25 th International Conference on Electricity Distribution Madrid, 3-6 June 2019 Paper n°2107 CIRED 2019 1/5 MODELLING OF SYNTHETIC POWER DISTRIBUTION SYSTEMS IN CONSIDERATION OF THE LOCAL ELECTRICITY SUPPLY TASK Jacob TRAN Christoph WIRTZ Pascal PFEIFER FGH e.V. – Germany FGH e.V. – Germany FGH e.V. - Germany jacob.tran@fgh-ma.de Christoph.wirtz@fgh-ma.de pascal.pfeifer@fgh-ma.de Dominik WURSTHORN Dr. Hendrik VENNEGEERTS Prof. Dr. Albert MOSER FGH e.V.– Germany FGH e.V. – Germany RWTH Aachen - Germany dominik.wursthorn@fgh-ma.de hendrik.Vennegeerts@fgh-ma.de am@iaew.rwth-aachen.de ABSTRACT With the increasing integration of renewable energy sources into the electricity supply system, planning and operation of power distribution systems has become more complex and challenging. As a consequence of this progress, new and innovative concepts for distribution system operation are in research. Therefore, detailed grid models are required. Due to different power demand and generation according to the region, these grids show a very heterogenous structure. Thus, this paper proposes an approach to synthesize grid models based on the real supply task of German grids. Publicly available data of geoinformation systems, demography, and power generations in combination with a probabilistic approach based on evaluations of real grid data is used to identify the electricity supply task in high spatial resolution. Considering the local supply task, the grid topology is simulated in a two-step heuristic, including a capacitated vehicle routing problem to determine the basic grid topology. The resulting model is able to generate German MV grids in high spatial resolution. Further, it is possible to identify and reproduce local characteristics. INTRODUCTION The energy transition in Germany has led to a significant restructuring of the energy supply system. This change is particularly evident on the generation side, where there is an increased expansion of renewable energies at the distribution grid level. In the future, however, consumer behavior will also change significantly, e.g. due to the electrification of mobility and an increased amount of home battery storage facilities. The generations and loads installed in the distribution system show a wide range of local variations. These differences are reflected in the energy supply task and result in varying grid structures. Considering these developments, planning and operation of power distribution systems are becoming more complex and challenging in the future. The development of new concepts and methodologies for improving planning and operation of power distribution system needs performance evaluation with realistic grid topologies. However, even though several small-scale grid data sets are publicly available, access to diverse and large-scale distribution power systems is limited. In addition, there is a risk that locally specific grid characteristics are neglected due to the lack of representativity of real grid samples. Synthetic grids can offer an alternative for this problem. Modelling synthetic grid aims to generate test systems that are completely fictitious but capable of representing local characteristics of actual real power grids. Hence, this paper presents an approach to generate models of present and future synthetic medium voltage (MV) grids based on both statistical evaluation analysis of real grid data and spatial high-resolution land cover and land usage data. This paper focuses on modelling MV grids of Germany. The Low Voltage (LV) grid is neglected and represented as aggregated load and generation at the MV-LV substation. MODELLING APPROACH Model Overview Based on a given MV grid area the presented model determines the grid topology according to the specific electricity supply task and the positions of the grid customers in the area. As data basis, the model uses multiple high-resolution spatial data, which contain information about land cover and use, power generations and loads, and population density. These data sources are aggregated in a so-called input database. In addition, the model also includes a scenario database as an input. The scenario database enables the integration of forecasts for future developments of renewable energy sources (RES) and loads into the procedure. Forecast data is often broken down into different spatial resolutions. For example, forecasts for numbers of grid customers or their installed capacities are usually related to a certain spatial area in the form of state or administrative boundaries. The model dissolves the forecast data and allocates the grid customers into the grid areas by using suitable distribution functions. After a preprocessing of the database, the first step of the model is to define a MV grid area. This represents the spatial boundaries of one MV grid. Using land cover and usage data as well as population density information the area is then spatially classified in order to determine possible locations for loads and generations. Next, the supply task is being defined. Using publicly available data for renewable generation and the previously determined