Tilman Wippenbeck

A toolbox for efficient parameter and structure variation of time domain power system simulation models in Simulink

Variation studies with Matlab/Simulink models are often limited to a handful of variation subjects due to the efforts involved. A toolbox for the efficient automation of parameter and structure variation of complex power system time domain simulation models in Matlab/Simulink has been designed and implemented. A requirement engineering process has been used to define necessary functionality. Several research use cases related to power systems have ensured the usage of best practices in the implementation. The toolbox proves to be beneficial for increasing efficiency in automated parameter and model structure variations and analyses of the results. The authors intend to make the source code and the documentation of the toolbox available for download under an open source license.

Voraussetzungen des Überstromzeitschutzes in wechselrichterbasierten gewollten Inselnetzen unter Berücksichtigung von Strombegrenzung und transienter Stabilität

Intentional islanding of electrical distribution systems raises potentials for customer backup power supply in future energy systems. An uncertain amount of additional expenses is needed to ensure an adequate and reliable protection system for the islanded mode of operation. Nowadays, customer installations are mainly equipped with low-cost overcurrent protection devices. Their reliable fault tripping is assumed but is not verified in intentionally islanded grids built by multiple inverter coupled distributed energy resources. It is unknown, whether critical influences and requirements need to be taken into consideration in addition to a sufficient dimensioning of the sources. Potentially influencing factors are investigated by means of time domain simulations of an islanded low voltage distribution system containing two grid building inverters. One and three phased faults are investigated in main and final circuits of customer installations. A sensitivity analysis varies up to 32 factors simultaneously from the areas grid structure and parameters, dimensioning of inverters, inverter neutral current injection capability as well as structure and parameters of LCL-filters, control circuits, current limiting mechanisms and anti-windup. For the first time, these variation studies consider an increasing degree of inhomogeneity of structure and parameter settings and exploit a significant bandwidth of values per setting. A rule based procedure is created as a precondition for parameterizing the overall model and the control circuits. Adequate variants of inverter current reference limiting mechanisms are identified by means of a developed test procedure. For the first time, according mechanisms are systematically derived for grid building inverters with neutral current injection capability. Reliable tripping of overcurrent protection devices is achievable for single wire final circuits with tripping times lower than 400 ms, when a successful fault ride through can be ensured for an adequate quantity of grid building inverters. The required installed rated inverter currents are in the range from 0.92 to 1.18 times the overcurrent protection device’s tripping current. Influencing factors are the current limiting mechanisms used, the type of overcurrent protection installed and the inhomogeneity of inverter control structures present. Faults in main circuits experience a critical influence by a loss of transient stability during the fault (desynchronization) due to the resulting longer prospective tripping times. A reliable protection tripping is then impeded by the

Modelling and sensitivity analyses of equivalent models of low voltage distribution grids with high penetration of distributed generation

This paper addresses the fault current contribution of low voltage (LV) grids with high penetration of distributed generation (DG) to large disturbances in overlaying transmission networks. In order to be able to analyse the impact of active distribution networks (ADNs) within a reasonable computational time, equivalent models need to be developed. Within this paper black box models of LV grids are proposed using methods of nonlinear system identification. The large variety of LV grids with regard to different grid topologies, r/x-ratios and DG penetration levels can be analysed in parameter studies using methods of sensitivity analysis. With the help of the results a limitation of the parameter space shall be achieved in order to reduce the total number of equivalent models that need to be included in future stability studies with strong contribution of ADNs.