Lecture 12 AI, Big Data for Energy and Climate and Power Sector Decarbonization

Gang He

November 25, 2024

Sample analytic questions

• What are the new advancements in AI and big data?
  
• How AI and big data could be used in facilitating clean energy transition and addressing climate change?
  
• Any good examples or success stories of AI for energy/climate change?
  
• What skills are need to work in this field?
  

Offshore wind turbine inspection using drones and AI

A local example

“AI is the new electricity”

Just as electricity transformed almost everything 100 years ago, today I actually have a hard time thinking of an industry that I don’t think AI will transform in the next several years.

–Andrew Ng

Milestones in AI

Perspective:

• Universe created:

  • 13.8 billion years ago

• Earth created:

  • 4.54 billion years ago

• Modern human:

  • 300,000 years ago

• Civilization:

  • 12,000 years ago

• Written record:

  • 5,000 years ago

• 1943: Neural networks
• 1957: Perceptron
• 1974-86: Backpropagation, RBM, RNN
• 1989-98: CNN, MNIST, LSTM, Bidirectional RNN
• 2006: “Deep Learning”, DBN
• 2009: ImageNet
• 2012: AlexNet, Dropout
• 2014: GANs
• 2014: DeepFace
• 2016: AlphaGo
• 2017: AlphaZero, Capsule Networks
• 2018: AlphaFold 1, BERT
• 2021: AlphaFold 2
• 2023: ChatGPT

Aartifitial intelligence, machine learning, and deep learning

AI Algorithms

How AI has transformed our society

Key enabling factors

• Technological maturity
  
• Availability and quality of data
  
• Growing importance of cybersecurity
  
• Training and re-skilling of energy sector professionals
  

Deep solar

AI to improve wind perfomance

AI saves data center cooling energy

Global inventory of solar PV

Machine learning for electricity access

AI for climate solutions

AI for GHG mitigation

AI ethics: why?

• Biases
  
• Privacy
  
• Mistakes
  
• Environmental impacts
  
• Other unaware impacts

AI ethics: principles

• Proportionality and do no harm
  
• Safety and security
  
• Fariness and non-discrimination
  
• Sustainability
  
• Right to privacy, and data protection
  
• Human oversight and determination
  
• Transparency and explainability
  
• Responsibility and accountability
  
• Awareness and literacy
  
• Multi-stakeholder and adaptive governance and colloboration

AI ethics: how?

• Still slow responding to the fast evolving challenges
• Education and research
  
• Assessment and communication
  
• Governance and cooperation
  
• Value alligned design

Summary

• AI and big data has transformed how we live, and how can we save the planet
  
• AI does not automatic add new insights, but it provides powerful tools to analyze increasingly available big data
  
• Learning those new tools and skills becomes necessary to work in the field
  
• The potential of using AI/big data to analyze our understanding of energy supply/demand and human behavior is enomorous.

Sample analytic questions

• How much solar and wind capacity need to be built?
  
• How cost delince of renewables and storage will change the capacity and generation mix of the power sector?
  
• How much new transmission capacity is needed to harvest the benefits of interconnection?
  
• Does CCS has role to play in the power sector decarbonization?
  

Power sector’s central role

Overarching strategy

• Electrification
  
• Decarbonization

Why power sector is special?

• Essential good
  
• Infrastructure
  
• Technology/network complexity: balance on real-time
  
• Supply/Demand inelasticity
  
  • Capital intensive
    
  • Investment takes time
    

U.S. electricity flow

U.S. transmission grid

Refresh the basics

• Energy and power
  
• kW, MW, GW, TW
  
• Heat rate and efficiency (Carnot, 1st, 2nd)
  
• Thermodynamics

Load factor and load curves

\(LF=\frac{Energy\ consumed}{Energy\ at\ peak\ demand}=\frac{Average\ power}{Peak\ power\ demand}\)

A low load factor means a “peaky” load shape

Supply curve and dispatch

Power system modeling family

Type	Production Cost (Unit Commitment and Dispatch)	Network Reliability (AC Power Flow, Dynamics)	Capacity Expansion
Generator Adequacy	Yes	No	Often
Flexibility Requirement	Yes	No	Somewhat
Transmission Adequacy	Partially	Yes	Typically No
Gen Contingencies	Somewhat	Yes	No
Transmission Contingencies	Somewhat	Yes	No
Frequency Stability	Somewhat	Yes	No
Voltage Stability, Voltage control	No	Yes	No
Examples	PROMOD, GE-Maps, PLEXOS, GridView	Positive sequence load flow (PSLF), power system simulator for engineering (PSSE)	NEMS, ReEDS, SWITCH, Grid-path, GenX, PyPSA, Haiku

Capacity expansion models

• Capacity expansion models simulate generation and transmission capacity investment, given assumptions about future electricity demand, fuel prices, technology cost and performance, and policy and regulation
  
• What mix of generators should we build to meet load?
  
• Does a policy affect cost of service regions and competitive regions in different ways?

Strength and limits

• Strength: Examine the impacts of power sector policies (or alternative technology/fuel trajectories) on the generation and capacity mix in the mid-to long-term
  
• Limits: Many do not have chronological unit commitment (i.e., every hour of the year chronologically); some use aggregate (model) plants for dispatch; transmission and power flow are a stylized representation (transport or DC)
  
• Example questions: Quantifying the impacts of environmental policies on generation and capacity? What are the cost implications of alternative pathways to a low greenhouse gas emissions future? How will alternative future prices of natural gas impact capacity investment? What is the change in consumption and expenditures? What are the efficiency and distributional effects of various policy designs?

Model comparison

SWITCH Model as an example

Key questions in model selection

• Spatial: 
  • Geographic: state, regional, national, international
    
  • Cost of service vs. competitive regions
    
• Temporal: time of day, seasons, annual, decadal
  
• Time steps: Building new capacity, dispatch
  
• Time horizon: near (2025), medium(2030), long (2050)
  
• Generating units: individual plants or model plants, representation of capital costs and other production costs
  
• T&D: pipeflow, transport, DC powerflow, aggregated
  
• Representation of renewables: RE technologies, availabilities, accessibility cost, variability
  
• Link with economy wide model, environmental constraints

Typical output: capacity/generation mix

Typical output: transmission expansions

Unit committment and network reliability

• Unit Commitment Model: Simulate detailed (hourly to sub-hourly) operation of a given system; Assess resource adequacy and other aspects of reliability of a system; Analyze the impact of changes in the system (e.g., retirement/addition of capacity) on system operation; Assess transmission congestion and locational marginal prices; Describe the daily pattern of emissions
  
• Network Reliability Model: Detailed simulations of the transmission network including dynamic events that can occur in seconds (and cause big problems); these models aren’t run on a day to day basis –they are only run to examine significant changes to an existing system

Emerging trends

• Smart grid
  
• Distributed generation
  
• Renewable penetration
  
• Efficiency improvement
  
• Electrifying
  
• DR/DSM

Summary

• Power sector’s central role in decarbonization
  
• Capacity expansion model to analyze the optimized investment decisions
  
• Decisions in the real world is much more complicated
  
• Emerging trends in the power sector

References

He, Gang, Jiang Lin, Froylan Sifuentes, Xu Liu, Nikit Abhyankar, and Amol Phadke. 2020. “Rapid Cost Decrease of Renewables and Storage Accelerates the Decarbonization of China’s Power System.” Nature Communications 11 (1): 2486. https://doi.org/10.1038/s41467-020-16184-x.

Johnston, Josiah, Rodrigo Henriquez-Auba, Benjamı́n Maluenda, and Matthias Fripp. 2019. “Switch 2.0: A Modern Platform for Planning High-Renewable Power Systems.” SoftwareX 10: 100251. https://doi.org/10.1016/j.softx.2019.100251.

Kaack, Lynn H, Priya L Donti, Emma Strubell, George Kamiya, Felix Creutzig, and David Rolnick. 2022. “Aligning Artificial Intelligence with Climate Change Mitigation.” Nature Climate Change, 1–10. https://doi.org/10.1038/s41558-022-01377-7.

Kruitwagen, L, KT Story, J Friedrich, L Byers, S Skillman, and C Hepburn. 2021. “A Global Inventory of Photovoltaic Solar Energy Generating Units.” Nature 598 (7882): 604–10. https://doi.org/10.1038/s41586-021-03957-7.

Lee, Mekyung, and Gang He. 2021. “An Empirical Analysis of Applications of Artificial Intelligence Algorithms in Wind Power Technology Innovation During 1980–2017.” Journal of Cleaner Production 297: 126536. https://doi.org/10.1016/j.jclepro.2021.126536.

Masters, Gilbert M. 2013. Renewable and Efficient Electric Power Systems. John Wiley & Sons.

Ratledge, Nathan, Gabe Cadamuro, Brandon de la Cuesta, Matthieu Stigler, and Marshall Burke. 2022. “Using Machine Learning to Assess the Livelihood Impact of Electricity Access.” Nature 611 (7936): 491–95. https://doi.org/10.1038/s41586-022-05322-8.

Rolnick, David, Priya L Donti, Lynn H Kaack, Kelly Kochanski, Alexandre Lacoste, Kris Sankaran, Andrew Slavin Ross, et al. 2022. “Tackling Climate Change with Machine Learning.” ACM Computing Surveys (CSUR) 55 (2): 1–96. https://doi.org/10.1145/3485128.

Yu, Jiafan, Zhecheng Wang, Arun Majumdar, and Ram Rajagopal. 2018. “DeepSolar: A Machine Learning Framework to Efficiently Construct a Solar Deployment Database in the United States.” Joule 2 (12): 2605–17. https://doi.org/10.1016/j.joule.2018.11.021.