Power system dynamic studies rely on model-based simulations. With increasing penetration of DG, modeling of a network that is not only geographically diverse but is also technologically mixed entails great efforts and accuracy for such studies. Despite the availability of highly powerful simulation tools and increased processing power, performing time domain simulations on such interconnected and complex system remains a great hurdle. Reducing the complexity of the model using aggregation can help with these problems. The purpose of this chapter is twofold. First, we go through the concept of network aggregation and introduce a white-box DE suitable for system studies with high penetration of distributed PV. Next, we introduce a sophisticated approach for solving the nonlinear optimization problem of parameter identification of the developed DE. This approach is based on a new optimization technique called mean-variance mapping optimization (MVMO). The parameters of the DE are identified by measuring the response of DE to certain disturbances and comparing them with reference signal sets. By using the proposed approach, we determine the parameters of our suggested DE that entails high accuracy. The accuracy of the DE is measured by calculating the root mean square error (RMSE) value. The time required to simulate faults and perform time domain simulations using a DE is considerably less than for a full-scale model. This has significance for system planners and researchers who want to analyze the effect of increasing penetration of DE into distribution grid on the system as a whole with regard to strategy development

Power systems are undergoing a historic structural and technological transformation. The increase of distributed generation (DG), recently mostly wind power park modules (WPPMs) and photovoltaic power park modules (PVPPMs), is already changing the way power systems are structured and operated. Distribution systems are turning from ‘passive’ into ‘active’ parts of the system (ADS). This structural and technological transformation influences the power system’s network fault response and stability. This thesis investigates the network fault response of transmission and distribution systems with very high penetration of distributed generation. The impact of DG on transient stability, large disturbance voltage stability, and frequency stability is analysed. The analysis is limited to symmetrical, three-phase network faults only but extended to adequately represent all voltage levels, including DG connected at low voltage (LV). Requirements for the response of DG to network faults are defined in grid connection requirement (GCR). The massive insertion of DG into distribution systems (DSs) leads to new challenges like the regular occurrence of reverse power flow (RPF) from the distribution to transmission level etc. Further, DG is more likely to be exposed to fault-induced delayed voltage recovery (FIDVR) events than transmission and sub-transmission connected generating facilities. The overall objective of this thesis is, therefore, to critically review the necessity and the specification of current and proposed grid connection requirements with regard to the network fault response of transmission systems with very high penetration of distributed generation and to propose changes to the specification where needed. The scientific contributions of this thesis are (1) a proposed comprehensive methodology of aggregation of DGs and dynamic equivalencing to derive highly accurate dynamic equivalent models of ADSs that considers the composition of ADS with DG classes in terms of their technology type and grid code performance (performance legacy); (2) a case study demonstrating that nowadays undervoltage protection schemes for small- and medium-scale DG connected in LV distribution networks may become a risk for power system stability; (3) identification of minimum requirements and improvement of existing grid connection requirements for the network fault response of DG to maintain power system stability.

This paper introduces an approach for determining the parameters of aggregated dynamic equivalents for active distribution system from a reference set of signals associated with various disturbances using mean-variance mapping optimization (MVMO) algorithm. As the penetration of renewable energy sources in the LV and MV network increases, it has become extremely important that we have models that mimic actual system response with least computational overhead for bulk transmission system studies. Full scale detailed, complex, interconnected models are most accurate but carry significant simulation time, making analysis unfeasible. Dynamic equivalents (DE) are simplified representations of larger models that mimic the dynamic response of those models. They can be used to replace the neighbor area in a study while the area of focus is modeled with great detail. This reduces the computational burden. Our test system is a small part of the German MV/LV network whose dynamic equivalent is identified and compared. The dynamic equivalent is a Western Electric Coordination Council (WECC) developed distributed PV model (PVD1). This identification is done with the use of MVMO algorithm that utilizes a single parent crossover and a unique mapping function. The results show an almost identical response with good normalized root mean square error (NRMSE) between the detailed and aggregated model. The MVMO shows fast convergence and accurate results.