Research And Implementation Of Balise Data Acquisition Analysis And Fault Prediction

In recent years,the rapid development of China’s high-speed railway industry has put forward higher requirements for the intelligence and automation of the train control system.As an important component of the train control system,the balise has the problems of large data volume,difficult manual verification,and cumbersome maintenance work.Aiming at the status quo of balise detection and maintenance,this thesis studies balise data acquisition analysis and fault prediction,realizes dynamic detection of transponder message data and monitoring of transponder fault status.Based on the improvement of the acquisition board equipment,this thesis optimizes the design of the acquisition and analysis system to improve the decoding efficiency and transmission capacity of the system on the premise of ensuring the accuracy of the analysis results.Improve the performance of the acquisition and analysis system through the optimized design of decoding,transmission and storage modules.The derivation of the verification synchronization process proves the feasibility of computing multiplexing,and a three-stage window moving scheme is proposed in combination with the look-up table method to improve the decoding efficiency.A parallel architecture of the decoding process is proposed for different devices.Use keep-alive mechanism,flow control mechanism and retransmission mechanism to improve UDP protocol and increase data transmission speed.Referring to the characteristics of the B+tree structure,the storage design is proposed to realize the classified storage of data.The system collects and analyzes the signal characteristics of the message,studies the relationship between the state of the balise and the signal frequency,uses the Gaussian Hidden Markov Model to describe the fault model of the balise,and realizes fault prediction.A baise fault model is established to analyze the fault state of the balise.The performance degradation of the key circuit of the balise is analyzed and verified by simulation,and the frequency point distribution is selected as an important index for analyzing the fault state of the resonant power amplifier circuit of the balise.According to the distribution characteristics of the frequency points,the Gaussian hidden Markov model is used to describe the fault,and the process of iteratively calculating the model parameters of the expectation maximization algorithm is derived and realized.By fine-tuning the capacitor to simulate circuit degradation,collecting balise message frequency points as training data to iteratively calculate model parameters,collecting balise data in typical conditions as test data,predicting fault status through Viterbi algorithm,and comparing and analyzing with the actual balise status.Integrating fault prediction and data acquisition,the system performance is tested through the laboratory test environment.The test results prove that the acquisition and analysis performance of the system is excellent,and the accuracy of the fault prediction part is high.