Skip to main content

Package with additional tools to the OrcFxAPI package

Project description

NsgOrcFx

Library of tools for the OrcaFlex API

This package wraps the original API from Orcina (OrcFxAPI) to include:

  • methods: pre- and post-processing tools such as line selection, load case generation, modal and fatigue analysis
  • coding facilities: auto-complete and hints with descriptions in IDE


All the attributes and methods from the source (OrcFxAPI) still accessible in the same way.


Installation:

pip install --upgrade NsgOrcFx

Example 1 - Auto-complete feature of IDE (e.g. VS Code and Spyder)

import NsgOrcFx

model = NsgOrcFx.Model()
line = model.CreateLine()

The data name may be found in the data attribute with the auto complete of the IDE (e.g., Visual Studio Code, Spyder, and PyCharm).

Screenshot of auto-complete with the 'data' component of objects (e.g., line.data.{data name})

In addition, a hint shows the description of the parameter (mouse cursor stopped in the data name).

Screenshot of hint with the 'data' component of objects (e.g., line.data.{data name})

In the exemple below, data names of general, environment, and line objects are accessed

model.general.data.ImplicitConstantTimeStep = 0.01 # data from general object
model.environment.data.WaveHeight = 5.0 # data from environment object
line.data.EndAConnection = 'Anchored' # data form the line object

The line could be alse located by name with the following method. Although it could be done with the original method (line = model['Line1']), the new method is recommended to allow the functionality of auto-complete (data attribute)

line = model.findLineByName('Line1')

A list of all lines in the model may be retrieved and then select the first one by

lines = model.getAllLines()
line1 = lines[0]

Example 2 - Reduced simulation time for irregular wave

import NsgOrcFx as ofx

model = ofx.Model()

# set irregular wave
model.environment.data.WaveType = 'JONSWAP'
model.environment.data.WaveHs = 2.5
model.environment.data.WaveGamma = 2
model.environment.data.WaveTp = 8

# set reduced simulation duration with 200 seconds
model.SetReducedSimulationDuration(200)

# save data file to check the wave history
model.Save('reduced.dat')

# after executing this code, open the generated data file
# then open Environment -> Waves preview, and set duration of 200s 
# click in View profile and observe that the largest event (rise or fall)
# is in the midle of the sea elevation history

Screenshot of Wave preview (Environment -> Waves preview -> View profile) for a simulation of irregular wave with reduced duration based on the largest rise/fall occurence

Example 3 - Generate load cases

import NsgOrcFx

model = NsgOrcFx.Model()
model.CreateLine()

# list of wave direction, height, and periods to define the Load Cases (LCs)
directions = [0, 45, 90] 
heights = [1.5, 2.0, 3.0]
periods = [5, 7, 9]

# Folder to save the generated files (LCs)
outFolder = 'tmp'

# Regular waves
model.GenerateLoadCases('Dean stream', directions, heights, periods, outFolder)


In case of irregular wave:

model.GenerateLoadCases('JONSWAP', directions, heights, periods, outFolder)


To run irregular waves with reduced simulation time, based on the occurance of the largest rise or fall in the specified storm period.

model.GenerateLoadCases('JONSWAP', directions, heights, periods, outFolder, reducedIrregDuration=200)

Example 4 - Calculating modal analysis and getting the normalized modal shape

import NsgOrcFx

model = NsgOrcFx.Model()
model.CreateLine()

modes = model.CalculateModal()

# mode shape index (0 for the 1st)
modeIndex = 0

# mode frequency
freq = modes.getModeFrequency(modeIndex)

# if normalize = True, the displacements will be normalized, so the maximum total displacements is equal to the line diameter
[arcLengths, Ux, Uy, Uz] = modes.GlobalDispShape('Line1', modeIndex, True)
print('Frequency = ', freq, 'Hz')
print(arcLengths, Ux, Uy, Uz)

Example 5 - Defining fatigue analysis and getting the fatigue life calculated

import NsgOrcFx

simFile = r'tests\tmp\fatigue.sim'
ftgFile = r'tests\tmp\fatigue.ftg'

# First, it is necessary a model with simulation complete
model = NsgOrcFx.Model()
model.CreateLine()
model.RunSimulation()
model.Save(simFile) 

# The fatigue analysis is defined, including the S-N curve based on the DNV-RP-C203
analysis = NsgOrcFx.FatigueAnalysis()
analysis.data.AnalysisType = 'Rainflow'
analysis.data.LoadCaseCount = 1
analysis.addLoadCase(simFile)
analysis.addSNCurveByNameAndEnv('F3','seawater')
analysis.addAnalysisData()
analysis.Calculate()
analysis.Save(ftgFile)

# Result of fatigue life in each node
lifePerNode = analysis.getLifeList()
print(lifePerNode)

Example 6 - Generates RAO plots from vessel type data

import NsgOrcFx as ofx

model = ofx.Model()

# Create a 'Vessel Type' object with default data
model.CreateObject(ofx.ObjectType.VesselType)

# Create RAO plots (amplitude and phase) and save to the defined folder
model.SaveRAOplots(r'tests\tmptestfiles')

 plot generated with SaveRAOplots() method

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

nsgorcfx-1.0.21.tar.gz (35.1 kB view details)

Uploaded Source

Built Distribution

NsgOrcFx-1.0.21-py3-none-any.whl (39.5 kB view details)

Uploaded Python 3

File details

Details for the file nsgorcfx-1.0.21.tar.gz.

File metadata

  • Download URL: nsgorcfx-1.0.21.tar.gz
  • Upload date:
  • Size: 35.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.12.0

File hashes

Hashes for nsgorcfx-1.0.21.tar.gz
Algorithm Hash digest
SHA256 de83006b7ae040a568b81046bb80ea57e7f12f1ed0036bedce0e4b77a6502718
MD5 d9dd9dae4c810833ffc1395191e51ad3
BLAKE2b-256 a8a023155d0bce191756c98690d5221608482acdb67ca5f7621dd98853b8040e

See more details on using hashes here.

File details

Details for the file NsgOrcFx-1.0.21-py3-none-any.whl.

File metadata

  • Download URL: NsgOrcFx-1.0.21-py3-none-any.whl
  • Upload date:
  • Size: 39.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.12.0

File hashes

Hashes for NsgOrcFx-1.0.21-py3-none-any.whl
Algorithm Hash digest
SHA256 4965903182e05ce88f28dd4ec621836cf6abbb3d8c7db12a607b94d318108ead
MD5 f25b25d4e8eefc16b210e273237afb6c
BLAKE2b-256 507ffa46599e154fb31a210f2d26e8e0bc4215567913ae341358211a429b985e

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page