Artificial neural network for state-of-charge estimation of Ni-MH batteries with photovoltaic power system

We developed a Ni-MH battery module and combined it with a photovoltaic panel as a hybrid power system for meteorological sensors. When the battery is charged with small current, such as less than 5 amperes (0.05 C), a conventional method on the state of charge (SOC) of the battery calculated from the summation of the current of the battery is inaccurate. For estimation of the SOC of the battery module, we therefore introduced an artificial neural network model in which the input data were the temperature and voltage of the battery module. It appears that the new estimation method is more accurate than the conventional one.