Application of data mining algorithm to detect defects on the catenary

The author investigates a very interesting problem which is to reduce the down costs of the train electric capture failure that happens due to the defect occurs in its suspended catenary line. When this type of incidence happens, the train stops, cost a fortune for the railway company and an enormous disagreement among the passengers. For this purpose, the signature of the contact force that manifests between the pantograph and the catenary line could be the key to reveal and maintain the health of this complex suspended line. This solution has a potential working business asset. This work includes theory, model design and implementation, data validity examination and data warehouse building and finally the data mining tasks. The first chapter is a brief introduction and forecast the contribution of the author. The second chapter is a brief description, the third chapter shows an analytical model of the catenary-pantograph system, the model in chapter III is implemented in chapter IV, which is done with Matlab (Simulink) with the help of a written Java program. Once the model is ready to run, the last chapter reveals the classifying algorithm as well as the results of the run sets.