Predictive diagnostics for electric and IC engine vehicles

Abstract

This work explores integrating Machine Learning (ML) and logic-based models for predictive diagnostics in Electric Vehicles (EVs) and Internal Combustion (IC) engine vehicles, using data from Volvo Eicher Commercial Vehicles Limited (VECV). The objective of this work is to develop a predictive analytics model to forecast potential faults, optimize vehicle parameters, and enhance maintenance strategies. Key use cases include cell imbalance monitoring, temperature monitoring of cells and motors, and engine oil pressure warnings, aiming to improve fault detection systems' reliability. Despite challenges like the availability of labeled data for rare events in limited numbers and the computational demands of deep learning models for real-time applications, this thesis establishes a foundation for future advancements in automotive predictive analytics.