Collaborative design model for a point absorber wave energy converter in irregular waves coupling GA-based geometric optimization and DRL-Based control

Ocean wave energy is a promising renewable resource that can be harvested by Wave Energy Converters (WECs). For a point absorber WEC, the energy capture performance is highly influenced by the hydrodynamics and implemented control strategy. Traditional sequential design optimizes the floater's geometry to obtain better hydrodynamics first, and refines the control strategy afterwards, ignoring the interaction between these two processes and failing to find the global optimal solution. To address this challenge, this study proposes a collaborative design model coupling geometric optimization based on Genetic Algorithm (GA) and control strategy based on Deep Reinforcement Learning (DRL) for a point absorber WEC in irregular waves. Through the proposed model, the optimal solution balancing hydrodynamics and controllability can be found globally in one strongly coupled process. The wave energy capture performance of the proposed model is compared to the traditional sequential optimization to show its capability and necessity. More results concerning the generalization of the proposed model in different wave conditions are discussed. Compared with previous co-design studies which optimized the geometry with passive control, the proposed co-design model trains the independent DRL-based control strategy for each geometry, which advances the co-design studies from a perspective of active control.