Turbulent shear stress
 

i simulated turbulent flow in a channel, i created a custom field functions and after 4 flow through time i averaged and plot turbulent stress u'u', v'v' and u'v,

-u'u' =(2 / density) * viscosity-turb * (dx-velocity-dx) - (2 / 3) * turb-kinetic-energy

-v'v' = (2 / density) * viscosity-turb * (dy-velocity-dy) - (2 / 3) * turb-kinetic-energy

-u'v' = (1 / density) * viscosity-turb * ((dx-velocity-dy) + (dy-velocity-dx))

i used kw-sst model

but i am getting large error as compare to experiment result , please suggest me where is mistake.

thank you

What variable or parameter are you comparing with experiment

I am not satisfied with the velocity stress, since k-w k-epsilon, or SA or kw std or keps (standard) all are designed to model Reynolds stress, which otherwise are computed using RSM model. I think you will not get something right with intentionally computing u'v' yourself. and also if your are using custom field function , first you will need to compute kinetic energy as well.

how can i compute u'v', is there any other method to calculate u'v', u'u' and v'v'.
please give me some detail

i want to compare with experiment result of average turbulent stress u'u', v'v' and u'v'

Try to use RSM model, if your computational power allows, otherwise try to use the custom FF of Kinetic energy then compute u'v', but I do not think this will give you some answers with k-w or k-epsilon , just try to do it. But first try to use RSM

actually in fluent there is kinetic energy as a variable is available can i use that in place of custom FF for kinetic energy? and if not then can you please tell me how can i create kinetic energy custom field function.

thank you

I think you should be able to use TKE=0.5*(mean(ux'ux')+mean(uy'uy')+mean(uz'uz'))