Sintonia de um controlador PID adaptativo para controle do superaquecimento através do algoritmo da colônia artificial de abelhas

Abstract

Classic approaches such as Ziegler-Nichols and Cohen-Coon methods have been used for years in tuning PID controllers, especially when there is not much information about the system subject to control. Besides these methods allow stable and robust control, the controller gains are never guaranteed for being optimal. When the system are nonlinear, of high order or have time delay, the conventional tuning methods sometimes fail to achieve satisfactory performance. Thus, other approaches have been used such as bio-based methods. The Artificial Bee Colony (ABC) algorithm were used to adjust the gainf of an adaptive PID controller to regulate the opening of an electronic expansion valve. A dynamic model obtained from experimental tests was used in the controller design. The controller effectiveness was evaluated through computer simulations and compared to classical methods. The experimental tests showed that the ABC adaptive controller could track the setpoint after a setpoint change from 10 °C to 7 °C. The control action was also compared to the SIMC method and the responses were similar, although the ABC method tracks the set point faster with a smaller overshoot and a smaller MAE (Absolute maximum error), RMSE (Root mean squared error) and IAE (Integral of absolute error) in two of the three tests.Comparing adaptive and non-adaptive controllers, the adaptive controllers were slower to track the setpoint but presented less overshooting and oscilations, in a general way, the non-adaptive controller was better than the adaptive one when comparing the performance indexes. The expected was that the adaptive was better. The results indicated that the proposed controller provides good disturbance rejection and set point tracking, and was able to control the superheating efficiently.