LA TRANSFORMATION DIGITALE DU MONDE DE L'NERGIE Data science in - - PowerPoint PPT Presentation

β–Ά
la transformation digitale du monde de l nergie data
SMART_READER_LITE
LIVE PREVIEW

LA TRANSFORMATION DIGITALE DU MONDE DE L'NERGIE Data science in - - PowerPoint PPT Presentation

Jan 11, 2023 847 likes β€’1.57k views

LA TRANSFORMATION DIGITALE DU MONDE DE L'NERGIE Data science in energy industry Data science in energy industry 40% of the mondial energy consumption Data science in energy industry 40% of the mondial energy consumption Data science in

slide-1
SLIDE 1

LA TRANSFORMATION DIGITALE DU MONDE DE L'Γ‰NERGIE

slide-2
SLIDE 2

Data science in energy industry

slide-3
SLIDE 3

Data science in energy industry

40% of the mondial energy consumption

slide-4
SLIDE 4

Data science in energy industry

40% of the mondial energy consumption

slide-5
SLIDE 5

Data science in energy industry

40% of the mondial energy consumption

slide-6
SLIDE 6

Data science in energy industry

slide-7
SLIDE 7

Data science in energy industry

slide-8
SLIDE 8

Data science in energy industry

slide-9
SLIDE 9

Data science in energy industry

Machine learning

slide-10
SLIDE 10

Machine learning challenges

slide-11
SLIDE 11

Machine learning challenges

80% of the job

slide-12
SLIDE 12

Machine learning challenges

20% of the job

slide-13
SLIDE 13

Machine learning challenges

20% of the job Suitable for competitions

slide-14
SLIDE 14
slide-15
SLIDE 15

Why ?

  • Critical role in energy efficiency
  • Optimize operations of chillers, boilers and

energy storage systems

  • Baseline for flagging potentially wasteful

discrepancies

β‡’ Forecasting the use of the electrical energy is the backbone of effective operations

slide-16
SLIDE 16

Forecasting building energy consumption

Com Competitio ion Da Data

  • Energy consumption historic for

~200 buildings

  • Temperature
slide-17
SLIDE 17

Forecasting building energy consumption

Com Competitio ion Da Data

  • Energy consumption historic for

~200 buildings

  • Temperature

Com Competitio ion Ob Objective

  • Forecast Energy consumtption

through different horizons

???

slide-18
SLIDE 18

Winner solution

slide-19
SLIDE 19

Winner solution

slide-20
SLIDE 20

Winner solution

slide-21
SLIDE 21

Feature engineering

slide-22
SLIDE 22

Feature engineering

y ~ ~ X

slide-23
SLIDE 23

Feature engineering

y ~ ~ X

Engineered features

slide-24
SLIDE 24

Feature engineering Cyclical time encoding

D= 20h

slide-25
SLIDE 25

Feature engineering Cyclical time encoding

slide-26
SLIDE 26

Feature engineering Cyclical time encoding

slide-27
SLIDE 27

Feature engineering Cyclical time encoding

slide-28
SLIDE 28

Feature engineering Cyclical time encoding

slide-29
SLIDE 29

Feature engineering Cyclical time encoding

???

slide-30
SLIDE 30

Feature engineering Cyclical time encoding

𝒕𝒋𝒐 πŸ‘π†π’Š πŸ‘πŸ“ & 𝐝𝐩𝐭 πŸ‘π†π’Š πŸ‘πŸ“

slide-31
SLIDE 31

Winner solution

slide-32
SLIDE 32

Winner solution

Boosted trees

slide-33
SLIDE 33

Boosted trees

BT BT Boosted trees You

slide-34
SLIDE 34

Boosted trees Decision trees

slide-35
SLIDE 35

Boosted trees Decision trees

Depth

slide-36
SLIDE 36

Boosted trees Decision trees

Prediction

slide-37
SLIDE 37

Boosted trees Boosting

Prediction

slide-38
SLIDE 38

Boosted trees Boosting

Prediction Reality

slide-39
SLIDE 39

Boosted trees Boosting

Error

  • Prediction

Reality

slide-40
SLIDE 40

Boosted trees Boosting

  • Reality

Prediction Error

slide-41
SLIDE 41

Boosted trees Boosting

Original data

slide-42
SLIDE 42

Boosted trees Boosting

Original data Decsion tree

slide-43
SLIDE 43

Boosted trees Boosting

Original data Decsion tree Error

slide-44
SLIDE 44

Boosted trees Boosting

Original data Decsion tree Error

slide-45
SLIDE 45

Boosted trees Boosting

…

Original data Decsion tree Error

slide-46
SLIDE 46

Boosted trees Boosting

…

Original data Decsion tree Error

Nb of trees

slide-47
SLIDE 47
  • Improve the state of the Art
  • Create a community
  • Provide a solution to a typical

Energy problematic βž” This solution can now be used in

  • ther context
slide-48
SLIDE 48

Why ?

  • Flexibility in energy management is

essential for secure supply and increasing the penetration of renewable sources.

  • Energy storage and local production can

increase smart building flexibility.

  • Time of use tariffs can incite use of

energy when it is the most available.

β‡’ Algorithms can help battery charging systems to be as efficient as possible

slide-49
SLIDE 49

Competition Description

Com Competitio ion Da Data

  • Actual Consumption and

Production (for 11 buildings)

  • Forecast for next 24h
  • Grid energy price (sell and buy)
slide-50
SLIDE 50

Competition Description

Com Competitio ion Da Data

  • Actual Consumption and

Production (for 11 buildings)

  • Forecast for next 24h
  • Grid energy price (sell and buy)

Com Competitio ion Ob Objective

  • Plannify a battery usage to save

money How to use the battery for the next 15 minutes ?

slide-51
SLIDE 51

Competition Results

Perf erformance Metric ic

slide-52
SLIDE 52

Competition Results

Perf erformance Metric ic Be Best Co Competiti tion score: drives 19% savings with a battery.

slide-53
SLIDE 53

Linear Programming

slide-54
SLIDE 54

Linear Programming

slide-55
SLIDE 55

Linear Programming

slide-56
SLIDE 56

Linear Programming

Iss Issue: Future consumption and prediction are unknown. We only have forecastings.

slide-57
SLIDE 57

Forecasting Error

slide-58
SLIDE 58

Scenario based stochastic programming

slide-59
SLIDE 59

Scenario based stochastic programming

slide-60
SLIDE 60

Scenario based stochastic programming

slide-61
SLIDE 61

Results

Scores es

Method Percentage of saving with a battery Our method 19,6 % 1st competition method 19,4 % 2nd competition method 19,2 % 3rd competition method 19,1

slide-62
SLIDE 62

Results

Scores es

Method Percentage of saving with a battery Our method 19,6 % 1st competition method 19,4 % 2nd competition method 19,2 % 3rd competition method 19,1 Want to go further ? https://github.com/kaizen-solutions/power-laws-optimization

slide-63
SLIDE 63
  • Algorithms driving 19%
  • f savings with a battery
  • Algorithms and

comparison code are on github

slide-64
SLIDE 64

Conclusion

Business needs

  • Business

context

  • True dataset

Open Sources

  • Understand

Solutions

  • Formation

Continuous Improvement

  • Compare

with existing

  • Community
slide-65
SLIDE 65

Conclusion

Business needs

  • Business

context

  • True dataset

Open Sources

  • Understand

Solutions

  • Formation

Continuous Improvement

  • Compare

with existing

  • Community
slide-66
SLIDE 66

Conclusion

Business needs

  • Business

context

  • True dataset

Open Sources

  • Understand

Solutions

  • Formation

Continuous Improvement

  • Compare

with existing

  • Community
slide-67
SLIDE 67

Any questions ?

slide-68
SLIDE 68

Winner solution

slide-69
SLIDE 69
slide-70
SLIDE 70

Data Collection Data viz - QC - Transfo Feature engineering Model building Final model Problem formulation

slide-71
SLIDE 71