Nuclear power plants are often portrayed as inflexible and incapable of integrating with intermittent renewables. German Green politicians such as Katrin Göring-Eckardt and Bernhard Herrmann argue that nuclear power "clog ups" the grid, and "crowd outs” solar and wind. Here I want to show some data indicating how nuclear output can coexist with, and even complement, variable renewables once we consider the statistical shape of demand and generation, and especially when we add in storage.
Modern reactors can easily load-follow rapidly enough to track hourly swings in demand and renewable generation. But running below nameplate capacity is less economical given nuclear power plant's high capital costs. Luckily nuclear and solar have low correlation, and even better, residual load net of nuclear and residual load net of solar also have low correlation. Intuitively, nuclear power plants keep the lights on at night, solar powers the midday peak.
Be it gets even better. Any intermittent renewable focused system must include storage. Buffers smooth the residual load after scaled nuclear, wind, and solar output are combined. Night-time nuclear surplus can be shifted into the morning ramp, while midday solar surplus can be released during the evening peak. The combined profile often requires less backup or storage than either an all-nuclear or an all-renewable system sized to meet the same annual energy target. In that sense the technologies are complementary rather than competitive; their joint effort means less energy is wasted, less batteries, peaker plants and hydrogen must be constructed.
The scenarios below apply inferred German weather and demand data from Energy Charts, with total demand scaled to the 1,000 TWh projection in Agora Energiewende’s 2045 outlook. The model highlights peak backup requirements rather than full chronological dispatch economics, letting you probe how different nuclear shares influence the magnitude and timing of those peaks.
Data: Underlying data is an estimate of historical weather patterns derived by taking renewable energy generation from energy-charts.info, and dividing by estimated installed capacity. Installed capacity is estimated by taking historical year-end values for each year and doing a linear interpolation. The resulting capacity factors are imperfect, but more or less correct. Load data is supposed to represent demand in a decarbonised Germany and is derived by taking the demand values from the same data source, increasing each day's demand by a factor corresponding to twice the idiosyncratic daily usage averaged over all years, and rescaling so it matches the projected 1000 TWh annual demand from Agora Energiewende's 2045 scenario.
Estimation of capacity: Two steps. One, find the total energy required, and rescale generation to perfectly equal demand. Second, find the distribution of nuclear to renewables. (The ratio of solar to offshore and onshore wind is taken fromt he Agora Energiewende target of Wind offshore: 70 GW, Wind onshore: 145 GW, Solar: 385 GW.) For the latter, installed renewables capacities are simply reduced to (1 - nuclear share), and the residual total generation is then filled with nuclear.
Storage potential and storage/backup consumption: These depict the instantaneous surplus or deficit between scaled generation and load. The displayed backup requirement reflects the maximum surplus or deficit (a peak-based sizing heuristic) rather than a chronological state-of-charge simulation, so it highlights extreme balancing needs instead of time-step-by-time-step storage behavior. Thus, it strongly underestimates demand. It also does not take round-trip efficiencies into account.