2023 ASEE Annual Conference & Exposition

Smart Grids represent one of the most suitable and relevant applications in a sustainable scenario, where it is of high importance to provide electricity to millions of customers using advance technology and efficient methods. Microgrids are key subsystems forming the Smart Grid and, as a consequence, the study and analysis of microgrids and the methods to improve its operation are crucial for the optimal performance of Smart Grids. Microgrid infrastructure may be seen as a combination of information technology and equipment, becoming a target for cybersecurity attacks, vulnerable both in software and hardware. The importance of microgrids as part of the smart power system makes them become a target for cybersecurity attacks. Abrupt disconnection from the smart grid, failure in equipment and the technical problems with the control and telecommunication elements are among the multiple issues that a microgrid faces in a cyber secure scenario. Thus, microgrids must become more and more resilient to cyberattacks. In research areas associated with Smart Grids, the U.S. National Science Foundation (NSF) has offered support for projects for workforce development and research in cybersecurity, state estimation and optimization in electrical microgrids through several programs. Further research is needed specially using real dataset. This paper focuses on the design of laboratory experiences on microgrid State Estimation, Optimization and Cybersecurity for a proposed Massively Open Online Course (MOOC). In a post-pandemic scenario, the design and implementation of MOOCs became a valuable tool to reach students and professionals around the world. For the present study, simulations for cybersecurity cover the utilization of real dataset associated with the electrical power system of the Dominican Republic by means of deep learning tools offered by the MATLAB software. Thus, students and instructors may be able to count on a resource to learn and expand their knowledge on microgrids, using real data. This paper presents the main ideas associated with the design of the MOOC, as well as some results obtained from those simulations implemented in MATLAB that are included as part of the MOOC and the results associated with reinforcement learning techniques for optimization in a microgrid that is operating in stand-alone mode. The cybersecurity simulations display the capacity of the MATLAB software to analyze sets of data to classify the different results into a variety of categories, training a neural network and allowing for the analysis of new data points in order to classify whether or not an electrical system is subject to a cyberattack. All this relevant research work has been funded by the Engineering Postdoctoral Fellowship eFellows program, administer by the American Society of Engineering Education, funded by the National Science Foundation (NSF). The MOOC is planned to be offered as a free resource for the community. The real datasets used for the Cybersecurity simulations will be available in an Open Science website.