Mining Haulage Road Maintenance Prediction Dashboard Utilising Machine Learning Techniques

This paper details the development of a predictive maintenance dashboard designed for haulage roads in surface mining operations using machine learning algorithms. We assessed various machine learning models, including Logistic Regression, Random Forest, and K-Nearest Neighbors (KNN), to determine their effectiveness in predicting the timing and location of necessary road maintenance. The research identifies the Logistic Regression model as the most effective, which was trained on data including vibration measurements, equipment counts, and speed metrics to evaluate road surface conditions. These inputs are processed through a logistic function to estimate the probability of required repairs. The predictions generated by the model are stored in an SQL database and visualized through Power BI, which offers a user-friendly interface. This interface enables operators, including those with limited experience, to interpret the data effectively and make informed maintenance decisions, thus improving operational efficiency and safety.