Jump to content

Electricity Systems: Difference between revisions

(add intro to problem areas)
 
(36 intermediate revisions by 7 users not shown)
''This page is about the intersection of electricity systems and machine learning (ML) in the context of climate change mitigation. For an overview of electricity systems as a whole, please see the [https://en.wikipedia.org/wiki/Electric_power_system Wikipedia page] on this topic.''[[File:ElectricitySystems.png|thumb|A schematic of selected opportunities to reduce greenhouse emissions from electricity systems using machine learning. From "Tackling Climate Change with Machine Learning."<ref name=":0">{{Cite journal|last=Rolnick|first=David|last2=Donti|first2=Priya L.|last3=Kaack|first3=Lynn H.|last4=Kochanski|first4=Kelly|last5=Lacoste|first5=Alexandre|last6=Sankaran|first6=Kris|last7=Ross|first7=Andrew Slavin|last8=Milojevic-Dupont|first8=Nikola|last9=Jaques|first9=Natasha|last10=Waldman-Brown|first10=Anna|last11=Luccioni|first11=Alexandra|date=2019-11-05|title=Tackling Climate Change with Machine Learning|url=http://arxiv.org/abs/1906.05433|journal=arXiv:1906.05433 [cs, stat]}}</ref>]]The energy supply sector contributes about 35% of human-caused greenhouse gas emissions,<ref>Bruckner T., I.A. Bashmakov, Y. Mulugetta, H. Chum, A. de la Vega Navarro, J. Edmonds, A. Faaij, B. Fungtammasan, A. Garg, E. Hertwich, D. Honnery, D. Infield, M. Kainuma, S. Khennas, S. Kim, H.B. Nimir, K. Riahi, N. Strachan, R. Wiser, and X. Zhang, 2014: Energy Systems. In: ''Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change'' [Edenhofer, O., R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, J. Savolainen, S. Schlömer, C. von Stechow, T. Zwickel and J.C. Minx (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.</ref> within which decarbonizing electricity supply plays an important role. In addition, many climate change strategies in sectors such as [[Buildings and Cities|buildings]], [[transportation]], and [[industry]] rely on low-carbon electricity. To reduce greenhouse gas emissions from electricity systems, it will be necessary to both transition quickly to low-carbon electricity sources (e.g., solar, wind, and nuclear) and to reduce emissions from existing electricity system operations in the meantime.
AI has been called the new electricity, given its potential to transform entire industries.<ref>{{Cite web|title=Andrew Ng: Artificial Intelligence is the New Electricity - YouTube|url=https://www.youtube.com/watch?v=21EiKfQYZXc|website=www.youtube.com}}</ref> Interestingly, electricity itself is one of the industries that AI is poised to transform. Many electricity systems are awash in data, and the industry has begun to envision next-generation systems (smart grids) driven by AI and ML.<ref>{{Cite journal|last=Ramchurn|first=Sarvapali D.|last2=Vytelingum|first2=Perukrishnen|last3=Rogers|first3=Alex|last4=Jennings|first4=Nicholas R.|date=2012-04|title=Putting the 'smarts' into the smart grid|url=http://dx.doi.org/10.1145/2133806.2133825|journal=Communications of the ACM|volume=55|issue=4|pages=86–97|doi=10.1145/2133806.2133825|issn=0001-0782}}</ref><ref>{{Citation|title=Machine Learning Techniques for Supporting Renewable Energy Generation and Integration: A Survey|url=http://dx.doi.org/10.1007/978-3-319-13290-7_7|publisher=Springer International Publishing}}</ref><ref>{{Cite web|title=How artificial intelligence will affect the future of energy and climate|url=https://www.brookings.edu/research/how-artificial-intelligence-will-affect-the-future-of-energy-and-climate/|website=Brookings|date=2019-01-10|language=en-US}}</ref>
 
AI hasand beenmachine calledlearning theare newoften electricity,discussed given its potential to transform entire industries.<ref>{{Cite web|title=Andrew Ng: Artificial Intelligence isin the New Electricity - YouTube|url=https://www.youtube.com/watch?v=21EiKfQYZXc|website=www.youtube.com}}</ref> Interestingly, electricity itself is one of the industries that AI is poised to transform. Many electricity systems are awashsector in data, and the industrycontext hasof begun to envision next-generation systems (smart grids) driven by AI and ML.,<ref>{{Cite journal|last=Ramchurn|first=Sarvapali D.|last2=Vytelingum|first2=Perukrishnen|last3=Rogers|first3=Alex|last4=Jennings|first4=Nicholas R.|date=2012-04|title=Putting the 'smarts' into the smart grid|url=http://dx.doi.org/10.1145/2133806.2133825|journal=Communications of the ACM|volume=55|issue=4|pages=86–97|doi=10.1145/2133806.2133825|issn=0001-0782|via=}}</ref><ref>{{Citation|title=Machine Learning Techniques for Supporting Renewable Energy Generation and Integration: A Survey|url=http://dx.doi.org/10.1007/978-3-319-13290-7_7|publisher=Springer International Publishing}}</ref><ref>{{Cite web|title=How artificial intelligence will affect the future of energy and climate|url=https://www.brookings.edu/research/how-artificial-intelligence-will-affect-the-future-of-energy-and-climate/|website=Brookings|date=2019-01-10|language=en-US}}</ref> which broadly refer to the concept of "intelligent" electric grids managed automatically in a data-driven manner. In particular, ML has been used to forecast electricity supply and demand, to improve power system optimization, and to improve system efficiency through applications such as predictive maintenance. In addition, ML has also been used to accelerate scientific discovery of clean energy technologies, and to gather electricity infrastructure data that may be useful for system planners and policymakers.<ref name=":0" />
Electricity systems are responsible for about a quarter of human-caused greenhouse gas emissions each year.<ref>IPCC. Climate Change 2014: ''Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change'' [O. Edenhofer, R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, J. Savolainen, S. Schlomer, C. von Stechow, T. Zwickel, J.C. Minx, (eds.)]. 2014.</ref> Moreover, as buildings, transportation, and other sectors seek to replace GHG-emitting fuels, demand for low-carbon electricity will grow. To reduce emissions from electricity systems, society must
* Rapidly transition to low-carbon electricity sources (such as solar, wind, hydro, and nuclear) and phase out carbon-emitting sources (such as coal, natural gas, and other fossil fuels).
* Reduce emissions from existing CO<sub>2</sub>-emitting power plants, since the transition to low-carbon power will not happen overnight.
* Implement these changes across all countries and contexts, as electricity systems are everywhere.
 
== Machine Learning Application Areas ==
ML can contribute on all fronts by informing the research, deployment, and operation of electricity system technologies. Such contributions include accelerating the development of clean energy technologies, improving forecasts of demand and clean energy, improving electricity system optimization and management, and enhancing system monitoring.<ref name=":0" /> These contributions require a variety of ML paradigms and techniques, as well as close collaborations with the electricity industry and other experts to integrate insights from operations research, electrical engineering, physics, chemistry, the social sciences, and other fields.
 
== Problem areas ==
Applications through which machine learning can help reduce greenhouse gas emissions in electricity systems are listed below. (For demand-side applications, see the pages on [[transportation]], [[Buildings and Cities|buildings and cities]], and [[industry]].)
 
=== Enabling low-carbon electricity ===
 
*'''[[Electricity Supply Forecasting|Supply]], [[Energy Demand Forecasting|demand]] and [[Energy Price Forecasting|price]] forecasting''': The supply and demand of power must both be forecast ahead of time to inform electricity planning and scheduling. In more volatile energy systems, also forecasting of prices becomes relevant to utilize flexibility effectively. ML can help make these forecasts more accurate, improve temporal and spatial resolution, and quantify uncertainty.
*[[Electricity Supply and Demand Forecasting|Electricity supply and demand forecasting]]
*'''Improving [[Power System Optimization|power system optimization]]''': Scheduling algorithms on the power grid have trouble handling large quantities of solar, wind, and other time-varying electricity sources. ML can help improve electricity scheduling algorithms, control storage and flexible demand, and design real-time electricity prices that reduce CO<sub>2</sub> emissions.
* Improving power system optimization
*'''Improving [[Power System Planning|system planning]]''': Algorithms for planning new low-carbon energy infrastructure are often large and slow. ML can help speed up or provide proxies for these algorithms.
* Accelerated materials science for clean energy technologies
*'''Informing [[Maximum Power Point Tracking|maximum power point tracking]]''': Maximum power point tracking refers to a variety of techniques that aim to maximize the power output of weather-dependent renewable energy generators, such as solar panels and wind turbines. ML can help model attributes of renewable energy systems or actively control these systems (e.g., by modulating wind turbine rotation speed) in order to improve power output.
* Optimizing variable generators
*'''[[Accelerated Science|Accelerated science]] for clean energy technologies''': Designing new materials is important for many applications, including energy storage via batteries or solar/chemical fuels. ML can help suggest promising materials to try, thereby speeding up the materials discovery process.
* System planning for clean energy technologies
*'''Informing [[Nuclear Fusion|nuclear fusion]] research''': Nuclear fusion has the potential to produce safe, carbon-free electricity, but such reactors continue to consume more energy than they produce. While basic science and engineering are still needed, ML can help inform nuclear fusion research in a variety of ways, e.g., by suggesting parameters for physical experiments or modeling the behavior of plasma inside reactors.
* Predictive maintenance and fault detection
* Accelerating nuclear fusion science
 
=== Reducing current-system impacts ===
 
*[[Methane Leak Detection|'''Methane leak detection''']]: In addition to the unavoidable climate impacts of burning fossil fuels, natural gas extraction sites, pipelines, and compressor stations leak methane, a powerful greenhouse gas. ML can help detect and prevent these leaks.
* Methane leak detection
*[[Power Grid Emissions Modeling|'''Power grid emissions modeling''']]: Reducing the emissions associated with electricity use requires understanding what the emissions on the electric grid actually are at any given moment. ML can help estimate and forecast emissions, and potentially model the uncertainty in these estimates.
* Modeling emissions
 
=== AdditionalGeneral-purpose areasapplications ===
 
*[[Energy Infrastructure Mapping|'''Energy infrastructure mapping''']]: There are many cases in which decision-relevant information about energy infrastructure -- such as the locations and sizes of solar panels, or the location of power transmission and distribution infrastructure -- is not readily available. ML can help map some of this energy infrastructure using satellite imagery.
* Data collection via remote sensing
*[[Predictive Maintenance|'''Predictive maintenance and fault detection''']]: Quickly detecting power system faults can help reduce power system waste or improve the utilization of low-carbon energy resources. ML can help detect faults in real time from sensor data, or even forecast them ahead of time to enable preemptive maintenance.
* Predictive maintenance
*[[Power System State Estimation|'''State estimation''']]: Many power distribution systems have few sensors, but are increasingly necessary to monitor due to the increase in rooftop solar power. ML can provide algorithms for understanding the state of distribution systems in "low-observability" scenarios where traditional state estimation algorithms may not suffice.
*'''[[Greenhouse Gas Emissions Detection|Greenhouse gas emissions mapping]]''': While some electricity system operators release publicly-available data on the emissions associated with fossil fuel generators, this data is not available in many cases. ML can help map greenhouse gas emissions using a combination of remote sensing and on-the-ground data.
*'''[[Non-Intrusive Load Monitoring|Non-Intrusive Load Monitoring (NILM)]]''': ML can be used to disaggregate the net-load measurements into their individual components. It finds application in disaggregating native load and PV generation measurements from the net-load measurements obtained at the feeder-lever or at the residential-level. Another aspect of energy disaggregation is non intrusive load monitoring, wherein the load measurement is disaggregated into different components representing the consumption of different household electric appliances.
 
== Background readingsReadings ==
 
=== Primers ===
 
*'''Greening the Grid toolkit''': A collection of readings, trainings, and other resources on power grids, renewable energy, energy storage, and electric vehicles. Available [https://greeningthegrid.org/toolkits here].
* '''"Nuclear fusion" in Nature Physics''': A collection of articles on the state of nuclear fusion research. Available [https://www.nature.com/collections/bccqhmkbyw/ here].
 
== Online coursesCourses and courseCourse materialsMaterials ==
 
*'''"Computational Methods for the Smart Grid"''': "[A]n introduction to recent advances in computational methods applied to sustainable energy and the smart grid... provid[ing] students with a broad background in state-of-the-art computational methods that repeatedly arise in these domains, such as machine learning, optimization, and control." Lecture slides, videos, and assignments available [http://www.cs.cmu.edu/~zkolter/course/15-884/ here].
*'''"Electric Power Systems" on Coursera''': A course covering the "standards and policies of the electric utility industry," including "basic vocabulary used in the business" and an introduction of "the electric power system, from generation of the electricity all the way to the wall plug." Enroll [https://www.coursera.org/learn/electric-power-systems here].
 
== Conferences, Journals, and Professional Organizations ==
== Community ==
 
=== Major conferences ===
 
 
 
*'''ACM e-Energy''': "The International Conference on Future Energy Systems (ACM e-Energy) aims to be the premier venue for researchers working in the broad areas of computing and communication for smart energy systems (including the smart grid), and in energy-efficient computing and communication systems". Website [https://energy.acm.org/eenergy-conference/ here].
*'''IEEE Power & Energy Society General Meeting''': One of the [https://www.ieee-pes.org/meetings-and-conferences IEEE Power & Energy Society's] flagship annual conferences, held in North America. Website [https://pes-gm.org/ here].
*'''Power Systems Computation Conference''': A biennial conference held in Europe, focused on computational power system methods. Website [https://pscc-central.epfl.ch/ here].
*'''IEEE Transactions on Power Systems''': Covers the "requirements, planning, analysis, reliability, operation, and economics of electric generating, transmission, and distribution systems for general industrial, commercial, public, and domestic consumption." Journal website [https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=59 here].
*'''IEEE Transactions on Smart Grid''': "[A] cross disciplinary and internationally archival journal aimed at disseminating results of research on smart grid that relates to, arises from, or deliberately influences energy generation, transmission, distribution and delivery." Journal website [https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=5165411 here].
*'''IEEE Transactions on Sustainable Energy:''' This journal is aimed at disseminating results of research on the design, implementation, grid-integration, and control of sustainable energy technologies and systems that can be integrated into the power transmission and/or distribution grid. Journal website [https://www.ieee-pes.org/ieee-transactions-on-sustainable-energy here].
 
=== Major societies andprofessional organizations ===
 
*'''IEEE Power & Energy Society''': "[T]he world's largest forum for sharing the latest in technological developments in the electric power industry, for developing standards that guide the development and construction of equipment and systems, and for educating members of the industry and the general public." Website [https://www.ieee-pes.org/ here].
*'''ACM Special Interest Group on Energy Systems and Informatics''': "ACM SIGEnergy is a professional forum for scientists, engineers, educators, and professionals for discussing energy systems and energy informatics. It brings together an inter-disciplinary group of computer scientists with diverse backgrounds [...] to discuss and address key challenges in future energy systems, and their impact on society." Website [https://energy.acm.org/ here].
 
== Libraries and toolsTools ==
 
*'''PowerTAC''': A power system simulation environment, available [https://powertac.org/ here].
*'''Energy Policy Simulator''': A tool to simulate the greenhouse gas emissions effects of various climate and energy policies, available [https://us.energypolicy.solutions/docs/ here].
*'''Optimal Power Flow (OPF) Sampler Package''': A Julia package to generate power grid data samples via optimal power flow methods, available [https://github.com/invenia/OPFSampler.jl/ here].
*'''Pyiso''': A Python client library for data from power grid balancing authorities in the United States, Canada, and Europe. Documentation [https://pyiso.readthedocs.io/en/latest here].
*'''List of Resources by ACM SIG Energy'''': A list of models, libraries, software and datasets curated by the [https://energy.acm.org/ ACM Special Interest Group on Energy Systems and Informatics], available [https://energy.acm.org/resources/ here].
* '''OPFLEarn.jl''': A [https://github.com/NREL/OPFLearn.jl Julia package] for creating datasets for machine learning approaches to solving AC optimal power flow (AC OPF).
 
== Data ==
''See the subpages listed [[Electricity Systems#Machine Learning Application Areas|above]] for application-specific datasets.''
 
=== ElectricityUnited system dataStates ===
 
*'''Public Utility Data Liberation (PUDL) Project''': A project integrating a variety of United States federal electricity data sources, available [https://github.com/catalyst-cooperative/pudl here] via GitHub.
*'''United States ISO/RTO data:''' Electricity market data (e.g., prices, supply, and demand) are available online for a number of US independent system operator/regional transmission organization regions, namely [http://www.caiso.com/TodaysOutlook/Pages/prices.aspx CAISO], [http://www.ercot.com/gridinfo ERCOT], [https://www.iso-ne.com/markets-operations ISO-NE], [https://www.misoenergy.org/markets-and-operations/ MISO], [https://www.nyiso.com/energy-market-operational-data NYISO], [https://dataminer2.pjm.com/list PJM], and [https://marketplace.spp.org/ SPP]. (These data can also be accessed via the [https://pyiso.readthedocs.io/en/latest pyiso Python library].)
*'''EuropeanPecan electricity market data from electricityMapStreet:''': LinksDisaggregated toenergy (and parsers for) European power systemwater data, available [https://githubwww.compecanstreet.org/tmrowcodataport/about/electricitymap-contrib#data-sources here] via(requires GitHub.login)
*'''United States Environmental Protection Agency's Air Markets Program data''': Datasets from the US EPA's emissions trading programs. For instance, the Continuous Emissions Monitoring System dataset (also available via the EPA's [ftp://newftp.epa.gov/DMDnLoad/emissions/ FTP site]) provides hourly emissions and generation for many fossil fuel generators in the United States. Available [https://ampd.epa.gov/ampd/ here].
* '''Utility Transition Hub™ Data''': Datasets on US-based utility finances, investments, operations, emissions and the utilities' alignment and commitments for the 1.5°C goals. Available [https://utilitytransitionhub.rmi.org/data-download/ here].
 
=== OtherEurope ===
=== Renewables forecasting contest data ===
 
*'''European electricity market data from electricityMap''': Links to (and parsers for) European power system data, available [https://github.com/tmrowco/electricitymap-contrib#data-sources here] via GitHub.
*'''SubseasonalRodeo''': "A benchmark dataset for training and evaluating ''subseasonal forecasting systems''—systems predicting temperature or precipitation 2-6 weeks in advance—in the western contiguous United States." Available [https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/IHBANG here].
*'''European electricity market data via Pyiso''': The [https://pyiso.readthedocs.io/en/latest pyiso Python library] provides data from European power grid control areas.
*'''American Meteorological Society 2013-2014 Solar Energy Prediction Contest''': Contest data for producing daily forecasts of solar energy, available [https://www.kaggle.com/c/ams-2014-solar-energy-prediction-contest here].
 
=== Demand dataIndia ===
 
*'''Rural Electricity Demand in India (REDI) Dataset''': "The dataset contains detailed information on electricity demand in rural India. The dataset covers 10,000 households and 2,000 rural enterprises across 200 villages in Bihar, Uttar Pradesh, Odisha, and Rajasthan." Available [https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/1ZNLUY here].
 
=== Other ===
*'''Pecan Street:''' Disaggregated energy and water data, available [https://www.pecanstreet.org/dataport/about/ here] (requires login)
 
*'''Project Sunroof by Google''': Detailed estimates of rooftop solar potential based on sunlight and roof space, available [https://www.google.com/get/sunroof/data-explorer/ here].
=== GHG emissions data ===
*'''EDP Open Data''': EDP, a multinational utility company, publishes open data on solar and wind power assets (including clean datasets put together for previous competitions). Available [https://opendata.edp.com/pages/homepage/ here].
*'''Ausgrid Data'''': Ausgrid, a DSO in NSW, Australia, [https://www.ausgrid.com.au/Industry/Our-Research/Data-to-share provides several datasets], among them 300 households, 225 substations and data about outages.
 
== Relevant Groups and Organizations ==
*'''Copernicus global methane data''': Dataset on global methane emissions from 2002 onwards from the European Space Agency, available [https://cds.climate.copernicus.eu/cdsapp#!/dataset/satellite-methane?tab=overview here].
*'''United States Environmental Protection Agency's Air Markets Program data''': Datasets from the US EPA's emissions trading programs. For instance, the Continuous Emissions Monitoring System dataset (also available via the EPA's [ftp://newftp.epa.gov/DMDnLoad/emissions/ FTP site]) provides hourly emissions and generation for many fossil fuel generators in the United States. Available [https://ampd.epa.gov/ampd/ here].
*See also: '''[[Remote Sensing Datasets|Satellite imagery datasets]]'''
 
* '''Electric Power Research Institute (EPRI)''': A non-profit organization that conducts research, development, and demonstration projects focused on electricity generation and delivery, and with over 1,000 member organizations around the world. Website [https://www.epri.com/ here].
=== Other ===
*'''The Global Power System Transformation Consortium (G-PST)''': An international consortium of the leading electricity system operators "to identify common, cutting-edge research questions that can inform large- scale national research and development investments." Website [https://globalpst.org/ here] and inaugural research agenda [https://globalpst.org/resources/ here].
 
* '''[[Accelerated Materials Science Datasets|Accelerated materials science datasets]]:''' Datasets that may be useful for research on solar fuels, next-generation battery conducting fluids, or other accelerated materials science applications in the electricity sector.
* '''[[Remote Sensing Datasets|Satellite imagery datasets]]:''' Datasets that may be useful for applications such as power plant emissions detection, power grid mapping, solar panel mapping, etc.
*'''Project Sunroof by Google''': Detailed estimates of rooftop solar potential based on sunlight and roof space, available [https://www.google.com/get/sunroof/data-explorer/ here].
 
== References ==
Cookies help us deliver our services. By using our services, you agree to our use of cookies.