Short-Circuit Fault Diagnosis and State Estimation for Li-ion Battery using Weighting Function Self-Regulating Observer

A wrong estimation of Li-ion battery state of charge (SOC) can seriously affect the safety, the availability and the reliability of electrical-powered systems particularly for embedded applications. One of the main cause is the occurrence of non-detected battery soft short-circuit (SC). Hence we propose to develop an observer that can simultaneously estimate the SOC and the SC current. To deal with the open-circuit voltage (OCV)-SOC non-linear characteristic, we have used the Takagi-Sugeno fuzzy model to decompose the characteristic into segments for different SOC ranges. The membership functions are optimised with a genetic algorithm. Further on each segment a proportional-integral observer is designed by solving Linear Matrix Inequalities (LMI). Thanks to the slowly inherent varying SOC, a weighting function self-regulating mechanism is used along with Takagi-Sugeno fuzzy model to merge the g observers. Intensive numerical simulations using real cell parameters show that this structure is able to accurately estimate the SOC and the short circuit current at an early stage.