| As an important part of aerospace research,the space environment simulation test is used to simulate a special space environment,so as to verify the system and check the defects of spacecraft equipment.Vacuum pump is an important equipment to guarantee the space environment simulation test,and is responsible for the acquisition and maintenance of high vacuum environment in the test container.However,when the vacuum pump fails,it will affect the normal progress of the simulation test,and even endanger the test safety when the fault condition is intensified.Therefore,it is particularly important to make a timely and accurate diagnosis of the vacuum pump operating state.In this paper,the vacuum pump in the aerospace research institute is taken as the research object,and conducts a diagnosis method study for the overload and wear fault of the vacuum pump.First,the design requirements of the vacuum pump signal acquisition system are analyzed,and the overall design of the signal acqusition system is completed.The selection of the acquisition device of the system hardware is discussed respectively,and the system acquisition program is written based on Lab VIEW software,and the signal acquisition system was built through the combination of software and hardware.According to the acquisition conditions,the experimental scheme of signal acquisition is established,and the multiple acquisitions of five different operating states of the vacuum pump are realized,which further verified the feasibility of the acquisition system and scheme.The combined noise reduction method based on ensemble empirical mode decomposition and wavelet packet threshold is used to perform noise reduction on the signal.The feature extraction methods in the time domain and the time-frequency domain are used respectively to construct the dimensionless feature vectors in the time domain and the energy feature vectors in the wavelet packet,and they are used to compare and analyze the different operating states of the vacuum pump.The results show that the two feature vectors constructed can effectively reduce the original signal data dimension,and both have a good performance to characterize the operating status of the vacuum pump.Back propagation neural network and support vector machine method are used for fault diagnosis and signal recognition research,and the parameters of support vector machine are optimized through particle swarm optimization algorithm.A variety of vacuum pump operating state diagnosis models are constructed,and the influence of different combination of feature vectors and pattern recognition methods on the recognition accuracy is studied and discussed.The data results show that the diagnosis model combining wavelet packet energy feature and particle swarm optimization support vector machine can obtain the best diagnosis and recognition effect,and its recognition accuracy rate can reach 96.4%,which meets the application requirements.Finally,an intercative fault diagnosis system for vacuum pumps is designed and written based on MATLAB software.Through the system function test,the effectiveness of the vacuum pump fault diagnosis system is verified,and the practicability of this research method is improved,which provides a reference for further research and engineering application of vacuum pump fault diagnosis. |
