Bayesian Exponential Random Graph Modelling for defining Sustainable Energy Districts

This research aims to define potential and sustainable energy districts based on a concept of optimal planning determined with a statistical network model. Sustainable energy districts might include local authorities and administrations that are willing to adopt a common plan for sharing energy services and goods focused on similar needs and available resources, thus contributing to sustainable development goals. Within the Italian legislation, the energy district defines single or collective local energy plans named Sustainable Energy Action Plan (SEAP) and Sustainable Energy and Climate Action Plan (SECAP). The planning activity of municipalities is time and cost cosuming and a collaborative way could increase its efficiency and sustainability. The sustainable energy districts are set in an intermediate level between the municipal and the regional ones that, at present, is not totally structured within the Italian energy sector. However, the sustainable energy district presents a set of social, economic, ecological, organizational potentialities related to the improvement or enlargement of actual formal relationships among municipalities and their utilities and regarding the sharing of goods and services within the local energy system. In this research, Statistical Network Analysis (SNA) is used with two purposes. The first consists in describing the main characteristics of the formal relations between municipalities set in a specific geomorphological context by using network statistics (density of the network, degree-based configurations, cliques, etc.). The second aim is to investigate the probability that a sustainable energy district could emerge from the current relationships in the energy sector. For this purpose, a fullyprobabilistic inferential approach for exponential random graph models is used assuming that the observed network structure is generated by local and governance processes that depend on relationships between actors. This methodology will be applied in a case study in the Alpine area: the South Tyrol region (Italy) and the results of the SNA models will be compared to the ones based on a territorial cohesion, highlighting similarities or dissimilarities in the two approaches. This paper is a first application of a Bayesian SNA approach with the aim to model sustainable energy districts. The results can be used by regional and local authorities for addressing new collaboration, new feasibility, and new sustainability of energy planning.