling only the least cost measure to be selected. Most DSM studies model the economic potential with the highest performing measures included that pass the economic cost test. The due diligence conducted during the CIMS modeling reveals that in most instances the highest performing measures are selected.
2.2.4 Achievable potential
Two achievable potential scenarios are modelled in this study, referred to as achievable scenario 1 DSM Status Quo and achievable scenario 2DSM Aggressive. These scenarios represent considerably different visions of how various policy instruments may
12 There is evidence that utilities commissions are beginning to force the internalization of greenhouse gas (GHG) liabilities such that these are now part of the real energy cost structure faced by utilities in their decisionmaking. We have included a price of $15 / t CO2e in the modelling that is incorporated into the energy prices based on the carbon content of the affected fuels. This price was chosen as this has already been approved in at least one jurisdiction for utility investment analysis (by the BC Utilities Commission for BC Hydro). It also reflects the commitment from the Canadian government to the Large Final Emitters (LFE) group that their GHG reduction cost compliance will not exceed this value. Note that this liability does not represent an estimate of the full externality cost of GHG or other emissions. It is simply a financial cost liability that is considered in a full calculation of LRMC, recognizing that all cost estimates have present and future uncertainties associated with them.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 10
be brought to bear on the residential, industrial and commercial/institutional markets during the study period.
Scenario 1: DSM Status Quo
DSM Status Quo assumes a continuation of approximately the current levels and types of market interventions by government and utilities. A scan undertaken during the study revealed that current annual energy efficiency expenditures by government and utilities amount to between $400 million and $500 million per annum. 13 Moreover, the majority of the energy efficiency program costs borne by utilities fall into the category of subsidies of one form or another. Not surprisingly then, most utility reported annual energy savings are attributed to the effect of these subsidies, in the vicinity of 75% of total reported savings. 14 Government program costs are more broadly distributed, among subsidies, energy performance standards development and administration, information and R&D. Consequently, reported energy savings from government initiatives are attributed more broadly to the foregoing mix of instruments, particularly due to energy performance standards.
In consultation with the CGA client group, the DSM Status Quo scenario was designed as a combination of subsidies and information/voluntary programs, with the major driver in the scenario assumed to be the subsidy instruments. Financial subsidy is a policy instrument designed to reduce the energy management investment cost to a level commensurate to the business and consumer hurdle rates. Subsidies for energy management continue to be a prevalent means of delivering DSM in Canada and elsewhere. As discussed in the Economic Potential report, there is a considerable gap between the social and private discount rates for energy management. Hence, the argument is that if a particular energy management measure passes a societal cost test, then it is legitimate to use subsidies to induce market takeup of the measure. 15
The inclusion of energy performance standards was considered for this policy mix, since they are certainly part of the current DSM landscape in Canada. Mandatory energy performance standards are presently focused on improving equipment performance levels, less so on building performance. It was posited that there remains a considerable upside for enhanced performance standards and, consequently, it was decided to include this policy instrument in the second, more aggressive scenario.
13 This estimate is based on a scan of the following documents: i) NRCan “Improving Energy Performance in CanadaReport to Parliament Under the Energy Efficiency Act Fiscal 200405, Appendix 1”. The estimate for federal expenditures is about $165 million per year. ii) Canadian Electricity Association and Natural Resources Canada, Description and Results of Energy Management ProgramsA survey Of Programs Operated By Electric Utility Companies in Canada, March 2003 and Update in October 2003. iii) Indeco in association with B. Vernon and Associates, DSM Best PracticesCanadian Natural Gas utilities Best Practices in DemandSide Management, undertaken for the Canadian Gas Association, 2005. 14 This is based on inhouse data/files plus a small selection of telephone conversations with gas and electric utility officials. 15 Another way of looking at this is that, if the cost of delivering the energy management measure is less than the social cost of the displaced energy form, then it is an economically legitimate investment from the standpoint of society.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 11
The subsidy schedule targeted the energy efficient technologies identified in the economic potential at rates consistent with current observed utility incentive levels (10% 35% of the measure cost). 16 The effect of the information/voluntary programs was modelled exogenously as a multiplier applied to the results based on utility and NRCan estimates of program effectiveness.
Scenario 2: DSM Aggressive
Scenario 2, DSM Aggressive, models the achievable potential to 2025 as a vision of how to more effectively address market barriers and failures and consequently expand and accelerate the energy management effect in the economy over this period. The scenario includes new and expanded policy instruments involving all levels of government, utilities and the private sector that can capture a greater array of options. It also assumes that policies could do more to address fundamental changes that need to be made regarding urban land use intensity and form which, in turn, will affect needed changes to foster sustainable infrastructure. The DSM Aggressive scenario comprises the following policy elements:
An aggressive application of energy efficiency standards, for both enduse equipment and buildings.
Subsidies to energy efficiency technologies. These are applied as a complementary instrument to subsidies. The same subsidy levels used for DSM Status Quo were applied but at a different rate of application. The technologies eligible for subsidy application fall into two categories: i) those that will be affected by the standards and ii) those that will not be affected by the standards.
The energy efficiency standards are introduced at varying schedules during the study period. Consequently, the subsidies are applied to the technologies to be affected by standards in year one of the study period and continue to be applied only until the technology is affected by the performance standard. The subsidies are applied to the technologies, not affected by standards, in year one of the study period and continue during the study period.
An aggressive subsidy policy directed to induce a greater market penetration of some renewable energy technologies for onsite applications, which would have an incremental fuel substitution effect towards renewable energy, relative to the reference case forecast. The focus is onsite renewables applications to replace secondary energy consumption of gas and refined petroleum products and to reduce electric power purchases. While this is characterized as an aggressive renewables policy, conceptually it corresponds well to efforts internationally. There are many examples worldwide of government, at all levels, instituting aggressive renewable energy policies and programs. For example, the California Public Utilities Commission (PUC) recently voted to adopt the California Solar Initiative (CSI), which will provide up to $2.9 billion in incentives toward solar development over 11
16 These represent energy efficiency investments whose life cycle cost of the investment is lower than the longrun cost of energy supply. This is roughly equivalent to targetting those investments that ‘pass’ a Total Resource Cost test.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 12
years. One of the goals this initiative is to install 3,000 MW of solar power capacity by 2017, making it the largest solar program in the U.S. 17
Application of marginal cost pricing in electricity. This seeks to simulate the effects of advancing from a monopoly average cost pricing regime for electricity to a regime that embodies marginal cost pricing. In practice, a marginal cost policy instrument could be manifested in a number of ways: regulators requiring this for electricity pricing, or some form of timeofuse pricing measured and reported on a real time basis. This policy only applies to electricity because the other energy prices already represent marginal cost pricing as their prices are determined in competitive markets. This policy is modelled in CIMS by revising the electricity price forecast used in the simulation. The same long run marginal electricity price forecasts are used as calculated for the economic potential.
A $15/tonne CO2e price adder for all fuels based on the carbon content of the affected fuels. This is representative of mechanisms that are starting to be used by energy utilities to price or cost GHG emission reductions for use in planning, acquisition, project development or operational decisions. These mechanisms include: i) government instituted “safety valves” or price assurance relating to CO2 regulation, ii) resource planning GHG “adders” and iii) energy acquisition GHG bid price adjustments
Changes to shares of projected housing types (low rise versus mid to highrise) to mimic the potential effects of aggressive urban land use policy instruments. The percentage of single detached dwellings was reduced in absolute terms by 25% in 2025. This considers the largely untapped area of land use as a means to reduce the environmental footprint of communities, particularly in the urban centres where 80% or more of the Canadian population resides. In terms of affecting reductions of energy consumption, sustainable land use policy instruments can generate the following possible outcomes: i) reduced average energy use per dwelling or building, ii) reduced transportation energy use. This scenario deals with the challenge of reducing average energy use per dwelling.
There is a wide range of possible policy instruments to affect land use change in municipalities, which taken together, can affect: i) the type and amount of land use, ii) the intensity of use within the land boundary, iii) the spatial distribution and location of use (e.g., degree of sprawl).
To summarize, the aggressive DSM scenario includes:
Energy efficiency subsidies. These are the same as scenario 1, except they are retargeted where regulation is applied to the same energy enduse.
Marginal cost pricing for electricity.
17 California PUC website, CSI includes $2.5 billion in rebates for existing homes, businesses and public buildings, to be managed by the PUC and funded through revenues collected from gas and electric utility distribution rates. The California Energy Commission (CEC) will manage another $350 million in rebates targeted for new residential construction, utilizing funds already allocated to the CEC to foster renewable projects between 2007 and 2011.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 13
A carbon liability. A shadow price of $15/tonne CO2e is applied to all energy price forecasts.
An aggressive schedule of legislatively backed advanced minimum energy performance targets for both equipment and buildings.
Renewable subsidies. These are targetted at the residential and commercial sectors and in particular solar hot water heaters, solar photovoltaic and geoexchange, which are subsidized at 30%, 40% and 15% of installed capital cost respectively.
Changes in the shares of projected housing types (low rise versus mid to highrise). The percentage of single detached dwellings was reduced in absolute terms by 25% in 2025.
Exhibit 2.2 summarizes how the mix of policy instruments was applied in both of the achievable potential scenarios; the dark shaded area indicates application of the instrument.
Exhibit 2:2: Summary of Policies Instruments Applied in Each Scenario
Policy Instruments Scenario 1 Scenario 2 Subsidiesenergy efficiency schedule Subsidiesrenewables schedule Information Regulation & Standards Marginal cost pricing Carbon liability Change in dwelling type shares
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 14
3. RESULTS
3.1 REFERENCE CASE FORECAST
Exhibits 3.1 to 3.3 show the national reference case scenarios for the commercial, residential and industrial sectors respectively. Since the CIMS base year in all regions is calibrated to within +/ 5% of the latest 2000 energy supply and demand data from Statistics Canada, the Reference Case forecast runs from 2000 to 2025.
The high level national results by sector are as follows:
Across all sectors, energy demand is forecast to increase by 23% amounting to an average annual increase of 0.85%. The forecast growth occurs despite a projected decline in energy intensities (energy demand per unit of output) in all sectors. The activity effects of economic growth offset the energy performance improvements. There is no significant change in fuel shares among the major energy forms used in these sectors.
Commercial/Institutional. Exhibit 3.1 shows a total energy demand increase of 353 PJ over the study period, amounting to an annual increase in consumption of 1.14%. The model results also show that the fuel shares remains relatively constant in the commercial sector, with natural gas increasing from 51% to 55% by 2025, and electricity’s share falling slightly from 42% in 2000 to 37% in 2025. Energy intensity shows a small improvement over time with an average annual change (or decrease) of 0.56%.
Residential. Exhibit 3.2 shows a total energy demand increase 279 PJ over the study period, amounting to an average rate of less than one percent annually. Once again the split between fuels remains relatively constant. The share of natural gas fluctuates around 48%, and the share of electricity rises slightly from 36% to 39%. Annual growth rates for both fuels are in the order of 1% annually whereas growth in refined petroleum products (RPP) is lower (0.36%) and other fuels (wood) decline about 0.8% annually. Energy intensity show an improvement slightly greater than the commercial sector and in the order of 0.59% annually;
Industrial. Exhibit 3.3 shows that in the industrial sector total energy demand rises from 2,714 PJ in 2000 to 3,296 PJ in 2025, or at a rate of 0.78% annually. In this sector natural gas and electricity both exhibit declines in their fuel share , although the absolute demand for both these fuels continues to rise throughout the forecast period. Refined petroleum products and the other fuels listed see a slight increase as a percent of the total energy demand. The industrial sector forecast represents manufacturing and metals and mineral mining, and does not include energy supply subsectors (upstream oil and gas, coal mining and electricity supply subsectors). Construction and forestry are also not included.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 15
Exhibit 3.1: Reference Case Energy Demand (PJ), Commercial Sector
2000 2005 2010 2015 2020 2025 Average Annual Change
Total Energy 1,075 1,130 1,192 1,275 1,352 1,431 1.15%
Electricity 448 462 477 500 519 540 0.75% Natural Gas * 548 584 626 680 732 785 1.45% Refined Petroleum
Products 79 85 88 95 101 106 1.15%
*Natural gas includes Propane.
Exhibit 3.2: Reference Case Energy Demand (PJ), Residential Sector
2000 2005 2010 2015 2020 2025 Average Annual Change
Total Energy 1,384 1,419 1,444 1,501 1,576 1,663 0.74%
Electricity 497 516 529 557 600 643 1.04% Natural Gas 659 676 692 722 753 795 0.75% Refined Petroleum
Products 132 135 134 138 142 145 0.36%
Wood 96 92 89 85 81 79 0.77%
Exhibit 3.3: Reference Case Energy Demand (PJ), Industrial Sector
2000 2005 2010 2015 2020 2025 Average Annual Change
Total Energy 2,714 2,785 2,931 3,053 3,154 3,296 0.78% Electricity 670 676 716 728 738 757 0.49% Natural Gas* 922 920 925 945 960 999 0.32%
Refined Petroleum Products 161 166 177 191 206 220 1.24%
Coal, Petroleum Coke, Waste Fuels, Off gases
463 514 567 607 653 700 1.67%
Wood Waste/ Spent Pulping Liquor
498 509 546 582 596 619 0.88%
*Natural gas includes propane and other liquefied petroleum products
3.2 ECONOMIC POTENTIAL RESULTS
The consulting team has diverging opinions concerning what is signified by the economic potential. Marbek generally accepts and uses the term ‘economic potential’ to represent an economic upset, a performance ceiling to which energy efficiency market interventions can be targeted. Conversely, MKJA prefers a term like technoeconomic potential, in recognition that the “economic potential” is usually not all economic for the individual investor or society, and does not in itself represent an economic performance ceiling. MKJA’s position is that the
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 16
analysis of economic potential rarely accounts for the different costs of competing technologies in terms of their risks or the quality of service.
Notwithstanding these differing views, the results indicate a significant potential for energy demand reduction. Exhibits 3.4 and 3.5 present the economic potential results for all sectors combined. In 2025, the total reduction in energy demand for all three sectors amounts to 918 PJ, a 14% reduction relative to the reference case. This savings impact is equivalent to about 60% of the total aggregate increase in energy consumption in the three sectors between 1990 and 2003. It also amounts to about $10.5 billion in operating savings for industry, businesses and consumers in 2025. relative to the reference case forecast of energy demand.
The economic potential scenario comprises a significant fuel substitution effect due to cogeneration applications in all three sectors, the largest application having been modelled in the commercial sector. As elaborated in the ensuing sections, when the sectoral cogeneration effect is netted out, the economic potential results are generally conservative when compared to recent DSM studies conducted in Canada.
Under the economic potential scenario nearly 40 TWh of electricity will be produced from cogeneration. Nearly 60% of the cogeneration load is attributed to the commercial sector, another 28% in industry.
About 50% of the total energy demand reduction in 2025 is attributed to electricity reduction. Of this amount, about 30% is due to added cogeneration supply. Natural gas savings represent about 28% of the total reduction in 2025 and represent a larger savings when the cogeneration effect is netted out.
Exhibit 3.4: All Sectors National Economic Potential Energy Demand Reduction by Milestone Year and Fuel (PJ)
2010 2015 2020 2025 Total Energy Demand Savings (PJ) 417.0 613.7 767.6 917.8 % Savings Relative to Reference Case 7% 11% 13% 14% Electricity (PJ) 184.1 285.1 379.3 466.4 % Savings Relative to Reference Case 11% 16% 20% 24% Natural Gas (PJ) 157.7 209.0 228.6 250.0 % Savings Relative to Reference Case 7% 9% 9% 10% Refined Petroleum Products (PJ) 21.6 29.0 39.0 47.6 % Savings Relative to Reference Case 5% 7% 9% 10% Wood Waste/ Spent Pulping Liquor (PJ) 39.5 57.7 69.0 76.7 % Savings Relative to Reference Case 6% 9% 10% 11% Coal, Petroleum Coke, Waste Fuels, Off gases (PJ) 13.9 33.0 51.8 77.1 % Savings Relative to Reference Case 2% 5% 8% 11%
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 17
Exhibit 3.5: National Economic Potential by Sector Share of Energy Reduction in 2025
3.3 ACHIEVABLE POTENTIAL RESULTS
The achievable potential is a measure of how a target market might respond to one or more market interventions designed to expand and accelerate market takeup of energy management measures. The rationale for market interventions is to address one or more barriers and failures which impede market takeup of these measures to the level of what is economically viable, today and in the future, when market circumstances are expected to change. As noted, two achievable potential scenarios were analyzed: DSM Status Quo and DSM Aggressive.
3.3.2 Overall Impacts
Exhibits 3.6 to 3.9 present the overall impact of the two scenarios. In 2025 the total reduction in energy demand ranges from 182 PJ to 647 PJ, a 2.9% to 10.1% range in energy demand reduction relative to the reference case forecast. The average annual growth rate in energy demand slows to 0.68% in scenario 1 and 0.36% in scenario 2, relative to 0.85% in the reference case. Using the projected energy market prices used in the Reference Case forecast, the achievable potential savings amounts to a range of $3.2 billion to $15.7 billion in energy operating cost savings in 2025 relative to the reference case forecast. The projected energy demand reduction under scenario 2 is equivalent to about 64% of the total aggregate increase in energy consumption in the three sectors between 1990 and 2003.
Residential, 24%
Industrial, 39%
Comemrcial , 37%
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 18
Exhibit 3.6: Total Enduse Energy Demand by Scenario, All Sectors
5,000
5,250
5,500
5,750
6,000
6,250
6,500
2000 2005 2010 2015 2020 2025 Year
Energy demand (PJ)
Reference Case Economic Potential Scenario 1 Scenario 2
Exhibit 3.7: Energy Demand, by Milestone Year: Achievable Potential Scenarios vs. Reference Case and Economic Potential
Annual Consumption (PJ/yr) All Sectors
Achievable Scenario Base Year Reference Case Economic
Potential 1 2
2000 5176 5176 5176 5176 2005 5335 5335 5335 5335 2010 5567 5150 5512 5441 2015 5829 5215 5719 5548 2020 6082 5315 5935 5627 2025 6389 5471 6207 5742
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 19
Exhibit 3.8: Energy Savings by Milestone Year: Achievable Potential Scenarios vs. Reference Case and Economic Potential
Annual Savings (PJ/yr) Savings as Percentage of Reference Case Demand Achievable Potential Achievable Potential Year Economic
Potential Scenario 1 Scenario 2 Economic Potential Scenario 1 Scenario 2
2010 417 55 125 7.49% 0.99% 2.25% 2015 614 110 281 10.53% 1.88% 4.82% 2020 768 147 455 12.62% 2.42% 7.49% 2025 918 182 647 14.37% 2.85% 10.13%
Exhibit 3.9: Comparison of Achievable Potential Scenario 2 Savings and 19902003 Energy Demand Growth
0
200
400
600
800
1000
1200
National growth in energy use from 19902003
Achievable 2 scenario savings
Energy
Dem
and (PJ)
3.3.3 Sector Contributions to Savings Potential
Exhibit 3.10 illustrates how the distribution of the DSM potential in 2025, among the three sectors, changes according to each of the achievable potential scenarios. It’s evident that industry’s share of the total energy demand reduction declines substantially as we move from the DSM Status Quo to the DSM Aggressive scenarios. Conversely, the share of this saving attributed to the residential and commercial sectors grows considerably; together these sectors represent 75% and 92% respectively, of the scenario 1 and scenario 2 energy demand reduction. This pattern is driven by the fuel substitution effects that occur as we move into an advanced, more complex policy mix. It also reflects differences in how the policies are targetted towards different sectors (for instance, the standards in scenario 2 are primarily directed at the residential and commercial sectors).
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 20
Exhibit 3.10: National Achievable Potential by Sector Share of Energy Reduction in 2025: Scenarios 1 and 2
Achievable 1
Residential 41%
Commercial/ institutional
34%
Industrial 25%
Achievable 2
Residential 53%
Commercial/ institutional
39%
Industrial 8%
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 21
3.3.4 Savings by Fuel: Achievable Potential
Exhibit 3.11 presents the distribution of the total achievable potential energy demand reduction in 2025 according to the types of fuel. The results show how different policy mixes can affect the energy demand reduction by fuel as the results are markedly different between the two achievable potential scenarios. For scenario 1 the largest energy demand reduction impact in 2025 is achieved in secondary natural gas enduses, representing 49% of the total savings, followed by electricity energy demand reduction , at 34% of the total. The results are largely reversed under scenario 2 where the largest energy demand reduction impact in 2025 is achieved in electricity reduction, representing 55% of the total energy demand reduction. The main driver contributing to this result is the considerable increase in cogeneration in the DSM Aggressive scenario. About 30% of the electricity reduction impact is due to added cogeneration supply.
Exhibit 3.12 summarizes the amount of additional electricity that is induced by the policies simulated in the two scenarios. As shown, the incremental cogeneration output ranges from 9.2 PJ to 61.7 PJ (2.6 TWh to 17.1 TWh). The upper value is equivalent to nearly 40% of the installed cogeneration capacity in Canada in 2003. 18 It is also about 40% of the economic cogeneration potential.
While more than 95% of the current installed cogeneration capacity is in the industrial sector, the commercial sector offers the highest potential for incremental cogeneration, in the range of 31% to 40% of the total for the two scenarios.
18 Mark Jaccard and Associates, Strategic Options for Combined Heat and Power in Canada, For Natural Resources Canada, August 2004, p.40. The installed capacity in 2003 was 6.8 TWe. Assuming an average capacity factor of 70% and an average heattopower ratio of 2.5, the amount of electricity currently produced is approximately 40 TWh and the amount of thermal energy produced is approximately 100 TWh per year. This amounts to approximately 6% of total electricity generation in Canada in 2003.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 22
Exhibit 3.11: All Sector Savings According to Fuel
Achievable 1
Electricity 34%
Natural Gas 49%
RPP 13%
Other 4%
Note: ‘Other’ includes: Coal, Petroleum Coke, Waste Fuels, Off gases, Wood Waste and Spent Pulping Liquor ‘RPP’ is Refined Petroleum Products
Achievable 2
Electricity 55% Natural Gas
22%
RPP 20%
Other 3%
Note: ‘Other’ includes: Coal, Petroleum Coke, Waste Fuels, Off gases, Wood Waste and Spent Pulping Liquor ‘RPP’ is Refined Petroleum Products
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 23
Exhibit 3.12: Added Cogeneration Generation by Sector
Achievable Potential scenario 1 scenario 2
Economic Potential
Additional Electricity Generated (TWh/year) Total 2.56 17.14 54.58 Residential 1.35 6.89 26.14 Commercial 0.42 3.43 6.17 Industrial 0.79 6.83 22.27 Additional Electricity Generated (PJ/year) Total 9.23 61.71 196.50 Residential 4.86 24.80 94.12 Commercial 1.52 12.33 22.20 Industrial 2.85 24.59 80.18
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 24
4. DISCUSSION
As noted, the results identify an achievable potential of between 2.9% to 10.1% range in DSM potential relative to the reference case forecast. The following discussion examines some of the dynamics affecting the outcomes and attempts to place the results in the context of findings from other studies.
Impact on Industry
In 2003 the industrial sector represented the largest percentage of Canada’s secondary energy consumption, 38% of the total (including transportation). Nevertheless, the achievable potential analysis reduction in energy demand for industry is considerably less than that of the residential and commercial sectors. On the surface, it would appear that the energy efficiency performance gains in the residential and commercial sector are not attainable in industry. That would be a misleading conclusion because, as shown in the last part of this section, other studies that focus solely on energy efficiency have shown significant economic and achievable potential in industry. Rather, it is important to understand that the modeling construct and dynamics of this study provide some insight into how a particular mix of policy instruments might affect industry, but in a more dynamic, less linear fashion than shown in some of the other energy efficiency studies.
We have seen from the analysis that, in a dynamic integrated modeling construct, industry could chose fuel substitution and cogeneration investments as alternative investments to energy efficiency or which could offset some of the energy efficiency gains. The key factors influencing the outcomes of the industry achievable potential results are:
Scenario 1 was largely driven by subsidies. It appears that, relative to the dynamics of the residential and commercial sectors, the reduced paybacks induced by the subsidies do not have the same effect for industry in addressing the gap between the social and private discount rates. This may be due to the typically higher hurdle rates that industry demands for energy efficiency investments.
In scenario 2, the application of standards in industry was limited and did not play the same role as building and enduse equipment standards do in the commercial/institutional and residential sectors. In addition, the renewable energy subsidies and the changes in building types (to mimic urban land use policies) had a far less application to industry than the other sectors.
In scenario 2, the marginal cost pricing instrument has a considerable effect on industry energy use dynamics as electricity prices increase relative to other fuels. This drives additional fuel switching for enduses where these fuels can be substitutes, particularly combustion. The enduse efficiency of electric heating is always higher than for direct combustion of fuels, resulting in additional secondary energy demand for the affected enduses. This drop in performance is particularly evident where there is a considerable switch to the utilization of wood waste in industry; wood waste use increases fairly significantly in scenario 2.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 25
Finally, it is also important to note that the study did not examine the energy management potential in the upstream oil and gas sector, which is an energy intensive and growing sector of the economy.
Why the DSM Aggressive Scenario Has a Large Fuel Substitution/Cogeneration Effect
CIMS simulates the competition of DSM technologies of different levels of efficiency and fuel type to meet a given energy service demand. The choice pathway has four options: i) choose a more efficient upgrade within the same fuel type, ii) choose a more efficient upgrade using a different fuel, iii) choose a base technology with a different fuel, iv) make no upgrade or change in fuel choice. The DSM technology competition is largely driven in the policies, but not exclusively so, by two main factors: changes in capital costs or changes to energy prices. Consequently, as the mix of policies assessed in CIMS varies, so does the impact weighted between energy efficiency and fuel substitution.
Therefore, we see that the policy mix in the DSM Aggressive scenario results in significant fuel switching and cogeneration. In particular, marginal cost electricity pricing and the carbon liability in scenario 2 affect different fuels unevenly. In response to the carbon liability, there is fuel switching to less carbon intense fuels (away from coal and oil), while marginal cost pricing encourages fuel switching away from electricity. Together, these two policy instruments contribute to a greater takeup of wood waste (‘hog fuels’) in the pulp and paper and wood and allied products sectors.
Similarly, the policies simulated in the DSM Aggressive scenario bolster the economic conditions for cogeneration, which has significant impact on the results. Marginal cost pricing for electricity, in particular, increases the differential between gas prices and electricity prices – which is critical to cogeneration development. It’s important not to let the current pricing conditions cast a shadow over the projected outcome in 2025. At the present time, high natural gas prices are making natural gas driven cogeneration less economic because they are reducing the “spark spread”, i.e. the cost differential between natural gas and electricity, so that selfgeneration becomes less cost effective. However, the simulation of the achievable potential includes policies that favourably influence the economics of cogeneration – marginal cost pricing for electricity in particular increases the differential between gas prices and electricity prices – which is critical to cogeneration development
Why the Sectoral Contribution Changes
We have seen an enormous shift in the allocation of the total savings between the DSM Status Quo and DSM Aggressive scenarios. While the magnitude of the overall savings increased, the share of energy savings that is attributed to industry is significantly smaller. This has occurred because the mix of policies simulated in the DSM Aggressive scenario: i) induce a greater degree of fuel substitution and cogeneration as noted above, and ii) in terms of enduse efficiency, are more conducive to performance improvements in the buildings sectors.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 26
In scenario 2, the application of standards in industry has limited application and cannot play the same role as building and appliance standards do in the commercial/institutional and residential sectors. The effect of the renewable subsidies are similar less pronounced in industry.
What CIMS Did Not Model
The achievable and economic potential scenarios were not run utilizing the CIMS’ energy price and macroeconomic feedback systems. This level of analysis was beyond the scope of the study. If CIMS had been allowed to iterate between the energy demand and supply sectors, we would have seen the impacts of reduced consumption of electricity on its cost of production, and hence its price. In turn, if the price change had been significant, the energydemanding residential, commercial and industrial models would have been rerun until a new energy supply and demand equilibrium was achieved. We speculate that if these macroeconomic feedbacks had been run, increased production costs in industry might have caused increased final prices and lower production demands, particularly for the scenario 2 policy mix. Ultimately, these dynamics would lead to additional secondary effects in the residential and commercial sectors.
Transportation Benefits: Location Efficiency
In scenario 2, we touched briefly on the possible energy reduction effect of advanced urban land use policies. This was modeled by changing dwelling shares running into the future, to reflect increased urban densities. There is a possible transportation dividend to be reaped from such policies. Numerous studies have been completed in the past 15 years on the energy and lifestyle cost savings of dense urban areas relative to sprawling urban areas – thus termed “location efficiency”.
Research has consistently shown savings of 20%40% in urban transportation energy as urban density doubles. For instance, if policy makers targeted a density of 10 people/hectare in 2030 Canadian urban areas, which would be a 43% increase in urban density compared with current patterns, this could result in a 10%20% annual reduction in urban transportation energy consumption. To put this into perspective, a 10%20% annual savings applied in 2003 in Canada’s urban areas would save roughly 100PJ200 PJ annually in passenger transportation alone.
System savings
The projected savings in electricity demand have been calculated at the customer level. However, these savings have a significant impact on capacity requirements to meet the demand. A unit of electric demand reduction is worth more (12% to 30% more depending on the generation mix) than a unit of additional supply in terms of generation capacity. 19
19 To meet electricity demand, you need to have a generation capacity that exceeds your demand by a minimum of around 12% for hydro generators to around 30% for coalfired generators to handle routine maintenance and down time of equipment.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 27
Comparison to other studies
There have been many North American studies in the past five years investigating the achievable potential for energy efficiency in various sectors. Comparisons of study outputs from one study to another are always difficult because of often different analytical constructs, modelling approaches, data sets and assumptions. Nevertheless, the comparisons provide another source of estimates to consider, and an indication of how the current study relates to other efficiency potential studies that have been undertaken.
The two achievable potential scenarios generate energy demand reduction reductions ranging from 2.9% to 10.3% in 2025, relative to the reference case forecast. These results are of the same order of magnitude generated by a 2005 U.S. study investigating energy efficiency potential in all sectors which produced a energy demand reduction range of 4% to 9%, also for two scenarios running to 2025. 20 The U.S. study used the National Energy Modelling System (NEMS) and considered a wide range of policy instruments, and like the current study, assessed the potential by directly representing policies in an energyeconomy model.
Exhibit 3.13 compares the range of achievable potential energy demand reduction from some recent demandside management (DSM) studies conducted in Canada, distinguished according to sector and fuel.
The comparison shows that, with the exception of the industrial results, the upper bound (the scenario 2 results) exceeds the upper bound of these recent DSM studies. Indeed, it is clear that the CIMS industrial results act to offset the performance from the other sectors when the overall reduction in demand is considered.
In interpreting this difference, it is important to bear in mind that scenario 2 as defined in this project includes price and regulatory instruments that extend beyond the scope of current utility programs. The analysis also incorporates landuse measures, cogeneration and renewables, and includes the interactive effects of the policies, including their impact on fuel switching.
Industry shows a lower potential for several reasons. First, the regulatory, land use and renewable subsidy policies are largely targetted to the residential and commercial sectors. Second, fuel switching to gas and the additional natural gas required to cogenerate (the cogeneration effect) simply outweighs the gains in energy efficiency gains in industry. Although there is fuel substitution and cogeneration in the other sectors, the other elements in the scenario induce significantly more efficiency over the long run.
20 Energy Information Administration, Office of Integrated Analysis and Forecasting, Assessment of Selected Energy Efficiency Policies, May 2005 U.S. Department of Energy Washington, DC 20585.
Demand Side Management Potential in Canada: Energy Efficiency Study –Summary Report–
Marbek/MKJA Page 28
Exhibit 3.13: Achievable Potential Performance Range From Recent DSM Studies
Savings Range Lower % Upper % Sector and Fuel
Other studies CIMS analysis
Other studies CIMS analysis
Residential Electricity 3 4.4 7.514 27 Natural Gas 2 5.6 37 11.8
CommercialElectricity 3 4.4 35 22.8 Natural Gas 3 3.5 610 11.5
Industrial Electricity 2 2.9 1525 14.3 Natural Gas 3 3.3 710 2.7