A K-STEP PREDICTIVE SCHEME FOR SPEED CONTROL OF DIESEL DRIVEN POWER-PLANTS

Abstract

An adaptive speed control scheme for diesel prime movers is presented in this paper. Diesel prime movers are characterized by nonlinear, time-varying parameters including a nonlinear input dead-time variation that introduces an unknown delay between the injection of fuel and the production of engine torque. Such plants are difficult to control by conventional proportional-integral (PI) controllers. It is shown that by assuming the existence of an additional system time constant, an explicit estimate of the plant parameters and dead-time may be obtained. The algorithm is fast converging. The resulting predictor can then be used to set up a k-step ahead predictive controller. Studies that compare the performance of the adaptive scheme to that of fixed, tuned PI controller are presented. The algorithm is shown to be flexible enough to operate under a wide range of droop settings and is equally applicable to various operating conditions.