residual spikes in fluent for airfoil case
 

Hi,

Im running a test on an transonic airfoil at M=0.73 and Re=6.5E6 im using the density based solver in FLUENT and the spallart-allmares turlence model. I've been having a problem getting the solution to converge even with a low Courant number of 4 (which is low since using implicit solver) and low URF's around 0.2. The force monitors of the drag, lift and momentum do not converge before the equation's residuals begin to spike. The force monitors are what im most interested in since im looking at performance of the airfoil. I built the mesh in ICEM and noticed that the best results in fluent are gotten when the spacing of the first node off the wall is at y+=10 (limit S-A). Also when more nodes are added in the y-direction the case diverges in FLUENT. The y-velocity is begins spiking early durring the itterations and then the energy, x-velocity, and continuity join in later on. However the trubulance residual continues to converge. I was wondering what could be cousing these spikes. Heres a picture of the residuals plot starting at the 2000 iteration point when things start to go bad and thanks for all help

http://dl.dropbox.com/u/29394155/res.jpg