Fault Diagnosis Research Of 25Hz Phase Sensitive Track Circuit Based On Fuzzy Neural Network

Track circuit fault may result in the slight impact on driving, or lack of the efficiency of railway station, or even the obstruction of the vehicle running and the occurrence of safety accidents. Due to the influence of equipment use and external environment as while as the complexity of track circuit, the cause of equipment failure is diversified, and random. As the diagnose of fault is a complicated and significant link, it is a common method that measuring parameters through apparatuses first, and determining the fault type on the basis of past experience on the scene. Only depending on the experience of on-site maintenance personnel is unable to precisely deal with the fault, so a scientific algorithm is needed to apply to fault diagnosis. The algorithm will reduce equipment failure time, establish and complete the maintenance support system of fixed equipment, strengthen the tracking management of equipment operation quality, reduce the testing difficulty and duplication of the failure in the fault field, provide the base for overhaul of equipment status repairing, thus improve transport efficiency and ensure the transportation safety. This dissertation mainly studied on the 25 Hz phase sensitive track circuit, which is mostly used in today's railway station. Focusing on the fault complex of 25 Hz phase sensitive track circuit, section analysis was performed in different states of the track circuit. Combining fuzzy logic system with artificial neural network, a new fault diagnosis system which can be applied to the 25 Hz phase sensitive track circuit was established.This dissertation mainly researched on the following contents. First, the working principle of the 25 Hz phase sensitive track circuit was expounded according to its characteristics. Combined with its three working states(idle state, car passing state and car parking state), the common fault of red light strip and bad shunt and causes of failure was analyzed. Four-terminal network model was established, and the maximum and minimum of voltage was calculated, which is used as the system output alarm.Secondly, according to the working status of the 25 Hz phase sensitive track circuit, three signal centralized monitoring real-time data was input into the system as input parameters, in order to perform the section analysis of the track circuit fault. In the process, establishing the intelligent algorithm combining the fuzzy logic theory and artificial neural network, based on fuzzy neural network, the fault diagnose model was set up, the model was proceeded to diagnose and analyze the fault, and carry out final output result, including the type, degree and cause of fault and suggested measures.Finally, according to the fault diagnosis system of the 25 Hz phase sensitive track circuit, C++ Builder was selected to build software development platform. The centralized monitoring data stored in the database, was connected to segment analysis module to process the corresponding intelligent analysis and fault diagnosis, then results would be carried out, as while as the output of fault alarm. This system will help the staff find out problems in time, and provide reliable information for fault handling and routine maintenance, and make the railway operations more efficient and secure.