compressibleTwoPhaseEulerFoam: order of the alpha-setup makes a difference
 

Hello everybody.
I try to simulate an air stream with waterdroplets into an air-filled volume. Therefore I edited the bubbleColumn and the fluidizedBed tutorials to my needs, but the result was not satisfying. So I made a few testcases and found a strange behaviour.
I set up a case with 2 identical phases. They have the same thermophysicalProperties and the same transportProperties. The velocity is the same as well. The despersedPhase is set to both. So I varied the alpha. At first alphaair was 1 in the internalField and 0 in the inlet. In the second case it was just the other way round, alphaair was 0 in the internalField and 1 in the inlet.
I expected to see the same result, because the two phases have exactly the same properties. But to my surprise alpha and u behaved differently. Not completely different, but significant.
I digged around in the sourcecode but couldn't find any explanation.
Could anyone give me a hint why the two cases react differently?
If you want to have a look at the cases, here the zipped cases (in zip or tar format):
https://dl.dropboxusercontent.com/u/...aVariation.tar
https://dl.dropboxusercontent.com/u/...aVariation.zip
Of course you have to compare alphaair of one case with alphawater with the other case.
I would appreciate any hint
Benedikt