Research Of CNC Maching Accuracy Estimation Technology

CNC is the key equipment of machinery industry, whose accuracy determines the quality of parts directly. In the home and abroad detection field of the CNC accuracy, Laser interferometer and ball-bar were applied constantly to measure CNC accuracy off line. A theory of detecting CNC maching accuracy based on multi sensor fusion was presented, which to detect CNC maching accuracy on line.Firstly, the structure of CNC was analyzed, and the key parts which influence the CNC maching accuracy were found. The coordinate system of CNC Established, contrast ideal trail and real trail of the tool in maching process, at last, the error type of CNC was achieved.Secondly, the proper sensor, data acquisition card, signal conditioning instrument were selected, and the positions where fix the sensor were defined. The platform of data acquisition was established, the software of data acquisition based on LabVIEW was compiled.At last, the orginal signal was amplified and filtered, because there was too much noise and the signal was too weak, and then the signal was analyzed in time domain, frequency domain and small wave. Characteristics of the signal were extracted. The characteristics were inputted to SOM net to training network, and the SOM net outputted the cluster result. Then, sample characteristics were inputted to SOM net which has been trained. And the SOM net outputted the cluster which the sample characteristics belong to.The result showed the characteristics which was extracted could reflect the accuracy of CNC, and the SOM net could distinguish the cluster result, and relizing to detect the accuracy of CNC on line, and improving working efficiency of CNC. It is worthy to be research deeply.