| With the rapid development of modern manufacturing and industrial automation,numerical control machine tools and machining centers has been widely applied. In order toensure that these automated and unattended processes are safe and reliable, a system is urgentneeded to monitor the cutting process on line and on real-time. As the direct executor of thecutting process, tool wear and breakage will directly affect the machining accuracy andefficiency, so the monitoring of tool condition becomes extremely important and it has played avery important role for the advancement of unmanned processing and automation development.Tool wear and breakage is inevitable in the cutting process and it will bring undesirableeffects for automated processing, such as affecting the quality of the workpieces surface,causing the machine damaged and threating the lives of the operator, resulting in incalculableloss. Research shows that CNC machine which is installed tool monitoring system can reduce25%of downtime, improve machine utilization by50%, improve the productivity and save theproduced costs. With decades of development, the depth and breadth of tool monitoringtechniques has reached a certain level, but there’s no one way that can be applied to a variety ofdifferent processing conditions to monitor all kinds of tools. That is to say, a variety of methodsare limited in applicability and still far from achieving automation control requirements. Thispaper mainly researches the tool wear monitoring system based on the parameters fusion ofcutting sound and cutting force, and contains the following related work:(1) The application range and characteristics of several commonly methods for toolcondition monitoring has been compared. The paper has summarized the research situation andresults of domestic and international, and analyzed the existing problems and future trends oftool condition monitoring technology. Based on this research, it has proposed the method tomonitor tool wear conditions based on the parameters fusion of cutting sound and cutting force.(2) Collecting sound signal using MEMS microphone and current signal using Hall sensor,extracting the characteristic parameter of sound signal, and proposing the scheme of indirectmeasurement for cutting force. The multi-sensor information fusion method based on neuralnetwork is introduced, and focused on the BP neural network. BP algorithm is theoreticallyderived. For the shortcomings of traditional algorithms, the improved BP algorithm is proposed.(3) In YCM-V116B vertical CNC machining center for the milling experiments,pre-processing for the experimental data, and extracting the characteristics parameters ofcutting sound signal and cutting force signal. Using these parameters as the input sigal of BPneural network, then designing the BP neural network which including the establishment ofinput layer, hidden layer and output layer, as well as the parameters selection when training theneural network. Through training this network, it obtains the characteristic fusion of cuttingsound and cutting force improves the accuracy and stability of recognizing tool wear.(4) After analyzing the functional needs of tool-wear monitoring system, ARM9S3C2410microprocessor is selected as the core of system, and we designes tool-wearmonitoring system surrounding the microprocessor, including the acquisition circuit of cuttingsound signal and current signal, through the A/D converter is sent to the processor. Designigthe power, reset, JTAG, serial and LCD display module, and making the designed tool wearmonitoring system to meet fully the functional requirements.(5) Building and designing the software platform. Firstly, it introducts briefly therelationship of the application, the driver and hardware, then builts the software platform onthe target machine. Finally, we finishes the GUI design based on QT/Embedded, and theapplication is transplanted to the target machine. It achieves basically the required softwarefunctional requirements. |
PDF Full Text Request