Evolutionary Sensor Placement for Spatiotemporal Modeling of Battery Thermal Process

Spatiotemporal modeling is critical to the simulation, optimization, and control of the thermal process of a lithium-ion battery, which is a typical kind of distributed parameter system (DPS). Data-driven spatiotemporal modeling methods are of practical interest to construct an analytical model of the thermal process of a lithium-ion battery, since they only need some sampled data rather than the structure descriptions or parameters of a DPS. How to sample data optimally for data-driven spatiotemporal modeling is still an open question. In this article, with the aim of minimizing the spatiotemporal modeling error, we propose a novel evolutionary algorithm to optimally place sensors for data sampling. First, an objective function that can quantify both the spatial error and the temporal error is designed. Additionally, a novel differential evolution algorithm with two kinds of encoding mechanisms (called DETEM) is proposed to optimize the objective function. Numerical simulations and experimental studies have shown that the proposed method is competitive. Besides, both the objective function and DETEM are critical to the proposed method. In summary, the proposed method provides an effective way to obtain the optimal sensor placement for spatiotemporal modeling of the thermal process of a lithium-ion battery.