System Design And Implementation Of Feature Database In Aviation AC Arc Fault

The speed and comfort of aircraft make it a popular transportation.In order to ensure the safety of life and property,and protect the image and interests of aircraft passengers,the safety,reliability,stability of aviation electrical systems is of paramount importance.The wires in the avionics system are densely distributed,and the insulation of wires may be damaged due to physical stress,chemical hydrolysis,and corrosion.Due to the large temperature difference between the ground surface and high altitude,during takeoff and landing,water droplets are generated inside the aircraft because of the liquefaction of water vapor.They are attached to the broken wires,resulting in Conductive loop between the wire and the wire,or between the wire and the aircraft shell,and then cause intermittent and continuous arc faults.If the arc fault occurs near a flammable location such as a fuel tank,the aviation accident is likely to occur.Therefore,effective detection and diagnosis of aviation arc fault signals and early isolation of arc faults are important prerequisites for the safe and reliable operation of aviation electrical systems.Based on the American aviation standard SAE AS5692 and the civil standard UL1699,and referring to the national standard GB / T 31143-2014,this paper designs an aviation arc fault simulation test platform,conducts point contact series arc fault test,vibration series arc fault test,and truncation test,wet arc track test and three-phase cross talk immunity test.Observe and analyze the current waveform distortion of each arc fault test.Starting from the time domain analysis,use mathematical statistical algorithms combined with mathematical statistical models to identify and diagnose arc fault signals.Proposed a method of feature extraction: high-order cumulants combined with autoregressive model parameters.The Mahalanobis distance is used to compare the difference between the normal current signal and the arc fault signal,and the test data is used to establish the identification threshold.In terms of time-frequency analysis,an aviation AC arc fault identification method combining Tsallis singular entropy(TSE)and extreme learning machine(ELM)with continuous wavelet transform is proposed.The continuous wavelet transform is performed on the current signal of the arc fault line,and the obtained time-frequency coefficient matrix is subjected to singular value decomposition(SVD)to obtain the Tsallis singular entropy of the measured line current signal,and the eigenvector is constructed.The extreme learning machine is used for the TSE.The eigenvectors are trained to obtain a classification model suitable for aeronautical AC arc fault detection.The model is used to identify and classify the feature data under different loads.The classification model is verified by a large amount of experimental data,and has a high recognition rate.In view of the different ways in which aviation AC arc faults occur,there are various types of arc fault tests involved.In order to conveniently compare and analyze the differences of arc fault characteristics under different test types and different loads,this paper designs an aviation AC arc feature database based on the GUI(Graphical User Interface Environment)technology of MATLAB platform.The feature database provides a human-computer interaction interface,and users do not need to input commands to perform feature extraction and recognition diagnosis on test data.It has guiding function and practical value for advancing the research work of aviation AC arc fault.