PLC5 Bumpless Transfer

[Alphite]

I didn't see an exact answer to what I was looking for, so here it goes.*

Right now a chilled water control valve is in PID auto the majority of the time (using a PLC5). When a button on the HMI is pressed, the valve should close until a heating cycle completes. At this point, it should revert back to PID control.

As it is, the valve closes correctly for the duration of the cycle, but when it switches back to auto, it's going to full open instantly (the current PID CV). What is the best way to implement a bumpless transfer, without having to receive the value from an HMI? I just want to hard-code the 100% closed value (4095). The main issue is that if the valve snaps back, it causes a rapid drain on the system and faults a chiller. I considered a second PID that's less reactive for the transfer, but bumpless seems like the correct, more elegant way?

Something else that was strange to me - with the current code, the CV of the PID is set to 4095 during the heating cycle (closed). However, our historian values show a moving CV, which gradually goes from 4095 to 0. How exactly is this happening, as the valve remains closed?

 

[Mickey]

See link...

At the end I say you may have to readjust the setpoint, this can be done slowly (ramping back to your desired setpoint)

http://www.plctalk.net/qanda/showpost.php?p=790354&postcount=8

 

[jamesau]

When a button on the HMI is pressed, the valve should close until a heating cycle completes. At this point, it should revert back to PID control.

Something else that was strange to me - with the current code, the CV of the PID is set to 4095 during the heating cycle (closed). However, our historian values show a moving CV, which gradually goes from 4095 to 0. How exactly is this happening, as the valve remains closed?

The PID's TIEBACK value should be set to 4095. When the HMI button is pressed, place the PID's MODE to Manual. When the heating cycle completes, place the MODE back to Auto.

As for the historian, you're either tracking a different variable or perhaps it is read from the PLC before it gets overwritten with your assignment.

If Ron B. visits this thread, he might have something more to add.

The above speaks to output tracking whereas Mickey is speaking to setpoint tracking; both may be used to effect 'bumpless transfer'.

Good luck and have fun.

 

[Ken Moore]

My first question is: How are you making the valve go closed via the HMI button? I suspect this is the issue.

If you have your PID setup with output tracking, and you place the loop in manual with a manual output value. When the PID is placed in Auto, it will start controlling from that output value.

 

[rupej]

Yeah, sounds like you're leaving the PID in auto when you close the valve, causing it to wind up to 100%. Putting the PID in manual and writing to the CV (same exact address used by the PID) to close it should automatically allow bumpless transfer when the PID is placed back into auto.

 

[Alphite]

Thank you everyone!

To answer some of the questions, there's basically one bit that's being set, and it controls whether or not the run executes (B3:0 - 13). I attached two screenshots.

The HMI button is switching on the bit for cold loop sanitization (B3:5 - 0).

Normally, there's hot water flowing in, and the heat exchanger cools to a setpoint. During a heat sanitization process, the heat exchanger is disabled for an hour.


As some of the comments mentioned, I think the PID is being left in auto the whole time, and the run in the WFI COLD program is overwriting it with 4095?

So I need to set the PID to manual, and set the CV to 4095 during this time? And then set the tieback to 4095 so that it comes out closed, and slowly ramps? I've seen some comments about different issues with using the manual mode, but I assume in this case it should be fine?