Validation

Validation is the step right after parameterization/calibration (the task to obtain parameters for a specific model from experiments).

The purpose of validation is to use the model and parameterization obtained for a set of exemplary cases (e.g. lab experiments) and apply the model to a new test case. The predicted results are then compared to the measured results for this new case. Hereby, it is important for the validation, that the measured data has not been used in the parameterization/calibration phase.

Delphin was validated several times, with focus on different aspects of the software:

Numerical Validation

• HAMSTAD Benchmarks 1 bis 5
  (transient heat and moisture transport problems)
• EN 15026:2007
  (steady-state heat and moisture transport)
• EN 10211:2007
  (steady-state thermal-bridge problems)

The most difficult HAMSTAD Benchmark 4 is included in the installation package as example project.

Validation of transport model

Test houses

• Talstrasse (Dresden)
• Herrenschießhaus (Nürnberg)
• Gemeinschaftshaus (Nürnberg-Langwasser)
• Umgebindehaus (Ebersbach)
• Fachwerkhaus (Niedersachsen/Edemissen)
• Bahnmeistergasse (Senftenberg)
• Handwerk 15 (Görlitz)

Dissertations

• Gregor Scheffler: Validation of hygrothermal material modelling under consideration of the hysteresis of moisture storage

Further, Delphin was and is used by several research institutes, and it is continuously tested and adapted for different application cases.

Benchmarks

Benchmarks are comparisons of different tools or methods with the aim of comparing/evaluating the individual strengths and capabilities for specific tasks. Hereby the tools and methods are expected to produce exactly the same results.

Typical aspects of benchmarks are:

• Performance
  (or how quickly are the desired results obtained)
  
  This type of benchmark is typically used for computational intensive simulation programs. For HAM models, multi-year simulations for 1D constructions, and 2D simulations in general are computational intensive.

• Usability
  (or how much effort is needed to create input files for simulation programs and obtain desired output)
  
  This is partially a subjective evaluation and depends on the individual skill on using a particular tool, but differences in order of magnitudes (10 minutes versus 2 hours) are still indicators for this type of benchmark.

Delphin was compared with other HAM tools in several benchmarks and (for example, using the HAMSTAD Benchmarks), and belongs to the fastest and user friendliest HAM simulation tools worldwide.