Reply To: Using PIDE for Fermenter Temperature Control

WhiskyjimJim Manley
Karma: 247
Rank: Jedi

Fred –

Thanks for the response.  See my answers below.

>>It would be difficult to say without seeing the response curves of your process. My gut is telling me that
>>the gains values you have may be part of the issue.

I’m working on getting a response curve now.

>>A proportional of 200 seems excessive, can I ask how did you arrive at these gain values?

I inherited the system from the company that did the initial coding and deployment.  When I asked them about how they got the gain values the response was “keep doubling the value until you see a response.”  It didn’t strike me as a sound engineering approach…

>>I would be curious to see in a trend what exactly is going on, perhaps the error on startup is
>>insurmountable for the controller to overcome due to the excessive proportional action,
>>and as you say, once the operator manually brings the temperature to within setpoint
>>“bumpless transfer” the controller can maintain.

I began to suspect this after reading your article “How to Implement a ControlLogix PID Controller,” which is what brought me to the forums.

>>Again with these gains that you have, I would like to see how it is maintaining.

As for maintaining, right now the system can’t maintain the set point.  A little background might be helpful here.  Forgive me if I cover material you may be familiar with.

The fermentation process generates heat.  It takes place is 16K gallon stainless steel vessels that are essentially convection cooled using a cooling coil through which is passed 44°F water.  The ingredients in the vessel are not mechanically agitated.  The only agitation of the ingredients is caused by the CO2 created as a byproduct of the fermentation.  The CO2 rises to the top of the tank and give the impression that the ingredients in the tank are simmering/boiling.  The room the fermentation vessels are in is NOT temperature controlled.   So, when it’s 100°F outside, the room temperature can be just as high.  (Got to love summer time in Texas.)

The current temperature control process does not start cooling the fermenter until the set point is reached.  By that time, the fermentation is in full swing.  For a 75°F set point, the system is never able to pull back the temperature to the set point and maintain it there.   This is another issue that I need to deal with once I can get the cooling loop to start properly.

>>Are there oscillations? Is there a significant phase shift between CV and PV?
>>Are you seeing any noise or amplification of noise?

No idea on these since I don’t have a trend on them yet.

>>What about deadtime and lagtime in your process?

I don’t have any definitive numbers but intuitively, these are going to be large.  I am effectively trying to maintain the temperature of a 16K gallon stainless container of a viscous mass that, generates its own heat and draws heat from its environment, without the benefit of being able to agitate the mass in order to distribute the heat.

>>Have you tried an Autotune?

No.  I’m not familiar enough with it to be comfortable trying it.  The PIDE is not set up to use Autotune at this point.  I can’t really do anything that would cause the PLC to reset during the work day since it would disrupt distillation.  That makes the distillers cranky.