Physics-based fuel consumption modelling, multi-source weather data, and route optimization algorithms for merchant ships.
Route analysis and performance monitoring belong where decisions are made. The Master and bridge team should have direct access to the models, the data, and the reasoning behind every recommendation.
Every calculation can be inspected, every assumption challenged, and every parameter calibrated against real operational data. No black boxes.
A reference implementation of the routing engine, physics models, and weather integration is published under Apache 2.0. The methods and assumptions are open to inspection. The full system, including calibrated vessel models and advanced modules, is developed privately for selected use cases.
Your vessel's data stays yours. The performance model lives on board or on your own server — no vendor lock-in, no data leakage, and the operators who navigate the vessel own and understand the tools they rely on.
From hydrodynamic resistance calculations and engine SFOC curves to real-time weather overlays and IMO CII compliance — all in one platform.
A* grid search with 9 pre-validated commercial strait shortcuts, GSHHS sub-km coastline polygons, course-change penalty, safety fallback routing, and multi-objective Pareto front for fuel vs. time trade-offs. VISIR Dijkstra available as optional secondary engine. Monte Carlo P10/P50/P90 envelopes quantify uncertainty.
Wind (GFS 0.25°), waves, currents, SST, swell, visibility, and ice from Copernicus Marine and NOAA. WMO-standard color ramps, forecast timeline (f000–f120), and animated particle flow. Pre-ingested into PostgreSQL for sub-second queries.
Holtrop-Mennen hull resistance, Blendermann wind loads, Kwon/STAWAVE-1 wave resistance, and quadratic SFOC curves. Calibrate against engine log data to match real-world performance. Four model curves — resistance, power, SFOC, daily fuel — with theoretical vs. calibrated comparison.
Upload CSV or Excel noon reports. Five KPIs, six interactive charts — fuel consumption timeline, RPM distribution, speed through water, event breakdown. One-click calibration bridge connects log data to the physics model.
NOAA GFS + Copernicus CMEMS. 7 layers, user-triggered resync, viewport-aware downloads, climatology fallback.
A* with GSHHS coastline, 9 strait shortcuts, course-change penalty, safety fallback, Pareto front, and 0.5 kt speed steps.
Holtrop-Mennen, Blendermann, Kwon, STAWAVE-1. Seakeeping motions, slamming, parametric roll criteria.
Upload noon reports, calibrate hull/wind/wave/SFOC factors. Match model to real-world performance data.
Correlated weather perturbations. P10/P50/P90 fuel and ETA envelopes for risk-aware voyage planning.
20 TSS polygons, 6 ECA boundaries, High Risk Area, Suez/Panama canal zones. Cost penalties in optimizer.
IMO CII calculator with A–E ratings. FuelEU GHG intensity, compliance balance, pooling scenarios, penalty estimator.
Time-based speed over ground charts, cumulative fuel curves, ETA comparison bars for all route variants.
Docker Compose deployment. PostgreSQL + Redis + FastAPI + Next.js. No cloud dependency. Runs on a laptop.
Click any screenshot to view full resolution.
Weather layers & WMO wind barbs (v0.1.5 — pre-dark-basemap)
The reference implementation has the following scope.
| Constraint | Detail |
|---|---|
| Forecast horizon | 5 days (GFS f000–f120). No historical reanalysis. |
| Wind source | NOAA GFS (0.25°). Free, no credentials needed. |
| Wave / current source | CMEMS. Requires a free Copernicus Marine account. |
| ERA5 | Not active in this release. CDS credentials accepted but ingestion is disabled. |
| Coverage | North Atlantic / NW Europe (ADRS 1+2). Configurable via area selection. |
| Data refresh | Manual — user triggers resync per layer. No background scheduler. |
| Vessel type | Calibrated for MR product tankers. Other vessel types require parameter adjustment. |
For voyages beyond the 5-day forecast window, the system uses climatological fallback values (monthly averages).
Pre-loaded weather data, engine logs, and noon reports. No API keys required for wind data.
mkdir windmar && cd windmar
curl -LO https://raw.githubusercontent.com/windmar-nav/windmar-demo/main/docker-compose.standalone.yml
curl -L -o .env https://raw.githubusercontent.com/windmar-nav/windmar-demo/main/.env.standalone
docker compose -f docker-compose.standalone.yml up -d
Then open http://localhost:3000
The publicly available release is intended for research and education. For operational evaluation, fleet-wide deployment, or custom vessel calibration, contact directly.
For consulting, audits, or technical discussion, reach out via the contact link below.