An intelligent broaching tool design method based on CBR and support vector machine
Chang Chen,Jiangtao Xu,Jing Ni,Guojin Chen,Zhengnan Lyu
DOI: https://doi.org/10.1016/j.aei.2024.102447
IF: 8.8
2024-03-07
Advanced Engineering Informatics
Abstract:Due to the numerous design parameters and complex relationships inherent in cutting tools, their design process relies heavily on expert design experience, which greatly hinders the design and production efficiency of cutting tool manufacturing enterprises. Case-based reasoning (CBR) and rule-based reasoning (RBR) methods have been widely used in the field of mechanical product design, both of which have improved product design efficiency to a certain extent, but still cannot meet the actual design and production needs, lacking research on the intelligent design process of complex cutting tools. In response to the above problems, a tool intelligent design method combining CBR and support vector machine (SVM) is proposed. Firstly, a comprehensive evaluation method is used to assess and quantify the relationship between design parameters, and based on the evaluation results, a hierarchical clustering method is used to divide the design parameters into different clusters. The tool design task is decomposed into multiple subtasks to reduce the complexity of design input data. Then, a numerical processing flow for design text parameters is proposed, and a hybrid retrieval method combining Euclidean distance and Manhattan distance is used to calculate case similarity. This method can handle both text and numerical design parameters in the design process. Secondly, a multi-classifier based on SVM and support vector regression machine (SVR) are used to predict the design results of each parameter of the tool. Finally, a broaching tool intelligent design cloud service system is developed. According to the actual test run results of the system, the mean squared error (MSE) of the prediction model loss is close to 0.01, proving that this method can improve the intelligent and automated design of complex cutting tools, and the design results are of certain reliability. Compared with traditional tool design methods, the design efficiency has been increased by about 98%, providing a new feasible solution for intelligent tool design, and has a certain reference value for the intelligent design of other types of mechanical products.
engineering, multidisciplinary,computer science, artificial intelligence