Optimizing the performance of the railway at AMIC by using data mining and advancedartificial intelligence techniques
This research project tackles a condition-based maintenance optimization problem in the railway equipment at ArcelorMittal Infrastructure Canada (AMIC). Logical Analysis of Data (LAD), a data mining technique, will be applied to exploit thousands of data records in order to identify and predict failure causes and degradation behavior of the equipment under study. New knowledge discovery approaches will be developed in order to deal with AMIC’s large scale databases. Mathematical programming and artificial intelligence techniques will be employed to develop these approaches. The results of this project will improve the railway performance, give an additional justification of the investments already done in sensors and measurements processes, and ensure no delay in providing the rail service at AMIC. AMIC is a world leader in steelmaking operations and mining business. Through an ongoing improvement process, AMIC seeks to improve the efficiency of its railway system aiming to provide a reliable and on time rail service. To achieve this improvement, tens of thousands of measurements data records of AMIC’s fleet of trains and their components, i.e. wagons, wheels, and bogies will be exploited to detect conditions of failure and degradation.