DeepGFM · Enabling Deep Grid Forming for Offshore HVDC Systems

Europe's offshore wind expansion demands robust high voltage direct current (HVDC) grids linking remote plants, energy hubs, and onshore grids. Today's grid following converters cannot stabilise weak grids, while the-state-of-the-art grid gorming (GFM) tightly couples AC-DC systems, heightening fault propagation risk. DeepGFM is developing a next generation HVDC GFM solution with three advances: (i) converter level GFM with embedded energy storage (supercapacitor/battery) to deliver improved inertia/frequency responses and absorb fault energy without HVDC braking chopper; (ii) dual port GFM control that ensures AC and DC side voltage source behaviours simultaneously, actively decoupling AC/DC dynamics during disturbances; (iii) a system wide coordination layer exposing operator tunable GFM parameters (inertia, droops, modes) with feasible region bounds and optimisation guided tuning.The work will progress from detailed functional specifications through converter and system level design to validation. Concepts will be proven via electromagnetic transient (EMT) simulations in PSCAD, then verified in real time using Controller Hardware in the Loop (CHIL) testing at Imperial College. During a secondment at the National HVDC Centre, industrial grade RTDS facilities will further validate DeepGFM on realistic meshed AC/DC topologies.Expected outcomes include: a validated storage augmented HVDC converter and dual port GFM control solution; a real time orchestration toolkit for scenario based dispatch; and open access releases of curated datasets, PSCAD models, CHIL test benches, and controller code (all DOI tagged). Findings and performance metrics will be channelled to industry and IEC SC8A working groups to inform emerging GFM and HVDC interoperability guidance. By enabling stronger and more resilient HVDC systems, DeepGFM will increase offshore wind hosting capacity, support European decarbonisation goals, and reinforce European leadership in HVDC technologies.