Fig 5 - uploaded by Yize Chen
Content may be subject to copyright.
Scenarios generated with J = 4, M = 4. (a) and (b) are the boxplots for generated scenarios' mean value (MW) and variance for four different generators respectively; in (c) we randomly select and plot the distinct generated scenarios for four generators.
Source publication
We present a method to generate renewable scenarios using Bayesian probabilities by implementing the Bayesian generative adversarial network~(Bayesian GAN), which is a variant of generative adversarial networks based on two interconnected deep neural networks. By using a Bayesian formulation, generators can be constructed and trained to produce sce...
Contexts in source publication
Context 1
... Fig. 5 we illustrate the simulation result using these 4 generators. In Fig. 5(a) and Fig. 5(b) we could observe that scenarios generated by generator 2 have much smaller variance and mean values compared with the other 3 genera- tors. Thus this trained generator could be used for generating scenarios for mild days. On the contrary, scenarios ...
Context 2
... Fig. 5 we illustrate the simulation result using these 4 generators. In Fig. 5(a) and Fig. 5(b) we could observe that scenarios generated by generator 2 have much smaller variance and mean values compared with the other 3 genera- tors. Thus this trained generator could be used for generating scenarios for mild days. On the contrary, scenarios generated by generator 1 represent most of the scenarios with frequent ...
Context 3
... Fig. 5 we illustrate the simulation result using these 4 generators. In Fig. 5(a) and Fig. 5(b) we could observe that scenarios generated by generator 2 have much smaller variance and mean values compared with the other 3 genera- tors. Thus this trained generator could be used for generating scenarios for mild days. On the contrary, scenarios generated by generator 1 represent most of the scenarios with frequent ramp events and ...
Context 4
... 3 genera- tors. Thus this trained generator could be used for generating scenarios for mild days. On the contrary, scenarios generated by generator 1 represent most of the scenarios with frequent ramp events and high generation output. Such differences of generated scenarios' dynamics are also depicted in the randomly selected output scenarios in Fig. 5(c). Wind Power [MW] Time [Day] Time [Day] Time [Day] Time [Day] Time [Day] Time [Day] Time [Day] Time [Day] (a) ...
Similar publications
Induction motors constitute the largest proportion of motors in industry. This type of
motor experiences different types of failures, such as broken bars, eccentricity, and inter-turn failure.
Stator winding faults account for approximately 36% of these failures. As such, condition monitoring
is used to protect motors from sudden breakdowns. This p...
In order to improve the limitation of the existing wind power grid connection quality assessment methods, a more reasonable combinatorial weight method considering the advantages of both subjective and objective is proposed. The objective function is to maximize the variance of the evaluation object under the combined weight, and to optimize the op...
Citations
... With the rapid development of artificial intelligence technology in power systems and energy industries [14][15], non-parametric approaches based on deep learning knowledge [16] have been widely used in scenario generation. Among them, Generative Adversarial Network (GAN) [17] including its improved models such as Wasserstein GAN [18][19], Sequence GAN [20][21], and Bayesian GAN [22][23], etc., is one of the most widely used scenario generation models in the field of deep learning. This family of models can learn probability distribution directly from historical data and generate new samples with similar probability distributions [24]. ...
With the high penetration of renewable generation systems in the power grid, the accurate simulation of the uncertainty in renewable energy generation is vital to the safe operation of the power system This paper proposes a novel controllable method for renewable scenario generation based on the improved VAEGAN model. The standard VAEGAN model is first improved using spectral normalization technique and the generator of GAN is trained using VAE. Then, the external and internal interpretable features in the latent space are learned as the controllable vector utilizing the principle of mutual information maximization. Finally, the renewable energy scenarios with overall features are generated using the external universal meteorological features, and renewable energy scenarios with specific features are generated by tuning along the internal interpretable feature of the controllable vector in the latent space. The proposed approach is used to produce real-time series data for renewable energy including wind and solar power. Experiments demonstrate that our method has a better performance in terms of controllable generation and enables the generation of preference patterns covering various statistical features.
... However, the quality of the generated scenarios in such approaches is highly limited by modeling and statistical assumptions [61], [62]. For instance, ARMA models face considerable limitations in modeling non-linear and non-stationary patterns due to their inherent assumptions of linear processes and data stationarity [63]. Wind power patterns, in particular, exhibit strong nonlinearity and non-stationarity [64], which become even more pronounced when dealing with very short timescales at long horizons (e.g., day-ahead). ...
In the rapidly evolving electricity market, wind energy portfolios are increasingly incentivized to actively participate in both the energy and reserve markets, driven by policy changes and advancements in wind farm control technology. This thesis empowers wind energy portfolios with the required decision framework to effectively engage in day-ahead energy and reserve markets. The first aspect of our framework is the day-ahead prediction of wind fluctuations at extremely short timescales, e.g., minutes or seconds, for optimal reserve scheduling, while also considering wind variability at rather longer timescales, e.g., hourly, for optimal energy scheduling. The second aspect of the framework involves developing a dedicated decision model that leverages the obtained information on wind uncertainty at both resolutions, to optimize the allocation of wind energy in day-ahead energy and reserve markets. Crucially, the decision framework also addresses the reliability of offered reserve services, ensuring system operators can confidently rely on them. Our results show that the proposed data-driven decision framework significantly improves both the profit and reserve reliability of wind energy portfolios.
... The scenarios can also be generated based on specific characteristics (e.g., high wind day, intense ramp events, or large forecasts errors) by using label information in the training process. In [22], Bayesian information is incorporated into the GANs to produce scenarios with different variance and mean value that capture different salient modes in the data. Even if wind and solar data are intentionally mixed, the generators can simultaneously distinguish and generate the respective wind and PV scenarios. ...
p>With the increasing penetration of renewable resources, such as wind and solar, the operation and planning of power systems, especially in large-scale integration, are faced with great risks due to the inherent stochasticity of natural resources. Although this uncertainty is anticipated, their timing, magnitude and duration cannot be predicted accurately. In addition, the renewable power outputs are correlated in space and time and bring further challenges in characterizing their behaviors. To address these issues, this paper provides a data-driven method to forecast renewable scenarios considering its spatiotemporal correlations based on generative adversarial networks (GANs), which has the ability to generated realistic samples from an unknown distribution making them one of the hottest areas in artificial intelligence research. We first utilize GANs to learn the intrinsic patterns and model the dynamic processes of renewable energy sources. Then by solving an optimization problem, we are able to generate large number of day-ahead forecasting scenarios. For validation, we use power generation data from NREL wind and solar integration data sets. The experimental results of this present research accord with the expectations.</p
... [30] extended the algorithm in [28] to the scenario forecasting problem and formulated an optimization problem aiming at generating scenarios conditioned on point forecasts. Apart from WGAN, variants of GANs [31,32] were also introduced for scenario generation task. The training of GANs is in a semi-supervised manner and doesn't need cumbersome manual labelings. ...
Because of environmental benefits, wind power is taking an increasing role meeting electricity demand. However, wind power tends to exhibit large uncertainty and is largely influenced by meteorological conditions. Apart from the variability, when multiple wind farms have geographical adjacency, their power generation also displays strong correlation. Thus, scenario generation considering the spatiotemporal relationships is a useful tool to model the stochastic process. In this work, we propose a wind power scenario generation framework based on the conditional improved Wasserstein generative adversarial network (WGAN). The framework includes a cluster analysis to establish the classification rule, a support vector classifier (SVC) to predict labels, a conditional scenario generation process based on improved WGAN, and finally a scenario reduction procedure. We demonstrate that the clustering analysis and SVC based labeling model can provide accurate classification results and the scenarios conditioned on input labels can not only follow the marginal distribution of each category but also capture the spatiotemporal relationships. We also illustrate that by adding a gradient penalty term to the discriminator’s loss function to enforce the Lipschitz constraint, the quality of scenarios is better than that of the existing method.
... Often, these techniques only allow for modeling either the temporal or the spatial relationship of renewable energy sources. More recently, research in scenario planning shows the strong capabilities of generative adversarial networks (GANs) modeling the temporal as well as the spatial relationship of, e.g., wind and photovoltaic (PV) farms [8,9,10] simultaneously. ...
... It is shown in [10], that GANs are capable of simulating scenarios conditioned on a previous forecast. In [8], Bayesian GANs create realistic scenarios for wind and PV simultaneously. In a sense, the approach in [8] is similar to ours, as we show the capability of GANs to simulate parks of different terrains together. ...
... In [8], Bayesian GANs create realistic scenarios for wind and PV simultaneously. In a sense, the approach in [8] is similar to ours, as we show the capability of GANs to simulate parks of different terrains together. ...
For the integration of renewable energy sources, power grid operators need realistic information about the effects of energy production and consumption to assess grid stability. Recently, research in scenario planning benefits from utilizing generative adversarial networks (GANs) as generative models for operational scenario planning. In these scenarios, operators examine temporal as well as spatial influences of different energy sources on the grid. The analysis of how renewable energy resources affect the grid enables the operators to evaluate the stability and to identify potential weak points such as a limiting transformer. However, due to their novelty, there are limited studies on how well GANs model the underlying power distribution. This analysis is essential because, e.g., especially extreme situations with low or high power generation are required to evaluate grid stability. We conduct a comparative study of the Wasserstein distance, binary-cross-entropy loss, and a Gaussian copula as the baseline applied on two wind and two solar datasets with limited data compared to previous studies. Both GANs achieve good results considering the limited amount of data, but the Wasserstein GAN is superior in modeling temporal and spatial relations, and the power distribution. Besides evaluating the generated power distribution over all farms, it is essential to assess terrain specific distributions for wind scenarios. These terrain specific power distributions affect the grid by their differences in their generating power magnitude. Therefore, in a second study, we show that even when simultaneously learning distributions from wind parks with terrain specific patterns, GANs are capable of modeling these individualities also when faced with limited data.
... Generative Adversarial Networks (GAN) is one of the training model frameworks to produce seemingly realistic data. The success of the GAN model in training two networks simultaneously makes this model accurate to use in the problem of detection of anomaly [26], [27], medical time series that have real value using conditional [28], Stock Market [29], portfolio management [30], [31], stock price manipulation detection [32], Estimated Renewable Energy Scenario [33], [34]. ...
The success of the work of Generative Adversarial Networks (GAN) has recently achieved great success in many fields, such as stock market prediction, portfolio optimization, financial information processing and trading execution strategies, because the GAN model generates seemingly realistic data with models generator and discriminator .Planning for drug needs that are not optimal will have an impact on hospital services and economics, so it requires a reliable and accurate prediction model with the aim of minimizing the occurrence of shortages and excess stock, In this paper, we propose the GAN architecture to estimate the amount of drug sales in the next one week by using the drug usage data for the last four years (2015-2018) for training, while testing using data running in 2019 year , the classification results will be evaluated by Actual data uses indicators of Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). From the results of the experiment, seen from the value of MAE, RMSE and MAPE, the proposed model has promising performance, but it still needs to be developed to explore ways to extract factors that are more valuable and influential in the trend disease progression, thus helping in the selection of optimal drugs
... This method can generate scenarios for spatiotemporally correlated multiple sites by learning the distribution of historical data; such scenarios can also be generated based on conditioned information, such as mean values, ramp events, and forecast errors in wind power. In [27], Bayesian GANs are constructed and trained to produce a set of renewable power scenarios with the same pattern as historical data. Even if wind and solar data are deliberately blended, the method can simultaneously distinguish and generate different scenarios, allowing better representation of renewable power production processes. ...
Wind power scenarios have a significant impact on stochastic optimization problems for power systems in which wind power is a significant component. Generative adversarial networks (GANs) are a powerful class of generative models, and can generate realistic scenarios for renewable power sources without the need for any modeling assumptions. However, the performance of GANs in generating scenarios can be further improved by modifying the way in which the Lipschitz constraint on discriminator network is imposed. Another critical problem of applying deep neural networks is overfitting, a phenomenon especially prone to appear on small training sets. In this paper, we propose an improved GAN for the generation of wind power scenarios. To improve the training speed, we use a gradient penalty term to enforce a Lipschitz constraint based on the output and input of the discriminator network. To improve the scenario quality, we further use a consistency term in the training procedure. Besides, the overfitting problem can be effectively alleviated by the enforced Lipschitz continuity. The proposed method is applied to actual time series data from the NREL wind integration dataset. The experimental results demonstrate that our method outperforms existing methods.
