No abstract

No abstract

<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>

<p>With the increasing penetration of renewable resources such as wind and solar, especially in terms of large-scale integration, the operation and planning of power systems are faced with great risks due to the inherent stochasticity of natural resources. Although this uncertainty can be anticipated, the timing, magnitude, and duration of fluctuations cannot be predicted accurately. In addition, the outputs of renewable power sources are correlated in space and time, and this brings further challenges for predicting the characteristics of their future behavior. To address these issues, this paper describes an unsupervised distribution learning method for renewable scenario forecasts that considers spatiotemporal correlation based on generative adversarial network (GAN), which has been shown to generate realistic time series for stochastic processes. We first utilize an improved GAN to learn unknown data distributions and model the dynamic processes of renewable resources. We then generate a large number of forecasted scenarios using stochastic constrained optimization. For validation, we use power generation data from the National Renewable Energy Laboratory wind and solar integration datasets. The simulation results show that the generated trajectories not only reflect the future power generation dynamics, but also correctly capture the temporal, spatial, and fluctuant characteristics of the real power generation processes. The experimental comparisons verify the superiority of the proposed method and indicate that it can reduce at least 50% of the training iterations of the generative model for scenario forecasts.<br></p>

No abstract