Disturbance Rejection Design For Some Typical Systems Based On The Equivalent-input-disturbance Compensation

Abstract:External disturbances are often encountered in industrial process control, such as electric power systems and mechanical systems. These disturbances cause not only the movement of the working point, but also bad dynamic and steady-state performance of the system. In order to reject the disturbances and to achieve good control performance, researchers have made a great effort on the study of theories and applications. However, there are still some drawbacks in the existing methods. For example, the trade-off between the control performance and the robust stability restricts the disturbance rejection effect; the inverse model of the plant may cause cancelation of unstable zeros and poles; some methods are difficult to realize due to the complexity of the coupling parameters design in the system. So, it is significant to find a simple, effective and practical disturbance rejection method.This paper focuses on the disturbance rejection problem for some typical systems based on the idea of the equivalent-input-disturbance (EID). Some methods are applied in this study, such as the pole placement method, multi-parameter optimization and Lyapunov quadratic stability (LQS) method. The main results and the contributions are listed as follows.(1) A disturbance rejection method for non-minimum-phase plants based on the pole placementA disturbance rejection method is presented based on the pole placement by analyzing the characteristics of a non-minimum-phase plant. The configuration of the system contains a disturbance estimator, a generalized state observer (GSO), a state feedback and an internal model. The GSO is employed to estimate the EID. The gains of the state feedback and the observer are designed separately. The stability condition is derived using the small-gain theorem and a controller design algorithm is further developed based on the pole placement method. Simulation results demonstrate that the presented method achieves good disturbance rejection performance for both minimum-and non-minimum-phase plants.(2) Robust disturbance rejection for a system with time-varying structured uncertaintiesTo solve the problem of parameters coupling in the configuration of the control system, an EID-based robust disturbance rejection method is presented for a system with time-varying structured uncertainties. A state-space model of the closed-loop system is derived using matrix transformation. The robust stability condition is presented in the form of an LMI using the LQS method. A parameter coupling design algorithm is further given for the controller design. Simulations and an experiment of a rotational-speed control system demonstrate that the method handles modeling uncertainties and rejects both matched and unmatched disturbances effectively.(3) Disturbance rejection for state time-delay systems based on the parameter optimizationA disturbance rejection method is presented for a state time-delay plant based on the parameter optimization. The time-delay observer is employed to reconstruct the state of the plant and to actively estimate the disturbances. A standard model for the stability problem of the time-delay system is derived. Then a stability condition and a parameter optimal design algorithm are given using a known stability criterion. Comparison simulations to the sliding-mode method demonstrate that the method handles time-delay and rejects any external disturbance effectively.(4) Aperiodic robust disturbance rejection for uncertain repetitive control systemsA repetitive control system (RCS) is often affected by both periodic and aperiodic disturbances. For an RCS that contains a strictly proper plant with time-varying uncertainties, an aperiodic robust disturbance rejection method is presented. An EID-based estimator is introduced into the RCS to yield a modified RCS. A robust stability condition for the system and a parameter coupling design for the controller are presented using the LQS method. In this system, the repetitive controller ensures tracking of a periodic reference input, and the disturbance estimator enables rejection of both unknown periodic and aperiodic disturbances. Simulation results demonstrate good performance of the method.(5) Decoupling control for multivariable systems based on the disturbance compensationA decoupling control method is presented for multivariable systems with disturbances based on the disturbance compensation. In this method, the undesirable coupling part in one loop is treated as the "disturbance" imposed on this loop. Then the "disturbance" and external disturbances are actively compensated for using the EID approach. The stability condition and controller design are given for each equivalent single loop. The control parameters in each loop can be designed independently. A practical example demonstrates good performance for both the decoupling control and the disturbance rejection.