Costs for HVDC underground

[lindnemi]

At the moment technology-data does not contain any costs for HVDC underground. The default assumption seems to be that cables are either submarine or overhead. For example in add_electricity:

pypsa-eur/scripts/add_electricity.py

Lines 374 to 385 in 4c1ec35

	costs = (
	n.links.loc[dc_b, "length"]
	* length_factor
	* (
	(1.0 - n.links.loc[dc_b, "underwater_fraction"])
	* costs.at["HVDC overhead", "capital_cost"]
	+ n.links.loc[dc_b, "underwater_fraction"]
	* costs.at["HVDC submarine", "capital_cost"]
	)
	+ costs.at["HVDC inverter pair", "capital_cost"]
	)
	n.links.loc[dc_b, "capital_cost"] = costs

On the other hand, the default assumption for Offwind DC connections is, that cables are either submarine or underground.

pypsa-eur/scripts/prepare_sector_network.py

Lines 471 to 481 in 4c1ec35

	connection_cost = (
	snakemake.params.length_factor
	* ds["average_distance"].to_pandas()
	* (
	underwater_fraction
	* costs.at[tech + "-connection-submarine", "fixed"]
	+ (1.0 - underwater_fraction)
	* costs.at[tech + "-connection-underground", "fixed"]
	)
	)

Relatedely the HVDC submarine cost is at around 1000€, while offwind-connection-submarine is around 2000€. Why is this discrepancy there?

So here are a few things we could do:

1. Add HVDC underground costs to technology-data. (And perhaps vice versa offwind-X-connection-overhead).
2. Think about harmonizing cost data for Offwind-connection/HVDC
3. Think about using the same default assumption for non-submarine DC lines (either overhead, or underground, consistently for Offwind-DC and Other-DC)

I made this observation, because in PyPSA-Ariadne we use NEP costs data, and the NEP does not contain any submarine costs, only underground costs. Our immediate fix will be to use the same costs for HVDC underground/ HVDC submarine/ Offwind-connection underground/offwind connection submarine.