As the share of renewable, converter-based generators increases, grid stability and reliability are becoming more critical than ever. Distributed and renewable generation from solar photovoltaic and wind power are the major drivers of this transformation, with other technologies such as battery storage playing a vital role to balance demand and supply. The components of such a mix include conventional grid-following converters (GFLC), grid-forming converters (GFMC), and synchronous generators (SG). The latter two types are capable of providing or emulating inertia, while GFLCs lack such ability by design.

Previous grid events have shown that grid stability is of significant importance and will become even more important when more renewable generation is added. Two events which had devastating effects on U.S. power system were the 2021 Texas power crisis and the 2003 East Coast blackout. Similarly, the interconnected network of Continental Europe faced system splits in 2021 and 2006.

Dynamic simulation tools aim to assist various stakeholders in avoiding events and enhancing grid resilience. However, model exchange and parameter studies can be challenging due to the proprietary nature of these tools. Open-source simulation tools are emerging, with the equation-based Modelica modeling language being used in various industries. Modelica collaborates with other open-source tools like the Functional Mock-Up Interface (FMI) for co-simulation. Models can be exported and compiled according to FMI, allowing dynamic coupling with other simulation tools. The exported model can be instantiated multiple times and used in parallel processing by employing high-performance computing resources or simply utilize multiple processing cores of local workstations to swiftly iterate through different scenarios or parameter settings.

This study employs the Modelica framework to conduct dynamic simulations on a medium-voltage regional distribution network model where demand is primarily served by local distributed generation. To increase scalability, the complexity of the distribution network model is reduced to an equivalent reduced-order model that preserves the dynamic behavior of the original detailed system.

We demonstrate the feasibility of such an approach by investigating sudden load disruptions of varying magnitudes across multiple generation mix scenarios. Results of the 703 simulations indicate the existence of a critical GFLC ratio. Notably, a higher GFLC penetration than 80 % resulted in the disconnection of all GFLCs when the power grid was exposed to load disruptions greater than 145 % of the base load. Further work will be needed to expand open-source modeling's capabilities for the electric power grid, but the effectiveness, transparency, and scalability of such open-source tool has been demonstrated.