Abstract
The transition from synchronous generator-based power systems to an increasingly higher share of inverter-based renewable generation raises concerns regarding voltage quality. Solar photovoltaics (PV) are expected to constitute a large share of renewable capacity expansion in Norway. As PV penetration in the grid increases, short-term power variability caused by rapid irradiance fluctuations may start to pose a threat to voltage quality, particularly in weak grids. A potential consequence of fast PV power fluctuations is flicker, a voltage phenomenon associated with visible variation in lighting installations, induced, for example, by mixed-sky conditions wherein fast-moving clouds rapidly change the incident irradiation on a collection of PV panels. Fast irradiance changes, at the scale of seconds, can induce equally fast voltage spikes and drops on the grid. Assessing flicker directly requires high-resolution voltage data, which is rarely available, thus motivating the need for alternative modeling approaches. This study investigates the impact of short-term variability from utility-scale PV plants of different capacities on flicker-related voltage changes in Norwegian high-voltage distribution grids. The study performs a quasi-static flicker assessment through power flow simulations in the CINELDI 22 kV reference grid. PV generation data is derived from high-resolution irradiance data and a wavelet-based variability model to simulate production and account for spatial smoothing in large PV plants. Plant capacities from 0.1-15 MW are examined at five specific buses in the reference grid, yielding 750 simulated voltage time series. Simulated voltage changes are evaluated against IEC flicker limits to estimate potential flicker severity. Moreover, short circuit levels are estimated for a collection of PV plant sizes to assess critical sizing of PV plants connected at medium-voltage levels. Results indicate that increasing PV capacity is associated with a higher percentage of time with voltage changes linked to flicker-related thresholds. In addition, both voltage levels and voltage-change magnitudes increased with increasing PV capacity. The simulations suggest that voltage sensitivity may depend more strongly on grid strength and injected active power than on irradiance variability itself. Flicker results should, however, be interpreted carefully, as the method of comparing against IEC curves is simplified and assumed a constant frequency of changes. Nevertheless, the method provides a simple and scalable framework for evaluating flicker-related voltage changes across multiple cases. The work aims to support grid operators and planners in assessing flicker risk and facilitate the integration of large-scale PV in Norway.