Research On Operation Monitoring And Fault Diagnosis Method For Test And Launch Control System Of Launch Vehicle

The ground test and launch control system of a launch vehicle mainly conducts the ground test of each subsystem before the launch,and evaluates the health and performance of each equipment of the rocket,in order to ensure the launch is all right.Because there are many subsystems in a launch system,and the ground testing procedure is very complex and changeable therefore numerous and scattered test data,it is difficult to entirely understand the testing status for operators,meanwhile,it is hard to determine and handle the faults of system.Therefore,it is urgent to develop an unattended monitoring system for the ground test and launch control system,which can automatically diagnose the fault and improve the reliability of the system operation.Given the above requirements,this paper firstly analyzes the operation mechanism and the requirement of fault diagnosis requirements for the ground test and launch control system,then an adaptive state-transition fault diagnosis modelling framework is proposed.The crucial problem of the proposed framework is how to identify the system operation state and the state migration sequence according to the system operation monitoring data.To solve this problem,the following researches are conducted.1.Feature extraction is carried out for the multi-dimensional timeseries signals obtained by the monitoring system of test and launch control system.The significant time-and frequency-domain features of analog signals are combined with digital signals to form the system operation features.Then,the Principle Component Analysis(PCA)method is used to reduce the dimensions of the system operation features,therefore a system operation state feature model,which can be constructed by analyzing the operation monitoring data of the test and launch control system.Then the fault diagnosis is carried out according to the accumulated sum of the model and inspection statistics.The simulation results show that although this method can realize automatic fault diagnosis for the test and launch control system,the accuracy of diagnosis needs to be improved.2.To enhance the accuracy of fault diagnosis,a feature model of the system operation state is constructed based on the long-short memory network-self-encoder method,which is used to encode the multiple channel monitoring data.Then the encode is used as the system’s operation state feature to establish the system’s state-transition model.The simulation results show that the fault diagnosis accuracy of this method is higher than the PAC-based method.3.Based on the above theoretical researches,the unattended system monitoring software is designed and implemented based on PYQT.