I'm currently doing two simulations. One of the Residual diverges in the beginning for around 100 iterations. Though it converges eventually till 0.001. Is it fine and acceptable? Another one diverges abit in the beginning and have zick zacks when the residual stays 0.001. I've checked the Cell Quality, whether I#m using the right Condition, checked the Wall Y+ and it's all fine. Are both the Simulations not acceptable? Are those Simulations, which Residuals have zick zacks of diverge few times but eventually converge acceptable?

Is this your own code or a commercial one? You really need to tell people on here how the residual is defined before we can comment on the convergence.

Have you tried getting the solution on a finer mesh? Is that solution the same as the one on the coarse mesh?

If it is your code, try to use double precision variables. The difference in the residuals is great. I had a similar problem in 2D and I was using single precision. The normalised residuals didn't get lower values than 0.001. With double precision the residuals reached 1.E-8 and in fewer iterations.

I also confirm the role of double precision, but what about that when in one radial direction the residuals are 100x higher (cylindrical coordinates)?

Is it in the radial direction, when you start calculating with your solver the variables? Because of the periodicity same errors occur at angle=0 and angle=2*PI, where these points are actually identical. These errors affect the "neighborhood" and sometimes huge residuals occur. 100x higher though is too large....

I am basically using "Star ccm+", so actuallz I've no idea what "code" do you all mean. So I guess it's a commercial one. My mesh quality is between 0.1 to 1, which should be fine cause it's being said that when the meshes are consider bad when the quality is lower than 0.001.

Well, if anyone knows about Star ccm+ , feel free to tell me how could I determine what kind of condition I have to choose or what value I could change. From what I see, cmu, c2e, under relaxation factor , turbulent viscosity ratio, turbulence intensity and such stuffs are pretty much what i could varify.

First of all, the pipe I've been simulating is 10cm diameter, 1,6meter length, normal straight pipe, K-epsilon 2layer, standard turbulent intensity and viscosity ratio. My Wall Y+ is between 12-28. Cell Quality 0.1-1. Inlet velocity 0.3m/s. Did i miss out something? It just oscilates when after it converges. Why?

By the way, it's water, and steady in time, stable condition.