Pre-trade screening, ISO 8217 compliance, stability prediction, and cost optimisation for marine fuel blends — powered by ML trained on lab-verified data.
FuelOS brings together spec compliance, predictive risk models, and blend optimisation into a single pre-trade workflow.
Screen blends against VLSFO and HSFO grade limits for RMG/RMK viscosity classes. Hard limits flagged with failed property breakdown.
Predict CII, SBN, TSP, and asphaltene compatibility using ensemble ML models. Classifies blends as LOW / MEDIUM / HIGH stability risk.
Enumerate candidate weight combinations across your cargo pool. Surface lowest-risk, lowest-cost blends that pass all spec limits.
Predict saturates, aromatics, resins, and asphaltene fractions from physical properties when direct measurements aren't available.
Every predicted property is labelled — measured, imputed, modelled, or missing — so you know exactly which values need lab confirmation.
Actionable screening decisions: ACCEPT / SCREEN LABS / DEFER / REJECT — with reasoning, risk score, and recommended lab tests.
Upload a PDF Certificate of Analysis. The system auto-fills cargo properties, reducing manual entry errors on lab-certified cargoes.
Save and reuse cargoes across sessions. Build a library of known streams and load them directly into the blend optimiser.
Enter real cargo properties below and run a live ISO 8217 screen against VLSFO grade limits. No login required.
Three steps, under two seconds.
Enter quality specs manually, upload a COA PDF to auto-fill, or load from your saved cargo library.
Select the ISO 8217 target grade and run Evaluate or Optimize. The engine screens every weight combination.
Each blend is classified ACCEPT / SCREEN LABS / DEFER / REJECT with a risk score, failed checks, and recommended lab tests.
Export a deal memo, commercial report, or CSV for your back office. Provenance labels stay attached to every number.
Each model is trained on lab-verified blending data and benchmarked against held-out test sets. Provenance labels show which values are measured vs. predicted.
| Model | Predicts | Type | Scope |
|---|---|---|---|
| Flash point | Flash point °C from density, viscosity, MCR | ML | Component & blend |
| Pour point | Pour point °C; cold-flow risk | ML | Component & blend |
| SARA | Saturates, aromatics, resins, asphaltenes wt% | ML | Component |
| MCR | Micro carbon residue from SARA & density | Hybrid | Blend |
| Stability (CII / TSP) | Blend stability risk: LOW / MEDIUM / HIGH | ML | Blend |
| ISO 8217 spec checker | Pass/fail per property, grade, year | Rule | Blend |
| Viscosity blending | Blended v50, v100 from component viscosities | Hybrid | Blend |
Request access or sign in to the full trading cockpit with blend optimiser and cargo library.