Optimal prediction horizon length in model predictive control to maximise energy absorption by a point absorber wave energy converter

Abstract

For real-time control in Model Predictive Control (MPC), predicting the future wave exciting force is essential. In MPC formulation, the prediction horizon length assesses how far in advance the prediction of wave exciting force is required. This paper proposes the optimal prediction horizon length in MPC to maximize energy absorption efficiency. The MPC formulation is tested numerically on a point absorber wave energy converter (PA-WEC) in regular waves. Heave oscillation is the only permitted motion of the floating body of the PA-WEC. The responses obtained by MPC are compared with those of the linear damping control. The analysis also indicates how the energy absorption efficiency depends on the prediction horizon length and sampling interval. Three time horizon lines have been investigated following the analysis of the time horizon with respect to frequency. After investigating absorbed power and energy absorption efficiency across three time horizon lines in relation to frequency, an optimal time horizon line is proposed. This line signifies the optimal prediction horizon length in MPC for the PA-WEC, ensuring higher energy absorption power and efficiency.