Abstract
The lithium-ion batteries are recurrently used in a variety of applications. To assure an effective battery utilization and longer life the Battery Management Systems (BMSs) are employed. Recent BMSs are becoming sophisticated and causes a higher consumption overhead on the battery. To enhance the BMS power efficiency, this work exploits the input signal non-stationary nature. The idea is to employ event-driven sensing and processing. In contrast to the traditional counterparts, the battery cells parameters like voltages and currents are no more captured periodically but are acquired based on events. It results in significant real-time data compression. Afterward, this non-uniformly partitioned information is employed by a novel event-driven Coulomb counting algorithm for a real-time determination of the State of Charge (SoC). The estimated SoCis calibrated by using an original event-driven Open Circuit Voltage (OCV) to SoC curve relation. The devised system comparison is made with the traditional counterparts. Results demonstrate a more than third-orderof magnitude outperformance of the proposed system in terms of compression gain and computational efficiency while assuring an analogous SoC estimation precision.