| The positioning accuracy is an important index in numerical control machine and a factor affecting machining accuracy. It can be influenced by thermal expansion resulting from the relative movement between screw and nun. Because of various machining conditions, the thermal expansion of screw is very complex. Combined with the feature of screw expansion, the thermal error model has been made by adapting a neural network method. Based on Windows 2000 operating system, the software of thermal error compensation techniques considering the thermal error of screw has been developed in the presence of Visual Basic 6.0 å’ŒAccess. The main contents of the thesis are as follows:(1) Discussing the influence by feed rate, preload, machining conditions and bearing mounting form in thermal expansion of screw.(2) Based on the theory of MBS, presenting the comprehensive errors model consisting of offset probe errors, geometric errors, screw thermal error and spindle error.(3) Introducing the theory of a neural network and structure and algorithm of RBF network. Establishing the parameter identification model of screw according to RBF method.(4) Developing the thermal error compensation module of screw and stating the calculation method of the software.(5) Identifying the thermal error parameter of screw in XHFA2420 machining center, it has been proved RBF network having the rapid learning rate and high approximation accuracy. At the same time, a series of experiments have been performed in MAKNIO machining center, it has been proved that the accuracy of the machine improving by 74.6% after errors compensation. The results verify the effectiveness of the method.
|
PDF Full Text Request