A Knowledge Management Approach to Early Breakdown Detection and Efficient Repair of Mining Trucks

摘要

Knowledge management is about using experience, knowledge and information and direct it towards the place where it is required. A current problem in modern companies is to collect information that is segregated in an organization, and find a strategy to direct this to the place where this information is required. In the maintenance process of electric drive dump trucks, there exist a great dependency of technical experts on technical information to repair faults to acceptable levels of quality and downtime. In order to improve this process, an integrated system has been developed that consolidates information from different sources. The system uses information from the maintenance process, root cause and fault tree analyses and troubleshooting guides, and combines this information with machine learning algorithms. The machine learning algorithms provide a preliminary diagnosis and the most probable root cause. The system also decides when to send this information to the field technicians.A support vector machine model is used to classify fault codes in order to detect breakdown detentions. Considering a highly imbalanced dataset, the model achieved a recall of 0.711 and a precision of 0.719.Through the first and last (4 months long) studied periods, the four most relevant fault codes presents better MTTR and a decrease of breakdown events.The combination of source information, machine learning algorithms and feedback from the technicians constitutes an expert system for the management of breakdowns on electric drive dump trucks. The availability of the right information at the right time provides improvements in terms of efficiency, reliability/quality and the overall maintenance management process.