Multidisciplinary System Design Optimization (MSDO) · Multidisciplinary System Design Optimization...
Transcript of Multidisciplinary System Design Optimization (MSDO) · Multidisciplinary System Design Optimization...
1 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Design for Value
Lecture 19
Karen Willcox
Olivier de Weck
Multidisciplinary System
Design Optimization (MSDO)
Select figures courtesy of Jacob Markish and Ryan Peoples. Used with permission.
2 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Today’s Topics
• An MDO value framework
• Lifecycle cost models
• Value metrics & valuation techniques
• Value-based MDO
• Aircraft example
• Spacecraft Example
3 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Optimal Design
• Traditionally, design has focused on performance
e.g. for aircraft design
optimal = minimum weight
• Increasingly, cost becomes important
• 85% of total lifecycle cost is locked in by the end of
preliminary design.
• But minimum weight minimum cost maximum value
• What is an appropriate value metric?
4 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Design Example• We need to design a particular portion of the wing
• Traditional approach: balance the aero & structural requirements,
minimize weight
• We should consider cost: what about an option that is very cheap to
manufacture but performance is worse?
aerodynamics?
• How do we trade performance and cost?
• How much performance are we willing to give up for $100 saved?
• What is the impact of the low-cost design on price and demand of
this aircraft?
• What is the impact of this design decision on the other aircraft I
build?
• What about market uncertainty?
structural dynamics?
manufacturing cost?
aircraft demand?
aircraft price?tooling?
environmental impact?
5 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost
Module
“Value” metricPerformance
Module
Aerodynamics
Structures
Weights
Mission
Stability & Control
Revenue
Module
Value Optimization Framework
Manufacturing
Tooling
Design
Operation
Market factors
Fleet parameters
Competition
“Optimal”
design
6 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Challenges
• Cost and revenue are difficult to model
– often models are based on empirical data
– how to predict for new designs
• Uncertainty of market
• Long program length
• Time value of money
• Valuing flexibility
• Performance/financial groups even more uncoupled
than engineering disciplines
7 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost Model
Need to model the lifecycle
cost of the system.
Life cycle :
Design - Manufacture -
Operation - Disposal
Lifecycle cost :
Total cost of program over
life cycle
85% of Total LCC is locked
in by the end of preliminary
design.
Cost
Module
“Value” metricPerformance
Module
Revenue
Module
8 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Lifecycle Cost
0
20
40
60
80
100
65%
Co
ncep
tua
l
desig
n
Pre
lim
inary
desig
n,
syste
m i
nte
gra
tio
n
Deta
iled
de
sig
n
Man
ufa
ctu
rin
g
an
d a
cq
uis
itio
n
Op
era
tio
n
an
d s
up
po
rt
Dis
po
sal
Time
Impact on L
CC
(%
)
85%
95%
9 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Non-Recurring Cost
Cost incurred one time only:
Engineering
- airframe design/analysis
- configuration control
- systems engineering
Tooling
- design of tools and fixtures
- fabrication of tools and fixtures
Other
- development support
- flight testing
Engin
eeri
ng
Toolin
gO
ther
10 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Development Cost Model
• Cashflow profiles based on beta curve:
• Typical development time ~6 years
• Learning effects captured – span, cost
11 )1()( tKttc
0
0.01
0.02
0.04
0.05
0.06
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
normalized time
Support
Tool Fab
Tool Design
ME
Engineering
no
rmalized
co
st
(from Markish)
11 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Recurring Cost
Cost incurred per unit:
Labor
- fabrication
- assembly
- integration
Material to manufacture
- raw material
- purchased outside production
- purchased equipment
Production support
- QA
- production tooling support
- engineering support
Labor
Mate
rial
Support
12 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Learning Curve
As more units are made, the recurring cost per
unit decreases.
This is the learning curve effect.
e.g. Fabrication is done more quickly, less
material is wasted.
n
x xYY 0
Yx = number of hours to produce unit x
n = log b/log 2
b = learning curve factor (~80-100%)
13 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Learning Curve
0.55
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50
Unit number
Co
st
of
un
it
b=0.9
Typical LC slopes: Fab 90%, Assembly 75%, Material 98%
Every time
production
doubles,
cost is
reduced by
a factor of
0.9
14 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
CASH AIRPLANE RELATED
OPERATING COSTS:
Crew
Fuel
MaintenanceLanding
Ground Handling
GPE Depreciation
GPE Maintenance
Control & Communications
Airplane Related Operating Costs
CAROC is only 60% - ownership costs are significant!
CAROC
60%40%
Capital
Costs
CAPITAL COSTS:
Financing
Insurance
Depreciation
15 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Value Metric
Need to provide a
quantitative metric that
incorporates cost,
performance and
revenue information.
In optimization, need to
be especially carefully
about what metric we
choose...
Cost
Module
“Value” metricPerformance
Module
Revenue
Module
16 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
What is Value?
• Objective function could be different for each stakeholder
e.g. manufacturer vs. airline vs. flying public
• Program related parameters vs. technical parameters
cost, price, production quantity, timing
• Traditionally program-related design uncoupled from
technical design
Customer
Value
Shareholder
Value
Product
Quality
Schedule
Cost
Economic
Value
Added
Demand
Revenue
EBIT
System
Design
Price
From Markish,
Fig. 1, pg 20
Customer value derived from
quality, timeliness, price.
Shareholder value derived
from cost and revenue, which
is directly related to customer
satisfaction.
17 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Value Metrics
performance
weight
speed
Traditional Metrics
cost
revenue
profit
quietness
emissions
commonality
...
Augmented Metrics
The definition of value will vary depending on your system
and your role as a stakeholder, but we must define a
quantifiable metric.
18 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Valuation Techniques
Investor questions:
• How much will I need to invest?
• How much will I get back?
• When will I get my money back?
• How much is this going to cost me?
• How are you handling risk & uncertainty?
Investment Criteria
• Net present value
• Payback
• Discounted payback
• Internal rate of return
• Return on investment
19 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Net Present Value (NPV)
• Measure of present value of various cash flows in different
periods in the future
• Cash flow in any given period discounted by the value of a
dollar today at that point in the future
– “Time is money”
– A dollar tomorrow is worth less today since if properly
invested, a dollar today would be worth more tomorrow
• Rate at which future cash flows are discounted is
determined by the “discount rate” or “hurdle rate”
– Discount rate is equal to the amount of interest the
investor could earn in a single time period (usually a
year) if s/he were to invest in a “safer” investment
20 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Discounted Cash Flow (DCF)
• Forecast the cash flows, C0, C1, ..., CT of the project
over its economic life
– Treat investments as negative cash flow
• Determine the appropriate opportunity cost of capital
(i.e. determine the discount rate r)
• Use opportunity cost of capital to discount the future
cash flow of the project
• Sum the discounted cash flows to get the net present
value (NPV)
NPV C0
C1
1 r
C2
1 r2
CT
1 rT
21 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
DCF example
Period Discount Factor Cash Flow Present Value
0 1 -150,000 -150,000
1 0.935 -100,000 -93,500
2 0.873 +300000 +261,000
Discount rate = 7% NPV = $18,400
22 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Risk-Adjusted Discount Rate
• DCF analysis assumes a fixed schedule of cash flows
• What about uncertainty?
• Common approach: use a risk-adjusted discount rate
• The discount rate is often used to reflect the risk
associated with a project: the riskier the project, use a
higher discount rate
• Typical discount rates for commercial aircraft programs:
12-20%
• Issues with this approach?
23 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Net Present Value (NPV)
0 (1 )
Tt
tt
CNPV
r
-1500
-1000
-500
0
500
1000
1500
1 3 5 7 9
11
13
15
17
19
21
23
25
27
29
Cashflow
DCF (r=12%)
Program Time, t [yrs]
Cashflow
, P
t[$
]
24 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Payback Period
• How long it takes before entire initial investment is
recovered through revenue
• Insensitive to time value of money, i.e. no
discounting
• Gives equal weight to cash flows before cut-off date
& no weight to cash flows after cut-off date
• Cannot distinguish between projects with different
NPV
• Difficult to decide on appropriate cut-off date
25 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Discounted payback
• Payback criterion modified to account for the time
value of money
– Cash flows before cut-off date are discounted
• Overcomes objection that equal weight is given to
all flows before cut-off date
• Cash flows after cut-off date still not given any
weight
26 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Internal rate of return (IRR)
• Investment criterion is “rate of return must be greater than the opportunity cost of capital”
• Internal rate of return is equal to the discount rate for which the NPV is equal to zero
• IRR solution is not unique
– Multiple rates of return for same project
• IRR doesn’t always correlate with NPV
– NPV does not always decrease as discount rate increases
NPV C0
C1
1 IRR
C2
1 IRR2
CT
1 IRRT
0
27 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Return on Investment (ROI)
• Return of an action divided by the cost
of that action
• Need to decide whether to use actual or
discounted cashflows
revenue cost
costROI
28 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Decision Tree Analysis (DTA)
• NPV analysis with different future scenarios
• Weighted by probability of event occurring
29 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Real Options Valuation Approach
In reality:
Cashflows are uncertain
Ability to make decisions as future unfolds
View an aircraft program as a series of investment decisions
Spending money on development today gives the option to build and sell aircraft at a later date
Better valuation metric: expected NPV from dynamic programming algorithm (Markish, 2002)
30 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Dynamic Programming:
Problem Formulation
• The firm:
– Portfolio of designs
– Sequential development phases
– Decision making
• The market:
– Sale price is steady
– Quantity demanded is unpredictable
– Units built = units demanded
• Problem objective:
– Which aircraft to design?
– Which aircraft to produce?
– When?
Dynamic Programming:Dynamic Programming:Problem ElementsProblem ElementsProblem ElementsProblem Elements
1. State variables stt
2. Control variables ut
3. Randomness
4 Profit function4. Profit function
5. Dynamics
• Solution: [ ]⎭⎬⎫
⎩⎨⎧
++= ++ )(
11),(max)( 11 ttttttutt sFE
russF
t
π
• Solve recursively
31 © Massachusetts Institute of Technology - Prof. de Weck and Prof. WillcoxEngineering Systems Division and Dept. of Aeronautics and Astronautics
32 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Dynamic Programming:
Operating Modes
How to model decision making?
33 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
BWB-450
BWB-250C
Fuselage bays
Inner wing
Outer wing
Example: BWB
• Blended-Wing-Body (BWB):
– Proposed new jet transport
concept
• 250-seat, long range
• Part of a larger family sharing common centerbody bays, wings, ...
34 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
BWB Example: Simulation Run
-4,000,000
-3,000,000
-2,000,000
-1,000,000
0
1,000,000
2,000,000
3,000,000
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
time (years)
cash
flo
w (
$K
)
0123456789
10111213141516
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
time (years)
op
era
tin
g m
od
e
0
20
40
60
80
100
120
qu
an
tity
dem
an
ded
per
yea
r
mode
demand
35 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
BWB Example:
Importance of Flexibility
-10
-5
0
5
10
15
20
25
3 5 7 11 18 28 44 69 108 171 270
initial annual demand forecast
pro
gra
m v
alu
e (
$B
)
dynamic programming
deterministic NPV
At baseline of 28 aircraft, DP value is $2.26B versus deterministic value of $325M
36 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Traditional Design Optimization
Objective function: usually minimum weight
Design vector: attributes of design, e.g. planform geometry
Performance model: contains several engineering disciplines
Performance
Model
Optimizer
Design
vector x
Objective
function J(x)
37 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Coupled MDO Framework
Performance
Model
Cost
Revenue
Optimizer
Valuation
Market
Price,
Demand
Cost
Design
vector xObjective
function J(x)
Objective function: value
metric, e.g. NPV
Simulation model:
performance and financial
Stochastic element
38 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Value-Based Optimization Results
Boeing BWB case study
475 passengers, 7800 nmi range
Baseline: optimized for minimum GTOW
Outcomes
Comparison of min GTOW and max E[NPV]
designs
Traditional NPV vs. stochastic E[NPV]
Effect of range requirement on program value
Effect of speed requirement on program value
39 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Different Objectives, Different Designs
New objective results in tradeoff:Lower structural weight, lower
cost
Higher fuel burn, lower price
Net result
2.3% improvement in value
Overall design very similarConstrained to satisfy design
requirements
Unable to move dramatically in design space
Minimum-GTOW planform
Maximum-value planform
40 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Deterministic vs. Stochastic Valuation
• Discount rates: 12% and 20%– Computational expense reduced, but NPV results are
negative
– E[NPV] for 12% design = 0.58% decrease
– E[NPV] for 20% design = 3.7% decrease
• High-rd drives design to reduce development costs
• Traditional NPV not
appropriate
– As valuation metric
– As optimization
objective
relative to
max-E[NPV] design
-10
-8
-6
-4
-2
0
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Program Year
Rela
tive C
ash
Flo
w
r_d = 12%
r_d = 20%
41 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Range Specification
• Comparison of min-GTOW and max-E[NPV] design
solutions
• E[NPV] for varying ranges relative to max-E[NPV] design
Design decisions based on understanding of E[NPV] impact
0
0.2
0.4
0.6
0.8
1
1.2
5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 10000
Range (nmi)
Rela
tive E
[NP
V]
max E[NPV]
min GTOW
42 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Speed Specification
• Comparison of E[NPV] and other metrics for varying
speeds
0.7
0.75
0.8
0.85
0.9
0.95
1
1.05
1.1
0.78 0.8 0.82 0.84 0.86 0.88 0.9 0.92
Cruise Mach number
Re
lati
ve
me
tric
va
lue
E[NPV]
M*L/D
M*P/D
M*P*R/GTOW
43 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Lecture Summary
• Designing for value is crucial: for a program to be
successful, we cannot focus exclusively on performance
• The definition of value is flexible, and will vary depending
on the application and on your interest as a stakeholder
• Financial metrics can be used to quantifying value, but
use caution in your choice of value objective function
• Cost and revenue are difficult to model – often use
empirical data
• It is important that uncertainty and risk are handled
appropriately
• There is much work to be done on this issue
44 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
References
Brealey, R. & Myers, S., Principles of Corporate Finance, 7th Edition,
McGraw-Hill, NY 2003
Markish, J., “Valuation Techniques for Commercial Aircraft Program
Design”, Masters Thesis, MIT Dept. of Aeronautics & Astronautics, June
2002.
Markish, J. and Willcox, K., “Value-Based Multidisciplinary Techniques
for Commercial Aircraft System Design”, AIAA Journal, Vol. 41, No. 10,
October 2003, pp. 2004-12.
Peoples, R. and Willcox, K., “Value-Based Multidisciplinary Optimization
for Commercial Aircraft Design and Business Risk Assessment,” Journal
of Aircraft, Vol. 43, No. 4, July-August, 2006, pp. 913-921.
Roskam, J., Airplane Design Part VIII, 1990.
Raymer, D., Aircraft Design: A Conceptual Approach, 3rd edition, 1999.
Schaufele, R., The Elements of Aircraft Preliminary Design, 2000.
45 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
System Cost Modeling
This section will:
- present a process for obtaining system
cost estimates, particularly for space systems
- provide cost-estimating relationships
- describe how to assess uncertainty in cost estimates
- provide a specific example: optical systems cost
dWo, 5-6-2002
46 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost Breakdown Structure (CBS)
Organizational Table that collects costs, covers:
- research, development, test and evaluation (RDT&E)
- production, including learning curve effects
- launch and deployment
- operations
- end-of-life (EOL) disposal
Space
Mission
Architecture
Program
Level Costs
Space
Segment
Launch
Segment
Ground
Segment
Operations
and Support
• Management
• Systems Eng
• Integration
• Payload
• Spacecraft
• Software
• “Systems”
• Launch Vhc
• Launch Ops
• S/C-L/V
integration
• Facilities
• Equipment
• Software
• etc
• Personnel
• Training
• Maintenance
• Spares
RDT&E
ProductionOperations
47 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost Estimating Methods
Basic techniques to develop Cost Models:
(1) Detailed bottom-up estimating
- identify and specify lower level elements
- estimated cost of system is of these
- time consuming, not appropriate early, accurate
(2) Analogous Estimating
- look at similar item/system as a baseline
- adjust to account for different size and complexity
- can be applied at different levels
(3) Parametric Estimating
- uses Cost Estimation Relationships (CER’s)
- needed to find theoretical first unit (TFU) cost
48 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Parametric Cost Models
Are most appropriate for trade studies:
Advantages:• less time consuming than traditional bottom-up estimates
• more effective in performing cost trades
• more consistent estimates
• traceable to specific class of (space) systems
Major Limitations:• applicable only to parametric range of historical data
• lacking new technology factors, adjust CER to account for new technology
• composed of different mix of “things” in element to be costed
• usually not accurate enough for a proposal bid
49 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Process for developing CER’s
Subsystem A
Subsystem B
Subsystem N….
Cost/parametric data
Constant Year Costs
Regression
Analysis
Preferred Form
(Cost Model Assumption)
Subsystem A
Subsystem B
Subsystem N….
$
Weight - kg
$ bAW
Computer
Software
Step 1
Develop Database
File
Step 2
Apply Regression
Analysis
Step 3
Obtain CER’s and
Error Statistics
Key statistics: R2,
Standard Error: RMS
50 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Adjustment to constant-year dollars
It is critical that cost estimated be based on a
constant-year dollar bases. Reason: INFLATION
Y Y NC R C
E.g. All costs are adjusted to FY92 (“Fiscal Year 1992”)
Past Years
Future Years
Use actual inflation numbers
Use forecasted inflation numbers
e.g. 3.1% yearly inflation in U.S.
See Table 20-1 on handout
92 93 94
1.040 1.037 1.034 1.115
FY FY FY
R Convert Oct-1991 cost
to Oct-1994 costs
1N
RATER i$ 1M in FY 1980 corresponds to
$ 2.948M in FY 2005 (projected)
51 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Case: Modular Optics Cost Modeling
• Investigate economical viability of modular
optics given performance constraints
• Focus on monolithic Cassegrain telescopes
versus Golay-3 design
• Use real data and experience from ARGOS
k
Lk
B
Dk
lk b
dk
sub-telescope plane
(input pupil)
combiner plane
(exit pupil)focal plane
(image pupil)
object
k-th aperture
beam
combiner
relay
optics
52 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Literature Search
Kahan, Targrove, “Cost modeling of large
spaceborne optical systems”, SPIE, Kona, 1998
Humphries, Reddish, Walshaw,”Cost scaling laws
and their origin: design strategy for an optical array
telescope”, IAU, 1984
Meinel, “Cost-scaling laws applicable to very large
optical telescopes”, SPIE, 1979
Meinel’s law: 2.580.37 [M$] (1980)S D
53 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost Modeling Approach
Monolithic System Modular Golay System
(A)Sub-telescope
Total Cost: total = CA + CB +CC
(B)
Relay optics and
combiner
(C)
Detector
Does not include
development cost
none
54 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost Estimation Relationships
A AC L D
Next Step: Need to determine coefficients based
on available commercial pricing data
(A) Optical Telescope Assy:
(B) Relay & Combining Optics:
(C) Detector: CCD:
Modular array telescope:
TFU
1 log 100% / / log 2
B
A AC L D N
B S
c c pixC L n
(includes learning curve)
Use ARGOS cost database
55 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Small Amateur Telescopes (I)
0 10 20 30 40 50 60 70 800
2000
4000
6000
8000
10000
12000
14000Telescope Comparison
Diameter Size[cm]
Price
[U
S$
]
DHQ f/5
DHQ f/4.5
D Truss f/5
Obsession f/4.5
Celestron G-f/10
56 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Small Amateur Telescopes (II)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3x 10
4
Telescope Diameter D [m]Te
lesco
pe P
urc
ha
se
Co
st
C [
20
01
$] Telescope Cost CER (Aperture): C=28917*D
2.76
1 2 3 4 5 6 7 8 9 100
1
2
3
4
5
6
7x 10
Power Coefficient
Cum
ula
tive
Weig
hte
d F
ittin
g E
rro
r
Telescope CER fitting8
Minimum yields best CER fit
exponent =2.76
57 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Professional Telescope OTA cost
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
3.5
4x 10
5
Aperture Diameter D [m]
OT
A C
ost [2
00
1 $
]
Ritchey-Chretien
Classical Cassegrain
Company: Optical Guidance Systems
(http:www.opticalguidancesystems.com)
CERs for
Ritchey-Chretien
Classical Cassegrain
Remarkable Result:
virtually identical
power law across
completely different
product lines.
2.80376000RCC D
2.75322840CCC D
58 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Relay and Combining Optics
Cost of relay optics depends on:
• number and type of actuators: ODL, FSM, active
translation stages, active pyramid mirrors
• sensing elements (CCD, quad cells etc)
• aperture magnification ma
• Quality requirements (RMS WFE)
As interim solution use ARGOS cost database:
Passive Optics: Collimators, Combiner, etc
Active Optics: FSM, Fold Mirrors etc...
ma = 10
$ 32,874.-
$ 18,859.-
About: $ 50,000.-
59 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
CCD Cost Models
0 500 1000 1500 2000 2500 3000 3500 4000 4500 50000
0.5
1
1.5
2
2.5
3x 10
4 Polynomial Fit of Compiled Data
Square Root of Numbers of Pixels
Price [
US
$]
Least Squares Polynomial Fit
Compiled Data
1.23
0.68CCD pixC n
It appears that CCD cost also depends on
factors other than raw pixel count. Need to investigate.
AP-9 (KAF-6303E)
60 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Break-Even Analysis
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.5
1
1.5
2
2.5
3
3.5
4x 10
5
Equivalent Aperture D [m]
Op
tical S
yste
ms C
ost [2
00
1 $
]Break Even Cost Analysis
Monolithic
Golay-3
ARGOS
Preliminary
Analysis
suggest
crossover
around 0.7 m
Assumptions:
N=3
npix = 2048
Relay optics
and combiner
costs fixed
MIT OpenCourseWarehttp://ocw.mit.edu
ESD.77 / 16.888 Multidisciplinary System Design Optimization
Spring 2010
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.