Test Cases should contain:
- A word description. This is typically a doc or a pdf file. It could also be a posting in this forum.
- Schematics, one-line diagrams, etc.
- Numerical input data provided in a simple form
- Computational results appropriate for comparing with output from other software
It should be explicitly and clearly stated what features of power system analysis tools are being tested by this test case. For example, the Comprehensive Test Feeder claims that “This tests the capability of a program to represent a wide variety of components in one circuit.” (http://ewh.ieee.org/soc/pes/dsacom/testfeeders/Comprehensive.htm).
Test cases may be defined for any of the following reasons:
- Validate tools for a common type of analysis
- Verify tools for a new type of analysis
- Provide stress tests for solution speed, ability to solve nonlinear problems, etc.
- Provide basic test cases for researchers to validate their work
- Provide a resource for students studying power system analysis
Input Data for Test Cases
Input data for the test cases should be provided in a common, simple form that most engineers should understand and does not require the purchase of software with costly licenses. Recommended forms include:
- Simple text files
- Comma-separated Variable (CSV) files
- Commonly-used spreadsheets
Do not require users of this data to key in voluminous amounts of data. Provide it in a computer-readable form. The simple text file format chosen should be easily parsed by computer.
Published results should be posted in an appropriate format (Tables, text, graphics) with sufficient accuracy to confirm the match between the published results and the results produced by the software under test.
Results should be confirmed by at least two members of the Task Force or Working Group providing the test case.
Provisional test case results may be posted for comment if only one member is able to execute the test case at the time it is posted. Once another entity has confirmed the results with a separate tool, the results can be designated as official published results.