Miroslav Stanković

The distribution network development planning in the Belgrade Distribution Area was preceded by extensive work on the development of simulation models of the network for different operating conditions. The models were created using the Tokovi snaga software package and represent the maximum non-coincident demand of 110/x kV substations during both winter and summer periods. Using these models, the existing network condition was analysed. The analysis included the assessment of loading levels of all medium-voltage network elements, voltage conditions throughout the distribution system, and verification of compliance with the N-1 criterion for all relevant network elements. The results of the analysis identified several network constraints in the Belgrade Distribution Area, including elements operating above the limits defined in the Distribution System Operation Rules, voltage deviations beyond the ±10% nominal voltage limits, contingencies of elements for which secure supply cannot be ensured, and power losses on all 10 kV, 20 kV and 6 kV feeders supplied from 110/x kV and 35/10 kV substations. In order for the existing distribution network to accommodate future loads, while also considering the integration of distributed energy resources, a load forecast for both existing and future electricity consumers was developed for each stage of the planning period (2026–2033, with an outlook to 2043). The load forecast was defined for several consumer categories: households, directly forecasted customers, and other commercial customers (customers connected at 35 kV, customers connected at 10 kV, small commercial customers, and public lighting). The category of directly forecasted customers includes existing consumers whose annual electricity consumption exceeded 2,000,000 kWh in the previous year or whose measured maximum demand exceeded 500 kW. It also includes future consumers with an approved coincident load above 500 kW and a confirmed network connection. For consumers whose future connection is uncertain or whose commissioning timeline is not yet defined, a separate group of forecasted loads was formed. This load forecasting approach made it possible to analyse the spatial distribution of future loads based on the relationship between electricity consumers and their supplying MV/LV substations or measurement points. Based on the load forecast and the condition of the existing network, a set of investments required to be implemented by the end of the planning horizon (2033, with an outlook to 2043) was defined to ensure that the future network meets all technical criteria. The proposed investments include the construction of new 110/x kV and 35/10 kV substations, as well as underground cable and overhead lines at voltage levels of 35 kV, 20 kV, and 10 kV. As a result, the target network configurations for 2033 and 2043 were defined. The implementation timeline for each investment was determined using a proposed investment prioritization methodology. The methodology considers investments aimed at relieving network constraints, replacing aging network assets, and ensuring compliance with the N-1 security criterion. In addition, investments were categorized according to the type of distribution network element involved, such as transformers (110/x kV and 35/10 kV), cable lines, bundled conductors, and overhead lines at voltage levels of 35 kV, 20 kV, and 10 kV. Each investment was assigned a weighting factor based on the reason for its implementation and the type of element being installed. This approach enables the definition of prioritized investment sets for each stage of the planning period. In addition to the network development planning process, further analyses were performed to assess the network capacity for integrating renewable energy sources, the potential for connecting electric vehicle charging infrastructure, and the expected impact of heat pumps on the distribution network.

Keywords: distribution networks, power flow modelling, load forecasting, network development planning