Surface Roughness Prediction Of Chip Removal Drilling Based On Process Parameter And The Signal Feature

Compared with ordinary drilling,chip removal drilling can improve the quality of drilling and the quality of surface roughness.But the chip removal drilling is also in semi-closed or close environment,the surface roughness of the hole is difficult to detect and analyze.This thesis intent to combine the drilling process parameters and the monitoring signal characteristics to carry out the research of surface roughness prediction of chip drilling.The main work carried out included the chip removal drilling monitor platform set-up,pretreatment of sensor signal,the influence of process parameters and the monitoring signal on roughness,the establishment and verification of the prediction model.(1)The chip removal drilling monitor platform set-up,the monitor signal acquisition and pre-processing.By building the drilling experiment platform,vibration signal,AE signal and surface roughness are received,and the trend term of the signal is eliminated by the least squares fitting method.(2)Pre-treatment of sensor signal.Aiming at the problem hard to extract drilling process features of vibration signal from background noise,a new method is developed in this thesis to enhance the signal feature by the wavelet packet division-spectral subtraction.Simulations and experimental results show that the method can be used to enhance drilling process features of the monitoring vibration signal,to reduce the influence of back-ground noise,and to establish the mapping model between drilling process and monitoring signal,which provides accurate data support for the follow-up study.(3)The influence of process parameters and the feature of monitor signal on surface roughness.According to the experimental data,this thesis firstly analyzes the influence of process parameters on the feature of monitor signal and the surface roughness.Then,the variance analysis method is used to study the significant interaction effects of the spindle speed,the feed rate and the interaction on surface roughness and monitoring signal.Finally,the correspondence between the feature of monitor signal and the surface roughness of the hole wall is analyzed.(4)The establishment and verification of prediction model.First of all,the number of output layers and output nodes of the neural network is determined.Then,aiming at the problem hard to determine the number of hidden layer nodes,dynamically adjust the number of hidden layer nodes on the background of the same approximation error,the optimal structure is determined by the coincidence of the output value and the predicted value.Finally,the validity of the prediction model is verified by the experimental data,in order to evaluate the effectiveness of the prediction model.Theoretical analysis and experimental results show that using the prediction model in this paper,can accurately the surface roughness.This method can overcome the problem of manual sampling detection like missing detection,low efficiency and so on.This new method offer a new method and the theoretical basis for surface roughness of the chip removal drilling.