Fault Diagnosis Method And Experimental Research Of Thruster Of Underwater Vehicle

The exploration of the oceans in countries around the world has never stopped.Underwater robots,as an important tool for marine resource exploration,are important tools for underwater exploration.Because the environment in which underwater robots work is extremely complex,ensuring their safety is an important prerequisite for the entire underwater robot to complete underwater operations.The thruster is an indispensable executive component of the underwater robot's movement,and it is the main power component that causes the underwater robot to produce motion,which leads to the failure of the thruster is one of the main sources of failure.It can be seen from this that it is important to ensure the safe and reliable operation of the thruster,and fault diagnosis of the thruster is an indispensable link.It is of great significance to study the fault feature extraction and enhancement in the fault diagnosis of the thruster.In order to carry out the experimental verification of thruster fault diagnosis method,the experimental carrier was built.In view of the direct power off during the debugging process of the experimental carrier will lead to the breakdown of the lower computer system,the shutdown monitoring program of the lower computer was designed.In view of the problem that the propeller will suddenly reach the highest speed after the lower computer is shut down,the power control circuit and program of the propeller are designed.The main work of this paper is as follows:Firstly,in order to facilitate the follow-up research,the experimental prototype is developed,which mainly includes the hardware circuit design and software framework design of the underwater robot prototype,including the power system circuit design,sensor circuit design and propeller drive circuit design.The design of the software framework mainly includes the design of data receiving and sending program,the design of monitoring program,the design of sensor command sending and data receiving response function,and the design of monitoring interface.Secondly,the methods of time-frequency boundary recognition and fault feature extraction are studied.From the point of view of time-frequency,instantaneous spectrum entropy and signal-to-noise energy difference are used to divide the time-domain and frequency-domain boundaries.The experimental data are processed by smoothing pseudo Wigner Ville distribution algorithm and absolute value operation,so as to obtain time-frequency boundary.Compared with the fault energy concentration area,the boundary between them is basically the same.The fault feature distribution can be obtained by fault extraction of fault boundary.Compared with that without boundary,the mapping relationship between time-frequency energy fault feature and fault degree is unique.The fault samples are constructed to classify the fault features accurately.Finally,the energy feature enhancement method of thruster fault signal is studied.The optimal wavelet decomposition scale is determined by the wavelet decomposition method of the maximum value of time-frequency power density spectrum,and the maximum value of wave peak area energy is determined by the modified Bayesian algorithm and time-domain convolution calculation for the obtained optimal wavelet decomposition scale,then the enhanced propeller fault feature is obtained.The optimal wavelet decomposition scale is calculated for different fault features.The experimental data are processed according to the above method,and the peak area energy of the fault feature enhanced by the method in this paper is always at the highest position,which is much higher than the peak area energy obtained by other methods.It is found that the optimal wavelet decomposition scale corresponding to different fault degrees is not the same.