Detailed dynamic modelling of large-scale power networks is becoming increasingly relevant as the share of inverter-based resources increases, introducing a multitude of power system stability problems. Faster system dynamics are often overlooked in conventional phasor-domain analysis; however, electromagnetic transient (EMT) models can capture instantaneous voltages and currents, enabling accurate representation of converter controls and protection systems. Even though plant-level modelling can be performed and modelling/grid code requirements met, large-scale EMT modelling for system-wide stability analysis remains a challenging task. With this backdrop, the paper proposes a systematic approach to electromagnetic transient modelling of large power networks for high solar/wind integration studies. The dynamic EMT modelling process follows a staged approach, starting with a reduced extra-high-voltage network (EHV) and gradually including selected 110 kV network zones, utilising the Kron method of network reduction. The dynamic modelling process also includes relevant data collection, model initialisation, parallelisation and validation steps. A future network model of the Irish grid is used as an example for the study.

The first stage of the dynamic EMT modelling process involves collecting the data requirements and component model details. Firstly, PV, wind, and conventional generation capacities, along with network details, are collected for the future 2033 Irish grid model, using publicly available datasets. This is followed by the development of detailed, dynamic models of all system components (conventional generators, PV/wind generation, HVDC, and loads) and their control structures. Snapshots of various test cases are extracted from a unit commitment and economic dispatch optimisation model using the same network for high- and low-generation, and loading cases.

The main modelling process includes three distinct components: initialisation, parallelisation, and validation. Initialisation of the power sources and loads is performed using the optimisation model, so the system can run without large transients at startup, thereby avoiding longer settling times. This is followed by parallel processing to reduce the computational burden imposed by a large number of inverter-based resources. This is achieved by using the Opal-RT real-time simulator using the Artemis toolbox. The system is broken down into multiple parts so that parallel processing can leverage reduced complexity by running across multiple CPU cores. The last step of the EMT modelling is the validation process, which includes validation of power flow and current injection during faults, using a publicly available PSS-E dataset published for the future Irish network.

To reduce the complexity of modelling a large power network, a staged approach is adopted in this paper. The Kron reduction technique is used to first model the EHV network of the Irish grid (220 kV and above) as a simplified system, with aggregated sources and load types. This is followed by the inclusion of selected zones of the 110 kV high-voltage network, where larger shares of renewable generation are expected. The entire HV network, consisting of 446 buses, 101 conventional generators and 210 PV and wind generators, is significantly reduced to a few selected 110 kV zones. This reduction enables the analysis of dynamic stability at important parts of the network without representing the entire HV system.