Research On Intelligent Monitoring Of Machining State For CNC Surface Grinder Based On Bayesian Network

A Bayesian Network model for monitoring grinding states of CNC surface grinder is set up. The thesis is based on and supported by National S&T Major Project of China (2009ZX04001-131). On the basis of analyzing the advantages of Bayesian Networks when researching problems of uncertainty, an intelligent model about workpiece quality, blunt level and contact of wheel of CNC surface grinder was established. Besides, the monitoring software was designed.The thesis brings out the background of the research and summarizes the grind monitoring methods. Then the thesis discusses building network, inference algorithms and parameter learning of Baysian Network. The thesis introduces generation principle, sources and impact factors of grinding acoustic emission. The scheme of the measuring system is established. Besides, the programming of signal preprocessing, time domain and frequency domain analysis has been given using MATLAB.The thesis analyzes the advantages of Bayesian Networks applied in grinding moitoring. Modeling for workpiece quality, blunt level and contact of wheel is introduced. Besides that, the produce of model inference and parameter learning are represented. Furthermore, the structure of monitoring system is introduced. The user interface and all functional modules of grinding monitoring software are designed through VC++.The thesis builds the sample database through experiments. The posterior probability distributions are given by network learning on database. The result of verify the accuracy of model indicates that Baysian Network is very efficient in predicting workpiece quality, identifying blunt level and contact of wheel.