5 replies to this topic

[LostInTheWoods]

We're working on bringing in a piece of equipment with a capability for automation. We have a time and temperature kill step. With our previous equipment, operators were manually recording this on a batch record. This equipment will run a PLC program and ensure that our temperature exceeds the minimum and stays up for the entire duration of the kill step.

 

In my reasoning, the fact that the program completes successfully could be evidence of CCP completion. It would also be less prone to random error, but perhaps more prone to systematic error (i.e. coding errors) that may occur. I believe that the software must be validated that it's performing correctly. Has anyone ever undertaken this validation? How have you documented it?

 

 

[kingstudruler1]

If all it is doing is automatically adjusting something to regulate temperature, I dont think the software needs to be "validated".   You still need something that indicates that the temperature has reached and maintained the critical limit.   Such as the data from the plc, a chart recorder or other external recorder, manual records, etc.  

[jfrey123]

Your CCP is time and temp, but is temp measured in the product or in the machine?  In my mind, you would need to validate that the new machine running at your already specified time and temp is able to hit the internal temp of the product which you've previously validated.  But once you've ensured the machine cooks the same way your old process did, add on a review of the logged data (print and sign would be fine) and let her run.

[LostInTheWoods]

Your CCP is time and temp, but is temp measured in the product or in the machine?  In my mind, you would need to validate that the new machine running at your already specified time and temp is able to hit the internal temp of the product which you've previously validated.  But once you've ensured the machine cooks the same way your old process did, add on a review of the logged data (print and sign would be fine) and let her run.

The temp is the temp of the product (which is liquid). It's measured by RTD's in the liquid.

 

The program has logic to it. By design (written internally), it:

• Heats to X temperature
• When temp is achieved, starts a timer.
• When CCP time is achieved, stops a timer and begins cooling.
• IF temperature unexpectedly drops (heating failure), or cannot achieve the temperature, it aborts, alarms out, and doesn't allow itself to proceed to the next step

 

I'd like to get to the point where process completion is itself evidence of CCP's being achieved. Is any testing required to prove that the program performs as designed? That it errors when it should?

[nwilson]

Can you trend all temperatures and times recorded manually and compare results?  You could also perform manual checks to start to make sure everything is lining up accurately and the data you are getting from the automated system is withing the set parameters you require for the CCP.  

 

I agree with @jfrey123 on printing out or downloading the data and signing off on review.  

[Bo16]

IQ, OQ, PQ packages are usually available from the supplier, some provide it free of charge or they may even be able to do it for you. to have them do it is usually costly, but you are dealing with CCPs and having it done professionally may be worth it.  We usually include the validation in the initial pricing quote for equipment like this.