This article examines the emerging challenges in modeling and analyzing the electric power system due to the widespread growth of variable renewable energy (VRE), particularly in the form of distributed energy resources (DERs), which are displacing traditional large power plants. Many of these resources are connected to the system through power electronic interfaces, also known as inverter-based resources (IBRs), which are reshaping the system dynamics and lowering the grid strength and inertia. Understanding the dynamic behavior of the power system should be critical to addressing the potential stability concerns, refining the grid requirements, and developing effective and reliable measures among many alternatives. However, conventional methodologies for resource integration and network expansion studies, as well as application-specific electromagnetic transient (EMT) studies, need to be improved. This article thus presents recent academic and industrial efforts to advance the existing approaches, especially by incorporating the uncertainty in model parameters of DERs, variability of VRE, and EMT dynamics of IBRs for the grid planning and operations studies such as the impact of DERs on load modeling and system-wide dynamic performance. In addition, this article showcases recent developments to expand the study boundaries by synergizing the strengths of the industry-accepted approaches along with real system studies for Korea’s electric power systems in particular.

This article proposes a deep neural network (DNN) framework for multivariate deterministic power forecasting in the context of the high penetration of variable and uncertain renewable energy sources. The deep learning model is organized based on the 1-D convolutional neural network to lessen the computational burden, typical of recurrent neural network based models, and combines WaveNet and EfficientNet to improve the forecasting accuracy. Motivated by the inefficiency that all the models conduct the same tasks in the popular ensemble approach, we also designed a feedforward error learning DNN, which computes the error of the basic model separately. We further incorporated embedded and filter methods for feature selection to enhance the model visibility and the utility of the framework. Comprehensive studies on the public load and PV datasets demonstrate that the proposed framework outperforms the conventional methods in applicability, computational efficiency, and forecasting accuracy.