Publications

Abstract : This paper elucidates an effective fault classification method for lithium-ion batteries employed in electric vehicles. Lithium-ion batteries are eccentric in nature with non-linear behavior which may lead to dreadful events, especially in the case of electric vehicles. This work takes into consideration the eventual, thermal runaway fault and the battery-life degrading faults such as the under-voltage and overvoltage faults. A mathematical model of the lithium-ion battery is deployed. The model is then exercised to collect sufficient data under conventional and non-conventional operating conditions. By supervised learning, a layered perceptron is trained, effectively enabling the system to efficiently classify the faults in lithium-ion batteries. Multi-Layer perceptron is employed in this work as the classifying algorithm which is trained using backpropagation. The validation of the trained algorithm is executed in a software workbench and, the discussions and results are presented.