Using distributed agents to optimize thermal energy storage

Abstract The Intelligent Building Agents Laboratory (IBAL) at the National Institute of Standards and Technology (NIST) will be used to develop and evaluate advanced control systems for commercial building heating, ventilation, and air conditioning (HVAC) applications. Advanced control techniques include model predictive control (MPC) and distributed optimization, both of which can utilize machine learning. This study is a first step in the development of more advanced techniques. MPC is used to generate an optimal schedule for the operation of a cooling plant consisting of ice-on-coil thermal energy storage and two chillers with different maximum capacities. The schedule is generated offline and then used in the IBAL. The performance of the actual system is used to validate the simulation. The simulation is then used to conduct studies relating to the effect of the time horizon and electrical tariff on the optimal schedule. The results demonstrate that an agent-based control architecture is effective in defining optimal control of thermal energy storage systems.