Grey wolf optimization based support vector machine model for tool wear recognition in fir-tree slot broaching of aircraft turbine discs

• Published in: Journal of mechanical science and technology 2022-12, Vol.36 (12), p.6261-6273
• Main Authors: Ying, Shenshun, Sun, Yicheng, Fu, Chentai, Lin, Lvgao, Zhang, Shunqi
• Format: Article
• Language: English
• Subjects: Accuracy; Broaching machines; Computing time; Condition monitoring; Control; Dynamical Systems; Engineering; Frequency domain analysis; Genetic algorithms; Industrial and Production Engineering; Intelligent manufacturing systems; Mechanical Engineering; Optimization algorithms; Original Article; Particle swarm optimization; Recognition; Redundancy; Signal processing; Spatial distribution; Support vector machines; Tool wear; Turbine disks; Vibration
• ISSN: 1738-494X; 1976-3824
• DOI: 10.1007/s12206-022-1139-x

Broaching tool condition monitoring is the basis of intelligent manufacturing of high-end broaching equipment. There are still technical bottlenecks in tool wear recognition accuracy and response speed. Aiming at the characteristics of complex cutter tooth shape and variable spatial distribution of turbine disc fir-tree slot broaching tool, a method of wear state recognition for broaching tool based on maximum relevance and minimum redundancy and gray wolf optimization algorithm is proposed. In the process of broaching, the broach vibration signals are collected in real time. The signal characteristics in time domain, frequency domain and time-frequency domain are extracted by signal processing technology, and the support vector machine (SVM) recognition model of broach wear state is established. The maximum relevance and minimum redundancy (mRMR) method is used to reduce the dimension of data, grey wolf optimization algorithm (GWO) is used to optimize parameters to improve the recognition accuracy of SVM. The experimental results show that the model can accurately recognize the wear state of fir-tree slot broach at different stages. In addition, grey wolf optimization-support vector machine (GWO-SVM) model shows higher accu-racy in classification than particle swarm optimization based support vector machine (PSO-SVM) and genetic algorithm based support vector machine (GA-SVM) models. Compared with PSO-SVM and GA-SVM models, the computational time of GWO-SVM is reduced by 54.2 % and 60.5 % respectively.