Wind Power Forecasting and Electricity Markets
To improve wind power forecasting and its use in power system and electricity market operations Argonne National Laboratory has assembled a team of experts in wind power forecasting, electricity market modeling, wind farm development, and power system operations. Argonne is the project lead, with support from the Institute for Systems and Computer Engineering of Porto (INESC Porto) in Portugal. On the industry side, Argonne is collaborating with the Midwest Independent System Operator and Horizon Wind Energy LLC. The project is funded by the U.S. Department of Energy’s Wind Power Program.
Background, Objectives, and Status
Wind power forecasting is an important tool for managing the inherent variability and uncertainty in wind power generation. Increasing the accuracy of forecasting can help to reduce the likelihood of an unexpected gap between scheduled and actual wind power generation, which can be extremely helpful for operators of power systems and wind power plants. Improvements in wind power forecasting and better use of the forecasts in operational decisions can, therefore, help facilitate a large-scale penetration of wind power into the electric power system. Substantial research efforts continue to focus on improving the accuracy and the robustness of wind power predictions, to increase their time horizon, and to consistently evaluate their performance. In addition, there is an urgent need for finding better ways to incorporate the output of advanced wind forecasting models into the operation of power systems and wind power plants.
The overall objective of this research project is to improve wind power forecasting methodologies and their use in power system operations. The project is structured into two parts. The first part concentrates on two areas: (1) surveying existing wind power forecasting methodologies and tools and (2) identifying strengths and limitations of different approaches. The results from the survey are documented in the report, “Wind Power Forecasting: State-of-the-Art 2009”. More recently, we have also completed “A Survey on Wind Power Ramp Forecasting”.
On the basis of the survey findings, the Argonne team is investigating new concepts and methods for improved wind power forecasting. The development and testing of novel statistical algorithms for point, uncertainty, and ramp forecasting is documented in a new report "Development and Testing of Improved Statistical Methods for Wind Power Forecasting". The software code developed in this project for statistical forecasting algorithms, ARGUS-PRIMA, is now available under a licensing arrangement.
The second part of the project addresses how operators of wind power plants and power systems can use advanced wind forecasting technologies to improve their operations, in the time span from several days ahead to real-time operations. Under this part, the team is developing improved methodologies and algorithms to incorporate the output of advanced wind energy forecasts into decision support models for wind power plant and power system operation as documented in the new report "Use of Wind Power Forecasting in Operational Decisions". More recently, the Argonne group has also started investigating the impact of wind power on system-wide emissions in the electric power grid, using the operational tools developed in the project.
Original project description
Software Code: ARGUS-PRIMA (Prediction Intelligent Machine)
Argonne National Laboratory and INESC Porto have developed novel computational learning algorithms for wind power point and uncertainty forecasting. Testing of the new algorithms on wind farm data from the Midwest show significant improvements in forecast performance compared to existing methods. The forecasting algorithms and test results are documented in the project report "Development and Testing of Improved Statistical Methods for Wind Power Forecasting", as well as in several journal and conference papers. The software code is now available under a licensing agreement to facilitate technology transfer to industry.
More information on ARGUS-PRIMA.
Analysis: Wind Power and System-Wide Emissions
In a new analysis, we discuss the environmental effects of incorporating wind energy into the electric power system. We present a detailed emissions analysis based on comprehensive modeling of power system operations for different wind penetration levels. We use the unit commitment and economic dispatch modeling framework developed in this project to simulate the operation of the power grid. We calculate emissions from the dispatch results from each power plant. The emissions model incorporates the effects of both cycling and startups of thermal power plants in analyzing emissions from an electric power system with increasing levels of wind power. Our results for the power system in the state of Illinois show significant emissions effects from increased cycling and particularly start-ups of thermal power plants. However, we conclude that as the wind power penetration increases, pollutant emissions decrease overall due to the replacement of fossil fuels.
Full paper: System-Wide Emissions Implications of Increased Wind Power Penetration
Two Reports on Wind Power Forecasting and its use in Operational Decisions
The main findings from this project have been documented in two new reports on wind power forecasting and its use in operational decisions. The full reports can be downloaded from the links below.
Within wind power forecasting, we have investigated improvements in the statistical algorithms used for wind power point forecasting, wind power uncertainty forecasting, and wind power ramp forecasting. For wind power point forecasting, we have focused on the training criteria used in the computational learning algorithms which convert weather forecasts and observational data to a point forecast for wind power. In particular, we focused on the use of training criteria from information theoretic learning (ITL). In contrast to the standard minimum square error (MSE) criterion, ITL criteria are not built on the assumption of a Gaussian distribution of the forecasting errors. We tested the new approaches on real-world data from wind farms in the U.S. Midwest. Our experiments show distinct advantages of using the new ITL training criteria compared to MSE, in terms of reduced wind power forecasting errors. For wind power uncertainty forecasting, we have developed two new methods for characterizing the wind power forecast uncertainty. The two methods are based on kernel density estimation (KDE) combined with either the Quantile Copula or Nadaraya Watson estimators. We developed time-adaptive versions for both KDE algorithms, which is an import contribution to the current state-of-the-art. The new algorithms were tested on dataset from the Eastern Wind Integration and Transmission Study (EWITS), as well as on two large-scale wind farms located in the U.S. Midwest. Results show that the new KDE algorithms tend to give a better match to the observed wind power distribution (i.e. better calibration) than traditional quantile regression. For wind power ramp forecasting, we have developed a new method for probabilistic ramp event detection, and we performed an extensive experimental evaluation using recently proposed ramp definitions. A detailed documentation of the proposed algorithms and test results is documented in the report “Development and Testing of Improved Methods for Wind Power Forecasting.”
In this project, we have also developed and successfully tested several methodologies and modeling tools for the use of wind power forecasting in operational decisions, from the perspectives of the system operator as well as the wind power producers. We have investigated how the different objectives of system operators and market participants may lead to different opinions on what constitutes a good wind power forecast, which in turn may influence the training criteria used for wind power point forecasting. We have also focused on the use probabilistic wind power forecasts in electricity markets. We have investigated the representation of wind power forecasting uncertainty in the unit commitment problem. Traditional deterministic unit commitment models use a point forecast for wind power output. In contrast, we have developed a stochastic alternative that represents forecast uncertainty by using scenarios that capture cross-temporal dependencies in the predicted wind power. Furthermore, we have investigated the use of probabilistic wind power forecasts to estimate dynamic operating reserve requirements. We have tested the new algorithms on several case studies on a small-scale hypothetical power system as well as on realistic data for the power system of Illinois. A comparison of a diversity of unit commitment and operating reserve strategies illustrate the potential advantages of using probabilistic forecasts in the scheduling of energy and operating reserves compared to the traditional deterministic approach. We have also developed a model for optimal trading of wind power under uncertainty in wind power and prices. The model has been tested on several case studies on both hypothetical and real-world data. The results show that the model can control the trade-offs between risk and return for wind farm owners depending on their risk preferences. We have also used the model to investigate how market design, in the form of potential deviation penalties, influences optimal bidding decisions, system imbalances, and wind power’s profitability. The proposed algorithms and test results are documented in the report “Use of Wind Power Forecasting in Operational Decisions.”
Development and Testing of Improved Statistical Methods for Wind Power Forecasting
Use of Wind Power Forecasting in Operational Decisions
Wind Power Forecasting: State-of-the-Art 2009
The objective of this report is to review and analyze state-of-the-art wind power forecasting (WPF) models and their application to power systems operations. We first give a detailed description of the methodologies underlying state-of-the-art WPF models. We then look at how WPF can be integrated into power system operations, with specific focus on the unit commitment problem. The report includes:
- A review of numerical weather prediction systems (meteorological systems for weather predictions) and a description of how their characteristics (spatial and temporal resolution) may affect the performance of the WPF models;
- A general presentation of WPF approaches;
- A detailed literature overview of various theoretical WPF approaches, including a description of some exemplary mathematical models;
- A review of commercial and operational WPF tools;
- A review of existing benchmarking results and an overview of the main conclusions;
- A review of approaches for estimating and representing forecast uncertainty;
- Synthesis of end-user requirements for forecasting tools, including input/output data, user interfaces, etc.;
- A review of how wind power is currently handled in power system and electricity market operations, with a focus on the electricity markets in the United States;
- A review of current and proposed approaches for including WPF into the centralized unit commitment problem;
- A set of alternative proposals for representing wind power and its uncertainty in unit commitment formulations; and
- Recommendations for improving WPF and its use in power system operations.
Full report: Wind Power Forecasting: State-of-the-Art 2009
Quick Guide: A Quick Guide to Wind Power Forecasting: State-of-the-Art 2009
A Survey on Wind Power Ramp Forecasting
This report presents an overview of current ramp definitions and state-of-the-art approaches in ramp event forecasting.
A Survey on Wind Power Ramp Forecasting
Silva C., Bessa R., Pequeno E., Sumaili J., Miranda V.., Zhou Z., Botterud, A., “Dynamic Factor Graphs - A New Wind Power Forecasting Approach,” Report ANL/ESD-14-9, Argonne National Laboratory, Sept. 2014.
Mendes, J., Sumaili J., Bessa R., Keko H., Miranda V., Botterud A. and Zhi Z., “Very Short-Term Wind Power Forecasting: State-of-the-Art,” Report ANL/DIS-14/6, Argonne National Laboratory, Dec. 2013.
Botterud A., Zhou Z., Wang J., Bessa R.J., Keko H., Mendes J., Sumaili J, Miranda V., “Use of Wind Power Forecasting in Operational Decisions,” Report ANL/DIS-11-8, Argonne National Laboratory, Sep. 2011.
Mendes J., Bessa R.J., Keko H., Sumaili J., Miranda V., Ferreira C., Gama J., Botterud A., Zhou Z., Wang J., “Development and Testing of Improved Statistical Wind Power Forecasting Methods,” Report ANL/DIS-11-7, Argonne National Laboratory, Sep. 2011.
Ferreira C., Gama J., Matias L., Botterud A., Wang J., “A Survey on Wind Power Ramp Forecasting,” Report ANL/DIS 10-13, Argonne National Laboratory, Dec. 2010.
Monteiro C., Bessa R., Miranda V., Botterud A., Wang J., Conzelmann G., “Wind Power Forecasting: State-of-the-Art 2009,” Report ANL/DIS-10-1, Argonne National Laboratory, Nov.2009.
Levin T., Botterud A., “Capacity Adequacy and Revenue Sufficiency in Electricity Markets with Wind Power,” IEEE Transactions on Power Systems, in press, Mar. 2015.
Jin S., Botterud A., Ryan S.M., “Temporal vs. Stochastic Granularity in Thermal Generation Capacity Planning with Wind Power,” IEEE Transactions on Power Systems, Vol. 29, No. 5, pp. 2033-2041, 2014.
Vilim M., Botterud A., “Wind Power Bidding in Electricity Markets with High Wind Penetration,” Applied Energy, Vol. 114, pp. 141-155, 2014.
Zhou Z., Botterud A., “Dynamic Scheduling of Operating Reserves in Co-optimized Electricity Markets with Wind Power,” IEEE Transactions on Power Systems, Vol. 29, No. 1, pp.160-171, 2014.
Jin S., Botterud A., Ryan S.M., “Impact of Demand Response on Thermal Generation Investment with High Wind Penetration,” IEEE Transactions on Smart Grid, Vol. 4, No. 4, pp. 2374-2383, 2013.
Zhou Z., Botterud A., Wang J., Bessa R.J., Keko H., Sumaili J., Miranda V., “Application of Probabilistic Wind Power Forecasting in Electricity Markets,” Wind Energy, Vol. 16, No. 3, pp. 321-338, 2013.
Botterud A., Zhou Z., Wang J., Sumaili J., Keko H., Mendes J., Bessa R.J., Miranda V., “Demand Dispatch and Probabilistic Wind Power Forecasting in Unit Commitment and Economic Dispatch: A Case Study of Illinois,” IEEE Transactions on Sustainable Energy, Vol. 4, No.1, pp. 250-261, 2013.
Bessa R.J., Miranda V., Botterud A., Wang J., Constantinescu E.M., "Time Adaptive Conditional Kernel Density Estimation for Wind Power Forecasting," IEEE Transactions on Sustainable Energy, Vol. 3, No. 4, pp. 660-669, Oct. 2012.
Botterud A., Zhou Z., Wang J., Bessa R.J., Keko H., Suamili J., Miranda, V., “Wind Power Trading under Uncertainty in LMP Markets,” IEEE Transactions on Power Systems, Vol. 27, No.2, pp. 894-903, May 2012.
Valentino L., Valenzuela V., Botterud A., Zhou Z., Conzelmann G., "System-Wide Emissions Implications of Increased Wind Power Penetration," Environmental Science and Technology, Vol. 46, No. 7, pp 4200-4206, 2012.
Bessa R.J., V. Miranda, A. Botterud, Z. Zhou, J. Wang, “Time-Adaptive Quantile-Copula for Wind Power Probabilistic Forecasting,” Renewable Energy, Vol. 40, No. 1, pp. 29-39, 2012.
Wang J., Botterud A., Bessa R., Keko H, Carvalho L., Issicaba D., Sumaili J., Miranda V., “Representing Wind Power Forecasting Uncertainty in Unit Commitment,” Applied Energy, Vol. 88, No. 11, pp. 4014-4023, 2011.
Bessa R.J., Miranda V., Botterud A., Wang J., “ ‘Good’ or ‘Bad’ Wind Power Forecasts: A Relative Concept,” Wind Energy, Vol. 14, No. 5, pp. 625-636, 2011.
Botterud A., Wang J., Miranda V., Bessa R.J., “Wind Power Forecasting in U.S. Electricity Markets,” Electricity Journal, Vol. 23, No. 3, pp. 71-82, 2010.
Botterud A., “Forecasting Renewable Energy for Grid Operations,” in Jones L. (ed) “Renewable Energy Integration: Practical Management of Variability, Uncertainty, and Flexibility in Power Grids,” pp. 137-148, Elsevier, 2014.
Wang J., Valenzuela J., Botterud A., Bessa R., Keko H., Miranda V., “Reliability Assessment Unit Commitment with Uncertain Wind Power,” in Pardals et al. (eds) “Handbook of Wind Power Systems: Optimization, Modeling, Simulation and Economic Aspects,” pp. 3-20, Springer, 2013.
Ferreira C., Gama J., Miranda V., Botterud A., “Probabilistic ramp detection and forecasting for wind power prediction,” in “Reliability and Risk Evaluation of Wind Integrated Power Systems,” pp. 29-44, Springer, 2013.
Ferreira C.A., Gama J., Costa V.S., Miranda V., Botterud A., "Predicting Ramp Events with a Stream-Based HMM framework," Lecture Notes in Computer Science, Vol. 7569, pp. 224-238, Springer-Verlag, 2012.
Sumaili J., Keko H., Miranda V., Zhou Z., Botterud A., Wang J., “Finding Representative Wind Power Scenarios and their Probabilities for Stochastic Models,” Proceedings 16th Int. Conference on Intelligent System Application to Power Systems, Hersonissos, Greece, Sept. 2011.
Sumaili J., Keko H., Miranda V., Botterud A., Wang J., “A Clustering-Based Wind Power Scenario Reduction Technique,” Proceedings 17th Power System Computation Conference (PSCC.11), Stockholm, Sweden, Aug. 2011.
Bessa R.J., Sumaili J., Miranda V., Botterud A., Wang J., Constantinescu E., “Time-Adaptive Kernel Density Forecast: A New Method for Wind Power Uncertainty Modeling,” Proceedings 17th Power System Computation Conference (PSCC.11), Stockholm, Sweden, Aug. 2011.
Botterud A., Zhou Z., Wang J., Valenzuela J., Sumaili J., Bessa R.J., Keko H., Miranda V., “Unit Commitment and Operating Reserves with Probabilistic Wind Power Forecasts,” Proceedings IEEE Trondheim PowerTech 2011, Trondheim, Norway, June 2011.
Bessa R.J., Mendes J., Miranda V., Botterud A., Wang J., Zhou Z., “Quantile-Copula Density Forecast for Wind Power Uncertainty Modeling,” Proceedings IEEE Trondheim PowerTech 2011, Trondheim, Norway, June 2011.
Botterud A., Wang J., Zhou Z., R.J. Bessa, H. Keko, Miranda, V., “Application of Probabilistic Wind Power Forecasting in Electricity Markets,” Windpower 2011 Conference and Exhibition, Anaheim, CA, May 2011.
Bessa R.J., Miranda V., J. Sumaili, A. Botterud, J. Wang, Z. Zhou, "Wind Power Forecast with Probability Density Estimation: A Tool for the Business," Windpower 2011 Conference and Exhibition, Anaheim, CA, May 2011.
Botterud A., Wang J., Bessa R.J., Keko H., Miranda V., “Risk Management and Optimal Bidding for a Wind Power Producer, ” Proceedings IEEE Power and Energy Society, General Meeting, Minneapolis, MN, Jul. 2010.
Bessa R.J., Miranda V., Principe J.C., Botterud A., Wang J., “Information Theoretic Learning applied to Wind Power Modeling, ” Proceedings 2010 IEEE World Congress on Computational Intelligence, Barcelona, Spain, Jul. 2010.
Botterud A., Wang J., Bessa R.J., Miranda V., “Designing Electricity Markets with Large Shares of Wind Power,” Windpower 2010 Conference&Exhibition, Dallas, TX, May 2010.
Botterud A., Wang J., Miranda V., Bessa R., Keko H., Akilimali J.S., ”Wind Power Forecasting and Electricity Market Operations,” presentation at IAWind 2010 Conference, Ames, Iowa, April 6, 2010.
Wang J., Botterud A., Miranda V., Monteiro C., Sheble G., “Impact of Wind Power Forecasting on Unit Commitment and Dispatch,” 8th Int. Workshop on Large-Scale Integration of Wind Power into Power Systems, Bremen, Germany, Oct. 2009 (revised version).
Impact of Wind Power Forecasting on Unit Commitment and Dispatch
Botterud A., Wang J., Monteiro C., Miranda V., “Wind Power Forecasting and Electricity Market Operations,” 32nd Int. Association for Energy Economics (IAEE) Int. Conference, San Francisco, CA, June 2009.
Presentation: Wind Power Forecasting and Electricity Market Operations
Miranda V., Bessa R., Gama J., Conzelmann G., Botterud A., “New Concepts in Wind Power Forecasting Models, ” Windpower 2009 Conference and Exhibition, Chicago, IL, May 2009.
Wang J., Shahidehpour M., Li Z., “Security-Constrained Unit Commitment with Volatile Wind Power Generation,” IEEE Transactions on Power Systems, Vol. 23, No. 3, pp. 1319-1327, 2008.
Bessa R.J., Miranda V., Gama J., “Entropy and Correntropy Against Minimum Square Error in Offline and Online Three-Day Ahead Wind Power Forecasting,” IEEE Transactions on Power Systems, Vol. 24, No. 4, pp. 1657-1666, 2009.
In the News:
For more information, contact:
Dr. Audun Botterud
Center for Energy, Environmental, and Economic Systems Analysis (CEEESA)
Energy Systems Division
Argonne National Laboratory
9700 S. Cass Ave., Bldg. 202
Argonne, IL 60439