Ship Kernels

A ship kernel is a data-driven vessel performance model.

Ship kernels are built from a ship's sensor data. They can model a ship's primary operational parameters in any given condition with much higher accuracy than classical performance models. As a ship's performance evolves, so does its ship kernel evolve. Our APIs enable them to be easily plugged into existing systems.

This flexibility enables them to be used for performance analysis, simulation, routing and numerous other use cases.

Primary operational parameters

The primary operational parameters are the drivers behind a ship's propulsion. It is the relationship between these parameters that reflect a ship's performance. Given the value of exactly one of these as input, called the entrypoint, a ship kernel will predict the values of the others as output.
Ship kernels can model the relationship between any of the following primary operational parameters:

  • Speed Over Ground (SOG)
  • Speed Through Water (STW)
  • Main Engine RPM (RPM)
  • Main Engine Shaft Power (ME Power)
  • Main Engine Fuel Consumption (ME FO Consumption).

Conditioning parameters

The conditioning parameters are the factors that influence the relation between the ship's primary operational parameters. These can be provided at the input and will influence the value of the predicted primary operational parameters.
Ship kernels can model the impact of any of the following conditions:

Weather conditions

  • Wind speed and direction
  • Wave height and direction
  • Current speed and direction
  • Sea surface salinity
  • Sea surface temperature

Operational conditions

  • Mean draft
  • Trim
  • Ship heading
  • Fuel specific energy


Not all of these parameters have to be present in order to build a ship kernel. At the expense of model accuracy, using a subset of them will also work.

Have a look at our API Reference for the naming we use for these parameters. Parameters that are required by all models are marked as "required".

Date parameter

A ship's performance is not static over time due to the effects of fouling, cleanings, hull degradation, etc. The date parameter can be provided to model the ship as it historically performed on the given date.

This allows for the analysis of the ship's performance over time.

Whatโ€™s Next

Get started with your first ship kernel by following the quickstart guide.