A Temperature-sensitive Battery Internal Resistance Model for accurate SOC estimation

— An estimated 75 billion connected devices are deployed worldwide. Most of these devices are deployed as Wireless Sensor Networks (WSN) with internet connected gateways or directly connected Internet of Things (IoT) devices. However, most of these devices are battery powered, have limited or no energy harvest capabilities, and are deployed under harsh environmental conditions. Therefore, estimating the remaining battery energy typically in the form of State of Charge (SOC) is crucial for device lifetime planning. As batteries are non-linear devices that are susceptible to performance changes due to ambient temperature changes, measuring the impact of temperature changes on battery’s internal resistance is crucial in accurate estimation. This research proposes a mathematical model extension to Tremblay’s model by integrating a temperature-sensitive battery resistive coefficient for Lithium-Ion batteries. The proposed model provides computationally efficient coefficients to map temperature variations, reporting a precision of 97% compared to the standard model. These enhanced model equations were then used to design an extended Kalman Filter (EKF) to estimate SOC of Li-Ion batteries at a range of temperatures, providing an RMSE of 0.983.