Skip to content

Technology mapping

Each PyPSA model needs to supply a technology mapping file that maps PyPSA's internal technology names to iampypsa's canonical names. This file also decides whether each cost parameter comes from the IAM, from PyPSA's own default, or is fixed to a specific value. The IAM and the PyPSA model typically won't agree on technologies, and even where they do, not every parameter should necessarily be taken from the IAM — this information lives in a single YAML file.

Processing

The technology mapping file is read by iampypsa.io.technology_mapping:

  • load_technology_parameters(path) — parses the YAML, returned as {"technologies": {...}}.
  • iam_name(tech, spec) — the IAM-side technology name an entry pulls values from (defaults to the model's own key if not given).
  • build_technology_sources(spec) — expands one entry into a {parameter: source} map across the seven standard parameters: investment, FOM, VOM, efficiency, lifetime, fuel, "CO2 intensity".

The YAML schema

Each top-level key is one of the model's own technology/carrier names. Its value is either a bare string (the source for all seven standard parameters), or a dict for finer control:

# Bare-string shorthand: every parameter comes from the IAM.
gas-ocgt: {source: IAM}

# iam_name: the IAM's name for this technology differs from the model's own.
solar: {source: IAM, iam_name: solar-pv}

# overrides: per-parameter exceptions, each either "IAM", "PyPSA", or a fixed value.
biomass-igcc-ccs:
  source: IAM
  overrides:
    CO2 intensity: {value: 0, unit: "tCO2/MWh_th", comment: "carbon-negative in REMIND, but that makes PyPSA unbounded"}

# A technology that's entirely PyPSA-sourced, with one parameter zeroed out because
# the IAM's own cost for the primary technology already includes it.
offwind-ac-connection-submarine:
  source: PyPSA
  overrides:
    investment: {value: 0, unit: "USD/MW/km", comment: "Connection costs already included in REMIND offwind investment"}

See examples/technology-mapping.example.yaml for a complete, annotated example covering every case above (IAM-sourced, PyPSA-sourced, iam_name remaps, partial and full overrides).

See pypsa-eur-iam's config/technology_mapping_REMIND.yaml for a working example used in the PyPSA-Eur coupling.

Where it's used

The same mapping serves two different consumers:

  • Cost sourcing — transforms/costs.py: build_iam_techdata and build_pypsa_techdata split the model's baseline cost table and the IAM's cost table according to each technology's build_technology_sources, and apply_overrides merges them back together, IAM values overriding baseline values only where source: IAM (or an explicit override) says so.
  • Capacity-target technology selection — transforms/capacities.py: build_capacity_targets and build_capacity_reporting_technologies use the mapping to decide which of the IAM's reported technologies should be summed into which of the model's carriers when building capacity targets (see Harmonising capacities for how those targets then get applied).

If a technology declares source: IAM (directly or via an override) for a parameter, but the IAM's own output has no matching data for it, build_iam_techdata raises an error.

Next