Generative Adversarial Networks for Operational Scenario Planning of Renewable Energy Farms: A Study on Wind and Photovoltaic

Schreiber, Jens; Jessulat, Maik; Sick, Bernhard

doi:10.1007/978-3-030-30508-6_44

“…There are multiple works that propose different versions of the different DGMs reviewed above for energy scenario generation. Applications of GANs and WGANs include wind power generation [17], [19], PV power generation [16], [18], [20]- [22], and residential loads [25]. VAEs were applied to learn the distributions of PV and wind power generation [15], concentrated solar power [23], and electric vehicle power demand [24].…”

Section: Dgm-based Scenario Generationmentioning

confidence: 99%

“…A popular modification of GANs are so-called Wasserstein GANs (WGANs), i.e., to use Wasserstein loss functions [14]. Applications of VAEs and (W)GANs include learning distributions of PV and wind power generation [15]- [22], concentrated solar power generation [23], electric vehicle power demand [24], and residential load [25].…”

Section: Introductionmentioning

confidence: 99%

Validation Methods for Energy Time Series Scenarios From Deep Generative Models

Cramer

¹

,

Gorjão

²

,

Mitsos

³

et al. 2022

View full text Add to dashboard Cite

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.

show abstract

“…In the literature on energy scenario generation, most authors evaluate the PDF on a linear scale (Gu et al, 2019;Jiang et al, 2018Jiang et al, , 2019Schreiber et al, 2019;Wang et al, 2018;Zhang et al, 2020;Wei et al, 2019;Qi et al, 2020;Pan et al, 2019;Zhanga et al, 2018) or the integral over the PDF, i.e., the cumulative distribution function (CDF) (Chen et al, 2018b;Jiang et al, 2019). However, the linear scaled PDF and the CDF can only show differences between historical and generated scenarios for values of high likelihood.…”

Section: Probability Density Functionmentioning

confidence: 99%

“…A popular modification of GANs are so-called Wasserstein GANs (WGANs), i.e., to use Wasserstein loss functions (Arjovsky et al, 2017). Applications of VAEs and (W)GANs include learning distributions of PV and wind power generation (Zhanga et al, 2018;Chen et al, 2018b;Jiang et al, 2018;Wei et al, 2019;Zhang et al, 2020;Chen et al, 2018a;Jiang et al, 2019;Schreiber et al, 2019), concentrated solar power generation (Qi et al, 2020), electric vehicle power demand (Pan et al, 2019), and residential load (Gu et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

Validation Methods for Energy Time Series Scenarios from Deep Generative Models

Cramer¹,

Gorjão²,

Mitsos³

et al. 2021

Preprint

0

View full text Add to dashboard Cite

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.

show abstract

“…Besides the generation of wind and PV scenarios, GAN-based scenario generation was also applied to residential load forecasts (Gu et al, 2019) and hydrowind-solar hybrid systems (Wei et al, 2019). Schreiber et al (2019) study different loss functions for GANs and found the Wasserstein distance to be superior to the binary-cross-entropy. Besides GANs, a popular type of DGMs are variational autoencoders (VAEs) (Kingma and Welling, 2014).…”

Section: Introductionmentioning

confidence: 99%

Principal Component Density Estimation for Scenario Generation Using Normalizing Flows

Cramer¹,

Mitsos²,

Tempone³

et al. 2021

Preprint

View full text Add to dashboard Cite

Neural networks-based learning of the distribution of non-dispatchable renewable electricity generation from sources such as photovoltaics (PV) and wind as well as load demands has recently gained attention. Normalizing flow density models have performed particularly well in this task due to the training through direct log-likelihood maximization. However, research from the field of image generation has shown that standard normalizing flows can only learn smeared-out versions of manifold distributions and can result in the generation of noisy data. To avoid the generation of time series data with unrealistic noise, we propose a dimensionality-reducing flow layer based on the linear principal component analysis (PCA) that sets up the normalizing flow in a lower-dimensional space. We train the resulting principal component flow (PCF) on data of PV and wind power generation as well as load demand in Germany in the years 2013 to 2015. The results of this investigation show that the PCF preserves critical features of the original distributions, such as the probability density and frequency behavior of the time series. The application of the PCF is, however, not limited to renewable power generation but rather extends to any data set, time series, or otherwise, which can be efficiently reduced using PCA.

show abstract

Generative Adversarial Networks for Operational Scenario Planning of Renewable Energy Farms: A Study on Wind and Photovoltaic

Cited by 13 publications

References 23 publications

Validation Methods for Energy Time Series Scenarios From Deep Generative Models

Validation Methods for Energy Time Series Scenarios From Deep Generative Models

Validation Methods for Energy Time Series Scenarios from Deep Generative Models

Principal Component Density Estimation for Scenario Generation Using Normalizing Flows

Contact Info

Product

Resources

About