Control Of The Electro-hydraulic Servo Systems With Unknown Dynamics

In recent years,the integration of cross-disciplinary work stimulates rapid development of new ideas,theories and technologies.The investigation of the electro-hydraulic servo systems control retains to be a hot topic in the field.However,as the key part of the electro-hydraulic servo systems,the prominent nonlinearities of the hydraulic servo valves and hydraulic cylinders lead to inevitable unmodeled dynamics in the control systems,which include parametric variations,modeling errors,disturbance,uncertain load and so on.The traditional control schemes,e.g.,PID and backstepping control,are difficult to tackle the problems of unmodeled dynamics.On the other hand,the “explosion of complexity” issue encountered in the backstepping control may be prominent when the controller for the high order systems is designed in backstepping procedure.Moreover,the backstepping method needs all of the information of the system states,which is restrictive in the practical industrial application.Hence,the design of novel controller for the electro-hydraulic servo systems should be furtherly studied to handle the aforementioned problems.The electro-hydraulic servo systems with unmodeled dynamics are studied in this thesis.The main work is as following:(1)Modelling of the electro-hydraulic servo system.First the studied servo system is depicted by a structural diagram.Based on the diagram,mechanical analysis is made and then the mathematical model of the servo system is established according to the electric circuit and the dynamics of the hydraulic components.Finally,we select appropriate state variables to formulate the state equations of the electro-hydraulic servo systems,which are in a strict-feedback form.(2)An adaptive backstepping control method for the electro-hydraulic servo systems based on neural networks.In this section,we introduce the radial basis function based adaptive neural networks to deal with the unknown dynamics and the derivative of the virtual laws in the backstepping.An adaptive law with ?-modification is designed to update the weight parameters for the neural networks online,which improves the convergence rate of the parameters.Finally,a novel backstepping control method for the electro-hydraulic servo systems based on the adaptive neural networks is proposed to address the effects of unknown dynamics and remedy the “explosion of complexity” issue in backstepping control.Theoretical analysis and experiments are carried out to validate the reliability and practicability of the proposed method.(3)A novel backstepping control scheme for the electro-hydraulic servo systems based on the unknown dynamics estimator.To improve the sluggish learning of neural networks,simplify the structure of controller and reduce the tuning parameters,a novel unknown dynamics estimator with guaranteed robustness is designed to replace the adaptive neural networks.The unknown dynamics estimator is used to estimate the unknown dynamics and the derivative of the virtual laws in the backstepping.Then a backstepping control scheme for the electro-hydraulic servo systems based on the unknown dynamics estimator is proposed to compensate the effects of unknown dynamics and improve the transient performance of control systems,which also reduces the calculation cost of controller.Theoretical analysis and experiments are carried out to validate the correctness and practicability in industrial application of the proposed method.(4)The unknown dynamics estimator based output feedback control for the electro-hydraulic servo systems.To furtherly improve the practicability in industrial application of the control algorithm and remove the requirement on the system states,a coordinate transform is employed to transform the original state space equation into a canonical form.A high order sliding mode observer that converges in finite time is used to reconstruct the transformed unknown system states.At the same time,the unknown dynamics estimator is utilized to estimate the lumped unknown dynamics after transformation.The unknown dynamics estimator based output feedback control approach for the electro-hydraulic servo systems is proposed.Theoretical analysis and experiments are carried out to validate the correctness and reliability in industrial application of the proposed method.(5)Prescribed performance guaranteed output feedback control method.To further simplify the structure of the proposed controller and reduce the difficulty of the controller design,we introduce a prescribed performance function based on equivalent coordinate transform and then incorporate the prescribed performance function with the controller design.All of the signals in the close loop system are retained in a prescribed bounded set by using the prescribed performance function,which further improves the transient system performance.