Fault diagnosis of lithium-ion battery sensors based on multi-method fusion

The lithium-ion battery serves as the nucleus of the new energy vehicle, playing a pivotal role in energy storage. The acquisition of sensor data from the battery holds paramount importance for the seamless functioning of new energy vehicles. Therefore, the real-time identification of faults in battery sensors becomes imperative to proactively prevent more severe lithium-ion battery failures. A proposed approach for typical fault diagnosis of battery voltage and current sensors involves an enhanced central differential multi-new interest adaptive traceless Kalman filter fusion Monte Carlo algorithm. This method compares residuals and thresholds to ascertain the occurrence of faults, enhancing robustness while minimizing estimation errors. Subsequently, the fault diagnosis for the battery temperature sensor is executed through the deep limit learning machine algorithm, coupled with wavelet energy spectrum fusion nonlinear ocean predator. This approach maintains an accuracy rate exceeding 90 %, even with adaptive sample size selection. In conclusion, a hardware-in-the-loop simulation verification platform utilizing the NI cRIO-9039 controller is established to confirm the algorithm's applicability in real vehicles.