Iterations resulting in !NUM results can occur. This document provides hints to minimize this behavior. If the K values all go to unity or if you get some negative numbers that cause the iterations to crash, then the problem is usually that the initialization was not done properly.
If you are having problems with the iterations it is due to:
1) wrong procedure - unlikely if you have followed the procedure carefully.
2) wrong formula somewhere - unlikely since the formulas are simple to program.
3) wrong copy and paste steps - be careful to follow the instructions about pasting VALUES. Other than that the procedure is pretty straightforward.
4) POOR INITIALIZATION - This is the most common problem, and the easiest to fix. Although it is blame one of the the first three possible problems, initialization is usually the problem. After the first coexisting compositions have been found for particular system, you rarely have to reinitialize the compositions in the activity coefficient table, but you almost always have to reintialize the Kold values when the compositions change.
INITIALIZATION. This discussion is constructed for a system where components 1 and 2 have limited miscibility. After you set up a system by changing the components, be sure to initialize Excel by starting with one phase nearly pure component 1, the OTHER phase nearly pure component 2, and add a trace (<1 mole%) of component 3. ONE MORE IMPORTANT POINT. After initializaiton the compostions, then update the Kold values by pasting VALUES of the Knew values. If you try to start off without updating the Kold values the very first step can be really way off, and the iterations may never recover.