The Surface Roughness Online Identification Research Based On Acoustic Emission Technology

Grinding is a dynamic and nonlinear machining process with many complicated influential factors. Exact mathematic model is difficult in creating by traditional method. The surface roughness online measure has been difficult.The surface roughness in-process intelligent identification system in grinding was constructed based on Acoustic emission (AE) signal. The grinding wheel wear model based on AE signal and the surface roughness online identification model were built.The acoustic emission signal denoising was studied during the grinding for the wheel wear. The method of extracting the acoustic emission signals which based on the statistics and analysis of the average wavelet decomposition coefficient was proposed and then the detailed characteristics of signals can be described..According to form factors and the existed mathematical model of the grinding surface roughness, we studied how the wheel velocity, the workpiece feed speed, grinding depth and force ratio under the different conditions affected the surface roughness, and summarized up the regular rules.IThe grinding wheel wear state recognition model was established based on the BP neural network. By the proposed method of extracting the acoustic emission signals, the detailed AE signals characteristics of grinding wheel wear are extracted and put as network input.The surface roughness in-process identification molde based on BP network was built by the grinding wheel wear conditions, the grinding wheel speed, the grinding depth. the workpiece feed speed and the grinding force ratio as inputs. Here, the grinding force ratio includes the unpredictable and difficultly measured surface roughness influential factors.The experimental system of the surface roughness in-process intelligent identification in grinding based on AE signal was set up. The experiment was conducted on the proposed model, and the results verified that the surface roughness in-process intelligent identification system based on AE signal in grinding developed in this paper was feasible.