Research On Milling Tool Condition Monitoring Based On Fusion Of Vibration Signals And Acoustic Emission Signals

Metal-cutting process monitoring is the key technology and the essential aspect of automation machinery manufacturing. Single-sensor monitoring has a number of shortcomings. This paper studies the intelligence fusion ways which is applied to both vibration signals and acoustic emission signals in the cutting process, so as to realize cutter wear on-line monitoring.In this paper, we have a detailed analysis of the technical aspects by using multi-sensor fusion technology in the process of milling tool wear monitoring.based on vibration sensors and broadband acoustic emission sensor which are used for signal acquisition.Firstly, we use graphical integrated development environment LabVIEW to develop the system as data acquisition platforms, in accordance with the experimental program, collect different multi-sensor signals in the different wear state cutting process.Secondly, we use Matlab to analyse the acquired multi-sensor signals, by comparing the different levels of tool wear signals, extract the eigenvalue related with the tool wear.On the part of time domain, we extract root-mean-square value as the eigenvalue; on the part of time-frequency domain, we process the frequency band energy of the cutting vibration signals and AE signals, and then obtain the sensitive frequency band features of tool wear.Thirdly, we use extracted eigenvalues mentioned above as the network input, tool condition as the network output, as for different sensor signals, we training different BP neural network. The network can be used to monitor the cutting process under the same tool condition, the output of each network has the probability of the tool condition. Then we use the D-S evidence theory to fuse the various networks'output information, and then get the overall probability of the tool condition.Fourthly, besed on the study mentioned above, we construct a muti-information processing platform based on LabVIEW, provide a convenient analysis tools for muti-information fusion technology in practical applications.Finally, through experiment, confirmed the milling integrated fault diagnosis model based on BP neural network and DS evidence theory of effective, and develop the tool wear monitoring techniques.