Adaptive learning of SOC applied on LiP battery pack

摘要

For most determinations of the State Of Charge (SOC), executed within a controller located in the battery pack, the SOC is a parameter that is updated on regular time bases and that, especially with battery technologies like NiMH and LiP, keeps track of the time the battery pack is not used. In most cases this is accomplished by using a real-time clock that determines the off-time and corrects the SOC at power up of the battery pack accordingly. There is a history that links the controller with the battery pack. This paper will present another methodology that, every time at startup, will determine the SOC by applying adaptive learning. The advantage is that the embedded controller does not require to be specifically 'initialized' when integrated into the battery pack and that the time when the battery pack is not being used (self-discharge of the batteries) does not need to be tracked. This methodology has been applied and tested on a specific application (Solar car participation on the World Solar Challenge with more than 3,000km over the Australian continent) with LiP batteries. Information of every individual cell was obtained by Sentinel System-on-Chip Battery Monitoring System combined with CAN-based embedded controller.