Characteristic Modeling And Adaptive Control For Multi-motor Driving Systems

The multi-motor driving servo systems play an irreplaceable role in driving the largepower or large inertia system. The controller design of multi-motor driving servo systembecomes difficult for the high order, nonlinear and complex model obtained by thetraditional method of dynamic modeling due to the backlash, friction and other nonlinearfactors. The characteristic model of multi-motor driving servo systems is established andthe adaptive controller based on characteristic model is designed in this thesis. Theverification experimental is carried out on the four-motor driving system experimentplatform. The main research contents of this thesis can be summarized as follows:(1) Dynamic model of multi-motor driving system with backlash is established, andthe method of eliminating the effects of backlash is studied. Two methods are proposed toeliminate the backlash based on the dynamic model: a dynamic backlash controller isdesigned to eliminate the backlash by applying the dynamic bias torque. An extended stateobserver is used to observe the backlash nonlinearity as an extended state, and the controllaw of the system is designed by using the fast recursive dynamic surface, which caneliminate the backlash and the convergence in finite time at the same time. Theeffectiveness of these methods is shown by simulation.(2) Characteristic model is established by using the recursive least squares methodwith forgetting factor based on the dynamic model. And the stability ranges of thecharacteristic coefficients of the closed-loop system are analyzed. The effects of inputsignal, backlash size and sampling time on modeling accuracy are discussed. Thesimulation result shows that modeling error increases with the increase of backlash orsampling time, the convergence speed of parameters is the fastest in tracking the sinusoidalsignal, but the modeling error is larger.(3) A composite adaptive synchronization controller is designed based oncompensation for characteristic model. The closed-loop system is guaranteed by the goldensection control, the steady-state accuracy is improved by the logic integral control, backlashnonlinear effect is also compensated, and the synchronization of the multi-motors is also considered in this controller. In addition, the modified golden segmentation adaptivecontroller is designed, and the tracking performance of the controllers is verified bysimulation.(4) The composition of hardware and software of the four-motor experimentalplatform is introduced, and three kinds of designed controllers in this thesis are verified byexperiment, mainly including the dynamic anti-backlash controller based on dynamicmodel, composite adaptive controller based on characteristic model and the improvednonlinear golden section adaptive controller based on characteristic model. Trackingperformance of the three kinds of controllers is analyzed, the experimental results show thatthe adaptive control based on characteristic model has higher tracking accuracy and betterdynamic performance.At the end of this thesis the research work is concluded and the future work isprospected.