Emerson Exchange 365

• Not Answered

Requirement to cascade PIC & SIC Fan speed control loops

Hi Experts,

There is requirement to cascade PIC & SIC loop & here to find all the best possible solutions please find the details below.

In scrubber area we have

1.Scrubber vent header pressure PIC-112 which controls control valve PV-108. (Direct loop PID IO in PT value PID IO OUT PV-108) range -50 to 50 mbar

2. We have SIC-142, SIC 143 fan speed controllers (with Lead lag pump set up for FA-01 & FA-02 pumps) range 0 to 2925 RPM

Currently these two loops are working independently. SIC loops are configured as PID bypass on just passing speed set point.

The issue is Vent header pressure constantly deviating from SP of -10mbar.

Scrubber inlet pressure regulated by fan spillback PCV PV108. Once PCV PV-108 is fully open or fully close, scrubber fan speed SP must be manually adjusted, or else vent header pressure will start to deviate from SP.

any suggestion would be helpful to achieve cascade loop for this.

I am attaching graphics screenshot for ref. In trend you can notice PIC SP -10mbar & PV-108 is fully closed but pressure is increased to 12 mbar as fan is running.

Regards:

Rakesh

• Delta V PID Loop
• DeltaV #DCS
• cascade loop
• PID
• batch
• Scrubber PIC SIC Loop
• deltaV #mass flow rate #pid #dcs #how to convert

Rakesh Kirasur

• 26 Sep 2024 2:19 AM
• Cancel

8 Replies

• Lou Heavner

  

  • 26 Sep 2024 1:31 PM

  You could consider a valve position controller. You take the output of the PIC loop (you can use the valve position readback if you have it) and make it the PV of a new PID loop named VPIC-xxx. The output of the loop would be a remote setpoint for the SIC. Often these VPICs are used to optimize the process and are configured as I-Only controllers with "relaxed" tuning. In this case, it looks like the valve caan close pretty quickly, so you may have to tune the VPIC fairly agressively. Still, using lambda methodology, tune the VPIC with a lambda value at least 3 times greater than the PIC to avoid loop interaction.

  • Cancel
• Rakesh Kirasur

  

  • 27 Sep 2024 2:55 AM

  In reply to Lou Heavner:

  Hi dear Lou Heavner,
  
  Thank you very much for suggestion we definitely implement this strategy by discussing with process team to understand clear requirements. I will let you know about the result once it is implemented & tested.
  
  Could you please elaborate this "output of PIC loop & make it PV of New PID"

  Also copying loop schematic/block diagram to understand it more clearly could please check & let me know if my understanding is correct.

  

  
  Regards:
  Rakesh

  • Cancel
• Lou Heavner

  

  • 27 Sep 2024 3:59 PM

  In reply to Rakesh Kirasur:

  I would show it like this:

  

  The Diamond T could be a switch or a splitter block, depending on how you want it to behave.  But you will want to set up the backcal functionality.

  • Cancel
• Lou Heavner

  

  • 27 Sep 2024 5:28 PM

  In reply to Lou Heavner:

  Actually, there is a mistake on the sketch. The valve position is the PV of the VPIC loop and not the SP. Sorry for that error. I will try to fix it later, but don't have time right now.

  • Cancel
• Rakesh Kirasur

  

  • 29 Sep 2024 8:13 AM

  In reply to Lou Heavner:

  Thank you, dear Lou Heavner, I will configure this loop on dev & test it to see how it will behave.

  • Cancel
• Lou Heavner

  

  • 30 Sep 2024 2:38 PM

  In reply to Lou Heavner:

  Here is a better drawing:

  

  The speed controller I/O is not shown for clarity.  This was copied from Control Studio.  This is all one module, but you would probably not put each of these PID loops in in their own module.  I am showing a splitter bock with its backcals.  You could use a switch, but you would need to switch the backcals as well.  The SP for VPIC-112 could be set to hold the valve nearly open or nearly closed or halfway depending on operational objectives.

  • Cancel
• Rakesh Kirasur

  

  • 1 Oct 2024 5:04 AM

  In reply to Lou Heavner:

  Hello
  
  Thanks a lot, I did same configuration along with splitter I used scalar block to convert 0 to 100% to RPM because my input to SIC's is in RPM o to 2925.
  Along with this solution we are considering one more solution. We will propose both solutions to client & will do test on both loops.
  
  Regards:
  Rakesh

  • Cancel
• Lou Heavner

  

  • 1 Oct 2024 1:30 PM

  In reply to Rakesh Kirasur:

  You might be able to avoid the scaler block by setting the output units on VPIC-112 to 0 - 2925 RPM. If you do that, make sure you also set the ARW high limit to 2925.

  • Cancel