KDS Offshore
Start a project →
Capability · 09 / 09 · Digitalisation

Digitalisation & Digital Twin

Physics-based twins from design to fleet operation.

Lead time · 6–16 weeksFixed-fee, fixed-scopePrincipal-ledAuthor · Sérgio Ribeiro e Silva, PhDLast reviewed ·
✦ TL;DR

A maritime digital twin is a physics-based predictive model — CFD-derived hydrodynamics coupled with onboard sensor streams — that answers operational what-if questions (trim, speed, route, retrofit, CII trajectory) without onboard trials. KDS Offshore builds twins to ISO/IEC 23247 + ISO 19030, calibrated against measurements with documented residuals. The same Ship@Sea physics that sized your vessel during design carries into operation as a calibrated predictor of resistance, fuel burn, and CII. Real-time voyage optimisation by Vectorized Simulated Annealing, published by Ribeiro e Silva & Bento Moreira (IST/CENTEC) at IMDC 2024 (paper 832) and ICCAS 2024.

Overview

What this service is

Most maritime "digital twins" sold today are dashboards layered on top of SCADA streams. Useful for situational awareness, but they cannot answer the question the operator actually has: "what if we change trim by one degree? What if we slow by half a knot? What if we re-route around this front? What if we retrofit this propeller?" Answering those needs a physics-based model — and that is the layer KDS Offshore builds.

We couple CFD-derived hydrodynamic models with onboard sensor streams to produce a twin that lives next to the data, not in place of it. The same physics model that sized the vessel during design carries into operation as a calibrated predictor of resistance, fuel burn, trim sensitivity, and CII trajectory. The architecture follows ISO/IEC 23247 (Digital Twin Framework) for the model / data / actor layering, and ISO 19030 (ship performance & fuel-consumption monitoring) for the in-service calibration loop. Documented residuals and uncertainty bounds replace black-box confidence.

The principal Sérgio Ribeiro e Silva (PhD IST, MSc UCL, 364+ Google Scholar citations, h-index 10) has been building physics-based seakeeping and manoeuvring simulators since his 2005 IST PhD on parametric rolling. The in-house Ship@Sea time-domain code is the operational descendant of that line of work, and it carries the same validation pedigree (parametric rolling: Ribeiro e Silva 2005, Ocean Engineering 2013). The twin you get is not a Series-A startup's first product. It is a 20-year hydrodynamic codebase wired to your sensors.

Reference work includes the SEAPOWER 1500 — a 15 m fully electric pilot boat whose digital twin was built before steel was cut (2025) — and the real-time Ship Operation Optimisation System (SOOS) that reduces fuel consumption and emissions on navigation and port calls. SOOS combines a CFD-derived calm-water power curve, semi-empirical wind loads, the Salvesen-1978 added-resistance-in-waves model, trim through centre-of-gravity, and a Vectorized Simulated Annealing weather-routing optimiser. On a 712 TEU geared containership in synthetic Atlantic conditions, SOOS sampled a population of 200 candidate routes, converged in under 100 epochs (capped at 750), and reduced voyage fuel by 8–9% versus the great-circle direct route, holding the saving even when an obstacle (island, marine corridor) was inserted in the search space. Published at IMDC 2024 (Amsterdam, paper 832) and ICCAS 2024 (RINA, Genoa). The optimiser has no hyperparameters the bridge needs to tune.

The methodology is published, not proprietary opacity. The two underlying papers document the optimisation algorithm, the hydrodynamic model, the regulatory context (CII compliance objective), and the validation case. Operators get a twin they can audit, classification societies get a model they can review, and lenders get a framework they can defend in a Poseidon-Principles disclosure. Black boxes do not survive an EEXI / CII conversation.

See our methods →

Outcomes

Numbers we expect to defend

Pre-build
twin delivered before steel was cut
SEAPOWER 1500 — 15 m fully electric pilot boat, SeaPower (2025)
8–9%
voyage fuel saved on a synthetic Atlantic case
712 TEU geared containership, SOOS programme, IMDC 2024 (paper 832) + ICCAS 2024
<100 epochs
optimiser convergence
Vectorized Simulated Annealing on 200 candidate routes; cap was 750 epochs
0
hyperparameters the bridge has to tune
operator-grade real-time decision support — by design
Minimum-viable
sensor set: GPS, IMU, fuel flow, shaft RPM, environment
simpler than vendor norm — fewer failure modes, faster retrofit
ISO/IEC 23247
+ ISO 19030 architecture standards
model / data / actor layering + in-service performance monitoring conventions
Deliverables

What you get

01
Digital Twin development for electric propulsion systems
02
LNG ship-to-ship Digital Bunkering
03
Route mapping from just-in-time weather forecasts
04
Applied software development
05
Offshore operations & port infrastructure
06
Physics-based digital twin (CFD-derived, calibrated against onboard sensors)
07
Real-time performance dashboard with documented residuals
08
Predictive trim, speed, and fuel-burn models across the operational envelope
09
Voyage optimisation deployment (SOOS, Vectorized Simulated Annealing weather routing)
10
Sensor architecture & instrumentation specification (minimum-viable set)
11
Fleet benchmarking framework with cross-vessel performance metrics
12
Operator decision-support tooling (trim advisor, speed-fuel curve, retrofit-impact estimator)
13
CII / FuelEU forecasting and rating-trajectory projection
14
EEXI verification package (CFD-validated power curves)
15
ISO 19030 in-service performance monitoring report
16
Peer-reviewed methodology backing every claim
Fit

When this is the right call

✓ When to use
  • A new build where the design-stage CFD model can carry into the operational twin from day one — no second model, no calibration gap.
  • An existing vessel needing a trim advisor, speed-fuel curve, or retrofit-impact estimator — the CFD baseline closes the loop the sensors cannot.
  • A fleet operator needing a single benchmarking framework across mixed-age vessels — same model architecture, vessel-specific calibration.
  • A pilot project preceding a multi-year fleet rollout — KDS ships a working twin in weeks, not months.
  • A class-society EEXI / EEDI verification needing CFD-validated power curves over the operational envelope.
  • An owner facing a CII rating downgrade who needs operational evidence of remedial action (SOOS deployment, trim optimisation, hull-cleaning trigger).
  • A charter-party negotiation where the operator needs to defend the projected fuel saving from a proposed retrofit, ahead of the asset change.
✗ When not to use
  • A pure data-engineering job (collect and visualise SCADA streams) — use a DAQ vendor instead.
  • A pure ML / regression model with no physics anchor — accurate inside the training envelope, dangerous outside it. We build hybrid models, not curve fits.
  • A vessel whose operator is not willing to share at least 90 days of operational data for calibration — the twin needs a calibration loop to be defensible.
Methodology

How we run a project

  1. 01 / 06

    CFD-derived baseline

    Resistance & propulsion curves from CFD (Simerics MP, OpenFOAM, or STAR-CCM+) across the operational envelope — calm-water, free-trim, free-sinkage, multiple loading conditions. Sanity-checked against Holtrop-Mennen empirical resistance. This is the twin's physics layer.

  2. 02 / 06

    Seakeeping & wave-added resistance

    Salvesen-1978 strip-theory added-resistance model coupled with the Ship@Sea time-domain seakeeping core. Wave forces and motion-induced added resistance feed the operational fuel-burn predictor under real metocean.

  3. 03 / 06

    Sensor architecture

    Specify the minimum viable sensor set — shaft RPM, torque (when available), GPS, fuel flow, IMU, environmental package — that drives a useful twin. We bias for simplicity; one extra sensor that adds nothing is one more failure mode in service.

  4. 04 / 06

    Twin integration & calibration

    Couple the CFD-derived model with live data streams. Calibration loop against onboard measurements over 90+ days, with documented residuals and uncertainty bounds. ISO 19030 conventions for performance-trend extraction.

  5. 05 / 06

    Voyage optimisation (SOOS)

    Where the route is non-trivial: SOOS deploys a Vectorized Simulated Annealing optimiser over 200 candidate route variants, converging in <100 epochs on commodity hardware. Pre-computed physics offline + cheap re-evaluation online means re-planning happens in seconds, not minutes.

  6. 06 / 06

    Decision-support layer

    Operator-facing outputs: trim advisor, speed-fuel curve, voyage comparison, retrofit-impact estimator, CII trajectory projection. The model answers questions; the operator decides. No autopilot, no automation surprises.

Tools & methods

Software stack we use

  • Simerics MP — CFD baseline (electric pilot-boat reference)
  • OpenFOAM / STAR-CCM+ — CFD (production resistance and propulsion)
  • WAMIT — diffraction / radiation, RAO generation for the seakeeping core
  • Ship@Sea (KDS proprietary, FORTRAN) — non-linear time-domain seakeeping
  • SOOS (KDS, peer-reviewed) — Vectorized Simulated Annealing weather routing
  • Salvesen-1978 — added resistance in waves
  • Holtrop-Mennen — empirical resistance sanity check
  • Rhino + Grasshopper / Orca3D — geometry
  • Python · NumPy · SciPy · pandas · scikit-learn — data processing & calibration
  • MATLAB / Simulink — control & state-space models
  • InfluxDB / Grafana / TimescaleDB — time-series storage & operator dashboards
  • ISO/IEC 23247 — digital-twin architecture conventions
  • ISO 19030 — in-service performance monitoring methodology
  • IMDC 2024 / ICCAS 2024 published optimisation methodology
Regulatory context

Frameworks we help you satisfy

  • ISO/IEC 23247 (Digital Twin Framework for Manufacturing / Marine)Model / data / actor architecture conventions for production-grade twins.
  • ISO 19030 — Ship Performance & Fuel Consumption MonitoringIn-service performance methodology; underpins the twin calibration loop and retrofit verification.
  • IMO CII (Resolution MEPC.336(76))Twin-derived operational efficiency feeds CII forecasting and rating-trajectory uplift.
  • IMO EEXI (Resolution MEPC.328(76))Existing Ship Energy Efficiency Index; CFD-validated twin supports verification at first survey.
  • EU MRV (Regulation (EU) 2015/757)Twin-derived fuel-burn predictions support MRV reporting and emissions verification.
  • FuelEU Maritime (Regulation (EU) 2023/1805) ↗Twin-projected voyage saving feeds FuelEU compliance forecasting.
  • IMO MSC.428(98) — Maritime Cyber Risk ManagementTwin integration with shipboard systems must satisfy the cyber-risk safety-management framework.
  • NIST SP 800-160 — Systems Security EngineeringReference for the twin's system-level architecture and trustworthiness.
  • IMO MSC-MEPC.2/Circ.12 — Autonomous Vessel (MASS) Trial FrameworkFor unmanned-vessel digital twins, operability evidence must satisfy the MASS evaluation process.
Vessel types

Where this discipline applies

  • Pilot boats and harbour craft (SEAPOWER 1500 — full design + operational twin)
  • Ferries, RoPax, and short-sea passenger
  • Container ships and feeders (SOOS-published validation case — 712 TEU)
  • Bulk carriers and tankers (voyage optimisation, EEXI verification)
  • Cruise vessels (trim and speed advisory, port-call optimisation)
  • Tugs, OSVs, and working boats (duty-cycle profiling)
  • Floating renewable-energy installations (wind, wave) — coupled monitoring
  • Mixed-age fleets needing cross-vessel benchmarking
Selected work

Where this discipline was used

·01SEAPOWER 1500 — CFD-derived digital twin, delivered pre-constructionSeaPower2025↗·02SOOS — real-time voyage optimisation (712 TEU containership, 8–9% fuel saved)KDS R&D · IST/CENTEC2024↗·03IMDC 2024 paper 832 — published methodology for fleet-grade optimisationInternational Marine Design Conference, Amsterdam2024↗·04ICCAS 2024 — SOOS programme presentationRINA, Genoa2024↗·05"Belize I" — performance-monitoring framework on remotorised catamaranNautiber2023↗
Common questions

FAQ

Is this just another SCADA dashboard?

No. SCADA dashboards visualise what the sensors measure. A physics-based twin can answer "what if we change trim by one degree?" without onboard testing — because the CFD-derived model fills the gap the sensors cannot see. The model lives next to the data, not in place of it.

How much sensor instrumentation do I need to install?

Less than most vendors will tell you. A typical setup runs on GPS, IMU, fuel flow, shaft RPM, and a small environmental package. The CFD-derived model infers what the sensors cannot directly measure (resistance components, propulsive efficiency at off-design points), so we can ship a useful twin with a minimum-viable sensor set and add instrumentation only where it genuinely improves calibration.

Can the twin run on existing vessels, or only new builds?

Both. For new builds we couple design-stage CFD into the operational twin from day one. For existing vessels we run a one-off CFD baseline against the as-built hull and then attach the live data stream. The retrofit path takes weeks, not months.

Will the twin replace our sea-trial programme?

No, and we will not pretend otherwise. Sea trials remain the authoritative measurement at the design point. The twin extends that point into a continuous prediction across the full operational envelope — speeds, drafts, headings, sea states the trial cannot economically cover. The twin is calibrated against the trial, not a substitute for it.

How does the twin feed CII forecasting?

The CFD-derived power curve plus the SOOS optimiser project fuel consumption over the planned voyage profile. Aggregated over an MRV reporting year, that produces a CII trajectory with confidence bands. The operator sees the rating they will hit if they change nothing, and the rating they will hit under each remedial measure (trim optimisation, hull cleaning, SOOS routing, retrofit) — quantified, before the year ends.

Who owns the model and the data?

The operator. KDS hands over the CFD baseline, the calibration scripts, the optimisation code, and the dashboards under a standard licence. We do not lock you into a SaaS we control — if you want to migrate the twin to a different infrastructure provider next year, the artefacts are yours.

How is this different from major OEM digital-twin offerings?

OEMs sell twins tied to their own equipment and their own platform — fine when the fleet is single-OEM, problematic when it is not. KDS is OEM-agnostic and physics-first: the model architecture is published (ISO/IEC 23247 + ISO 19030, IMDC 2024 / ICCAS 2024 methodology), works across mixed-OEM fleets, and is auditable by the operator, the class society, and the financier. No vendor lock-in.

Research

Reference publications

  1. [1]Ribeiro e Silva, S., Bento Moreira, M. (2024). An optimisation-based approach to reduce fuel consumption and emissions from shipping navigation. 15th International Marine Design Conference (IMDC 2024), Amsterdam — paper 832. ↗
  2. [2]Ribeiro e Silva, S., Bento Moreira, M. (2024). An integrated real-time Ship Operation Optimisation System (SOOS) to reduce fuel consumption and emissions from shipping navigation and port calls. ICCAS 2024, RINA, Genoa.
  3. [3]Ribeiro e Silva, S., Varela, J. M. (2022). Ship Gyroscopic Roll Stabilisation. OMAE 2022, ASME, Hamburg. Paper OMAE2022-7953 — BEM with speed corrections + ST methodology informing performance-monitoring twin calibration on small craft.
  4. [4]Ribeiro e Silva, S. et al. (2013). Prediction of parametric rolling in waves with time-domain non-linear strip theory. Ocean Engineering, Vol. 72 — validation pedigree of the Ship@Sea code that anchors the operational twin.
  5. [5]Ribeiro e Silva, S. (2005). Parametrically excited roll in regular and irregular head seas. International Shipbuilding Progress, Vol. 52 — foundational time-domain seakeeping methodology in continuous KDS use.
Other capabilities

Eight other disciplines, one office

✦ Brief this service · Digitalisation

Tell us about the vessel.

A 30-minute call with a principal. No deck, no sales engineer. Two engineers, your problem, an honest answer about whether KDS is the right firm.

Start the conversation