Influence of Uncertainties of Renewable Energy Sources on Providing Flexibilities for the Superimposed High Voltage Grid

摘要

Since the last years, the number of renewable energy sources (RES), especially photovoltaic (PV) systems and wind turbines, has increased in distribution grids. The financial stress of many conventional power plants is another side effect of the increased RES capacity. In high voltage (HV) grids with an increasing number of decommissioned conventional generation units, there could be a rising need for alternative flexibility providers regarding the active and reactive power control. The flexibility is necessary to avoid thermal as well as voltage limit violations in the HV grid. It is widely discussed that this flexibility could be provided through the interconnection points at the substations from the underlying medium voltage (MV) and low voltage (LV) grids, where most of the RES capacity is installed and innovative solutions such as remote control of distributed generators to integrate these RES could be used. Nevertheless, many MV and LV grids themselves are facing technical problems such as overvoltages, which mean that a provision of flexibility for the superimposed grid might be limited. When providing flexibility from the MV and LV grid to the superimposed grid, the uncertainties of RES feed-in become another crucial aspect. Forecast errors of PV and wind generation resulting from the difficulty to precisely predict the weather condition have to be considered. Therefore, the distribution system operator (DSO) can only provide the flexibility with a certain probability. The aim of this paper is to assess the influence of these uncertainties on the provision of flexibility for the superimposed HV grid. First, the forecast errors of PV and wind generation with regards to their forecast horizon and the current feed-in are analysed. This also includes the linear and nonlinear correlations between generation units of the same technology at different locations as well as alternative technologies. These correlations are modelled using a copula, which is a method used in quantitative finance to model the correlation between multivariate variables. Based on the analysis, a Monte-Carlo-Simulation is developed, with which for each draw a consistent set of feed-in with the superimposed errors is taken out of the copula in order to model the feed-in of RES in the considered MV und LV grid. The draw is the input for the estimation of the flexibility potential by coordinating all degrees of freedom under consideration of technical constraints. The estimation of the flexibility potential is achieved by an optimization method called Optimal Power Flow (OPF). In order to consider the interdependency between the active and reactive power exchange at the HV/MV substation, the applied methodology further scans the flexibility potential successively, which can be provided to the HV grid without leading to criticalities in the own grid. The estimation is performed for each draw in order to obtain the relation of the flexibility potential and a certain probability. The outputs show for a certain quantile the maximum provision of flexibility potential for the superimposed HV grid. The methodology is applied to a rural MV/LV grid with a high share of RES for a situation with a high generation of RES and a low load. The results show that the technical flexibility potential is constrained by the forecast errors of PV and wind generation, when the DSO wants to provide the flexibility with a high reliability. Besides, it is found that with the same reliability the DSO can provide more flexibility with a decreasing forecast horizon for this considered situation.