The Safe Interaction of Nuclear with a Renewable Rich Power System

The power system has already become more dynamic, and the future promises much higher amounts of stochastic power flow as wind, solar and demand fluctuate, despite a present abeyance in new requests for wind parks due to high interest rates. Nevertheless, the role of nuclear is playing an ever-increasing role in providing carbon-free generation and aiding with stability. Nuclear power plants clearly do not work in isolation, but are an integral and increasingly important part of a highly interconnected power system. It is imperative for the nuclear operators to know what the future holds and ensure the nuclear plants are able to safely cope in this new context. The geographically-dependent scenarios for wind and solar integration have been adequately modelled in terms of time series, e.g. [1], and so the relationship between large wind parks and nuclear plants in terms of frequency and indeed a host of other oscillations, the impact of progressive voltage rise or drop as wind shifts across Finland and the ability of the grid and nuclear to handle severe outages during such transients warrant investigation.
The COSI project, e.g. [2], in SAFIR succeeded in creating a co-simulation platform that combines electrical grid models with thermomechanical models in APROS. A reasonably detailed model of the 400 kV grid in Finland was also successfully created and integrated with the co-simulation platform. Fixed power system scenarios were explored (minimum summer and maximum winter without wind) and both the grid and Loviisa nuclear power plant (NPP) seemed to cope well with common but severe fault simulations. However, some bench-mark testing at the end of COSI revealed that the transient response of the new grid model was highly inaccurate when comparing with published frequency event data from Fingrid. SINARP2023 has improved the grid modelling, in terms of adding new connections, placing series compensation devices, and most significantly but still work-in-progress, changing the model-ling of voltage sources (in the swing bus and on the Swedish or Estonian ends of the HVDC connections) to synchronise machines that represent the inertia more correctly. This will still need work in SINARP2024, as well as continuing to develop more realistic load scenarios.

The main measurable results will include plots and analyses of the vital component responses in a nuclear plant to faults (short-circuits, open circuits, etc.) in the grid while the grid is in a transient state (and, for very little extra effort, plots representing what is happening in the external grid):

Due to N-1/N-2 failures, e.g., a grid fault when HVDC connections are down or a large generator is out of service

·        while undergoing a general voltage increase or decrease, due to:

o   wind conditions changing across the country

o   generators sequentially tripping

o   major load changes, including scenarios with high activation of consumer and P2X flexibility

·        and, for various base scenarios in both summer and winter

o   high wind / low hydro and carbon-free generation

o   low wind / high hydro andc arbon-free generation

o   taking into account the direct impact of the following more specific factors

§  low inertia vs. high inertia

§  weak grid (low short-circuit power) vs. strong grid

§  high vs. low share of converters

·        investigate the impact of oscillations, both natural and forced, on the nuclear plants

To achieve such results, we will have to liaise with the development of and utilise:

·        platform-compatible models of the net effect of wind parks (single machine equivalents)

·        adequate HVDC models to model the interconnection with neighbouring countries

·        refine the load flow values through the 400 kV nodes in the grid model

·        refine the modelling of the nuclear plants under investigation in terms of:

o   their connectivity to the grid

o   their own protection and backup arrangements

Co-simulation platform development. It connects electrical simulations in Simulink and thermodynamical simulations in Apros NUclear

To address discrepancies observed in the co-simulations, we should adjust the grid-only model to better align with the power output and setpoint voltage used in the co-simulations. While key parameters such as harmonics, voltage stability, and reactive power differences remain points of interest, particular attention should be given to harmonics behavior. These may increase as the integration of converter-connected equipment grows but could decrease as grid codes tighten and converter technologies improve.
The coincidence of a potential turbine shaft torsional oscillation mode with one of the subsynchronous frequencies measured in the grid is a significant observation. These harmonics were reproduced only via current sources at wind parks, suggesting a need for further investigation into wind park behavior and its interaction with the system. The stronger oscillations observed in high wind generation scenarios highlight the importance of careful parameterization of the grid stabilizer and wind turbine/converter models to accurately capture both transient and steady-state behavior.
Initial observations from last year’s co-simulations—such as the ripple or measurement spike caused by mismatched initial conditions in Simulink and Apros—further underscore the challenges of coupling different simulation tools. Similarly, the steady-state offset between co-simulations and Simulink-only results points to model parameter mismatches that require iterative adjustments and careful investigation.
The swing bus in the transmission system model also remains a critical focus, as its apparent power distribution enforces mismatched active and reactive power across several buses. This discrepancy, combined with the lack of inertia in certain scenarios, impacts frequency stability and transient dynamics, especially in high renewable penetration cases. Addressing these challenges by refining swing bus modeling and enhancing inertia representation in synchronous machines will be essential for future studies.
Despite these challenges, the SAFER-SINARP project concludes with a robust 400 kV grid model and a co-simulation platform successfully linking the grid model, the internal electrical network of Loviisa NPP, and the thermodynamic APROS software. This combination of tools has deepened our understanding of system dynamics and lays the groundwork for further improvements in future projects.