Challenge Overview
Welcome to the OpenWarp Test Comparison Justification contest. As part of this contest, we would like you to check out the comparisons between the outputs that we carried out between our app and its origin and fix mismatches that we found in the comparison or justify the results by giving a reasonable explanation for the mismatch.
The contest will be reviewed by our client. Apart from carrying out bug fixes, the submission is also expected to contain reasoning / explanation for the behaviour of OpenWarp and thus, our client will verify if the reasoning / explanation provided is reasonable or not. Please read the contest details below for more information
Project Overview
OpenWARP allows Wave Energy Conversion (WEC) companies to create a software model to simulate the movement of waves and determine the amount of power that the WEC device would output in the modelled environment. It consists of a web application that can mesh and simulate a body of an offshore structure.
Contest Details
Our OpenWarp application can be thought of as a fork of the Nemoh application - an app that computes the impact of waves on offshore structures. Specifically, the OpenWARP fork adds a GUI, meshing capability, and simulation capability improvements to Nemoh.
We recently ran three test cases both in Nemoh and OpenWarp using the same inputs. We then created a comparison script to compare the outputs which we will share with you in the contest forums. The expectation was that the output values from OpenWarp should not differ by more than 0.5% of the output value from Nemoh. However, this has not been the case for all output values. We would like you to go through the test cases, understand them, go through the output values which differ by 0.5% and:
- Determine if the difference is due to a bug, in which case, you need to fix the bug
- Determine if the difference is due to a feature of OpenWarp and thus provide a suitable explanation / reasoning for the same.
The files that you will need to compare are:
/Results/RadiationCoefficients.tec: This file contains the added mass and damping forces for the radiation problems. A = added mass, b = radiation damping. E.g. 'A 1 1' is the added mass coefficient of body 1 in direction 1., 'A 1 2' is the added mass coefficient of body 1 in direction 1. In the headers 'I=40' refers to and correcponds with the frequencies specified in the Nemoh.cal or the OW GUI. Any 'beta' terms refer to wave angle as entered in the input files.
/Results/DiffractionForce.tec: This file contains the diffraction force for the diffraction problems. Diffraction forces follow a similar convention to the information above.
/Results/ExcitationForce.tec: This file contains the excitation force for the diffraction problems. Diffraction forces follow a similar convention to the information above.
/Results/IRF.tec: This file contains the infinite frequency added mass and the impulse response function for the radiation force. Diffraction forces follow a similar convention to the information above. Of particular interest in this file are the infinite frequency added mass.
You can find more details on the Nemoh website and in the OW theory documentation - http://lheea.ec-nantes.fr/doku.php/emo/nemoh/start
Files that will be shared
- The Test Case results in both Nemoh and OW. There are three test cases and each contains the results from the test cases with the files mentioned above. Additionally, each test case (in the Nemoh folder) contains the configurations or input values used for the test case in the Nemoh.cal file. The same input was used for executing the test cases in OpenWarp too
- The comparison script used to determine the different in the output values. You should only concern yourselves with differences that exceed 0.5%. That is, values in OpenWarp results that exceed the corresponding Nemoh results by more than 0.5%. The comparisons carried out use the above files. That is, you need not compare the files manually. Instead use the script shared.
Final Submission Guidelines
For instances where the difference was due to a bug, fix the bug and submit a git patch of your changes.
For instances where there is a scientific explanation for the difference (not a bug but a feature), please submit a word or text document with the explanations.
Also, Submit plots/comparisons of the added mass, radiation, damping, and excitation force for the same degrees of freedom that are shown on the Nemoh website. These are the dominant degrees of freedom for the test cases. This could be automated in the comparison script.