Multiple Fault Diagnosis and Prognosis of Chinese Train Control Systems

Abstract

In railway transportation, prevention or reduction of delays due to any equipment failure is becoming ever more crucial. This work presents a D-matrix-based multiple fault diagnosis of Chinese train control system equipment. It integrates rough set and Huffman coding-based heuristic functions to obtain the near-optimal test sequences. This facilities to construct test sequences with the balance between precision and minimum expected testing cost. By embedding D-matrix with reliability parameters and probability of missing detection and false alarm, a continuous time Bayesian network-based prognosis algorithm is used to compute the failure probability distribution with time and evaluate the failure trend. A CTCS equipment is used to demonstrate the efficiency of the proposed multiple fault and prognosis algorithm.