Low Reynolds k-e model

Frank

I am using low Reynolds k-e models (Abe,Kondoh and Nagano, 1994, International Journal of Heat and Mass Transfer, Vol.37, No.1, pp. 139-151). However, it predicts 10-15 percent higher than the theoritical values (friction factor using Blasius equation)for the straight squart ducts. Does anybody have the experience about this: the model itself or my program? However, my program predicts well for laminar flow, even using this low Reynolds k-e version. Looking for help from you.