Estudo comparativo entre diferentes controladores PID baseados em técnicas de predição aplicados a processos com atraso de transporte

Abstract

The dead-time is the time difference between the manifestation of the input signal and the output signal of the system, present in some industrial systems. Considering that PID controllers (Proportional, Integral and Derivative) are predominant in the industrial environment, the conventional PID controller was studied at a first moment, since, for small transport delays, the control systems can be controlled adjusting the derivative gain of the controller; in a second moment, we studied the PID controller with Smith Preditor (SP), which uses the mathematical model of the system plant to predict the delay and thus reduce its effect on the control loop. However, in the industry, many controllers already have a fixed gain structure of a conventional PID controller that can not be modified, so, in a third moment, were studied some PID controller tuning techniques based in prediction, among them, a) The PID Approximation from the Smith Preditor, which uses a control loop equivalent to the Smith Preditor to approach a conventional PID; b) The PID setting with Internal Model Controller (IMC), in which it connects the model that describes the process with the controller structure; c) The PID setting per pole allocation, similar to the IMC method, but allowing the poles to be looped in the system at desired locations; d) The PID adjustment with two degrees of freedom (2DOF), which consists in placing a filter in the reference of the control loop. The adjusted controllers were applied in a fluid level didactic workbench, the PID control with SP obtained the best results in closed loop, followed by controllers based on prediction techniques, followed by conventional PID controllers.