Semester of Graduation

Spring 2025

Degree

Master of Science (MS)

Department

Electrical Engineering

Document Type

Thesis

Abstract

This thesis investigates the application of deep learning models for State of Charge (SOC) estimation in Battery Management Systems (BMS) for electric vehicles (EVs), focusing on optimizing EV range, lifespan, and performance while addressing challenges like range anxiety. The study explores three deep learning architectures—Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), and Gated Recurrent Unit (GRU)—each designed to capture complex temporal dependencies in battery data. The LSTM model is trained on EV battery data, including voltage, current, temperature, and SOC, providing a strong baseline for SOC estimation. The BiLSTM model enhances accuracy by processing data in both forward and backward directions, thus capturing contextual relationships within the data more effectively. The GRU model offers a simpler design that maintains reasonable accuracy, focusing on computational efficiency for real-time BMS applications. A hybrid deep learning model is also introduced, combining BiLSTM layers with fully connected layers (FCLs) to leverage both bidirectional processing and complex pattern recognition. This hybrid approach effectively integrates past and future time steps, achieving superior SOC prediction accuracy across various test scenarios, especially under complex driving cycles and varied environmental conditions. The models were evaluated on a Tesla battery dataset, with performance comparisons demonstrating the hybrid model’s effectiveness in adapting to dynamic conditions. By advancing robust SOC estimation methods, this research supports the development of efficient BMS solutions that contribute to the adoption of EVs and the advancement of sustainable transportation.

Date

1-7-2025

Committee Chair

Meng, Xiangyu

Share

COinS