POEM

Performance pOlar Exchange forMat

C++ library and Python bindings defining an exchange format for performance polar in marine engineering, wind assisted vessels performance prediction, weather routing optimisation, embedded control systems, training simulators.

• Links
• Introduction
• POEM in a nuttshell
• Python interface
  • Requirements
  • Install from Pypi
  • Install from sources
  • Basic usage
• Integrating the C++ library
  • Requirements
  • CMake
  • Basic usage
• Current limitations

Links

• GitHub repository: https://github.com/D-ICE/poem
• POEM Documentation: https://dice-poem.readthedocs.io/en/latest/
• D-ICE ENGINEERING: https://www.dice-engineering.com/en

Introduction

• what is a VPP
• what is a digital twin
• what is a performance polar
• the different type of polars
• how a polar is used

POEM in a nuttshel

• why poem
• what is poem
• is poem for me

Python interface

The pypoem package that wraps poem is pypoem. It does not currently give acces to the whole C++ API but feature are progressively ported to Python.

Requirements

pypoem is tested with Python version >= 3.9.

Install from Pypi

pypoem is available on Pypi, so it can be installed as:

$ pip install pypoem

Install from sources

pypoem relies on scikit-build-core for the packaging. Source installation is realised by running the following command int the repository root directory:

$ pip install .

Basic usage

Here is a simple example of the creation of a MPPP Polar with one PolarTable and its writing to a netCDF file compliant with POEM file format thanks the of POEM library.

from pypoem import pypoem
import numpy as np

# Declaring three dimensions
STW = pypoem.make_dimension("STW", "kt", "Speed Through Water")
TWS = pypoem.make_dimension("TWS", "kt", "True Wind Speed")
TWA = pypoem.make_dimension("TWA", "deg", "True Wind Angle")

# Creating a DimensionSet
dimension_set = pypoem.make_dimension_set((STW, TWS, TWA))

# Creating a DImensionGrid and filling the dimensions with values
dimension_grid = pypoem.make_dimension_grid(dimension_set)
dimension_grid.set_values("STW", np.linspace(8, 20, 13))
dimension_grid.set_values("TWS", np.linspace(0, 40, 9))
dimension_grid.set_values("TWA", np.linspace(0, 180, 13))

# Creating a Polar of type MPPP (Motor only Power Prediction)
polar_MPPP = pypoem.make_polar("MPPP", pypoem.MPPP, dimension_grid)

# Creating a PolarTable from the MPPP Polar
BrakePower = polar_MPPP.create_polar_table_double("BrakePower", "kW", "Brake Power", 
                                                  pypoem.POEM_DOUBLE)

# Generating a dummy NDArray with ones and the same shape as the DimensionGrid
# Note that NDArrays must be arranged in a row major order with respect the DimensionSet
# of the DimensionGrid. If you think in terms of nested for loop with Dimensions, 
# the last Dimension (here TWA) is moving fastest (most inner loop).
brake_power_data = np.ones(dimension_grid.shape())

# Setting the BrakePower PolarTable with the array
BrakePower.set_values(brake_power_data)
assert(np.all(brake_power_data == BrakePower.array()))

# Writing to netCDF
pypoem.to_netcdf(polar_MPPP, "polar_basic.nc")

• Reading a POEM file
• More complex example

Work In Progress

Integrating the C++ library

Work In Progress

Requirements

Work In Progress

CMake

Work In Progress

Basic usage

Work In Progress

Current limitations

POEM has been primarily developed under Linux OS. Port to other platforms is expected to come in the future but is not a current top priority. Contributors for a Windows port are welcome :)

Latest standard is available at: https://dice-poem.readthedocs.io/en/latest/