Skip to content

TYNDP Transmission Capacity Trajectories

PyPSA-AT derives per-corridor NTC (Net Transfer Capacity) values from ENTSO-E's Ten-Year Network Development Plan (TYNDP) reference grid and investment dataset. These values are applied in two complementary ways: as per-snapshot flow-limit constraints that cap net power flows on every modelled interconnector during optimisation, and as minimum expansion lower bounds that prevent the solver from building less transmission than TYNDP-mandated targets.

This page covers transmission capacity trajectories only. For generator capacity trajectories (onshore wind, solar PV, batteries, hydrogen electrolysis), see Open-TYNDP Capacity Trajectories.

Raw Data

Both input files are part of the Open-TYNDP dataset, downloaded by the retrieve_open_tyndp rule (rules/open-tyndp/retrieve.smk):

File Contents TYNDP term
Investment Datasets/GRID.xlsx, sheet Electricity Per-border capacity increases for investment projects, with a BORDER column tagging each row as "Real" or "Concept" Investment Dataset
Line data/ReferenceGrid_Electricity.xlsx, sheet 2030 Installed NTC per border in the 2030 reference grid, columns Summary Direction 1 / Summary Direction 2 Reference Grid

The dataset is retrieved from the Open-TYNDP Google Cloud Storage archive maintained by Open Energy Transition. The version entry tyndp,2024,archive in data/versions.csv points to https://storage.googleapis.com/open-tyndp-data-store/2024.

Building the Trajectory Table

Rule: build_tyndp_transmission_trajectories in rules/open-tyndp/build.smk (line 222)
Script: scripts/pypsa-at/build_tyndp_transmission_trajectories.py
Output: resources/tyndp_transmission_trajectories.csv

The script constructs four annual snapshots that represent the NTC capacity on each border at each planning horizon:

Year Derivation
2025 Equal to the 2030 reference grid — projects assumed already built by 2025
2030 Summed NTC from ReferenceGrid_Electricity.xlsx per border
2040 2030 base plus all "Real" capacity increases from GRID.xlsx
2050 2040 values plus all "Concept" capacity increases from GRID.xlsx

The distinction between "Real" and "Concept" projects reflects TYNDP maturity levels: Real projects have a firmer status (permitting or under construction), Concept projects are longer-horizon proposals. Stacking them cumulatively gives a plausible NTC trajectory through 2050.

direct_capacity and indirect_capacity carry the NTC in each direction of a border (e.g. FR→DE and DE→FR independently). Both columns are always present in the output CSV.

Node mapping

TYNDP uses its own node codes (AT00, DE00, ITA0, ITSI, etc.). These are mapped to PyPSA-AT location strings via TYNDP_TO_PYPSA_LOCATION_TRANSMISSION in mods/constants.py. Key rules:

  • Most countries have a single node: AT00 → AT, DE00 → DE, FR00 → FR.
  • Countries with sub-national TYNDP nodes are collapsed to their PyPSA-AT cluster code: ITSI → IT1 (Sicily), GBNI → GB1 (Northern Ireland), DKE1 → DK1 (East Denmark).
  • Countries not modelled in PyPSA-AT (Turkey TR00, Ukraine UA00/UA01, Egypt EG00, etc.) map to None and their border rows are dropped before writing the CSV.
  • The function map_tyndp_nodes (scripts/pypsa-at/build_tyndp_transmission_trajectories.py, line 113) raises a ValueError if any unknown node code appears — this guards against silent data drift in new TYNDP releases.

Border direction is canonicalised so that from_node < to_node lexicographically. Where the raw data has the opposite ordering, direct_capacity and indirect_capacity are swapped. This ensures the CSV has a unique, predictable row per corridor.

Application 1: Per-Snapshot NTC Flow Constraints

Function: add_cross_border_flow_limits in mods/constraints.py (line 156)
Called from: scripts/pypsa-de/additional_functionality.py
Enabled by: mods.tyndp_cross_border_flow_limits.enable: true in config/config.at.yaml

During solve, two per-snapshot linopy constraints are added for every corridor in the trajectory CSV:

# direct direction
sum(Line-s fwd) − sum(Line-s rev) + sum(Link-p fwd) − sum(Link-p rev) ≤ direct_capacity

# indirect direction
sum(Line-s rev) − sum(Line-s fwd) + sum(Link-p rev) − sum(Link-p fwd) ≤ indirect_capacity

Constraint names follow the pattern tyndp_ntc_flow_dir-{A}-{B}-{year} (direct) and tyndp_ntc_flow_indir-{B}-{A}-{year} (indirect). Corridors where no matching active AC Lines or DC Links exist in the model are silently skipped with a warning.

The shared helper compare_tyndp_and_model_borders (mods/utils.py) validates that every model cross-border corridor is covered by the trajectory CSV — missing coverage raises a ValueError. Corridors defined in TYNDP but absent from the model are accepted (a warning is logged).

Time-varying, not a GlobalConstraint

Each constraint generates one inequality per snapshot. The constraints are not registered as GlobalConstraint network objects — they operate at the snapshot level, not as annual aggregates.

Application 2: Minimum Expansion Lower Bounds

Function: apply_tyndp_transmission_lower_bounds in mods/transmission_lower_bounds.py (line 157)
Called from: mods/network_updates.py:modify_prenetwork (line 461)
Active for years listed in: mods.tyndp_lower_bounds.years in config/config.at.yaml

This function ensures the optimizer cannot build less cross-border transmission than the TYNDP NTC target for that planning horizon. It operates on the pre-solve network (modify_prenetwork step), before linopy constraints are constructed.

For each corridor the function:

  1. Reads the target: NTC_target = max(direct_capacity, indirect_capacity) — the dominant transfer direction is used as the single floor value.
  2. Computes installed_cap: the sum of all Line and Link capacities on the corridor, weighted by s_max_pu / p_max_pu to convert rated capacity to active transfer capacity. Both extendable components (contributing their current s_nom_min / p_nom_min) and non-extendable components (contributing their fixed s_nom / p_nom) are included.
  3. Derives shortfall = NTC_target − installed_cap. When shortfall ≤ 0, the target is already met and the corridor is skipped.
  4. When shortfall > 0 and extendable capacity exists, all extendable lower bounds are scaled proportionally by cap_factor = (total_ext_cap + shortfall) / total_ext_cap, preserving the relative distribution across parallel components.
  5. When no extendable capacity is currently present (s_nom_min = p_nom_min = 0), the shortfall is distributed as a raw lower bound per component such that their effective capacity sum equals the shortfall.

The function validates that cross-border DC Links are modelled as bidirectional pairs with symmetric lower bounds, and raises a ValueError if either condition is violated.

Why apply lower bounds separately from flow constraints?

Per-snapshot flow constraints (Application 1) cap power flows during dispatch but do not force the model to install capacity. Lower bounds on s_nom_min / p_nom_min complement them by ensuring that the installed grid itself meets TYNDP targets, regardless of whether the dispatch constraint is binding.

Configuration

mods:
  tyndp_cross_border_flow_limits:
    enable: true          # adds per-snapshot NTC constraints during solve

  tyndp_lower_bounds:
    years:                # apply s_nom_min / p_nom_min lower bounds for these horizons
      - 2030
      - 2040

Setting tyndp_cross_border_flow_limits.enable: false removes the per-snapshot constraints entirely. Setting tyndp_lower_bounds.years to an empty list disables all lower-bound enforcement.

Interaction with Other Transmission Constraints

PyPSA-AT applies multiple transmission constraints in sequence. Their interaction is:

Constraint Mechanism Scope
Global volume cap (v1.25) electricity.transmission_limit → prepare_network.py Total installed grid volume ≤ 125% of reference
TYNDP flow limits add_cross_border_flow_limits (during solve) Per-snapshot net flow per corridor
TYNDP lower bounds apply_tyndp_transmission_lower_bounds (pre-solve) Per-corridor minimum installed capacity

All three can be active simultaneously. The volume cap constrains the global ceiling; the TYNDP lower bounds set per-corridor floors; the flow constraints cap dispatch-time flows independent of installed capacity.

See Transmission Limits for details on the global volume cap.
See Open-TYNDP Capacity Trajectories for generator capacity bounds from the same dataset.