The Research On Cutting Quality Prediction Model Of Abrasive Water Jet Based On SVM

The newly emerging abrasive water jet cutting technology is not only environmental friendly, but also no selection for objects’ material, so it has been applied in many fields. But the fact that the cutting quality is hard to control precisely affects its broader application and promotion.The abrasive water jet cutting process is complex. There are many parameters, and the researches on the influencing cutting quality ability are most with qualitative analysis. In order to identify the primary and secondary factors, the correlation analysis was carried out to quantitatively study the orthogonal experiment data. The result is that the coefficients between cutting parameters and cutting quality are0.96for the diameter of abrasive focusing nozzle,0.63for the standoff distance,0.61for the workpiece thickness,0.5for jet traverse speed. It is benefit for adjusting the parameters to control cutting quality.A precise prediction model was established to set parameters quickly and control the cutting quality effectively. The common regression equation and BP neural network modeling methods are difficult to build prediction model for the small sample data of orthogonal experiment. The SVM which can make accurate forecasts in complex nonlinear environment was brought to establish the cutting quality prediction model of abrasive water jet. The verification experiments shows that the prediction accuracy of the model is high, and the maximum value of the relative tolerance is4.9227%, the minimum value is0.8866%, and the average value is2.8696%. The best cutting quality occurs when the diameter of abrasive water jet is0.15mm, the standoff distance is8mm, and the workpice thickness is1mm. The prediction model can set main parameters and control the cutting quality, which has certain significance for developing the abrasive water jet technology.The leveling system based on the electro-hydraulic proportional technology was designed. It can guarantee the cutting quality of wedge workpiece. The mathematical model was built and corrected with PID controller. The simulation results show that the system has higher precision and better stability, and the steady-state error for step response is0.005. Therefore the system can meet the application requirements, and has practical significance for controlling the cutting quality of wedge workpiece. It can broaden the abrasive water jet cutting technology application.