Font Size: a A A

Research And System Development Of The Deep Hole Drilling Monitoring Technology Based On SVM

Posted on:2011-12-03Degree:MasterType:Thesis
Country:ChinaCandidate:L L WeiFull Text:PDF
GTID:2121360305970862Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
The deep hole drilling has been widely used in the mechanical manufacturing. However, because of its difficult process, complex system consititution, difficult process quality control and high failure rate, the deep hole drilling has been a bottleneck in mechnical machining. Therefore, it is essential to research on tool wear and chip removal condition real-time monitoring in deep hole drilling.Taking BTA deep hole drilling as objective, the power and oil pressure are extracted as the monitor signals according to the characteristic of deep hole drilling. The real-time monitoring system, which is used to monitor block chips and tool condition in deep hole drilling, is established in this paper.The monitoring system hardware circuit is designed by way of modularization. The hardware system includes the clock circuit, power supply circuit, the reset circuit, the memory extended circuit, the data acquisition circuit, the serial interface circuit and simulation interface. This monitor system based on TMS320F2812DSP has the powerful functions of the real-time acquisition, analysis and recognition of the condition signal.Considering that the BTA system is complex and power and oil pressure signals are random and nonstationary, wavelet transform is used to multi-resolution analysis the signals in different frequency bands, then the relation between power signal and tool condition is researched as well as oil pressure signal and chip removal. The results show that power signal frequency band energy and smooth characteristics are closely related with tool wear as well as the peak and pressure changing rate of oil pressure signal low-frequency component with chip removal condition.Considering that the relationship between eigenvector and working condition is complex and uncertain, the SVM (Support vector machine) models for tool wear and chip remove condition recognition in deep hole drilling are established. Mapping from power signal characteristics to tool wear condition, as well as oil pressure signal characteristics to chip remove state, is realized in this paper. The monitoring results show that the method has high recognition accuracy and strong anti-jamming capability, and the sample does not require great.The serial communication PC system is developed on the circumstances of LABVIEW, so the software system is user-friendly and the functions of the DSP data acquisition, display and storage are realized. The low-level system programs, upper application programs and the writing FLASH of the DSP monitoring system are completed by using CCS integrated development environment. On-line real-time monitoring experiments show that the whole system has good accuracy, reliability and real-time property, and meets the deep hole drilling on-line real-time monitoring requirements basically.
Keywords/Search Tags:BTA deep hole drilling, tool condition monitoring, support vector machine (SVM), DSP, wavelet transform
PDF Full Text Request
Related items