In reply to Len Laskowski:

I received this great response from David Sheppard that I wanted to add for everyone's benefit:

The scenario describe as I see it is a large matrix with 32 steps and 60 valves. It is similar to a 45 step project we did for the Cokers at BP Whiting. There were over 30 valve outputs and they spanned multiple SLSs.

 One approach is a very large matrix.  It has one module and it drives all the decisions.

The module becomes very large and has so many inputs (to decide what step to move to)

And many outputs, that it becomes very large.  It is also difficult to simulate / make online changes to test the large matrix.

This would look like:

The approach that was used at Coker broke this down.

There was the sequence module that simply took the inputs and decided what step.

There was NO valve information in that module.

Separately, each valve would have its information in the module that drove the specific valve.

If step = 2 through 5 or step = 7, then Open the valve, else close the valve.

This was a simple way to move the logic to the valve itself.  It was easy to force items, make changes, etc.