Adaptive model predictive control based power system control for all-electric ships

The wide penetration of integration and the development of electrification make all-electric ships gradually become the future direction of ship development. Since the randomly changing sea state environment will produce uncertainty disturbances to the ship load power, it makes a serious challenge to maintain the shipboard power quality. In this paper, an adaptive model predictive control method is proposed to control the bus voltage, which is combined with online prediction algorithm to eliminate the control errors caused by random disturbances and improve the bus voltage control performance. An adaptive kernel based online sequential extreme learning machine algorithm is used to forecast the ship load power, making the model predictive control adaptable to random disturbances. The proposed method is verified on an equivalent model of a nominal shipboard power system with various types of sea state scenarios. Simulation results proved the algorithm accuracy and the advantages of the proposed bus voltage control method.