Research On Tool Wear Condition Monitoring Based On Workpiece Surface Texture

Tool condition monitoring based on workplace surface texture integrates the technology of vision and texture analogy to research tool wear condition. The paper is aimed to research the technology profoundly by analyzing the vision features of workpiece surface texture on the various tool wear conditions. Owing to the particular advantages of Random Field texture models in aspect of analyzing texture, the paper establishes Markov Random Field texture model to extract the feature parameters of workpiece surface texture images. Based on the extracted feature parameters, tool wear conditions are evaluated. Each chapter research contents are as follows:The first chapter talks about the significance and background of the research. The theory of tool condition monitoring is introduced. Based on the large national and foreign documents, the develop survey of tool condition monitoring technologies and the main tool condition monitoring technologies are introduced. The application and advantages of too condition monitoring based on workpiece surface texture are presented. The main research contexts and innovation are presented.The second chapter analyzes the characters of rack face, flank face and boundary wearing. The tool wearing process and the selecting principles of blunt tool criterions are researched. The factors effecting workpiece surface texture are analyzed. The rationality and feasibility of tool condition monitoring based on workpiece surface texture are represented.The third chapter gives the expatiation about the concept of texture and the main technologies of texture analysis. Combining the workpiece surface texture images, gray co-occurrence matrix is researched. The characteristics showed in the normalization gray co-occurrence matrix are analyzed.The fourth chapter introduces the theories of Vision Labeling and Markov Random Field. The equivalence between Markov Random Field and Gibbs Random Field. The Gibbs-MRF and Gauss-MRF texture models are established to be applied to analyze workpiece surface texture.The fifth chapter researches the parameter estimation methods of Gibbs-MRF and Gauss-MRF texture models. The consistency of parameter estimation of Gauss-MRF model is demonstrated. The effects of small angle rotation of texture images and noise to the parameterestimation of Gauss-MRF model are researched.The sixth chapter analyzes the workpiece surface texture images which are obtained by cutting experiment. The feature parameters of Gibbs-MRF model are analyzed. Three parameters which attribute the vertical and cross direction of texture are presented as the feature parameters to evaluate the degree of tool wear. Based the three parameters, recognition parameter p is presented. The meanings of parameter p is represented. Based on the feature parameters of Gauss-MRF model, relative distance is presented to evaluate the degree of tool wear.The seventh chapter summarizes the main fruits and prospects the future of the research.