A low order computer model for adaptive speed control of diesel driven power-plants

An adaptive control scheme, based on a low-order model of a diesel driven power plant, is used for the speed control of the prime-mover. By using an explicit identification of the delay, it is shown that a low-order identification model can prove to be adequate even when the actual plant delay is time-varying. It is shown both by the frequency response studies and by observing the output error variance that the model can be used to obtain an accurate prediction and good control in the speed loop of the plant. The performance is compared with a fixed parameter PI (proportional-integral) controller tuned to the plant, and a significant improvement is observed in both peak overshoots and settling times. The identifier/controller is robust enough to operate under the effect of flexible couplings, though large disturbances may be imposed. Correction terms may be used to account for droop factors. This, in practice, can result in significant computational gains without a significant loss of accuracy.<>