Tools
Tools
Dinakaran, Balaji (2024) Revolutionizing Fleet Efficiency: The Integration of AI Predictive Maintenance. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (428kB) | Preview |
Preview |
PDF (Configuration Manual)
Download (1MB) | Preview |
Abstract
Vehicle maintenance faces significant challenges due to unplanned downtime and inefficient resource allocation, which traditional reactive methods are inadequately addressed. These conventional approaches wait for breakdowns or some vehicle related issues before initiating repairs which in turn results in increased downtime, higher costs, and reduced efficiency. This paper proposes an AI-driven predictive maintenance model to proactively predict maintenance needs. The model utilizes a comprehensive dataset of 50,000 vehicle entries involving both categorical and numerical data of vehicle specifications, maintenance history, and operational metrics. This study conducts a comparative analysis to identify the best-performing approach for prediction using machine learning supervised algorithms like Gradient Boosting, Random Forest, Decision Tree and Logistic Regression and deep learning like Minimal Gated Unit (MGU), Gated Recurrent Unit (GRU), and Long Short-Term Memory (LSTM) are all components of deep learning architectures. The MGU, GRU and LSTM types of Recurrent Neural Network (RNN) cells used for processing sequential data. The initial results demonstrate significant reductions in unplanned downtime and cost savings with the best model achieving an accuracy of 97% which highlights the transformative potential of AI-enhanced predictive maintenance in improving fleet vehicle reliability and operational efficiency.
Actions (login required)
 |
View Item |