The COVID-19 pandemic represents not just a global health crisis, but may signal the beginning of a new era of economic activity, the potential consequences of which we currently do not fully understand. In this context, the mid-to-long-range impacts of the pandemic on the energy sector have been studied extensively in the last few months. Despite these efforts, the pandemic still raises many open questions concerning the long-term operation and planning of power systems. For instance, how will the pandemic affect the integration of renewable energy sources? Should current power system expansion plans change in light of the COVID-19 pandemic? What new tools should be provided to support system operators during global health crises? It is the purpose of this paper to better understand the many aspects of these open questions by reviewing the relevant recent literature and by analyzing measured data. We point out the main challenges that the pandemic introduced by presenting patterns of electricity generation and demand, frequency deviations, and load forecasting. Moreover, we suggest directions for future research that may assist in coping with the mentioned challenges. We hope that this paper will trigger fruitful discussions and encourage further research on these important emerging topics.

Advanced machine learning techniques have recently demonstrated outstanding performance when applied to power quality disturbance (PQD) classification. Nevertheless, a possible problem is that power experts may find it hard to trust the results of such algorithms, if they do not fully understand the reasons for their outputs. In this light, this article suggests a method that explains the outputs of PQD classifiers, using explainable artificial intelligence (XAI). The method operates as follows: first, various XAI techniques and classifiers are combined and scored based on their explanations during the validation step. Then, the best combination of classifier and XAI technique for each disturbance is used on the testing set, such that the classifier outputs are more transparent. To accomplish these steps, a definition for the correct explanation in PQD is given. Also, to determine the quality of an explanation for a certain output, we propose an evaluation process to measure the explainability score for each XAI technique and classifier. By means of this approach, the PQD classifier outputs are optimized to be both accurate and easy to understand, allowing experts to make informed and trustworthy decisions.

In this work, we review papers that employ game theoretic tools to study the operation and design of modern electric grids. We consider four topics in this context: energy trading, energy balancing, grid planning, and system reliability, and we demonstrate the advantages of using game-theoretic approaches for analyzing complex interactions among independent players. The results and conclusions provide insights regarding many aspects of design and operation, such as efficient methodologies for expansion planning, cyber-security, and frequency stability, or fair-benefit allocation among players. A central conclusion is that modeling the system from the perspective of one entity with unlimited information and control span is often impractical, so correct modeling of the selfish behavior of independent players may be critical for the development of future power systems. Another conclusion is that correct usage of incentives by appropriate regulation or sophisticated pricing mechanisms may improve the social welfare, and, in several cases, the results obtained are as good as those obtained by central planning. Using an extensive content analysis, we point to several trends in the current research and attempt to identify the research directions that are currently at the focus of the community.

Non-intrusive load monitoring is an algorithm or process that disaggregates the total power in a facility to identify consumption of individual appliances. In this paper, a new algorithm is proposed to classify events of appliance states based on modification of the Cross-Entropy method. The main contribution is a formulation and solution of the problem with the Cross-Entropy method as a constrained optimization problem. This new technique is called the Modified Cross-Entropy (MCE) method. The proposed algorithm is simple, as it operates in real time, using low rate sampling of the active power. In addition, there is no need to train the system, or to use complex hardware. The algorithm is tested using the REDD and AMPds datasets, and the results are compared to several state-of-the-art techniques.