Research On Gear Pump Degradation State Recognition Based On Arithmetic Optimization And Gated Recurrent Neural Network

As the most widely used rotary pump,external gear pumps have many unique advantages over other types of rotary pumps.However,as the working hours increase,the performance of the gear pump will inevitably decline due to wear and tear,which will affect the safe and stable operation of the hydraulic system.Therefore,identifying the degradation state of the external meshing gear pump can not only maintain the gear pump in time,ensure the safety of the hydraulic system,but also effectively reduce maintenance costs and production losses.Firstly,the wear degradation mechanism of the external gear pump is analyzed and studied.By adopting the accelerated degradation method,a corresponding gear pump accelerated degradation test bench is established to simulate the wear and degradation process of the gear pump,and the corresponding degradation data is collected to provide data support for subsequent state identification.Secondly,in order to eliminate the noise in the degraded data,this paper innovatively introduces successive variational mode decomposition(SVMD)to reconstruct the vibration signal of the external meshing gear pump with noise reduction.SVMD has the core advantage of not needing to preset the number of decomposition modes and bandwidth balance parameters.By constructing artificial simulation signals and comparing them with variational mode decomposition(VMD),the superior noise reduction performance of SVMD is verified.At the same time,feature extraction is performed on the reconstructed vibration signal based on the time domain,frequency domain,and time-frequency domain,and then the degraded features are normalized to eliminate the differences between different units and scales among the features.Furthermore,this paper selects the gated recurrent unit(GRU)neural network to identify the degradation state of the external meshing gear pump,and compares it with the identification results of the back propagation neural network and the recurrent neural network to verify the identification advantages of the GRU neural network.Finally,in order to further improve the accuracy of gear pump degradation identification,an innovative arithmetic optimization algorithm(AOA)combined with GRU neural network identification method is proposed.By comparing with the GRU neural network model optimized by particle swarm optimization,the final identification results show that the recognition accuracy of the gated recurrent network model based on arithmetic optimization is higher and the effect is better.