Temperature anomaly detection in Battery Energy Storage System (BESS) using machine learning methods

Authors:

Abstract
The common problem with Microgrid (MG) is the imbalance bet ween electricity supply and demand caused by renewable energy intermittency. Battery Energy Storage System (BESS) helps mitigate that problem by storing the energy during low demand and release it during peak. BESS provides a way to enhance the durability of the MG. However, the BESS in operation can suffer from excessive heat due to internal processes inside the battery. Continuous temperature stress can reduce the performance of BESS. It can lead to hazardous consequences, such as the risk of explosion. Hence, it is necessary to detect the temperature anomalies within the system. Recently, the data-driven method employed to detect battery anomalies includes unsupervised learning methods. In this work, the model obtains using the supervised learning method such as random forest (RF), support vector machine (SVM), and artificial neural network (ANN). Supervised learning provides a more accurate and fast learning model by incorporating prior domain knowledge of the system compared to the unsupervised one. Instead of using the battery temperature as in the previous works, the main feature of battery modeling in this research is the battery temperature rates of change. Therefore, it allows more depth modeling of battery thermal behavior as temperature rates of change governs the temperature dynamics. The result revealed that RF is slightly more accurate than ANN and SVR from the metrics model root mean square error (RMSE) and mean absolute error (MAE). But the plot of the prediction result and measurement data shows ANN is the fittest model compare with the other two. Thus, RMSE and MAE cannot be the primary reference to justify the model performance. Based on the quantitative and qualitative examination ANN model was selected to deploy in anomaly detection. The battery thermal model successfully locates some possible anomalies of battery cell temperature.

This publication has 13 references indexed in Scilit: