Optimizing Mining Track Equipment Undercarriage Shoe Life Using Convolution Neural Network

The health of track undercarriage equipment in the mining and construction industry requires a high degree of oversight in enhancing equipment availability and reliability to improve the usage efficiency at minimal operational cost. There have been a great number of conditions monitoring techniques employed to monitor the wear rate of track shoes of these undercarriages. The most current and more advanced one is the ultrasonic wear indicator. Although this monitoring equipment gives accurate results, the challenge is the involvement of experts that go through vigorous training with voluminous training manual, and spending a lot of time on inspecting the track equipment. To minimize some of these challenges, this paper has proposed an Artificial Intelligent (AI) system that employs Convolution Neural Network (CNN) with track shoe images to predict the wear rate in a safer, efficient and faster manner. The system was developed, trained and evaluated on eight hundred and seventy-six (876) image data set. The results indicates that our system out performs other techniques in terms of efficiency and speed while at the same time eliminating the need of experts.