Wednesday 25 February 2009

There are Simultaneous Equipment and Recipe procedural elements

This was recently stated on the Part 5 Blog
"In some cases it may be sufficient to allow the recipe phase to directly control the equipment modules. In other systems where complex equipment modules exist, it may be necessary to implement some level of state model based procedural control in the equipment between the recipe phase and equipment modules in order to better deal with the underlying equipment complexity. Again, it’s an implementation decision left to the developer and thus does not restrict creative efforts."

I don't agree. The Equipment Phase must exist 'in' the equipment. Allowing "the recipe phase to directly control the equipment modules" is not a concept I can subscribe to.

I am assuming in the following a phase level interface for simplicity, but it could be Operations or higher.

The Recipe Phase is that procedural control that speaks to the Equipment Phase, it may just be one step in an operation, which interfaces with the Equipment Phase but does not actually control the equipment module (or Unit) - that is the job of the Equipment Phase. But in my view (and others) both the recipe phase and it's corresponding Equipment phase exist

To further illustrate, suppose we have a PC based batch manager executing the Recipe Procedure (so the PC is the recipe controller) and it speaks to the equipment controller (eg PLC or DCS controller) then the Operation and its steps and Transitions are coded PC, whilst the steps and transitions in the Equipment Phase are coded in the Controller.
These corresponding phases speak to each other via a Phase Logic interface and some data transfer such as recipe parameters

Now, it may be that there are applications of PC Batch managers where some of the equipment control steps and transitions are coded in the PC. That just means that the recipe and equipment control are not nicely separated in the implementation, but they still both exist.
And the S88 standards are not supposed to be about actual implementations, just the concepts and models.

No comments: