Predictive control of drum water level based on ant colony optimization algorithm

Drum water level is one of the main control parameters for turbo-generator unit. It is significant for drum water level researching as its many influence factors and false water level easily to bring about. The predictive control based on ant colony optimization (ACO) is used for the control of drum water level in this paper. The SVM with RBF kernel is used for training regression vector in order to accomplish water level dynamic systems modeling. ACO used for continuous function optimization is established. And the optimization of non-convex objective function is solved by the algorithm. Then the best control series can be obtained. From simulation results it shows that the algorithms have good control performance such as the rise time is 6*T_S, overshoot is as small as 3.0192%, much small steady-state error and so on.