Concept selection & screening
Field-development screening and layout
A client brings well locations and what they expect from the prospect. We return a laid-out field over the real terrain or bathymetry, engineering-screened, rendered as visuals, with the engineering exposed through a small API. Onshore first, then offshore.
It is built to sit before the high-end tools, not instead of them — the deterministic screening pass that tells you which concepts are worth the expensive analysis.
What it does
Layout over real ground
Assets are placed over terrain or bathymetry from verified public sources, with routing that respects the surface. Wells, manifolds, trees, jumpers, flowlines, pipelines, umbilicals and the host all live in one declarative model — so the picture and the engineering read from the same file rather than drifting apart.
Engineering behind the visuals
The screening pass runs deterministic checks, not decoration: cable sizing to IEC, erosional velocity limits per API RP 14E, and pipe-size sweeps, each emitting a one-page summary per layout iteration. Every figure traces back to the same layout model the visuals are drawn from.
Historical analogs, ranked
Past field developments are queryable as a ranked analog set — by region, water depth and asset type — so a concept can be positioned against what has actually been built rather than against an assumption.
Iteration that stays cheap
Four entry points — layout, screen, visualize,
analogs — over a YAML model with overrides. Change the configuration,
re-screen, re-render. Concept selection typically runs five or six models per prospect with
ten to twenty iterations shown; the cost of one more iteration is what decides whether that
is comfortable or painful.
Renewables in the same model
Wind turbines, offshore substations, and array and export cables use the same schema as the oil-and-gas assets, so a mixed or converted development is one model rather than two.
What has actually been built
Onshore, deepwater-FPSO and floating-wind demonstration fields all run end to end — layout, screening, scene export and animation — and are render-verified on a headless pipeline, so they regenerate reproducibly rather than existing as saved pictures.
The engine is open source and MIT licensed. The terrain, bathymetry and past-project sources are public and catalogued with provenance, so the inputs can be checked as readily as the outputs.
Honest limits
This is a screening tier. It is designed to narrow a field of concepts quickly and defensibly, and it is not a substitute for detailed flow assurance, dynamic analysis or certified design. A screening estimate is not a certified deliverable, and nothing here is presented as one.
The historical analog set currently covers the Gulf of Mexico with provenance enforced on every record; other regions are being added, and the coverage is stated rather than implied. Where the public data cannot support a conclusion, the output says so — that admission is the point, not a caveat bolted on afterwards.