Aplicación de control de matriz dinámica a planta piloto.
Loading...
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
En el presente documento se detalla el desarrollo y la implementación del controlador Control de Matriz Dinámica (DMC) para el control del lazo de flujo de la planta piloto del Laboratorio de Control del Departamento de Ingeniería Eléctrica de la Universidad de Concepción. Para la identificación paramétrica del proceso se utiliza el Filtro de Kalman Extendido (EKF). Además, se desarrolla una interfaz gráfica para facilitar la interacción entre el operador de la planta y el algoritmo. Así, el presente proyecto consiste en estimar los parámetros de un proceso sujeto a ruido, para posteriormente aplicar control.
En primer lugar, se realiza un estudio bibliográfico del EKF. Luego, dicho algoritmo se implementa en una simulación en Matlab, con el fin de estimar los parámetros de un proceso de primer orden con retardo cuya medición, o salida, es afectada por ruido blanco. Al finalizar la sintonización del filtro se prosigue con la siguiente etapa.
En segundo lugar, se hace el estudio bibliográfico del DMC. Luego, es implementado en una simulación en Matlab, cuya entrada es la salida del EKF, es decir, los parámetros estimados del proceso de primer orden con retardo. Para finalizar esta etapa, se sintoniza el controlador para obtener una respuesta adecuada.
En tercer lugar, se implementa el controlador en un PLC, teniendo presente que estos son limitados en cuanto a su poder de procesamiento, y se desarrolla una interfaz gráfica. Por ende, tanto el EKF como una parte del algoritmo del DMC son ejecutados paralelamente al controlador en Matlab, siendo necesaria la comunicación entre el PLC y Matlab para la lectura y escritura de datos. Esta sección es probada y simulada utilizando Simulink para simular la planta piloto, Matlab para ejecutar el EKF y una parte del DMC, y RSLogix 5000 para simular el controlador, es decir, la parte final del DMC.
Finalmente, se hacen pruebas en la planta piloto y se validan las simulaciones. Se encuentra que el controlador regula de manera adecuada el proceso y que el filtro es capaz de estimar los parámetros del proceso con bajo error RMS.
The present document details the development and implementation of the Dynamic Matrix Control (DMC) controller for the flow loop control of the pilot plant in the Control Laboratory of the Department of Electrical Engineering at the University of Concepción. The Extended Kalman Filter (EKF) is used for the parametric identification of the process. Additionally, a graphical interface is developed to facilitate the interaction between the plant operator and the algorithm. Thus, the present project consists of estimating the parameters of a process subject to noise, in order to later apply control. First, a literature review of the EKF is conducted. Then, this algorithm is implemented in a Matlab simulation to estimate the parameters of a first-order process with delay, whose measurement or output is affected by white noise. After tuning the filter, the next stage is carried out. Second, a literature review of the DMC is conducted. Then, it is implemented in a Matlab simulation, where the input is the output of the EKF, i.e., the estimated parameters of the first-order process with delay. To conclude this stage, the controller is tuned to obtain an adequate response. Third, the controller is implemented in a PLC, keeping in mind the processing power limitations of these devices, and a graphical interface is developed. Therefore, both the EKF and part of the DMC algorithm are executed in parallel to the controller in Matlab, making it necessary to establish communication between the PLC and Matlab for reading and writing data. This section is tested and simulated using Simulink to simulate the pilot plant, Matlab to execute the EKF and part of the DMC, and RSLogix 5000 to simulate the controller, i.e., the final part of the DMC. Finally, tests are conducted on the pilot plant, and the simulations are validated. It is found that the controller adequately regulates the process and that the filter can estimate the process parameters with low RMS error.
The present document details the development and implementation of the Dynamic Matrix Control (DMC) controller for the flow loop control of the pilot plant in the Control Laboratory of the Department of Electrical Engineering at the University of Concepción. The Extended Kalman Filter (EKF) is used for the parametric identification of the process. Additionally, a graphical interface is developed to facilitate the interaction between the plant operator and the algorithm. Thus, the present project consists of estimating the parameters of a process subject to noise, in order to later apply control. First, a literature review of the EKF is conducted. Then, this algorithm is implemented in a Matlab simulation to estimate the parameters of a first-order process with delay, whose measurement or output is affected by white noise. After tuning the filter, the next stage is carried out. Second, a literature review of the DMC is conducted. Then, it is implemented in a Matlab simulation, where the input is the output of the EKF, i.e., the estimated parameters of the first-order process with delay. To conclude this stage, the controller is tuned to obtain an adequate response. Third, the controller is implemented in a PLC, keeping in mind the processing power limitations of these devices, and a graphical interface is developed. Therefore, both the EKF and part of the DMC algorithm are executed in parallel to the controller in Matlab, making it necessary to establish communication between the PLC and Matlab for reading and writing data. This section is tested and simulated using Simulink to simulate the pilot plant, Matlab to execute the EKF and part of the DMC, and RSLogix 5000 to simulate the controller, i.e., the final part of the DMC. Finally, tests are conducted on the pilot plant, and the simulations are validated. It is found that the controller adequately regulates the process and that the filter can estimate the process parameters with low RMS error.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Electrónico/a.
Keywords
Reguladores eléctricos, Algoritmos