Data in PyPSA-Eur 41-node

library(reneuro)
library(dplyr)

This article shows the processed data that enters the 41-node model — regions, network, weather, demand and capacities — drawn live from the shipped pypsa_eur_41 and its geoscale. Everything upstream (the raw sources, the NUTS datasets and their aggregation arithmetic) lives in the data-sources article.

The regions

PyPSA-Eur clusters its substation network to 41 onshore regions — most countries are one node; the largest split into several. The polygons ship with the model as the bus geoframe of pypsa_eur_41_gs:

library(ggplot2)
gs <- pypsa_eur_41_gs
geo <- geoscales::geoscale_geometry(gs, "bus")
cc <- geoscales::geoscale_geometry(gs, "country")

ggplot() +
  geom_sf(data = geo, fill = "wheat", colour = "white", linewidth = 0.4) +
  geom_sf(data = cc, fill = NA, colour = "grey35", linewidth = 0.3) +
  theme_void() +
  theme(plot.background = element_rect(fill = "white", colour = NA))

The 41 model regions (PyPSA-Eur onshore clustering). Clusters never cross country borders.

The network

Transmission enters the model as trade corridors between region centroids — rated capacities from the summed thermal limits of the underlying lines:

ctr <- suppressWarnings(sf::st_centroid(geo))
xy <- as.data.frame(sf::st_coordinates(ctr))
xy$bus <- geo$bus

tr <- objs[cls == "trade"]
routes <- unique(do.call(rbind, lapply(tr, function(t) {
  mw <- t@capacity$stock[1]
  if (is.null(mw) || is.na(mw)) mw <- t@capacity$cap.lo[1]
  cbind(t@routes[1, , drop = FALSE], MW = mw)
})))
seg <- routes |>
  left_join(xy, by = c(src = "bus")) |>
  rename(x0 = X, y0 = Y) |>
  left_join(xy, by = c(dst = "bus")) |>
  rename(x1 = X, y1 = Y)

ggplot() +
  geom_sf(data = geo, fill = "wheat", colour = "white", linewidth = 0.3) +
  geom_segment(data = seg, aes(x0, y0, xend = x1, yend = y1,
                               linewidth = MW / 1e3),
               colour = "dodgerblue", alpha = 0.8) +
  geom_point(data = xy, aes(X, Y), size = 0.7, colour = "grey25") +
  scale_linewidth(range = c(0.2, 2.2)) +
  labs(linewidth = "GW") +
  theme_void() +
  theme(plot.background = element_rect(fill = "white", colour = NA))

The 41-node network: buses and trade corridors, line width by rated capacity.

Weather: capacity factors per region

The model’s VRE profiles are per-region hourly capacity factors (the weather objects, 8,760 hours of 2025). Their annual means show what each region’s blended resource offers:

ggplot(cf_map("W_ONWIND")) +
  geom_sf(aes(fill = cf), colour = "white", linewidth = 0.3) +
  energypal::scale_fill_energy_c("windatlas", name = "mean CF") +
  theme_void() +
  theme(plot.background = element_rect(fill = "white", colour = NA))

Mean onshore-wind capacity factor of the blended regional profile (W_ONWIND), 2025.

ggplot(cf_map("W_SOLAR")) +
  geom_sf(aes(fill = cf), colour = "white", linewidth = 0.3) +
  energypal::scale_fill_energy_c("solaratlas", name = "mean CF") +
  theme_void() +
  theme(plot.background = element_rect(fill = "white", colour = NA))

Mean solar capacity factor of the blended regional profile (W_SOLAR), 2025.

The blending is optimistic by construction — every region offers its whole eligible area at close to its best-sites capacity factor. The wind and solar articles document the quality-aware layer built to refine this.

Demand

Measured 2025 load, distributed to regions by the GDP/population proxy:

dem <- bind_rows(lapply(objs[cls == "demand"], function(d) d@demand)) |>
  summarise(TWh = sum(demand) / 1e6, .by = region)
ggplot(geo |> left_join(dem, by = c(bus = "region"))) +
  geom_sf(aes(fill = TWh), colour = "white", linewidth = 0.3) +
  scale_fill_viridis_c(option = "C", trans = "sqrt") +
  theme_void() +
  theme(plot.background = element_rect(fill = "white", colour = NA))

Annual demand per region, TWh (measured ENTSO-E 2025, distributed below country level by proxy).

sum(dem$TWh)
#> [1] 3162.754

Capacities and potentials

Per carrier: the existing fleet (stock for the non-extendable, cap.lo floors for the extendable), and the buildable potential (cap.up):

cap <- bind_rows(lapply(objs[cls == "technology"], function(t) {
  d <- t@capacity
  tibble(tech = t@name,
         stock_GW = sum(d$stock, na.rm = TRUE) / 1e3,
         floor_GW = sum(d$cap.lo, na.rm = TRUE) / 1e3,
         potential_GW = sum(d$cap.up, na.rm = TRUE) / 1e3)
}))
cap |>
  filter(stock_GW + floor_GW + potential_GW > 0) |>
  mutate(across(where(is.numeric), \(x) round(x, 1))) |>
  arrange(desc(pmax(stock_GW, floor_GW)))
#> # A tibble: 17 × 4
#>    tech            stock_GW floor_GW potential_GW
#>    <chr>              <dbl>    <dbl>        <dbl>
#>  1 E_ONWIND             0      236.         9718.
#>  2 E_CCGT               0      227.            0 
#>  3 E_SOLAR              0      189.        14163.
#>  4 E_NUCLEAR          140.       0             0 
#>  5 E_HYDRO            109.       0             0 
#>  6 E_COAL              77.4      0             0 
#>  7 E_LIGNITE           61.8      0             0 
#>  8 E_OFFWIND_AC         0       48.7         525.
#>  9 E_ROR               47.3      0             0 
#> 10 E_BIOMASS           25.1      0             0 
#> 11 E_OIL               17.1      0             0 
#> 12 E_WASTE              8.1      0             0 
#> 13 E_OCGT               0        6.1           0 
#> 14 E_GEOTHERMAL         0.9      0             0 
#> 15 E_OFFWIND_DC         0        0           290 
#> 16 E_OFFWIND_FLOAT      0        0          2348 
#> 17 E_SOLAR_HSAT         0        0         12302

Potentials are flat area sums (eligible km² × deployment density) — the data-sources article documents how they and the profiles are made, and the known caveats.

The regions

The network

Weather: capacity factors per region

Demand

Capacities and potentials

See also