Model-based predictive control to minimize primary energy use in a solar district heating system with seasonal thermal energy storage.

• Primary energy use is minimized in a district heating network with seasonal storage. • Predictive control formulation was applied to optimize variable speed pump use. • Pumping power savings of 47% are obtained. • Energy costs and greenhouse gas emissions are reduced by 38% and 32%, respectively. This paper investigates the development and assessment of a model-based predictive control strategy for the district heating system at the Drake Landing Solar Community (DLSC), in Okotoks (Alberta, Canada). Thermal energy is collected by solar thermal collectors and stored seasonally by means of a borehole field. Two water tanks are used as short-term storage, acting as a central unit connecting solar collectors, long-term storage and a district loop. The DLSC has succeeded in using solar energy collected during the summer to provide nearly all the heating needs of this 52-home community in winter, with solar fractions consistently over 90%. The proposed predictive control strategy aims to minimize primary energy consumption while maintaining the same solar fraction. This simulation study –based on model calibrated with on-site measurements– focuses on the optimization of circulation pump speed to manage energy exchange between long-term and short-term storage systems. Minimizing pumping electricity use is a critical aspect of the community environmental impact, since fossil-fuel thermal plants are prevailing in Alberta. Simulation results indicate that the proposed strategy would save, on an annual basis, about 47% of total pump electricity use. This would result in savings in terms of cost (38%) and greenhouse gas emissions (32%). [ABSTRACT FROM AUTHOR]