Skip to content

Latest commit

 

History

1 Commit

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 

Repository files navigation

Maritime Fuel Optimization

A deterministic decision-support prototype for maritime fuel, carbon, and compliance-cost scenarios

Python · SciPy · Numerical optimization · FuelEU Maritime · EU ETS · Scenario analysis

Illustrative 2025 modeled-cost comparison

Overview

Maritime Fuel Optimization began as a team prototype in 2024. It explores how ship operators could compare fuel strategies while accounting for fuel cost, greenhouse-gas intensity, carbon allowance exposure, voyage scope, and onshore power supply (OPS) at berth.

The original experimental repository has been preserved privately. This public case study presents the problem, verified engineering approach, and selected sanitized results without distributing source code, third-party calculators, or regulatory documents.

The problem

Fuel choice is not a single-variable decision. A lower-emission fuel can have a higher purchase price, while a conventional fuel may create additional FuelEU or EU ETS exposure. Voyage geography and berth operations change the modeled scope again.

The prototype evaluates those interacting costs as one optimization problem:

flowchart LR
    A[Voyage energy profile] --> D[Deterministic optimizer]
    B[Fuel and carbon assumptions] --> D
    C[FuelEU and EU ETS rules] --> D
    D --> E[Fuel-mix recommendation]
    D --> F[Cost breakdown]
    D --> G[Emissions and compliance indicators]
Loading

What the model evaluates

  • Intra-EU and extra-EU voyage energy
  • Fuel mass, energy density, and purchase cost
  • Well-to-wake greenhouse-gas intensity
  • Tank-to-wake CO2 allowance exposure
  • FuelEU target milestones from 2025 to 2050
  • EU ETS phase-in for maritime emissions
  • MDO-at-berth versus OPS scenarios
  • HFO, VLSFO, MDO, biodiesel, LNG, and e-methanol assumptions

Engineering approach

The cleaned model uses a deterministic multi-start SLSQP optimizer with bounded energy shares and an explicit sum-to-one constraint. It replaced several exploratory scripts and a nondeterministic evolutionary solver.

The completed private implementation includes:

  • One portable Python package and command-line interface
  • Versioned, human-readable assumptions
  • Deterministic scenario generation
  • Automated tests for scope, phase-in, cost reconciliation, OPS, and optimizer repeatability
  • A reproducible matrix of 40 scenarios across milestone years, berth choices, and fuel configurations

Illustrative result

The bundled 2025 profile uses 14.5 million MJ of intra-EU energy, 52 million MJ of extra-EU energy, 14.5 million MJ at berth, a carbon price of €90/tCO2, and no OPS.

Configuration Optimized energy mix Total modeled cost
VLSFO baseline 100% VLSFO €1,504,957
VLSFO + biodiesel 96.5% VLSFO / 3.5% biodiesel €1,449,847
LNG 100% LNG €1,282,199
E-methanol 100% e-methanol €5,005,059
Flexible four-fuel mix 100% LNG €1,282,199

Within these illustrative prototype assumptions, the flexible scenario was 14.8% lower cost than the VLSFO-only baseline. This is evidence that the model runs and reconciles—not a prediction that LNG is universally optimal.

The sanitized result table is available in data/2025-scenario-summary.csv.

What was learned

  • Regulatory scope and phase-in rules should be explicit model inputs, not scattered constants.
  • Determinism matters when comparing scenario outputs or presenting results.
  • Lifecycle intensity and direct-emission factors answer different questions and should remain separate.
  • A technically plausible optimization can still be misleading if fuel prices, eligibility, and regulatory exclusions are not carefully sourced.

Limitations

This is a research and portfolio prototype—not regulatory, legal, engineering, financial, procurement, or operational advice.

  • Fuel prices and lifecycle factors are illustrative 2024 prototype inputs.
  • Methane and nitrous oxide ETS exposure from 2026 is not modeled.
  • Pooling, banking, borrowing, RFNBO incentives, exemptions, verifier workflows, and ship-specific eligibility are outside scope.
  • OPS is simplified as electricity cost with zero onboard emissions; grid lifecycle intensity is not included.
  • Real decisions require current attributable data and independent review.

Official references

Repository notice

This repository is a public portfolio case study. The production source and its historical artifacts remain private. See NOTICE.md for usage terms.

About

Portfolio case study for a maritime fuel-mix, FuelEU and EU ETS optimization prototype

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors