Open Collections

Abstract

In this thesis, we tackle the challenge of modeling and optimizing fuel consumption for heavy trucks on mining roads. Our aim is to develop a model that predicts fuel usage and minimizes costs through better road design. The research involves advanced mathematical modeling and a review of fuel consumption models used over the past two decades. We introduce a novel quartic-polynomial fuel consumption model. To enhance computational efficiency, we simplify this model to a quadratic form, which still closely matches the real-world fuel consumption data of heavy-duty mining vehicles. Through extensive numerical experiments, we establish that the quadratic model achieves a computational speedup of 4 times while maintaining an approximation error below 2%. This confirms the accuracy of the model and its practical utility for road design applications. Additionally, we approximate the quartic model to Mixed Integer Linear Programming (MILP) model which further optimize the computational efficiency. The MILP model shows promise for handling larger road networks by effectively balancing computation time and accuracy. We chose to approximate the original model to quadratic and MILP forms due to the availability of efficient commercial solvers such as Gurobi and CPLEX, which enhance the practicality of these models for large-scale applications. The study uniquely integrates vehicle dynamics with road design parameters, providing a comprehensive approach that reduces construction costs and fuel consumption. This integration enables mining companies to lower fuel expenses, enhance operational efficiency, and reduce their environmental footprint.