Multi-mode power-to-heat-to-power system with AI based control and dispatch for simultaneous multi-energy service provision.

Phasing out fossil fuels while integrating massive shares of variable renewable energy (VRE) has created urgent challenges: how to store energy at scale, provide heating and cooling, and keep the grid stable. Power-to-Heat-to-Power systems (PTES) stand out as the only single technology that can address all these needs. Yet today’s PTES concepts are limited: confined to two temperature levels and struggling with inefficient dynamics that hinder real-world deployment.FLEXPTES breaks this barrier, proposing a novel multi-optional PTES architecture operating across four temperature levels, unlocking new thermodynamic pathways for simultaneous delivery of electricity, heat, cold, and grid stability services. The system will be designed, dynamically modelled, and validated through scaled component testing. From this foundation, fast, modular surrogate models will be built to capture system dynamics. These surrogates will then power a cutting-edge AI framework, combining predictive control with reinforcement learning for real-time, efficient, and flexible dispatch.The outcome is a storage technology capable of delivering electricity, heat, cold, and grid stability simultaneously with high efficiency, unprecedented response time, and low computational overhead. Beyond direct operation, FLEXPTES will have a transformative impact on long-duration storage, enabling day-to-day multi-energy services in renewable markets, while also revolutionizing future dispatch planning and storage expansion modelling through AI integration. In doing so, the project directly supports the sustainability goals of clean and affordable energy while contributing to advancement in energy infrastructure and supporting climate action.