Tool Vibration Analysis Based On Wavelet Packet Technique And Surface Roughness Prediction

The surface roughness has a great influence on parts usability, if it can be predicted on-line in cutting process, cutting parameters can be adjusted in time, it can improve produce quality and production efficiency. In this paper, the effect of tool vibration on surface roughness is considered and surface roughness prediction model based on cutting theory is established.In this paper, the influence factors of workpiece surface roughness are analyzed, among them, tool vibration is the main dynamic factor and it’s influence mainly focuses on cutting depth direction. The formation mechanism of workpiece surface topography is studied, the actual surface roughness profile is expressed as superposition of the ideal profile and high frequency vibration between tool and workpiece. Assumed the spindle rotation frequency is lower than tool vibration, it has less effect on surface roughness, therefore, tool vibration is the main research object.The direct measurement method is used to analyze tool vibration signal. Firstly, tool vibration signals are collected by acceleration sensor, pre processed through software programming and converted to displacement signals in the frequency domain. Secondly, tool displacement signals are processed by wavelet packet decomposition and reconstruction, the frequency which takes large energy is selected. Lastly, the signals spectrum are analyzed using the Fourier transform, the main frequency can be selected among them, so tool vibration equation can be expressed as the sum of some cosine equations.By solving trajectory equation of each point in the tool, workpiece surface topography simulation model is established. Tool vibration equation is taken into the model, the simulation model is graphed by MATLAB software, surface roughness value can be conveniently figured out through the3d model.Compare the predicted roughness value with the actual measured value, result shows that the prediction accuracy is above90%. It means workpiece surface roughness value can be correctly predicted using this method and tool vibration equation solved by wavelet packet analysis and spectrum analysis is correct.