Self-optimising sampled-data control system employing power-constraint equalisation

In this paper, a sampled-data control system is continuously optimised by a sampled-data self-optimising loop which operates on the controller gain. The action of the optimising loop is to maintain the controller output power at its maximum permissible level. The optimising-loop controller is actuated by the error between the desired system-controller output power and the system-controller output power obtained, and does not require the perturbation action normally associated with hill-climbing techniques. A method of adjusting the optimising-loop gain, to suit various operating conditions, is proposed, and a model-reference technique is employed to minimise the error in the measurement of the controller power over finite intervals. Algorithms have been developed to facilitate the transient analysis of controller power, following changes in the controller gain, and to determine the mean-square error in the measurement of the controller power over finite intervals.