ARMONI·K

Many industrial and tertiary buildings still operate with static HVAC schedules, fragmented monitoring systems and reactive maintenance strategies, leading to unnecessary energy consumption, operational inefficiencies and avoidable CO₂ emissions. ARMONI·K addresses this challenge through a closed-loop optimisation architecture combining digital twins, IoT data integration and Reinforcement Learning-based control.

The platform integrates real-time sensor data, Building Management Systems (BMS) and EnergyPlus simulation models to create physically consistent digital twins capable of predicting thermal and operational behaviour. These simulations provide the training environment for lightweight RL agents that optimise HVAC, ventilation, lighting and equipment scheduling dynamically according to occupancy, comfort requirements, operational constraints and energy costs.

ARMONI·K delivers two complementary operational layers:

• An AI analytics copilot that assists facility managers through anomaly detection, operational insights and energy optimisation recommendations.

• An autonomous RL control engine capable of executing real-time optimisation strategies directly on existing infrastructures and edge devices.

The solution is interoperable by design and compatible with FIWARE ecosystems, NGSI-LD data models and standard MQTT/REST interfaces, enabling scalable deployment across municipal, tertiary and industrial building portfolios.

Expected impacts include:

• measurable reductions in energy consumption and CO₂ emissions,

• improved operational efficiency,

• increased energy flexibility,

• optimisation of thermal and electrical loads,

• and accelerated digitalisation of building and industrial operations.

We are looking for manufacturing SMEs, industrial facilities and technology providers interested in validating AI-driven energy optimisation solutions in real operational environments within collaborative European innovation projects.