A systems approach to the optimal operation of HVAC processes in buildings

A systems approach is developed for predictive operation of HVAC processes in buildings, which is described with an objective function, a state-space model, and a number of constraints. The state-space model is transformed from z-transfer functions that are identified in real-time. The system constraints are described mathematically so that they can be easily applied. Dynamic programming techniques are combined with on-line simulation to enhance efficiency in searching for the optimal operation strategies. Simulation studies show that the techniques are computationally efficient. Results also indicate that the predictive control of floor heating systems may lead to significant savings in operating cost.