Passive predictive controllers for wave energy converters: Design, implementation, and evaluation on a real-device benchmark

Abstract

This paper presents a unified framework for passive predictive control of a wave energy converter (WEC), evaluated within a nonlinear benchmark platform. Two receding-horizon controllers are proposed: a nonlinear model predictive controller (NLMPC), which directly optimises the power take-off (PTO) force under explicit passivity constraints, and a predictive passive damping controller (PPDC), which optimises a non-negative damping coefficient BPTO(t)≥0, yielding structural passivity PPTO=BPTOv2≥0 by construction and reducing optimisation complexity. Both controllers share a causal online estimation and autoregressive prediction pipeline requiring no upwave measurements. Evaluated across three irregular JONSWAP sea states and 20 wave realisations per condition, the NLMPC achieves higher gains in short- and medium-period seas (＋31.9% and ＋29.9% over an optimally tuned constant-damping baseline), while the PPDC prevails in the long-period regime (＋25.9% versus ＋17.0%). A covariance-based power decomposition shows that PPDC gains arise primarily from operating at lower mean damping levels. The results reveal a fundamental trade-off between control authority, structural passivity, and computational complexity in passive predictive WEC control.