Research On Recognition For Machine Tool Types, Working State And Machining Parameter Based On Acoustic Signal | Posted on:2007-12-21 | Degree:Master | Type:Thesis | Country:China | Candidate:G L Yang | Full Text:PDF | GTID:2121360185959798 | Subject:Mechanical Manufacturing and Automation | Abstract/Summary: | PDF Full Text Request | There are noises when machine tool works. These machining noises can be regarded as ideal signal to monitor machining because it is easy to get without any influence on machining process. But how to get the characteristics for recognition is urgent to be resolved. In this paper, for the first time we apply acousitic signals' frequency–domain analysis, wavelet packet analysis and BP artificial neural network to recognize typical machine tool types, working state and its machining parameter based on the study on recognition using acousitic signals home and abroad. The main work done is as follows:1. The frequency corresponding with the maximum magnitude of signal's Power Spectrum in frequency-domain and the energy ratio of low frequency band to high frequency band in time-frequency-domain are extracted as the characteristics to recognize the machine tool types and working state.2. The characteristics mentioned above are input to a BP artificial neural network designed. The neural network is optimized through changing the number of nerve cell of connotative layer, the ratio of recognition for machine tool types and working state is above 95%.3. Based on the above recognition, another BP artificial neural network is designed to recognize the shaft speed of the drill press. This network's input is the characteristics extracted from the signal by wavelet packet analysis. The characteristics are those equal frequency band energy under certain layers decomposition. The ratio of recognition is also above 95%.4. A system of recognizing machine tool types, working state and its machining parameters based on Acousitic is built up by use of VC++ and Matlab.
| Keywords/Search Tags: | Acousitic signal, characteristics, BP neural network, wavelet packet, machine tool types, working state, machining parameter, recognition | PDF Full Text Request | Related items |
| |
|