Lead PI: Dr Peng Wu, University College London
Summary:
This project addresses the critical safety challenge of integrating Battery Energy Storage Systems (BESS) into Dynamic Positioning (DP) vessels operating in uncertain marine environments. Current systems struggle to balance stochastic environmental loads against strict redundancy requirements, risking battery health and operational safety.
We will develop a physics-informed energy management framework combining a degradation-aware digital twin with constrained deep reinforcement learning. Leveraging existing validated ship models and industrial datasets, this six-month project will deliver a redundancy-aware AI agent capable of guaranteeing power availability during critical missions, and minimising battery degradation under stochastic environmental uncertainties.
