Research On Fault Diagnosis Technology For Jointless Audio Frequency-shift Modulated Track Circuit

As the key equipment of train operation control system,track circuit has an important impact on the efficiency and safety of railway transportation.However,the current of track circuit has many faults,and the research methods for the track circuit faults are not perfect,which seriously restricts the construction and development of railway.Therefore,scientific,effective,timely track circuit fault diagnosis has important significance.The track circuit compensation capacitor disconnection and insulation damaged section are selected as research directions,and the research contents of this thesis are as follows.Firstly,the ZPW-2000A track circuit is taken as the research object,and this thesis describes structure and working principle of track circuit,on this basis,established an equivalent model of two-port network theory for track circuit.Then the accuracy of the model in unoccupied and occupied states is verified,the influence of different state modes of track circuit on the locomotive signal induced voltage amplitude is analyzed,and the overall scheme of fault diagnosis is designed.Secondly,the induced voltage signal obtained by equivalent circuit model simulation is decomposed by the variational mode decomposition method.The number of modes and penalty factor parameters of variational mode decomposition are determined by kurtosis,correlation coefficient and running time.At the same time,three different signal decomposition methods are compared with the variational mode decomposition method.The performance of variational mode decomposition is analyzed,then the feature extraction of signal is completed by using fuzzy entropy.Thirdly,the basic principles of extreme learning machine and quantum genetic algorithm are briefly summarized.The quantum genetic algorithm is used to optimize the connection weights between input layer and hidden layer and neuron thresholds in hidden layer of extreme learning machine,and a classification model of extreme learning machine optimized by quantum genetic algorithm is established.Then the quantum genetic algorithm is optimized from the aspects of population coding and initialization strategy,quantum crossover and mutation strategy,dynamic adjustment strategy of quantum rotation gate.The fault diagnosis of track circuit is realized by optimizing extreme learning machine based on improved quantum genetic algorithm,and compared with five different classification and recognition algorithms.And then the classification and recognition of feature extraction data is completed.Finally,according to the designed fault diagnosis model of track circuit,the structure and function of fault diagnosis system of track circuit are designed by using the hybird programming technology of MATLAB and C#,and the modules of fault diagnosis system of track circuit are realized,which achieve the design requirements.