Ecowende offshore wind: wildlife-aware curtailment screening#
News trigger, decision and provenance#
On 2 November 2026, Eneco reported that Ecowende’s last turbine had been installed on 30 October 2026. Its announcement identifies 52 turbines, approximately 760 MW offshore wind capacity, adaptive shutdown control, AI-assisted species-recognition radar, a bird corridor and seven red-blade experiments. These are developer-reported facts, not monitoring results. Eneco explicitly says collision-reduction effectiveness remains under study.
Waardenburg Ecology’s Maasvlakte II bird-radar shutdown protocol documents a different real site where 3D radar counts of birds at rotor height trigger turbine stops. It provides a method precedent, not Ecowende detections, trigger values or effectiveness. TNO’s Eemshaven bat-radar pilot motivates targeted rather than blanket shutdown; its data are not assumed to apply to Ecowende. The WOZEP public data portal is a candidate for future measured data, subject to dataset-level access and licensing checks; no connector or observation is claimed in this contribution.
Business question: How much synthetic energy is lost under a broad night shutdown versus an illustrative radar-triggered policy? At the same activity-exposure-proxy reduction, what is the theoretical minimum energy loss of an ideal continuous policy? This is an engineering screening and monitoring investment concept, not a wildlife conservation or permit-compliance assessment.
Input classification#
Input |
Classification |
Provenance and interpretation |
|---|---|---|
Installed turbines: 52; plant capacity: approximately 760 MW |
Developer-reported |
Eneco, 2 Nov 2026; capacity is approximate, not a metered output or turbine performance curve |
14 days hourly wind speed |
Synthetic |
Deterministic 8.8 + sine + cosine m/s example, no real weather |
Cubic power curve |
Synthetic physical surrogate |
Assumed 3 m/s cut-in, 12 m/s rated, 25 m/s cut-out; not a Vestas OEM curve |
Hourly radar-like activity index |
Synthetic |
Arbitrary index with deterministic nighttime peaks, not observed counts |
Nighttime 18:00-05:59 shutdown |
Synthetic policy comparison |
Not Ecowende’s actual practice |
Activity threshold >=3.0 |
Synthetic policy comparison |
Not a published radar or regulatory threshold |
Energy-minimum policy |
Synthetic mathematical bound |
Ideal foresight and fractional-hour curtailment; not practical real-time dispatch |
There are no Ecowende measured wind-speed, production, radar, animal-collision or financial records in this example. All derived numerical energy results are scenario calculations, not actual lost production or verified wildlife benefits.
Method, units and baseline#
The baseline is unconstrained synthetic energy output. For interval (t) of length (\Delta t) hours, modelled available power (P_t) (MW), arbitrary positive radar activity (a_t) (index) and shutdown fraction (f_t) in [0,1], the method computes:
Baseline energy: (\sum P_t \Delta t) MWh.
Curtailed energy: (\sum P_t f_t\Delta t) MWh.
Operating exposure proxy: (\sum_{P_t>0}a_t (1-f_t)\Delta t), in index-hours.
The proxy is not collision probability, prevented mortality or an ecological risk score. The zero-generation case assumes the rotor is stopped and carries no proxy exposure. The fraction of proxy avoided is relative to the no-curtailment baseline, not a quantified survival benefit.
The reusable min_energy_loss_schedule solves the continuous fractional-knapsack LP: minimize curtailed MWh while reducing this arbitrary proxy by a specified share. Sorting active intervals by (P_t/a_t) yields the exact energy minimum for this linear surrogate, assuming full knowledge of future index values. A fraction of one interval means the rotor hypothetically stops for part of that constant-condition hour. This does not solve binary turbine stop/start optimization or impose restart dynamics, uncertainty, spatial patterns, species-specific thresholds or legal restrictions.
Only compare the threshold case against the mathematical bound at the same proxy reduction; the blanket case generally delivers a different proxy reduction. No equivalence of real-world animal protection is established.
Reproduce and validate#
uv sync --locked --all-packages --group docs --python 3.13
uv run --no-sync pytest methods/tests/test_wildlife_curtailment.py tests/test_gallery_navigation.py
uv run --no-sync python examples/analysis/plot_ecowende_wildlife_curtailment.py
uv run --no-sync ruff check .
uv run --no-sync ruff format --check .
uv run --no-sync ty check
uv run --no-sync sphinx-build -W --keep-going -b html docs docs/_build/html
uv run --no-sync python scripts/check_docs_links.py docs/_build/html
The offline runnable example plots the trade-off of three synthetic policies and a perfect-foresight proxy frontier. The method validates nonnegative finite inputs and energy conservation. The unit tests cover hand-calculated energy/proxy accounting, zero power, fractional optimum, targets, monotonicity and invalid inputs.
Critical limitations: The 14-day profile is not representative of a year; radar accuracy, species response, actual turbine rotor speed, wind wakes, sensor false detections, shutdown delay and losses, and economics are absent. Do not annualize or claim avoidance effectiveness. Future work needs licensed measured time-stamped radar detections, SCADA power and controls, independent field verification, permit rules and collision outcomes.