- incorporate the minus sign of the darcy relation only at the velocity calculation. I.e take it away from kGradpNormal.
Therefore, >< for upstream decision has to be switched: in boxdarcyfluxvariables and in the smooth upwinding of mpnclocalresidualmass
- correct some spelling
- split the main body in two function: calculate gradients and calculate velocities. 
- calculate volumeFlux from velocity instead of calculating it from scratch: less error prone. 

boxelementvolumvariables: 
set the evaluation point to zero. This is only used for semi-smooth methods. If this is not set to zero, strange segfaults may occur because the evaluation point is *not* set

*Introducing* boxforcheimerfluxvariables: 
It is now possible to calculate velocities according to the Forchheimer relation in stead of exclusively the Darcy relation. This flux calculation can be used without changing the actual residual calculation: just derive the fluxvariables from boxforchheimerfluxvariables (or use them directly as your fluxvariables)

The only other necessary ingredient for using the Forchheimer relation is putting a forchheimer coefficient into your spatialparameters class (see the forchheimer tests in test/boxmodels/mpnc)

mpncfluxvariables: 
derive from Basefluxvariables Property in stead of DarcyFluxVariables. This way the velocity-potentialGradient relation can be chosen in the problem. 

mpncproperties: 
introduce new Property for Darcy / Forchheimer relation switch

mpncfluxvariables:
derive from  a property allowing for the setting of velocity relation (Darcy/Forchheimer) in the problem

mpncvolumevariables, mpnclocalresidualmass, mpncwritercommon:
correct typo / naming

simpleh2o: 
added heatcapacity function, 
removed ";"

text/boxmodels/mpnc: 
added two tests for the forchheimer relation. One test for single phase flow and one test for two-phase flow.

git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@8864 2fb0f335-1f38-0410-981e-8018bf24f1b0