Degradation and State of Health Prediction of a Battery Used in a Microgrid in Real-time

Battery degradation is a complex physicochemical phenomenon intricately influenced by operational parameters and environmental factors. This paper delves into the degradation of lithium-ion batteries within microgrid systems, utilizing historical aging data from the University of Wisconsin-Madison to train a degradation model employing neural networks. The dataset is transformed to align with the battery pack rating pertinent to this investigation. The study identifies and examines several contributing factors to battery degradation, with a specific focus on temperature, charging current, discharging current, and depth of discharge. Emphasizing the significance of these parameters, particularly the latter three, the research endeavors to predict battery degradation in diverse operational conditions within a Microgrid system. This study modeled a Microgrid configuration capable of operating in islanded and grid-connected mode featuring solar PV, wind, a diesel generator, and a battery. Real-time simulations of different operational scenarios are conducted using the digital real-time simulator. Additionally, the battery's operational status is sent to a controller to predict its degradation and state of health. The controller used here is a Raspberry Pi, and communication is facilitated through UDP protocol.