Deep Majukar , Anuradha Ajit , Rutuparna Sonna , Triveni Dhamale

Predictive maintenance plays a critical role in ensuring the reliability and efficiency of electric vehicles [EVs], particularly in preventing unexpected motor failures. This paper reviews the application of deep learning techniques, specifically Convolutional Neural Networks [CNN] and Long Short-Term Memory [LSTM] networks, for predictive maintenance in EVs. We analyze sensor data from electric motors. We look at how these models, combining L1 Regularization, Logistic Regression, and Random Forest, improve fault detection accuracy. It is preceded by a data analysis and followed by a discussion on machine learning and deep learning models used. A comparison of different models is done and CNN + LSTM emerges as the best possible solution, as it can capture spatial and temporal patterns in the data. Finally, we have the challenges and limitations of these models and give directions for future work including real-time monitoring systems and digital twin technologies

https://doi.org/10.58357/jedcs.2025.v4.i1.11