the difference between simulation and hotbox measured U-value

[Jose Li]

Hello all,

Last month one of my customer have tested a 80 series high insulation window in IFT  Rosenheim, Germany.

The results is 1.5 under china environment condition. And the simulated value is 1.6 using Therm/WIndow 6.3.

So I wonder is there any comparison data between software simulated and hotbox measured value?

How accuracy the simulated value will be?

Thanks  

[D. Charlie Curcija/LBNL]

THERM and WINDOW simulation models are generally accurate to within 1%. However, considering that we normally model "drawings" and not actual products (meaning that drawing is usually not a perfect representation of the actual product), it is accepted that simulation and test results should be accurate to within 10%. NFRC has CPD no has several hundred thousand validations between simulation and testing (yes, this is not a typo) and all products have validated to within 10%. In few cases, there were non-validation issues, but in all of cases it was uncovered that problems were due to testing issues, which were later corrected. This said, if you are trying to match test results done by European laboratory with EN simulation models (ISO/EN 10077), most likely the results will not validate due to gross simplifications in glazing model and to lesser extent in frame models used in that standard. What you should do is model in THERM and WINDOW using ISO 15099 model and apply CEN boundary conditions. That should validate to within 10%.

Charlie