The application of genetic algorithms to parameter estimation in lead-acid battery equivalent circuit models

This thesis summarises the research work in the development of the battery status estimation algorithm. A model was developed to describe the process of battery discharge. Genetic Algorithms were used as a tool to identify the parameters of the battery, including the internal resistances, SOC, and capacity. Simulation results show that the model is able to adequately simulate the battery discharge process. The aforementioned models were extended to a further investigation of the batteries state of health. There is a link between the status of battery health and the internal resistance. Six batteries were discharged and charged to simulate the capacity loss occurs in normal operation, which is related to the state of health, The parameter estimation was able to adequately distinguish between different state of health. These results indicate that the internal resistance increases when the state of health drops. This progress is at first slow when the battery is new but the becomes faster when the remaining capacity of battery drops to about 75% of the initial. It is found in the thesis that the value of internal resistance is increased by 25% approximately when the state of health is brought down to about 50%.