Hello, I am a database developer/Mech. Engineer and amateur mathemagician and my problem is this;

I have a simple convergence loop that is used in a heat transfer problem of radial geometry to pedict the interface and outer surface temperatures in an insulation system with a varying thermal conductivity. I am following the ASTM C-680 procedure and this method works fine for plane insulation systems but blows up on radial insulation systems. I think this problem is due to the log function that is required in the heat transfer equations of a radial system.

Basically the way this works is as follows;

The program makes a geuss at what the outer surface and interface temperatures are in a system of 3 layers. (2 interface temperatures and the outer surface temp.) This geuss is based on the hot face temperature. The program then executes in a do...until loop which simply compares the previously calculated temperatures with the newely calculated temperatures until their difference is less than 0.01. In the radial system calculations the intial guess temperature runs through the loop and then in the next iteration a new temperature is calculated based on the previous temperature and for some reason on this second iteration the temperatures go up into the thousands of degrees when it should be at most a couple of hundered degrees.

Does anyone know of a way to "damp" this problem and achieve convergence?

In C++ I don't have this problem but I have to use Access on this project.

Thank you very much.