Design of controllers based on active disturbance rejection control (ADRC) and its integration with model predictive control (MPC)
- Autores: Blanca Viviana Martínez Carvajal
- Directores de la Tesis: Javier Sanchis Saez (dir. tes.), Sergio García Nieto (dir. tes.)
- Lectura: En la Universitat Politècnica de València ( España ) en 2023
- Idioma: español
- Tribunal Calificador de la Tesis: José Luis Guzmán Sánchez (presid.), Francesc Xavier Blasco Ferragud (secret.), Gloria Gutiérrez Rodríguez (voc.)
- Programa de doctorado: Programa de Doctorado en Automática, Robótica e Informática Industrial por la Universitat Politècnica de València
- Materias:
- Enlaces
- Tesis en acceso abierto en: RiuNet
- Resumen
Numerous and varied contributions based on the ADRC are currently available. On the one hand, some works address the ADRC methodology. Still, only some offer a comprehensive explanation of its design and application aimed at those researchers who are starting to explore this control strategy. On the other hand, the ADRC tuning and the ADRC-based composite control are open research areas. One of the discussions that remain active in the literature is related to how to select the LADRC main parameters so that closed-loop stability is achieved with appropriate disturbance rejection and robustness, mainly when the ADRC is used to control processes approximated by more straightforward representations such as the first-order plus delay (FOPDT) model. Likewise, the active estimation of uncertainty and disturbances has made integrating the ADRC topology with advanced control techniques, like Model-Based Predictive Control (MPC), attractive. The major challenge in realising this combination lies in how to formulate the control loop so that the ADRC disturbance rejection mechanism transforms the behaviour of the controlled system into that of a simplified desired plant, thus relaxing the requirement for a detailed model while directly considering the constraints on the loop variables.
This thesis presents three contributions to ADRC knowledge to address the challenges mentioned above. The first is a guide for designing and applying linear controllers using conventional active disturbance rejection control. This guide offers a review of the theoretical foundation of the ADRC. It condenses in an algorithm the steps for designing these control loops to facilitate their implementation according to the problem formulation in the disturbance estimation and rejection framework and the empirical selection of their gains. The second contribution of this dissertation is a set of tuning rules for computing the three distinctive parameters of the ADRC with which the state observer and control law gains are designed. These rules allow tuning the ADRC to control an approximate process using a first-order plus delay model and offer different sets of parameters according to a desired level of robustness. This contribution is based on developing multi-objective optimisation design procedures focused on controlling a group of nominal FOPDT plants. The results of these procedures were fitted to the tuning formulae provided. The third contribution is a new control architecture that combines the disturbance rejection mechanism of the ADRC and the receding horizon strategy of the MPC. In this loop, a predictive control law governs a first-order plus integrator plant enforced on the real process subject to constraints. The above is possible by compensating for the mismatch between the real and desired plants and incorporating the ADRC compensation term in the constraints formulation of the predictive controller. The loop is intended to provide a solution to control constrained systems for which no nominal model has been identified.
This dissertation addresses researchers interested in exploring active disturbance rejection control and those considering this technology as one of their main lines of research. The contributions of this dissertation serve those new to the study of ADRC, controller designers seeking to implement linear ADRC by considering the disturbance rejection response of processes approximated using first-order plus delay models, and researchers open to discussing the potential benefits of combining ADRC with advanced techniques such as MPC.
