Paper Conference

Proceedings of eSim 2006: 4th Conference of IBPSA-Canada

     

DYNAMIC DATA-DRIVEN GRAY-BOX MODELS OF COMPONENTS OF HVAC SYSTEMS

Nabil Nassif, Radu Zmeureanu

Abstract: Operating performance of HVAC systems could be improved through real-time control and optimization. Therefore, component models are required to the estimate performance of HVAC systems. It is of practical importance to develop a simple, yet accurate and reliable dynamic model that will better match the dynamic behavior of local control process and overall system over the entire operating range. Dynamic data-driven models based on gray-box approach are proposed in this paper. Model parameters are tuned by using genetic algorithm in order to minimize the errors between measured and estimated performance data. Dynamic models such as the zone temperature model, the cooling coil model and the fan model are developed and validated against real data gathered from an existing HVAC system. The validation results show that the datadriven gray-box component models produce better accuracy compared with measured data. These models have a great potential for different applications especially for real-time control and fault detection.
Pages: 118 - 125
Paper:
esim2006_118_125