State of charge estimation of lead acid batteries using neural networks

State of charge estimation of lead acid batteries is critical for applications that rely on backup power like UPS sytems for telcom operators, data centers, renewable energy installations. Estimating the state of charge of a lead acid battery is a tough job mainly because there is no clear way to predict the behavior of the batteries. One of the common methods of estimation of state of charge of a lead acid battery is by means of measuring the terminal voltage of a lead acid battery. This is not very accurate, but due to its ease of use, it is widely adopted. The current reasearch into estimating the state of charge of a battery is by means of using intelligent control techniques. This paper proposes a methodology by using neural networks. By relying on the parameters of terminal voltage of the battery, load current, temperature of the battery, specific gravity of the electrolyte a methodology is proposed. An Adaptive Neuro Fuzzy Inference System is used here. This system improves the accuracy of the existing methods by using specific gravity when the battery is discharging. The developed topology was tested by using sample data obtained from a manufacturer's datasheet and the benifits of multiparameterized approach was vidicated.