Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive...
Transcript of Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive...
![Page 1: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/1.jpg)
© 2017 Heroes
Werner Vink | Engineer IoT
Case study: Integrating IoT in your
maintenance practice
Heroes – Think Digital
![Page 2: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/2.jpg)
© 2017 Heroes
Mission
Don’t add Digital. Become Digital.
Think Digital
Vision
Coach companies to become Digital
winners. Scout and Build smart digital
solutions.
2
![Page 3: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/3.jpg)
© 2017 Heroes
Market Prospects 2025
Internet of Things
LogisticsPrecise location
services for track
and trace
DevicesManagement and
control applications
RetailInsights about
shopping behaviour
HealthcareRemote monitoring
of patients health
MachineryConnected machines
for real-time
monitoring
Building
ManagementCreating smart
buildings/cities
4 BILLIONConnected people
$4 TRILLIONYearly revenue
opportunities
25+ MILLIONApps
25+ BILLIONEmbedded and
intelligent systems
25+ TRILLIONGB’s of data
3
*Source: Mckinsey 2016
![Page 4: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/4.jpg)
© 2017 Heroes
Where is the value potential?
Internet of Things
4
*Source: Mckinsey 2016
Interoperability required to
capture 40% of total value
< 1 % of data currently used,
mostly for alarms or real-time
control; more can be used for
optimization and prediction
2 x more value from B2B
applications than consumer
• Developing world: 40%
• Developed world: 60%
![Page 5: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/5.jpg)
© 2017 Heroes
B2B Opportunities
Internet of Things
5
Quality control
• e.g. reducing physical inspections
Asset tracking
• e.g. alarms using geofencing
Supply chain
• e.g. schedule tweaking
Use cases that will enhance the process of management in
regard to the total cost of ownership
Enhancing efficiency
• e.g. optimizing use of energy
Predicting and avoiding
• e.g. reducing downtime risk
Inventory management
• e.g. real-time status
![Page 6: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/6.jpg)
© 2017 Heroes
Technology Value Chain
Internet of Things
6
Devices Gateways Ingestion Automate BI & DC Report & Act
Event Producers & Gateways Ingestion & Transformation Report, Predict, Act
Event
Cloud
functions
![Page 7: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/7.jpg)
© 2017 Heroes 7
Internet of Things
A world connected
Use Case #1
Cutting out operational inefficiencies
Predictive Maintenance
![Page 8: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/8.jpg)
© 2017 Heroes
Predictive MaintenanceInternet of Things
8
• Sustainable System Maintenance
• Enabling machine learning to predict system behaviour
• Stimulates data-driven decision-making
• Improve asset lifespan and utilization
• Value Proposition
• Improve maintenance effectiveness
• Increasing system operational efficiency
• More flexibility in maintenance planning
![Page 9: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/9.jpg)
© 2017 Heroes
Predictive MaintenanceInternet of Things
9
• Replacing filter in air handling unit (AHU)
• Step 1: What is the preferred moment of replacement?
• Step 2: Which data is available that says something about the filter?
• Step 3: Which theoretical background has this data?
• Step 4: From the data, can a model be obtained that predicts filter behaviour?
![Page 10: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/10.jpg)
© 2017 Heroes
Predictive MaintenanceInternet of Things
10
• Step 1: What is the preferred moment of replacement?
• According manufacturer 10.920 hours of operations (+/- 65 weeks)Preferred moment
of replacement
![Page 11: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/11.jpg)
© 2017 Heroes
Predictive MaintenanceInternet of Things
11
• Step 2: Which data is available that says something about the filter?
First period
(26 weeks)
Second period
(23 week)
![Page 12: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/12.jpg)
© 2017 Heroes
Predictive MaintenanceInternet of Things
12
• Step 3: Which theoretical background has this data?
• At a constant debit and volume the pollution of a filter can be described according the law of Boucher’s
log(P/Pi) = -J*VL
P [Pa] = pressure over filter on time x
Pi [Pa] = initial pressure when clean
J [-] = Boucher’s ratio of filter pore blocking
VL [m3] = Volume of air passing
Variable to be
predict
P(t) = c0 + c1.eJ.V.t
C0 [Pa] = initial pressure of the clean filter clean
C1 [-] = Constant to determine
J [-] = Boucher’s ratio of filter pore blocking
![Page 13: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/13.jpg)
© 2017 Heroes
Predictive MaintenanceInternet of Things
13
• Step 4: Can a model be obtained from the data to predict filter behaviour?
• Through 2th order polynomial regression an average model is obtained over both data periods.
• First period (blue line):
R-square = 90%
• Second period (red line):
R-square = 85%
• Average regression function (black line):
P(t) = Pi + 2e0,08t
Prediction: +/- 8.900 hours (53 weeks)
250 Pa is reached.
![Page 14: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/14.jpg)
© 2017 Heroes
Predictive MaintenanceInternet of Things
14
• Conclusions
• Filter needs to be replaced around 250Pa.
• Current filter replacement at +/- 4.000 hours (80-90 Pa)
• Model predicts 250Pa is reached at +/- 8.900 hours of operation.
• Recommendations
• Dataset contains P < 100Pa. Extending the dataset with P > 100 Pa gives
outcome in obtaining a more accurate model
• Current replacement cycle can be optimized from 2 times to 1 time per year
• Prediction gives outcome to increase flexibility in maintenance planning
![Page 15: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/15.jpg)
© 2017 Heroes 15
Internet of Things
A world connected
Use Case #2
Making your data future proof
Semantic Data Modelling
![Page 16: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/16.jpg)
© 2017 Heroes
Semantic Data ModellingInternet of Things
16
• Workload in a typical IoT project
• 60% time spend on data extraction and understanding
• 30% time spend on performing value added analytics
• Value Proposition
• Bringing down the 60% spend on data extraction and understanding
• Focus on what matters: creating valuable analytics
![Page 17: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/17.jpg)
© 2017 Heroes
Semantic Data ModellingInternet of Things
17
RMT_SPZone_1_temperature
MainMeter
AHU-1_SA
DAT
Bld321_Elec_Meter
RHT_765SUPTEMPRTU_5-HeatStage1
OAT
OSA_temp
regelafsl2
Aanvoertemp gkw 20
Vallei OS1 GRFMET 5
3TS02ITAS
Temp_034
PT03-HP
• Data from control systems and IoT devices
• Lack of uniformity in naming the data
• Lack on control system documentation
![Page 18: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/18.jpg)
© 2017 Heroes
Semantic Data ModellingInternet of Things
18
• At Heroes we work according Haystack
• Open Source initiative from Virginia, USA (2014)
• Fast growing world wide community
• Sponsors: Intel, Siemens, KNX, Tridium, Arup
• How does it work?
• Adding extra (meta) data using predefined tags
• Comparable with the # of social media
• Why?
• Shortens the analysis time
• Simplifies the scaling of algorithms
• Broad integration possibilities: i.e. Python, Node.JS, C++, C#, Java
Historical Data
+
=Scalable Analytics
![Page 19: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/19.jpg)
© 2017 Heroes
Semantic Data ModellingInternet of Things
19
• Equips
• Heat pumps
• Heat Exchangers
• Gas-fired boilers
• Chillers
• Etc.
• Points
• Sensors
• Set points
• Control signals
![Page 20: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/20.jpg)
© 2017 Heroes
Semantic Data ModellingInternet of Things
20
Equip tags: ahu, hvac, equip
Point tags:
outside
air
filter
delta
pressure
sensor
point
Point tags:
return / discharge
air
temp
sensor
point
Point tags:
outside
air
temp
sensor
point
Point tags:
heatWheel
cmd
point
Point tags:
return
water
valve
cmd
point
Point tags:
return
water
temp
Sensor
point
![Page 21: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/21.jpg)
© 2017 Heroes
Semantic Data ModellingInternet of Things
21
Total 15 AHUs
216 points
Relevant for 8 AHUs, directly in one overview
Data water valve
Filter > 10%
Filter < 90%
Data heat recovery
Find periods where
both is true
![Page 22: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/22.jpg)
© 2017 Heroes
Semantic Data ModellingInternet of Things
22
• Advantages of Semantic Data Modelling
• Structured database; effective queries
• Decreases labour intensity of analytics
• Eases the scalability of algorithms and analytics
• Advantages of Project Haystack principles
• Predefined tags for different data sources
• Fast growing open source community
• Actively sharing ideas and code
![Page 23: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/23.jpg)
© 2017 Heroes 23
Internet of Things
A world connected
Exploring your business opportunities
Getting started with IoT
![Page 24: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/24.jpg)
© 2017 Heroes
What is your current practice?
Imagine that your remote monitoring could
automatically identify and fix potential
problems
before they happen.
Imagine if you could instantly access data from
facilities anywhere in the world and make mission-
critical decisions more intelligently than ever
before
Remote Monitoring
Predictive Analytics
Connected
Imagine that within your connected facilities
thousand of devices are monitored and no/less
physical inspection is needed.
Internet of Things
24
![Page 25: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/25.jpg)
© 2017 Heroes
Internet of Things
Dare to start, but start wisely
25
Explore IoT opportunities
Build your IoT practice
Secure your first IoT PoC
Get executive support
Grow your IoT practice
Idea analysis &
prioritization
![Page 26: Case study: Integrating IoT in your maintenance practice€¦ · 12/6/2018 · Predictive Maintenance Internet of Things 8 • Sustainable System Maintenance • Enabling machine](https://reader033.fdocuments.net/reader033/viewer/2022050610/5fb1cc23a8080d20e315f569/html5/thumbnails/26.jpg)
© 2017 Heroes
Heroes B.V. | Verenigde Naties 1, 3527 KT Utrecht | www.heroes.nl
https://www.linkedin.com/company/heroesthinkdigital
https://twitter.com/heroes_2017