Control strategy for performing predictions for a semi-active compensation system

Abstract

The heave compensation system can decouple the vertical movement of a load from the vertical movement of a ship to reduce the impact of waves on offshore operations. In this paper, we present a new prediction and control strategy for semi-active heave compensation system to isolate the load motion from the ship heave motion. The novelty of the present system is reflected in three folds as follows. 1) We propose an improved autoregressive (AR) method optimized by the genetic algorithm (GA) to predict a ship's heave position. 2) We present an appropriate fitness function and adaptive probability functions to find the best prediction parameters in the current environment, deal with slow changes in the waves and enable the GA continuously modify the prediction model to improve the prediction accuracy in the entire working process. 3) According to the dynamic model of the system, we design an adaptive backstepping sliding mode controller to suppress the influence of nonlinear factors and disturbances. Our experiments on the Stewart platform satisfactorily validate the present control strategy and the prediction compensation performance of the entire system.