West Virginia University
Machine Learning Maintenance Costs Prediction Model for Heavy-duty Alternative Fuel and Diesel Vehicles
Abstract
dc:description.abstractMedium- and Heavy-Duty Vehicles (MHDVs) are essential to global logistics and transportation. Around 77% of these vehicles are diesel-powered due to its powertrain mechanical characteristics that make them reliable and energy efficient. However, the environmental implications of diesel combustion, particularly its direct contribution to air pollution and greenhouse gas (GHG) emissions, have increased the interest in alternative fuel vehicles (AFVs). These alternatives include Natural Gas (NG), Propane, Hybrid Electric Vehicles (HEVs), and Battery Electric Vehicles (BEVs), each presenting unique benefits and challenges in terms of required fueling infrastructure, energy efficiency, environmental impact, and total cost of operation (TCO). Maintenance costs are a significant portion of the TCO for MHDVs and vary across different fuel types, vehicle vocations, and operational conditions such as weather and road conditions in the area that these vehicles operate. Machine Learning (ML) algorithms can analyze historical maintenance data and operational parameters to identify patterns and predict vehicle cost per mile with 95% accuracy. This predictive capability enables fleet management companies to schedule maintenance more efficiently, reduce unplanned vehicle degradation, and manage the lifecycle costs of their fleet more effectively. Moreover, the transition to alternative fuels in HDVs underscores the necessity for tailored maintenance strategies that address the specific needs of these vehicles. Additionally, a user-friendly graphical interface was developed to make the predictive maintenance cost ML model accessible for MHDV fleet or vehicle owners with no need of programming or specific software knowledge.
Degree
thesis:*- Name thesis:degree_name
- MS
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Mechanical and Aerospace Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Fernandez Villar, Esteban
- Contributors dc:contributor
-
- Gregory Thompson
- Arvind Thiruvengadam
- Rasik Pondicherry
- Hailin Li
- V’yacheslav Akkerman
Subjects
dc:subject × 5Identifiers
dc:identifier.*- Identifier
- https://researchrepository.wvu.edu/etd/12432