Benchmarks

To ensure credibility as well as reproducibility for the software, several measures are taken. The most important information is listed below:

Model Validation

TESPy has been used to model several research and engineering applications. In the paper on integration of generic exergy analysis in TESPy [9] three models have been built from literature sources: A solar thermal power plant, a supercritical CO2 Brayton cycle as well as a refrigeration machine using air as working fluid.

For the solar thermal power plant we have created a full model of the plant using a standard industry software in parallel. The comparison showed identical results. For the other two applications we have compared the results of the TESPy model with the data published in the respective research paper and found very well matching results. Differences can be explained by different implementations of the fluid property back end.

Finally, in the extension of the exergy analysis to chemical exergy [10] we have also compared results of the CGAM process [11] modeled in TESPy with a full model using industry software and with the data provided from literature as well [12].

The code for the full models is accessible open source on GitHub:

• So called “Solar Energy Generating System”

• Supercritical CO2 power cycle

• Air refrigeration machine

• CGAM process

Unit testing

On top of the full model validation, the software includes full unit testing. Here, single features of the software are tested by comparing the result the software provides with the result we would expect when manually modeling that feature. For example, we set up a turbine and check, whether the isentropic efficiency specification in the TESPy component matches the results that is expected doing the same process manually. This is done for all modules of the software.

Continuous Integration

TESPy has a Continuous Integration pipeline. The unit tests and the full model tests are automatically run whenever changes to the source code of the software are made. By this we can ensure, that changes in the code do not break with existing features or invalidate results of the existing models.

For more information on how to run the tests please see the how to develop section.