Butterfly Hysteresis Modeling And Adaptive Control Of Dielectric Elastomer Actuated Systems

With the continuous development of material science,the types of smart materials are gradually enriched,mainly forming actuators of rigid smart materials represented by piezoelectric materials and magnetostrictive materials and flexible smart materials represented by dielectric elastomer materials and IPMC materials.With the continuous enrichment of smart material actuators,the hysteresis nonlinearity embodied in different types of smart material actuators is also different.Unlike the single-loop hysteresis phenomenon in traditional rigid smart materials,the double-loop(butterfly hysteresis)dynamic hysteresis characteristics of dielectric-flexible actuators reflect a non-smooth nonlinear relationship between the input and output signals with memory characteristics.The presence of hysteresis nonlinearities often results in inaccurate actuator positioning and,when these actuators are used in closed-loop control systems,can cause system oscillations and even destabilize the system.Therefore,they need to be compensated and eliminated using modeling and control methods.The main research of the thesis includes.(1)For dielectric elastomer materials exhibiting butterfly hysteresis characteristics under unbiased periodic input signals,a butterfly relay operator is proposed as the basic operator for constructing a butterfly Preisach model,and the denstiy functions/weights of the developed model are defined forward,which facilitates the construction of unique inverse models and controller design.A piecewise butterfly Preisach model is proposed to improve the modeling accuracy further to illustrate the variety of asymmetric nonlinear hysteresis effects with different input amplitudes.To verify the effectiveness of the developed modeling approach,the established dielectric elastomer drive system experiments and the experimental results illustrate the effectiveness of the proposed model.(2)The butterfly hysteresis operator,including the butterfly play operator,the butterfly Krasnosel'skii-Pokrovskii kernel,and the butterfly asymmetric shifted operator,is further investigated for the butterfly hysteresis embodied by the dielectric elastomer driver to describe the butterfly hysteresis effect.To further improve the modeling accuracy,a butterfly hysteresis structure is developed,including the butterfly hysteresis operator and a neural network for the corresponding weights of the butterfly hysteresis operator and the unmodeled dynamic prediction.The validation of the proposed structure illustrates the effectiveness of the proposed model.(3)In order to mitigate the butterfly hysteresis characteristics and creep characteristics in dielectric elastomer drives,a novel butterfly hysteresis hidden inverse compensation output feedback control algorithm for butterfly hysteresis with creep characteristics is proposed in this paper.The implicit inverse compensation is an online decoupling mechanism that obtains an approximation of the actual control signal from the designed temporary control signals of hysteresis and creep,where the actual control signal is coupled.For the first time,a butterfly-shaped hysteresis characteristic model with creep is constructed for a dielectric elastomer drive.Finally,the tracking error is achieved by applying initialization techniques to set initial values in the adaptive law and virtual control signals.The test results show the effectiveness of the proposed control scheme.