Wireless connection timeout fault diagnosis of Chinese train control system using adaptive neuro-fuzzy inference system

摘要

Train-Ground wireless connection timeout is one of the common faults in Chinese Train Control System Level-3, and it will trigger a full-service brake application on high speed train. The paper presents a fault diagnosis method base on adaptive neuro-fuzzy inference system. First, a partition of universe and granularity for input and output are given. Moreover, a model is built and then trained by the diagnosis data in the knowledge base. The results of numerical simulation and experiments indicate that the model can provide high diagnosis precision of different fault types.