Some Studies on Uncertainty and Disturbance Estimation Based Techniques for Control System Applications

Abstract

Many physical plants in real world are multivariate and highly nonlinear. Accurate newlinemathematical models of physical plants are usually not available in industrial newlinecontrol. There are often problems such as parametric uncertainties, unmodelled dynamics, newlinenonlinearities, time delays, cross couplings and external disturbances. It is newlinechallange for control practioners to design accurate control strategy despite such undesirable newlinecharacteristics of systems. For continuously modulated control, a feedback newlinecontroller is used to automatically control a process or operation. Linear controllers newlinelike PID controllers are commonly used and are successful. However, it can lead to newlineinstability, due to the presence of time delays, parametric uncertainties and frequent newlineload changes. Model-based controller design algorithms involve a system model that newlineis deriving from the system input-output data. Due to plant-model mismatch and newlineparametric uncertainty in the model obtained, it has limitations in practical implementations. newlineThere are multiple uncertainties in the operating environment of various applications newlinelike marine vehicles, nuclear reactors, robotic systems, etc. Such multivariable newlinesystems are subjected to external perturbations and are mostly unmeasurable. To address newlinethem, a variety of techniques have been developed for robust control like High newlinegain control, Integral control, H-infinity control, Sliding mode control, etc. When a newlinedisturbance is measurable, the feedforward strategy could attenuate or eliminate the newlineinfluence of disturbance. However, the external disturbance cannot be directly measured newlineor it is too expensive to measure. The method to deal with this problem is to newlineestimate the disturbance or its influence from measurable variables. Then, a control newlineaction based on the disturbance estimate can be used to compensate the effect of newlinethe disturbance. As a result, modern advanced robust control algorithms include a newlinecompensator in addition to disturbance estimating techniques. newlineThe conventional sliding mode control is robust compen