Research On Fault Diagnosis Of Piston Head Crack In Axial Piston Pump Based On Dynamic Simulation

With the rapid development of high-speed railway,point machine is widely used in railway trunk lines.Zt-7yb axial piston pump in point machine is an important hydraulic component of point machine.Its working environment is bad,in the long-term work,prone to failure,serious harm to railway safety.Plunger head crack is one of the common faults of this type of axial piston pump.In this paper,a rigid-flexible fluid coupling dynamic model is established to diagnose the plunger head crack fault.The main research work of this paper is as follows:（1）the swash plate axial piston pump structure and main components,analysis of the axial piston pump working principle,the movement of the piston,the flow characteristics of the piston pump,the vibration path of the axial piston pump.（2）ADAMS and AMEsim are used to build rigid-flexible coupling dynamics model of axial piston pump.The 3d model of vertical axial piston pump is built in Inventor.Based on the analysis of the constraints between the moving parts of the piston pump,the rigid body dynamics model of the vertical axial piston pump was established in ADAMS.At the same time,the end cover,pump body and front pump body are flexible,and the rigid-flexible dynamics model is established.The hydraulic model of axial piston pump is established by using HCDAXPP0 super component in AMEsim and 3D import and 1D modeling mode to set leakage parameters.The hydraulic model outputs the fluid pressure,and the fluid pressure is imported into the dynamics model of axial piston pump,and the rigid-flexible dynamics model of axial piston pump is established.In the co-simulation,the dynamics model and hydraulic model adopt the same speed,the same sampling frequency and the same sampling time.The kinetic model,hydraulic model and co-simulation model are verified by analyzing the movement of the plunger,the flow rate,hydraulic pressure and oil pressure of the plunger pump.（3）Establish the fault dynamics model.The fault mechanism of plunger head crack was analyzed,and the plunger head in 3d model was modified to simulate the fault,forming two working conditions of normal and fault.The marker points of vibration measurement in dynamic model were set up,and VMD signal decomposition method and probabilistic neural network classification method were introduced.The two conditions of dynamic model and hydraulic model of the joint simulation,produces vibration acceleration data,using the decomposition of VMD,generate the IMF component,the IMF component calculated separately for each sample skewness,kurtosis,waveform,peak,pulse index,margin index statistics,the IMF component statistics calculate the characteristic value of matrix,Compose feature vector,normalize feature vector,use probabilistic neural network,carry on training and test.（4）Design and complete the piston pump experiment under normal and piston head crack conditions.After installing the sensor and connecting the acquisition instrument on the fault test bench of the piston pump,the experimental data of vibration acceleration on the center of the end cover and the body of the front pump were collected under normal and fault conditions.（5）the application of simulation samples and sample decomposition of VMD,generate the IMF component,the IMF from samples of each component of the skewness,kurtosis,waveform index are calculated respectively,peak value,pulse,margin index statistics,the IMF component statistics calculate the characteristic value of matrix and characteristic vector of the normalized eigenvector.Will get the samples of simulation part as the training sample,will get the sample as a test sample of real experiments,all experimental samples to labels,using probabilistic neural network diagnosis experiment sample is failure,then the real working condition of the experimental sample compared with the diagnosis,diagnostic accuracy,the calculation of diagnostic accuracy was 93.75%.Therefore,the correctness of rigid-flexible fluid dynamics model of piston pump under normal and fault conditions is verified,and the fault data is supplemented for this type of fault.