Toy model: the price of transporting electricity from Spain to the Netherlands by a dedicated HVDC line

The CAPEX cost model of (Härtel et al, 2017) gives a very rough estimate on the cost of VSC HVDC power lines. This provides a first toy model on the business case for transporting e.g. Spanish solar or wind power directly to the Netherlands, as a device to compensate for variability and other constraints. These estimates should by no means be taken as indicative of the final result, which is largely determined by political factors such as land ownership, international relations and utility regulation - rather, they serve as a yardstick for the amount of resources involved.

For a single line with one node at each end, the linear model estimates the cost at

\begin{equation} C_{est} = B_p \cdot l \cdot p + \lceil p / p_{ref} \rceil (B_l \cdot l + B_0) + 2 (N_p \cdot p + \lceil p / p_{ref} \rceil \cdot N_0), \end{equation}

where the parameters were estimated as follows:

Symbol Meaning Estimated value
\(B_p\) Length- and power-dependent branch cost 0.97 M€/(GW km)
\(B_l\) Length-dependent branch cost 0.68 M€/km
\(B_0\) Fixed branch cost 0.7 M€
\(N_p\) Power-dependent node cost 80.88 M€/GW
\(N_0\) Fixed node cost 28.83 M€
\(l\) Cable length 1000 km
\(p\) Power capacity -
\(p_{ref}\) Reference power capacity 2 GW

We opt for the submarine cable cost estimates here; these are in the same ballpark as estimates for underground cables. It is true that overhead lines are far easier to repair in case of breakage and enjoy significant cost reductions compared to this baseline, but they suffer from much more severe political concerns. Any realization would of course be a mixture of the three cable types.

If we fill in the variable \(l\) and the constants, the formula becomes:

\begin{equation} C_{est} = 1.13 \times 10^3 p + 7.10 \times 10^2 \lceil p / p_{ref} \rceil \end{equation}

The average Dutch total power consumption is on the order of 100 GW. Bringing in a significant factor - say, 10% - of that through such a power line would in this model correspond to a cost of roughly 14 G€, or about € 3/W.

For comparison, NREL estimates a CAPEX of roughly € 0.85/W for utility-scale PV and quotes CAPEX figures of roughly € 3.4/W for offshore wind. The numbers for Borssele 1+2 are on the order of € 2/W.

All this seems to suggest that, at least on a resource level, it would be expensive but not impossible to transport a significant amount of e.g. Spanish solar power to the Netherlands as part of the energy mix. The main issue is of course to complement the engineering perspective with an understanding of the politics involved.