Estratégia de controle distribuído aplicado em plantas Ball Balancer
Carregando...
Arquivos
Data
2021-05-04
Autores
Silva, Rafael Máximo da
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Nesta dissertação foi realizado um estudo sobre aplicação de controle distribuído em plantas Ball Balancer no Matlab® e Simulink®, bem como a implementação em bancada com dois Ball Balancer, usando técnicas clássicas de controle. Para obter o controle cooperativo das plantas, foi considerada a possibilidade de chaveamento entre controladores em cada instante de amostragem, para o sistema que está em pior condição em relação ao ponto de operação desejado. Para avaliar isso, considerou-se que o pior caso é a planta que tem a maior norma no sinal de controle, ou seja, a maior resposta ao controle Proporcional, Integral e Derivativo (PID). Também, foi utilizada a estratégia de obter o índice de desempenho com menor esforço computacional, calculando o valor absoluto do erro do sistema. Inicialmente, as plantas foram analisadas, modeladas e identificadas, tendo como referência o módulo didático 2D Ball Balancer da QUANSER®. Posteriormente, o controle individual da planta em nível de simulação foi realizado no Matlab® e Simulink®. Por fim, foram comparados os resultados entre o controle considerando o seletor de chaveamento automático de controle e sem o seletor atuando em período determinado, mostrando uma melhor resposta com a estratégia do seletor. Na implementação foi utilizado o Arduino Mega com shield Ethernet, para o controle das duas plantas e comunicação com supervisório Elipse E3®. Para visualização dos gráficos, e ajustes de ganhos do controlador PID, foi confeccionada uma interface gráfica em Elipse E3®, que se comunica com o Arduino através do protocolo Modbus TCP. O vídeo do sistema de controle com dois Ball Balancer implementado em laboratório, sem levar em conta o índice de desempenho, pode ser visto no link https://youtu.be/MuFVLBfAST0, enquanto que o controle que leva em consideração o índice de desempenho se encontra no link https://youtu.be/7EdA0D4ycvY.
In this dissertation, a study was carried out on the application of distributed control in Ball Balancer systems in Matlab® and Simulink®, as well as the implementation in workbench with two Ball Balancer, using classical control techniques. To obtain cooperative control of the plants, the possibility of switching between controllers at each sampling instant was considered, for the system that is in a worse condition concerning the desired point of operation. To evaluate this, it was considered that the worst case is the plant that has the highest modulus in the control signal, that is, the highest response to Proportional, Integral, and Derivative (PID) control. Also, the strategy of obtaining the performance index with less computational effort was used, calculating the absolute value of the system error. Initially, the plants were analyzed, modeled, and identified, using the QUANSER® 2D Ball Balancer didactic module as a reference. Subsequently, the individual control of the plant at the simulation level was carried out in Matlab® and Simulink®. Finally, the results were compared between the control considering the automatic control switch selector and without the selector operating in a determined period, showing a better response with the selector strategy. In the implementation, the Arduino Mega with Ethernet shield was used to control the two plants and communicate with the Elipse E3® supervisory. In order to visualize the graphs and adjust the gains of the PID controller, a graphic interface was made in Elipse E3®, which communicates with the Arduino through the Modbus TCP protocol. The video of the control system with two Ball Balancers implemented in the laboratory, without taking into account the performance index, can be seen at the link https://youtu.be/MuFVLBfAST0, while the control that takes into account the performance index is found at the link https://youtu.be/7EdA0D4ycvY.
In this dissertation, a study was carried out on the application of distributed control in Ball Balancer systems in Matlab® and Simulink®, as well as the implementation in workbench with two Ball Balancer, using classical control techniques. To obtain cooperative control of the plants, the possibility of switching between controllers at each sampling instant was considered, for the system that is in a worse condition concerning the desired point of operation. To evaluate this, it was considered that the worst case is the plant that has the highest modulus in the control signal, that is, the highest response to Proportional, Integral, and Derivative (PID) control. Also, the strategy of obtaining the performance index with less computational effort was used, calculating the absolute value of the system error. Initially, the plants were analyzed, modeled, and identified, using the QUANSER® 2D Ball Balancer didactic module as a reference. Subsequently, the individual control of the plant at the simulation level was carried out in Matlab® and Simulink®. Finally, the results were compared between the control considering the automatic control switch selector and without the selector operating in a determined period, showing a better response with the selector strategy. In the implementation, the Arduino Mega with Ethernet shield was used to control the two plants and communicate with the Elipse E3® supervisory. In order to visualize the graphs and adjust the gains of the PID controller, a graphic interface was made in Elipse E3®, which communicates with the Arduino through the Modbus TCP protocol. The video of the control system with two Ball Balancers implemented in the laboratory, without taking into account the performance index, can be seen at the link https://youtu.be/MuFVLBfAST0, while the control that takes into account the performance index is found at the link https://youtu.be/7EdA0D4ycvY.
Descrição
Palavras-chave
Controle Distribuído, Ball Balancer, Controle proporcional integral e derivativo (PID), Software SCADA, Arduino, Distributed Control, Ball Balancer, proportional integral and derivative Control (PID), SCADA software