Analysis And Research Of Signal Equipment Fault Diagnosis Based On Intelligent Learning Algorithm

Railway signaling system is composed of a variety of control equipments and is affectedby capacity utilization, the environment and other factors. As a result the faults of the signalsystem are characterized by randomness and fuzziness. Under the guidance of the high-speed,heavy-load and high-density development direction in China’s railway, there is a activedemand for establishing and perfecting railway signal equipment maintenance and monitoringsystem and gradually achieving signal equipment zero fault, in order to ensure the safety ofrailway transportation.Currently, in China microcomputer monitoring system is adapted to collect the real-timeinformation of railway signal equipments, which can provide a reliable basis for under-standing the status of equipment operation and for analyzing the causes of faults. However, acomprehensive monitoring software, which integrates both condition monitoring and faultdiagnosis, is needed when signal equipment failure occurred. Thus, the establishment of therailway signal equipment fault diagnosis expert system has the very realistic significance forfault location and maintenance schedule.Combined with the present situation of railway signaling equipment fault diagnosistechnology, firstly the common faults of railway signal equipment are listed and summarizedand the basic methods of fault diagnosis are introduced with comparing to each other.Secondly, in combination with the characteristics of railway signal system in China, aknowledge base building model based on C4.5learning algorithm is designed, and theinference engine, explanation facility and human-computer interface in fault diagnosis expertsystem are researched and designed. Finally, with a electric circuit fault example, the taskperformances of the main modules in the designed fault diagnosis expert system is verified.Aiming at implementing the diagnosis functions, a large amount of related literatures andtechnical data are researched, and knowledge representation and knowledge acquisitionmethod, which is based on decision tree learning algorithm, are mainly researched, with thepractical application situations of intelligent learning algorithm in other fields are fullyanalyzed. Basing on the new model, a compex fault diagnosis algorithm, which makes fulluse of the advantage of decision tree that integrates knowledge representation and acquisition,is designed. With C4.5learning algorithm used in building the knowledge base of expertsystem, the difficulty in knowledge acquisition is overcome. Finally, during the verifyingphase, with Visual C++6.0as development environment, each function module of expertsystem is designed for implementing railway signal equipment fault diagnosis. Theexperimental result shows that the proposed method can realize automatic knowledgeacquisition and lays the foundation for developing practical application system.