Font Size: a A A

Cutting Tool Wear Real-Time Monitoring Technology In Aeroengine Parts Machining

Posted on:2012-03-09Degree:MasterType:Thesis
Country:ChinaCandidate:T WangFull Text:PDF
GTID:2232330371458203Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
With the acceleration of national industrialization, the number of companies who recognize the importance of industrial automation has increased. As a pillar industry in Liaoning province, industrial automation and unmanned technique of manufacturing is an urgent need. Conventionally, cutting tool wear is identified by the operator. It is subjective and becomes bottleneck of the development of modern manufacture. Feature extraction and fault diagnosis algorithm of tool wear have been designed in this paper. And the algorithm has been implemented in the on-line monitoring equipment. It provides a new solution for the unmanned fault identification of tool wear.(1) Design the feature extraction algorithm of tool wear According to the characteristics of time-varying and non-stationary of the acoustic emission signals collected from the sensors,we have selected wavelet packet and empirical mode decomposition algorithm which are appropriate to analyze the time-varying signal. Respectively, the wavelet packet and empirical mode decomposition algorithm are used to decompose the acoustic emission signals and calculate the energy of each component as the feature vector.(2) Design the fault diagnosis algorithm of tool wear To identify the faulty mode of tool wear accurately, fault identify model is designed with Support Vector Machine (SVM) which possesses excellent pattern recognition performance. After train the model designed above with the vectors extracted, it possesses the ability of fault identifying. Then it can be employed into fault detection on line.(3) Implement the feature extraction and fault diagnosis algorithm of Tool wear in DSP According to the needs of real-time detection, faulty feature extraction and identifying algorithms are developed based on DSP. After code optimization, efficiency and timeliness of algorithm are both increased. Experiment result shows that the model designed can extract faulty feature and identify the tool wear effectively, which satisfies the demand of design.
Keywords/Search Tags:Tool wear, Fault diagnosis, Wavelet package, EMD
PDF Full Text Request
Related items