| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.35 MB | Adobe PDF |
Advisor(s)
Abstract(s)
Este trabalho de pesquisa e desenvolvimento tem como fundamento principal o Conceito
de Controlo por LĂłgica Difusa. Utilizando as ferramentas do software Matlab, foi possĂvel
desenvolver um controlador com base na inferĂȘncia difusa que permitisse controlar
qualquer tipo de sistema fĂsico real, independentemente das suas caracterĂsticas.
O Controlo LĂłgico Difuso, do inglĂȘs âFuzzy Controlâ, Ă© um tipo de controlo muito
particular, pois permite o uso simultĂąneo de dados numĂ©ricos com variĂĄveis linguĂsticas
que tem por base o conhecimento heurĂstico dos sistemas a controlar. Desta forma,
consegue-se quantificar, por exemplo, se um copo estĂĄ âmeio cheioâ ou âmeio vazioâ, se
uma pessoa Ă© âaltaâ ou âbaixaâ, se estĂĄ âfrioâ ou âmuito frioâ.
O controlo PID Ă©, sem dĂșvida alguma, o controlador mais amplamente utilizado no
controlo de sistemas. Devido à sua simplicidade de construção, aos reduzidos custos de
aplicação e manutenção e aos resultados que se obtĂȘm, este controlador torna-se a primeira
opção quando se pretende implementar uma malha de controlo num determinado sistema.
Caracterizado por trĂȘs parĂąmetros de ajuste, a saber componente proporcional, integral e
derivativa, as trĂȘs em conjunto permitem uma sintonia eficaz de qualquer tipo de sistema.
De forma a automatizar o processo de sintonia de controladores e, aproveitando o que
melhor oferece o Controlo Difuso e o Controlo PID, agrupou-se os dois controladores,
onde em conjunto, como poderemos constatar mais adiante, foram obtidos resultados que
vão de encontro com os objectivos traçados.
Com o auxĂlio do simulink do Matlab, foi desenvolvido o diagrama de blocos do sistema de
controlo, onde o controlador difuso tem a tarefa de supervisionar a resposta do controlador
PID, corrigindo-a ao longo do tempo de simulação. O controlador desenvolvido é
denominado por Controlador FuzzyPID.
Durante o desenvolvimento prĂĄtico do trabalho, foi simulada a resposta de diversos
sistemas Ă entrada em degrau unitĂĄrio. Os sistemas estudados sĂŁo na sua maioria sistemas
fĂsicos reais, que representam sistemas mecĂąnicos, tĂ©rmicos, pneumĂĄticos, elĂ©ctricos, etc., e que podem ser facilmente descritos por funçÔes de transferĂȘncia de primeira, segunda e
de ordem superior, com e sem atraso.
This development and research thesis have as principal concept the Fuzzy Logic Control. Using software tools of Matlab was possible to develop a controller based on Fuzzy Inference which allowed controlling any kind of real physical system, regardless their characteristics. The Fuzzy Logic Controller is a very particular type of controller, as it allows the simultaneous use of numeric data with linguist variables that is based on heuristic knowledge on the controlling systems. Thus, it enables to quantify, for example, if a glass is âhalf fullâ or âhalf emptyâ, if a person is âtallâ or âshortâ or if it is âcoldâ or âvery coldâ. The PID Controller is, without doubt, the most widely used controller in the control systems. Due to its simplicity of construction, the low cost implementation, maintenance and the good results that are obtained, this controller makes the first option when you intend to implement a control loop in a given system. Characterized by three adjusting parameters, which are proportional, integral and derivative components, those together allow efficient tuning of any type of system. In order to automate the process of tuning control and using what Fuzzy and PID Controllers best offer, grouped the two controllers, where together, as will be presented further, the obtained results meet the defined objectives. With the suport of Matlab Simulink, was developed a block diagram of the control system, where the Fuzzy Controller has the task of supervising the response of the PID Controller, correcting it during the simulation. The developed controller was named by FuzzyPID Controller. During the pratical development was simulated the response of several systems to unit step input. The systems studied are mostly real physical systems, which represent mechanical, thermal, pneumatic, electrical, etc, and they can be easily described by a transfer functions of first, second and higher order, with and without time delay.
This development and research thesis have as principal concept the Fuzzy Logic Control. Using software tools of Matlab was possible to develop a controller based on Fuzzy Inference which allowed controlling any kind of real physical system, regardless their characteristics. The Fuzzy Logic Controller is a very particular type of controller, as it allows the simultaneous use of numeric data with linguist variables that is based on heuristic knowledge on the controlling systems. Thus, it enables to quantify, for example, if a glass is âhalf fullâ or âhalf emptyâ, if a person is âtallâ or âshortâ or if it is âcoldâ or âvery coldâ. The PID Controller is, without doubt, the most widely used controller in the control systems. Due to its simplicity of construction, the low cost implementation, maintenance and the good results that are obtained, this controller makes the first option when you intend to implement a control loop in a given system. Characterized by three adjusting parameters, which are proportional, integral and derivative components, those together allow efficient tuning of any type of system. In order to automate the process of tuning control and using what Fuzzy and PID Controllers best offer, grouped the two controllers, where together, as will be presented further, the obtained results meet the defined objectives. With the suport of Matlab Simulink, was developed a block diagram of the control system, where the Fuzzy Controller has the task of supervising the response of the PID Controller, correcting it during the simulation. The developed controller was named by FuzzyPID Controller. During the pratical development was simulated the response of several systems to unit step input. The systems studied are mostly real physical systems, which represent mechanical, thermal, pneumatic, electrical, etc, and they can be easily described by a transfer functions of first, second and higher order, with and without time delay.
Description
Keywords
LĂłgica difusa PID Matlab Simulink Função de transferĂȘncia Sistemas fĂsicos Fuzzy logic Transfer function Physical systems
Citation
Publisher
Instituto Politécnico do Porto. Instituto Superior de Engenharia do Porto