A battery storage system for fault tolerance

Many a times the failure of a control system is attributed to malfunctioning or non-functioning of battery. Even while not in active use a battery discharges and the health of battery deteriorates. The task of identifying non healthy battery for replacement or an uncharged battery to put on charge is very essential and is performed manually which is not only time consuming but if not performed timely can lead to system breakdown. In the present paper we describe models of SoC and SoH of battery using a neuro-fuzzy and regression techniques respectively to model the charging process of the battery. We also present a configuration using models that connects only healthy batteries to an application to support the fault tolerance feature.