Hardware in the Loop Wave Energy Converter Control Under Control Faults and Model Mismatch

Ocean wave energy conversion is operated in a challenging environment featuring massive and impulsive stresses and is subjected to large amounts of biofouling and marine growth. As most wave energy converter technologies are off-shore, operation, maintenance, and repair costs may be significant. Therefore, any cost-effective control and operation of Wave Energy Converters (WECs) must be robust to faults and changes in the operational environment. This paper presents a Bayesian policy gradient-based WEC control approach that is responsive and adaptive to faults in controller feedback, controller actuation, and changes in the plant model. The controller is tested via Hardware in the Loop real-time testbed. The results show that the proposed control is capable of adapting its policy to the loss of information from one or two sensors and recovering nearly full power operation. In the case where sensor loss occurs, after the loss of plate excitation force estimation and displacement sensors, the power decreases to 215.2 kW from 252.7 kW before recovering to 238.4 kW mean power over 20 minutes. Moreover, the Bayesian RL-based control model proves its adaptability by compensating a large portion of power loss due to PTO failures and rapid changes to the dynamic system because of non-destructive damage to the float structure.