Simplified Simulation of Electric Vehicle Battery packs Using deep learning based Reduced Order Models

Accurate yet computationally efficient simulation of electric vehicle (EV) batteries is essential for system design, performance evaluation, and real-time control applications. This paper presents a simplified simulation framework for lithium-ion EV battery systems using reduced order models (ROMs) that balance fidelity and computational speed. The EV system simulation environment has been developed in MATLAB/Simulink environment to analyze Battery Management System (BMS) control system design and EV system performance. This simulation platform consists of BMS and other important EV controller models and high-fidelity plant models for battery and powertrain systems, which are modelled using Simulink and Simscape library blocks. Performing virtual testing, control design and system-level analysis with such high-fidelity plant models poses significant computational challenges and poor simulation performance. The focus of this paper is to enhance system level simulation performance by leveraging ROM modelling using Artificial Intelligence (AI) techniques. In this study, the high computational high-fidelity EV battery plant model, which was originally modelled using Simscape battery library block has been replaced with trained Neural State Space (NSS) ROM model using MATLAB/Simulink tool. This low-order nonlinear NSS ROM model is developed using deep learning technique that serves as a surrogate for a high-fidelity battery model. Trained ROM model has been deployed in Simulink system level Simulation platform and compared it against the high-fidelity model. Simulation results show that the ROM model captures the essential behaviour of the battery with significantly lower computational cost compared to the high-fidelity Simscape model. The proposed approach significantly reduces simulation time while maintaining acceptable accuracy for various drive cycles and operating conditions. Validation against high fidelity model demonstrates the potential of ROM for BMS and large-scale vehicle simulations. This work highlights the practicality of ROM model as essential approach for advancing EV battery analysis, design optimization, and control strategy development.