Research On The Online Detection System For Active Redundant Objects

As an essential part of aerospace products,the reliability of electronic equipment has a significant impact on the safety of aerospace products.One of the most important factors affecting the reliability of electronic equipment is the surplus of activities,which is a common focus in the current aerospace industry,and also an urgent problem to be solved.Therefore,it is of great practical significance to detect the surplus of sealing equipment,find out the surplus in time,and solve the problem before the equipment leaves the factory.In this paper,the common detection methods of superfluous objects at home and abroad are studied.At present,the detection of superfluous objects is mostly based on the principle of particle impact noise detection(PIND),but the detection of whether there is active superfluous objects in the tested equipment is mostly determined by the detection personnel observing the detection waveform with their own naked eyes.This detection method is greatly influenced by subjective factors and external factors,and the detection results are not allowed It is predicted that the rate of missing detection is high.This paper designs and builds an on-line detection system of active surplus,which still adopts the particle impact noise detection method as the principle,and designs an electric turntable to realize the speed and angle control so as to control the stress of the equipment to be tested,which provides an efficient and reliable test platform for further exploring the detection principle of surplus.The platform includes electric turntable device,data acquisition device,hardware circuit control module and upper computer software control module.It can realize the functions of electric turntable working in different modes,high-speed signal data acquisition,redundant signal data processing and system control through upper computer software.The BP neural network model is selected for the identification method of active surplus material.By extracting the characteristic values of signal frequency domain,wavelet packet energy entropy transform and Hilbert Huang transform as the input of neural network,the accuracy of automatic identification of active surplus material signal,metal material and non-metallic material is effectively improved.The final equipment can detect the active surplus in the electronic equipment with the overall dimension less than 0.027 cubic meters and the weight less than 10 kg,which is difficult to be directly recognized by the human eye,and realize the automatic identification of the active surplus at the same time.This study is of great theoretical and practical value for improving the reliability of aerospace electronic equipment and ensuring the working life and performance of aerospace models.