A new combined clustering method to Analyse the potential of district heating networks at large-scale

• Published in: Energy (Oxford) Vol. 156; pp. 73 - 83
• Main Authors: Marquant, Julien F., Bollinger, L. Andrew, Evins, Ralph, Carmeliet, Jan
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01-08-2018; Elsevier BV
• Subjects: Clustering; Combined clustering; Complexity; Computer applications; Computing time; Density; Distributed energy systems; Distribution management; District heating; Energy; Energy distribution; Energy hubs; Energy technology; Formulations; Genetic algorithm; Genetic algorithms; Homogeneity; Integer programming; Linear programming; MILP energy optimisation; Networks; Optimization; Renewable resources; Scale models; Spatial distribution; MILP energy optimisation; Distributed energy systems; Genetic algorithm; Combined clustering; Energy hubs
• ISSN: 0360-5442; 1873-6785
• DOI: 10.1016/j.energy.2018.05.027

For effective integration of large amounts of renewables and high-efficiency energy technologies, their benefits have to be quantified. Network-level energy optimisation approaches can determine the optimal location of generation technologies within a region and the optimal layout of energy distribution networks to link them. Mixed-integer linear programming (MILP) formulations are generally employed and this is often a burden for large scale models as the computational time drastically increases with the problem size.
Most methods used to reduce the complexity of MILP problems focus on the temporal scale or use aggregated demand profiles for the spatial dimension. There is a lack of a method addressing the spatial complexity to assess the potential of interlinked energy networks at large scale. Therefore, this paper introduces a new combined clustering schema enabling quantification of the potential of district heating networks based on results from building scale energy optimisation problems and taking into account building characteristics.
A city-scale case is divided into multiple districts based on the output of a density based clustering algorithm. The parameters taken into account by the clustering method are the cluster density, homogeneity index and load magnitude. The analysis of the clustering map along with building characteristics of each cluster reveals the required characteristics for the installation of a district heating network or distributed energy systems.
•Multi-scale hierarchical approach for optimal system design and operating strategy.•Combined clustering method for distributed energy system optimisation.•Aggregation schema to enable optimisation of urban energy systems at large scale.•New modelling formulations reducing computational time while preserving accuracy.•District heating networks allow a decrease of system's costs and environmental impact.