Motion Friction Compensation And Edge Tracking Force Control Of X-Y Table

Numerical control (NC) machine tool's performance is one of the main factors fordetermining the NC machine tool quality. In this dissertation, the research based on NCmachine performance developing trend is carried out. The main research work is splitinto three aspects as follows: friction compensation in NC machine tool's high precisioncontrol, force control in NC machine tool's intelligentized research, and control softwareprogram and intelligent control experiment research aimed at X-Y NC system.Friction is the main factor for system performance deterioration, and thus thereasonable solution of the friction especially nonlinear friction in the system has becomethe hot point problem in current research. Aimed at friction compensation problem, twoaspects of work have been accomplished: friction model-based friction compensation andnon-model-based friction compensation. In model-based friction compensation, staticexponential model and dynamic LuGre model for friction compensation research areselected and the friction compensation controller using Backstepping method is designed.As a result, friction infection is restrained and fine compensation effect is guaranteedwhen model parameters are changed. Considering that friction is a kind of unknownphysical phenomenon, it is difficult to attain accurate friction model. In non-modelcompensation methods, friction effect is regarded as bounded disturbance. Theuncertainty upper boundary is acquired through robust adaptive control method and theeffect due to uncertainties is further compensated. Another method is by using the meritof fuzzy system which needs no system model. By using fuzzy system approximatefriction disturbance effect and relative velocity, friction is estimated and furthercompensated by adaptive online learning and fuzzy nonlinear treatment function, andthus system performance is greatly upgraded.In machining system work pieces inevitably are contacted and force is thereforeproduced. If this force is controlled improperly, it not only can not attain control neededbut can also lead to damage and even to the damage of the system itself. So at this timeacting force control is very important. Aimed at edge tracking force control withuncertainties, we develop three force control methods: To compensate uncertainties usingneural network, force control loop adopting rule self-regulating fuzzy control. Thesystem has adaptive ability while system spindle contacts with work piece whosestiffness change range is big. It has been proven through simulation result that thismethod has good robustness and tracking ability. To develop adaptive fuzzy and CMACparallel force control strategy, CMAC neural network has realized object inversedynamic model, and at the same time adaptive fuzzy realization feedback control makessure that the system is stable and disturbance is constrained. This makes spindle toolwhich contacts with work piece have strong adaptive function force control. To develop aforce/position adaptive control strategy based fuzzy CMAC. In initial stages, feedbackcontroller FC plays the main role and Fuzzy CMAC are continually trained by means ofFIE output signals and at last it replaces feedback controller. This control strategythrough learning and control at the same time has good performances. This avoidsadopting direct feedback error in training which may lead to super saturation regulation.Based on WINDOWS operation system, the control software was programmed byVisual C++ language;through using of open motion control card, and adoptingmodularization design idea, user tend to second development. By using neural networkcomplicated curve neural network interpolation is done, and by using neural networkinterpolation and using motion control card realization interpolation. This method hasbeen demonstrated to be effective and practical by experiment results. Eventually,MATLAB engine combined with motion control card function is adopted, through themedium of Visual C++ program and fuzzy toolbox in NC system. And thus NC systemfuzzy PID control is appropriately realized, program work is simplified, and very goodcontrol purpose is attained.