Ple4Win is a finite element program for the simulation of the behaviour of (partly) buried pipelines under given conditions. The results are obtained using computer calculations, mathematical models, empirical data, approximations and other methods. To ensure that the results generated this way are of any use they have to be compared with known, real-world situations. This process is known as Verification.
Testing is also done during and after development, it is performed to eliminate errors (“bugs”) in the program’s code. Testing focuses on specific issues and on individual problems. Note that verifying is not testing, although there is a certain overlap. Verification sometimes exposes unwanted side effect of changes that in itself have been tested correctly (“collateral damage”).
How to verify
Verification is basically done by comparing results of calculations with pre-existing results. Each version of the Ple4Win program is verified prior to release, as is each patch of a version that might be necessary to release.
There are two different verification situations:
An initial verification (e.g. the first version of the program or new functionality) can only be a comparison of the results with external results, such as measurements or analytical modelling. Note that initial verification can also occur if the program/functionality already exists but new external results/measurements become available.
All subsequent verifications (e.g. new program versions or updated functionality) are performed by comparing the results to the results of the previous version. This way a verification chain is established, results can be tracked back all the way to the initial verification. Note that the comparison can be automated, however, the interpretation and weighing of the differences found has to be done manually.
What to verify
Globally, the Ple4Win verification can be subdivided into six categories:
- Analytical models
- Pipeline codes
- Experimental test data
- Thematically subdivided previous version results
- Functional requirements
- Verification of the educational version
Categories 1. to 3. are only used for initial verification, once results exist these become part of category 4. and future versions use those. Verifications using these categories have to be done by hand.
Category 4. is used for subsequent verification. These verifications are largely automated and form the bulk of the verification effort. More than 300 projects/situations are verified per release.
Category 5. is used in both initial and subsequent types of verification, it can be semi-automated.
Category 6. is a special case, it is a subset of categories 1. , 3., 4. and 5. tailored to the educational (free to use) version of Ple4Win.
1. Analytical models
Analytical models are situations where results can be obtained using known mathematical formulae. These typically are manual calculations in the form of plain text calculations or Excel spreadsheets. Most of these models cover only a specific part of the program’s functionality; some of these topics are described below.
The analytical model is recreated in a Ple4Win project database and the calculation results are compared with those of the manual calculations.
Analytical models can be further subdivided, some important sub-categories are:
1a. Basic physics
This topic consists of simple configurations up to somewhat more complex calculations which are based on well-established standard physics. Examples are basic supported beams, cantilever beams or linear elastically supported beams.
Although tees are also mentioned in pipeline codes, analytical models are available that allow a much wider range of tee configurations to be checked. The focus lies on the flexibility characteristic and the stress intensification factors. In total 6 tee types and over 50 different tee configurations are considered, from simple unreinforced tees to branch welded-on fittings.
1c. Primary elements
Primary elements are elements in a free pipeline span above or below ground level at supports in case only force-induced loadings are applied. The determination of whether an element is considered to be primary can be done analytical. In several situations where primary elements occur the number and position of these elements are compared.
1d. Mitre bends
For mitre bends (segmented bends), both single-kink and with multiple kinks, analytical models are available for comparison. Note that for the mitre bends experimental test data is also available, see here.
1e. Stress – Strain relation
Depending on the chosen type of stress – strain relation (diagram) a simple check can be performed to determine whether the calculated stress and strain for an element is positioned correctly on the given curve.
2. Pipeline codes
Most pipeline codes such as ASME, ANSI, PC ’72 and NEN 3650 prescribe certain calculation methods for pipeline calculations. These are usually based on analytical models, experimental data and proven best practices. Pipeline codes cover general pipeline calculations, but also specific situations such as implosion or axial buckling strain checks. The soil model wizard used to convert soil profiles into soil parameters is also based on pipeline codes.
3. Experimental test data
3a. TNO pipeline tests on straight pipes and smooth bends
The Dutch TNO organisation performed a number of tests with full-scale pipe configurations, both with and without (simulated) soil, and both with elastic and plastic deformations. With different loads the displacements, ovalisations, strains and much more were measured and collected in a number of reports. TNO also developed some calculation methods for the tests that were performed.
The TNO tests have been recreated as Ple4Win projects and the results of those calculations have been compared to both the measurement results and the calculation results from the TNO models.
3b. Mitre bends
In Great Britain Jones, Kitching and Bond performed a number of tests on full-scale mitre bends, both single-kink and with multiple kinks, with a total bend angle of either 90° or 180°. Only bending and/or internal pressure loads were applied, all pipes were without soil. Mainly displacements, ovalisations and strains were reported. The authors also developed a non-disclosed mathematical model for mitre bends of which the results were published.
These Jones, Kitching and Bond tests have also been recreated as Ple4Win projects and the results of those calculations have been compared to both the measurement results and the model results given in the publications of the two authors.
4. Thematically subdivided previous version results
This category contains Ple4Win projects of the latest released version of the program. As described above these projects are the references against which the next version will be checked.
For clarity, the projects have been grouped in “themes”, such as:
- Based on analytical models,
- Based on pipeline codes prescribed calculation methodology,
- Based on measurement results,
- Using material non-linear behaviour,
- Focussing on Steel-in-Steel,
- Having (table) supports,
- Considering different soil curves (reaction vs. displacement),
- and more…
5. Functional requirements
The purpose of verifications of this category is to check functional results (not the calculation results) of specific user actions. The program’s reaction is compared to a detailed description of what should be happening in specified situations. This verification ensures that actions such as opening or saving a project file function correctly and as expected; data listing, visualisation and export are smooth and intuitive; warnings and errors are correct and understandable; and more.
Note that these verifications change more often than those in the other categories, as the user interface and the program’s behaviour are constantly changing to a lesser or greater degree due to new requirements and trends with computers and operating systems.
6. Verification of the educational version
This category contains a limited subset of the thematical and functional verification mentioned above. The focus lies on the check of the differences between the normal “professional” version and free “educational” version of Ple4Win. Most verifications are about the limitations of the educational version; however, some verification of the calculation results is also performed.