Design and connection optimization of a district cooling network: Mixed integer programming and heuristic approach

In densely populated areas, district energy systems can bring significant savings and an overall reduction of CO2 emissions. In particular, district cooling is a valid alternative to conventional cooling technologies. Among the issues related to district cooling there is the relative smaller temperature difference between supply and return respect to district heating. This increases the design and operational costs. Consequently a strong focus is needed on the optimization of district cooling networks. In this paper a mixed integer linear programming model (MILP) and a heuristic model have been developed and implemented for the topology optimization of district cooling networks. The models allow to estimate: (i) the optimal design, (ii) the set of buildings that is convenient to connect to a district cooling network. They have been applied to three different case studies and the main results obtained by the two methods have been assessed and compared. The heuristic method proved to be faster, especially for complex networks. On the other hand, the MILP is more precise and the maximum difference among the models in terms of objective function is about 1.3%. A graph clustering method has also been implemented to improve the performances of the heuristic approach by facilitating its convergence.The models through the optimizations can bring savings up to 3.16% of the total life-cycle cost depending on the case study. They could therefore represent a potential tool that supports decision makers in the planning phase of district cooling networks.