Extended operating range for reduced-order model of lithium-ion cells

• Battery-management systems require accurate cell models over a wide operating range.
• A previously reported physics-based model is accurate over only a narrow range.
• We use model blending to extend the operating window of the physics-based model.
• We show that the blended model is stable and accurate over a broad operating range.

In a previous paper, we developed a method to produce a physics-based one-dimensional discrete-time state-space reduced-order model (ROM) of a lithium-ion cell. The method relies on linearizing the standard porous-electrode equations around a fixed state-of-charge (SOC) and operating temperature setpoint. The ROM is able to track a highly dynamic input accurately near the linearization setpoint, but its performance degrades as either the cell's SOC or temperature move away from this linearization point.This paper describes a way to extend the accuracy of the ROM over a wide range of SOCs and temperatures using a model-blending approach. Our results demonstrate that the approach accurately models the cell's voltage and internal electrochemical variables over a wide range of temperature and SOC, with little added computational complexity.