Correlation between vibration signal and surface quality based on recurrence analysis during surface burnishing process

Surface burnishing process (SBP) is the use of balls to plastically deform the sample surface to finishing processing, but its internal dynamic behaviors are poorly studied. Here we reported that the vibration signal performances extracted from the SBP system based on recurrence plot (RP). A novel denoised method based on EEMD and power spectrum was introduced. The correlation between the vibration signal and the surface quality was explored. The experimental results showed that via SBP, the surface roughness parameters Ra and Sa of the sample were reduced to 0.066 mu m and 0.217 mu m, respectively. And the cross-sectional hardness for region B in test 7 reached 206 HV. The proposed denoised method achieve effective information as much as possible from the original signals. The threshold for constructing a recurrence plot (RP) was determined by a fix recurrence rate RR = 0.1. The RP changed dramatically depending on the different paths created by the different loading force directions. The cross paths of the loading force (i.e. region B) corresponded to a more uniform RP and a more stable system. Under the interaction of burnishing depth, spindle speed and feed rate, the evolution trend of cross-sectional hardness was opposite to the recurrence quantitative analysis (RQA) parameters: DET, LAM, and RPDE. The largest cross-sectional hardness corresponded to the least DET, LAM, and RPDE. It indicated that the sample surface quality was reflected by the recurrence characteristics of vibration signals. This research provides a way for using processing system signals to predict the sample surface quality.