Temperature control on desuperheater 1

[marshalls]

Hi all, first time posting here. Not sure if this is the correct forum for this question, but there's no controls engineering forum. I work in EPC, and am working on a functional description for a desuperheater control loop. The way the system works is as follows:
Superheated steam passes through a letdown station (PCVs) and is dropped to 120 psig. It then passes through a desuperheater, where superheat is removed by condensate. Pressure is measured at the outlet of the DS. The condensate flow is controlled by a cascaded temperature and flow loop, with the temperature measured at the outlet of the DS and the flow measured upstream of the condensate flow valve.
Currently we show the FIC as being a feedforward controller. I don't have much experience with feedforward control, but to me this just looks like a cascaded loop. So my question is, does there appear to be any feedforward control in the system I described?
Thanks in advance!

 

[georgeverghese]

A feedforward FIC requires an algorithm to be set up in the DCS to be the remote SP for the condensate FIC. In this case, it would be simple, since condensate flow is directly and linearly proportional to HP steam flow. A feedforward arrangement would enable the condensate FIC to adjust condensate flow before the downstream TT on the sat steam line picks up the temp signal. But you need another FT on the superheated steam line to feed the algorithm to make it work. If you have no plans for this FT, then agree this loop can only be a cascade loop.

 

[marshalls]

Ok, that makes sense. The feedforward control in the scenario you described would come from the superheated steam flow rate, and there would be a combination of feedforward/feedback control on the condensate flow. I think our P&ID just needs updated. We actually do have a flow meter in the superheated steam line, but we don't show it having anything to do with this loop. Thanks for the response!

 

[crshears]

Welcome to the fora [foray?], marshalls!

For future reference, you can find the Control Systems Engineering forum here: [/

https://www.eng-tips.com/threadminder.cfm?pid=830indent

]

Not sure how pertinent the question is, but IIRC the time constant in desuperheaters can vary depending on type, i.e. contact or spray; which one is used in your application?

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[marshalls]

Hey crshears, got a 404 error on that link haha.
If I understand your question, it's contact type. Basically a packed bed that the condensate is being sprayed across. We're still designing the processes, so not quite at the stage of figuring out time constants or tuning loops yet.

 

[crshears]

Sorry about that, US! Try this:

https://www.eng-tips.com/threadminder.cfm?pid=830

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[georgeverghese]

A lot of the response lag in a TT - TIC - TCV loop is in the thermowell the TT sits in. I have used special thin wall externally finned thermowells in the past for these applications rather than the regular plain thick walled type to cut down on the response lag. Thermocouple type TT is also quicker to respond compared to RTD.

 

[marshalls]

A finned thermowell is a good idea for this application. Thanks!

 

[hacksaw]

Depending on your desuperheater, the thermowell needs to be placed 50+ pipe diameters down stream, to insure condenste evaporation/mixing.

You don't need special thermowells, traditional designs work fine. The finned designs are better suited for high precision, gas flow metering. They do offer improved response time and better accuracy at low temperatures. As long as your main run piping is insulated, accuracy is a non-issue for traditional designs. The calcs ar easily performed if you provide your data.

 

[marshalls]

Thanks for the tip hacksaw - we're just starting to specify these, so we haven't started weighing our options yet.