A Temperature Prediction Method of Valve-regulated Lead-acid Battery

To accurate predict valve regulated lead acid battery (VRLA) temperature changes to prevent overheat without information of battery’s mechanical structure, a thermal model for VRLA is developed. Four thin film platinum resistances (RTDs) temperature detectors are located on the battery’s core, case’s surface, terminal and outer surface of insulating material for extracting thermal parameters with using multi-objective genetic algorithm. Nine tests are completed to verify the model accuracy is acceptable for long-term prediction, 30000 seconds. A model-based prediction with RMSE of about 0.5°C and a maximum error of less than 1°C is created. The prediction method can be done without requiring any thermocouple or RTDs on battery after the parameters are identified. The method can also be used to estimate the battery’s internal temperature to prevent some tragic situations such as thermal runaway. The proposed model can run on microcontroller after the parameters have been obtained which means that it can be embedded into a battery management system for continuous temperature prediction and internal temperature estimation.