Validation Methods for Energy Time Series Scenarios From Deep Generative Models

The design and operation of modern energy systems are heavily influenced by time-dependent and uncertain parameters, e.g., renewable electricity generation, load-demand, and electricity prices. These are typically represented by a set of discrete realizations known as scenarios. A popular scenario generation approach uses deep generative models (DGM) that allow scenario generation without prior assumptions about the data distribution. However, the validation of generated scenarios is difficult, and a comprehensive discussion about appropriate validation methods is currently lacking. To start this discussion, we provide a critical assessment of the currently used validation methods in the energy scenario generation literature. In particular, we assess validation methods based on probability density, auto-correlation, and power spectral density. Furthermore, we propose using the multifractal detrended fluctuation analysis (MFDFA) as an additional validation method for non-trivial features like peaks, bursts, and plateaus. As representative examples, we train generative adversarial networks (GANs), Wasserstein GANs (WGANs), and variational autoencoders (VAEs) on two renewable power generation time series (photovoltaic and wind from Germany in 2013 to 2015) and an intra-day electricity price time series form the European Energy Exchange in 2017 to 2019. We apply the four validation methods to both the historical and the generated data and discuss the interpretation of validation results as well as common mistakes, pitfalls, and limitations of the validation methods. Our assessment shows that no single method sufficiently characterizes a scenario but ideally validation should include multiple methods and be interpreted carefully in the context of scenarios over short time periods.

[1]  Stefan Feuerriegel,et al.  Short-Term Dynamics of Day-Ahead and Intraday Electricity Prices , 2017 .

[2]  Weiping Zhu,et al.  Data-Driven EV Load Profiles Generation Using a Variational Auto-Encoder , 2019, Energies.

[3]  S. Feuerriegel,et al.  Contract Durations in the Electricity Market: Causal Impact of 15 Min Trading on the EPEX SPOT Market , 2017 .

[4]  Don H. Johnson,et al.  Gauss and the history of the fast Fourier transform , 1984, IEEE ASSP Magazine.

[5]  Chongqing Kang,et al.  GAN-based Model for Residential Load Generation Considering Typical Consumption Patterns , 2019, 2019 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT).

[6]  Birgitte Bak-Jensen,et al.  Modeling Daily Load Profiles of Distribution Network for Scenario Generation Using Flow-Based Generative Network , 2020, IEEE Access.

[7]  Qian Ai,et al.  Typical wind power scenario generation for multiple wind farms using conditional improved Wasserstein generative adversarial network , 2020 .

[8]  Elsa Arcaute,et al.  Multifractal methodology , 2016, 1606.02957.

[9]  Wei Hua,et al.  Optimized Operation of cascade reservoirs Considering Complementary Characteristics between wind and photovoltaic based on variational auto-encoder , 2018 .

[10]  H. Stanley,et al.  Multifractal Detrended Fluctuation Analysis of Nonstationary Time Series , 2002, physics/0202070.

[11]  Shakir Mohamed,et al.  Variational Inference with Normalizing Flows , 2015, ICML.

[12]  Wojciech Zaremba,et al.  Improved Techniques for Training GANs , 2016, NIPS.

[13]  E. Bacry,et al.  The Multifractal Formalism Revisited with Wavelets , 1994 .

[14]  Yang Long,et al.  Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models , 2021, IEEE transactions on pattern analysis and machine intelligence.

[15]  S. Havlin,et al.  Detecting long-range correlations with detrended fluctuation analysis , 2001, cond-mat/0102214.

[16]  Henrik Madsen,et al.  Integrating Renewables in Electricity Markets: Operational Problems , 2013 .

[17]  T. Brown,et al.  Synergies of sector coupling and transmission reinforcement in a cost-optimised, highly renewable European energy system , 2018, Energy.

[18]  Werner Römisch,et al.  Scenario Reduction Algorithms in Stochastic Programming , 2003, Comput. Optim. Appl..

[19]  E. Lorenz,et al.  Short term fluctuations of wind and solar power systems , 2016, 1606.03426.

[20]  Hu Wei,et al.  Short-term optimal operation of hydro-wind-solar hybrid system with improved generative adversarial networks , 2019, Applied Energy.

[21]  A. Conejo,et al.  Decision making under uncertainty in electricity markets , 2010, 2006 IEEE Power Engineering Society General Meeting.

[22]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[23]  D. Witthaut,et al.  MFDFA: Efficient multifractal detrended fluctuation analysis in python , 2021, Comput. Phys. Commun..

[24]  Yi Chai,et al.  Forecasting Spatio-Temporal Renewable Scenarios: a Deep Generative Approach , 2019, ArXiv.

[25]  Daniel Kirschen,et al.  Model-Free Renewable Scenario Generation Using Generative Adversarial Networks , 2017, IEEE Transactions on Power Systems.

[26]  E. Bacry,et al.  Singularity spectrum of fractal signals from wavelet analysis: Exact results , 1993 .

[27]  R. Weron Electricity price forecasting: A review of the state-of-the-art with a look into the future , 2014 .

[28]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[29]  Huahua Chen,et al.  Variational Autoencoder With Optimizing Gaussian Mixture Model Priors , 2020, IEEE Access.

[30]  Yee Leung,et al.  Multifractal temporally weighted detrended fluctuation analysis and its application in the analysis of scaling behavior in temperature series , 2010 .

[31]  M. Rosenblatt Remarks on Some Nonparametric Estimates of a Density Function , 1956 .

[32]  Petre Stoica,et al.  Spectral Analysis of Signals , 2009 .

[33]  S. Wallace,et al.  Evaluation of scenario-generation methods for stochastic programming , 2007 .

[34]  Yi Chai,et al.  Scenario Generation for Wind Power Using Improved Generative Adversarial Networks , 2018, IEEE Access.

[35]  Wei Hu,et al.  Optimal configuration of concentrating solar power in multienergy power systems with an improved variational autoencoder , 2020 .

[36]  Artur M. Schweidtmann,et al.  Model-based bidding strategies on the primary balancing market for energy-intense processes , 2019, Comput. Chem. Eng..

[37]  Léon Bottou,et al.  Towards Principled Methods for Training Generative Adversarial Networks , 2017, ICLR.

[38]  E. Bacry,et al.  Wavelets and multifractal formalism for singular signals: Application to turbulence data. , 1991, Physical review letters.

[39]  Florian Ziel,et al.  Variance Stabilizing Transformations for Electricity Spot Price Forecasting , 2018, IEEE Transactions on Power Systems.

[40]  John D. Hunter,et al.  Matplotlib: A 2D Graphics Environment , 2007, Computing in Science & Engineering.

[41]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[42]  Ali Borji,et al.  Pros and Cons of GAN Evaluation Measures , 2018, Comput. Vis. Image Underst..

[43]  Qiuwen Zhang,et al.  A Modified Multifractal Detrended Fluctuation Analysis (MFDFA) Approach for Multifractal Analysis of Precipitation in Dongting Lake Basin, China , 2019, Water.

[44]  Feng Liu,et al.  A Conditional Model of Wind Power Forecast Errors and Its Application in Scenario Generation , 2017, 1708.06759.

[45]  Bernhard Sick,et al.  Generative Adversarial Networks for Operational Scenario Planning of Renewable Energy Farms: A Study on Wind and Photovoltaic , 2019, ICANN.

[46]  Aaron C. Courville,et al.  Improved Training of Wasserstein GANs , 2017, NIPS.

[47]  David Pfau,et al.  Unrolled Generative Adversarial Networks , 2016, ICLR.

[48]  Pan Li,et al.  Bayesian renewables scenario generation via deep generative networks , 2018, 2018 52nd Annual Conference on Information Sciences and Systems (CISS).

[49]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[50]  E. Parzen On Estimation of a Probability Density Function and Mode , 1962 .

[51]  Michael Baldea,et al.  Challenges in process optimization for new feedstocks and energy sources , 2018, Comput. Chem. Eng..

[52]  P. Welch The use of fast Fourier transform for the estimation of power spectra: A method based on time averaging over short, modified periodograms , 1967 .

[53]  Jeff Dean,et al.  Time Series , 2009, Encyclopedia of Database Systems.

[54]  Soumith Chintala,et al.  Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks , 2015, ICLR.