Whereas all previous presentations on the Heiss Dampf Reaktor hydrogen distribution experiments E11, concerning data versus code predictions, concentrated on the blind posttest efforts, this presentation focuses on the results of the comparisons with parametric, best-estimate, open posttest predictions for experiments E11.2 and E11.4 with the containment analysis computer codes RALOC, WAVCO, CONTAIN, MELCOR, and GOTHIC.

The results of these comparisons show the following after correcting a number of deficient input parameters previously supplied by the Kernforschungszentrum Karlsruhe/Heiss Dampf Reaktor Project as specifications):

E11.4:

1. Standard lumped-parameter codes are able to predict H2 mixing and distribution phenomena when H2 is injected into a well-mixed atmosphere in lower zones of the containment with excellent agreement in most of the important quantities.

2. A few discrepancies remain, dependent on the codes' modeling methodologies and the impact of incorrect specifications.

E11.2:

1. Accounting for the corrections substantially improves the agreements compared to the blind posttest predictions.

2. However, concerning the predictions of the thermal stratification pattern and the H2 distribution, more or less large discrepancies still remain.

3. Parametric changes of input parameters lead to improvement of agreement in some quantities but at the same time worsen others.

4. "Innovative" concepts of changing certain input parameters beyond current practice improve the quality of the predicted H2 concentrations.