Research On Fault Diagnosis Technology Of Photoelectric Pod Based On Sensor Fusion

Airborne optoelectronic pods are widely used in aviation reconnaissance,measurement and positioning,and strike effect evaluation as repeated mounting equipment.They often face harsh conditions such as overload shock,vibration shock,and sudden temperature change during use in the airborne environment.,it will affect the working state and service life of the internal components,and increase the probability of failure during use.Bearings are one of the core components of optoelectronic pods,and their trouble-free operation plays a key role in equipment reliability and stability.Due to the limitations of the environment and operating conditions,timely and efficient monitoring of bearing working conditions and fault prediction and diagnosis are very important to ensure the reliable and stable working state of the optoelectronic pod.At present,the commonly used method is the routine inspection during the regular maintenance process,and the evaluation is based on the experience and knowledge of experts,and the situation of untimely fault diagnosis or excessive maintenance often occurs.Based on the systematic analysis of the causes and characteristics of bearing faults in optoelectronic pods,this paper proposes,analyzes and studies a fault detection method for optoelectronic pod bearings based on the fusion of sound and vibration sensors,aiming at the shortcomings of traditional diagnosis methods.Feature extraction and fusion to explore the bearing fault diagnosis method based on sensor fusion,to provide support for improving the reliability of optoelectronic pods and reducing maintenance and maintenance costs.The main contents of this article are as follows:1)According to the common problems of bearings,the causes and characteristics of rolling element,outer ring or inner ring faults are studied,suitable sound sensors and acceleration sensors are selected according to the fault cycle characteristics of bearings,and signals based on FPGA double AD sampling are designed.The acquisition system collects the sound signal and vibration signal of the optoelectronic pod.2)Data analysis is carried out for the collected signals,which shows the importance of time-frequency analysis.Then preprocessing is carried out based on the acquired signal,and three methods are used to analyze the bearing sound signal and vibration signal,and finally FFT and wavelet packet transform are selected as the preprocessing method of bearing signal through comparison.3)For a large number of signal processing problems,a feature extraction method of fusion convolutional neural network(fused CNN)is proposed.The sound signal is input into a one-dimensional convolutional neural network(1DCNN)through Fourier transform,and the vibration signal is transformed through wavelet packet.Input the two-dimensional convolutional neural network(2DCNN),and realize the feature extraction and fusion of different sensor information by determining the appropriate convolutional neural network parameters.4)For the fault classification problem,based on the fusion CNN combined with the fuzzy classification method,the feature extraction and fusion of the bearing signals collected by the two sensors can be realized,and the decision values corresponding to the two sensors can be obtained.The inference obtains the global decision,and finally the classification result of the bearing signal is obtained.