from the Forum department...
Shrink/Swell effect of boiler steam drum
Engineering and workplace issues. topic
Posted by Mikas on 22 February, 2009 - 1:41 pm
Hello to all,
Recently, me and my colleagues had very interesting discussion about shrink/swell effect of boiler steam drum. Shrink and swell effect is always present when there are sudden changes in steam or feedwater flow. On most drums, level is measured by differential pressure transmitters. Dp measurement must include steam table correction on pressure and temperature in order to have accurate reading. Now if pressure is changing inside a drum, can effect of shrink/swell be reduced in measurement reading if pressure and temperature correction is used? In other words, if correction is used will dp transmitter show less shrink/swell level change?

I think not, becaase when pressure changes, correction will help to get more accurate reading of actual level inside the drum, but because of presence of steam/water mixture,shrink and swell will have same impact on reading no matter if correction is used.

What do you think about this?

Posted by Bruce Durdle on 22 February, 2009 - 8:13 pm
A few idle thoughts ...

The water in a drum is usually below saturation temperature because relatively cold feedwater is added to the saturated water-steam mixture coming from the tubes. So the density of the water in the drum will be that of the unevaporated water at drum pressure.

The shrink/swell effect is caused by small pressure changes causing bubbles in the tubes to reduce (shrink) or increase (swell) as a result of small changes in pressure - so that the mixture of steam and water in the tubes remains at the saturation temperature for the pressure at each location.

Because the relative densities of water and steam are very different, a small change in the amount of steam present can have a big change on the density of the mixture.

My feeling is that the actual change in drum level due to shrink/swell effects is much more than anything predicted by looking at the drum density alone. In other words, if you implement full pressure correction on the drum level reading you will get at most only a very small reduction in the amount of shrink or swell.

Feel free to disagree!


Posted by Ben Janvier on 22 February, 2009 - 10:23 pm
Swelling is caused by a volume increase of the steam phase in the risers of a boiler. This increase in volume can come from two sources:
1- increase in firing rate (steam volume under the drum increases and drum swell)
2- Drum pressure drops causing the steam volume to increase because PV=nRT.

In other words, the shrink-swell effect is not a measurement problem but a disturbance to the drum level controller that you need to "tune out". Reducing the effect on shrink-swell can be done through tuning, shrink-swell feedforward, by eliminating the steam FF signal,...

Ben Janvier, senior control consultant

Posted by Mikas on 23 February, 2009 - 8:37 am
That's one point of view. I wasn't able to notice shrink and swell effect when there were steam output lfoe changes (when unit's power is increasing or decreasing). Many people say that shrink/swell is due to pressure change inside the boiler. The question is how big pressure change needs to be in order to notice shrink/swell efect?

All changes notmally going by some ramp (not step) function, and the real question is the following:
What is the most likely cause for not spotting shrink/swell effect, the fact that load is changing with some gradient or that measurement has pressure correction implemented.

I don't have a feeling of orders of magnitudes.
For example, if normal operating pressure is 140 bar inside a drum and level is in range about 400 mm H2O (-200 to 200 mm with an ofset), how big change in pressure needs to be to spot shrink/swell.

Of corse, one way would be to make step test which is not feasible (can be very costly).

Posted by Sivaram on 23 February, 2009 - 1:44 am
The "shrink/swell" effect in boiler level is actual and not apparent.
when the steam drum pressure is reduced, the water in most of the tubes vaporize and increases the level in the boiler. Similarly when the pressure increases the water flow from steam drum downwards through the tubes increases and this reduces the level.
So, providing compensation for pressure changes may not improve the level control. Providing feedforward compensation for water flow will give better control.

Posted by elavazhagan.s on 23 February, 2009 - 10:06 am

In simple, when drum pressure changes, density compensation will help to get more accurate reading of 'Actual Level inside the Drum' including variation due to shrink and swell effect.

Note the compensation using drum pressure for Drum Level DP measurement is only for Density Compensation related to Drum Level Measurement.

Thus to reduce shrink and swell effect, the various control strategies are used (Two Element control using steam flow as feedforward , Three Element Controls using fast feedwater control loop) with proper tuning.

Hope this clarifies your query.

S. Elavazhagan.

Posted by Mikas on 23 February, 2009 - 2:19 pm
Our discussion was actually an introduction to the the following:
Is three element control strategy really necessary to fight disturbance that is generated by shrink/swell effects, or dp measurement with good pressure correction will do the trick?

My main argument is that pressure correction in dp measurement cannot hide shrink/swell because otherwise, single element control would be sufficient.
Problem of tuning three element control scheme would be reduced to tuning of single control loop with dp measurement (included pressure correction).
Of course, someone may say that shrink/swell effect is not the main reason for three element control, but that's another story.


Posted by Ben Janvier on 23 February, 2009 - 10:38 pm
"Is three element control strategy really necessary to fight disturbance that is generated by shrink/swell effects"

No. Not at all. In fact, using steam flow as a feedforward signal can cause your boiler to trip and does not help the level controller during transcient operation. For example, when the firing rate of a boiler is increased, the steam flow will increase, the level increases (swells) and the feedforward signal is asking for more water... meanwhile the level controller output is on its knees trying begging for less water. This is why you will some times see large filter time on steam flow measurement. They use steam flow feedforward for long term mass balancing. They will also use shrink-swell feedforward signal rather than steam flow for drum level feedforward.

"dp measurement with good pressure correction will do the trick?"

No. dP compensation is just there to keep the true level calibrated so that when your setpoint is 1.0 the true level is at 1.0.

" My main argument is that pressure correction in dp measurement cannot hide shrink/swell because otherwise, single element control would be sufficient."

Yes you are correct. Shrink-swell is a firing rate and drum pressure disturbance to level. This means that the level valve need to compensate for this.

"shrink/swell effect is not the main reason for three element control, but that's another story."

The main reason for three element control is that too many people do not have the tools (and the guts) to bump the level controller of a 300 MW boiler to find the circulation dynamics of a boiler and to derive Lambda tuning parameters for their FW FIC and Drum Level LIC. With robust and fast LIC, steam feedforward can be eliminated (I usually keep the olf guy with a 2 minute filter - just to keep everybody happy)

Ben Janvier, Enero Solutions

1 out of 1 members thought this post was helpful...
Posted by Phil Corso on 23 February, 2009 - 3:51 pm
See Control List Thread "Drum Level Control",
Thread #959513894.

Regards, Phil Corso (

1 out of 1 members thought this post was helpful...
Posted by Phil Corso on 24 February, 2009 - 10:04 am
Correction to my earlier post...

The Thread No. is: 949513894

Regards, Phil Corso (

Posted by Bala on 24 February, 2009 - 9:00 am
I feel, during sudden raise of steam flow due increase in power demand the actual level inside drum is slightly pulled up, is called swelling effect. Under this condition the level at both extreme end may not be the same. The converted saturated steam at the top of water level will always have a pulling effect due to the discharge from drum through a small area drum internals such as demister, pipes .... at high pressure. Slight swelling effect will always be there.

I have never heard about shrinking effect.

Using pressure for compensation for accurate measurement, since the density is proportional to pressure.
Pressure = Density x height

It will be difficult to measure, on line density for steam and water

Posted by control guru on 26 October, 2010 - 11:43 am

Posted by PV on 25 June, 2012 - 4:24 pm
I am modelling the level for a steam drum, to determine the actual volume of it: volume=holdup/density.

The steam inside the SD is at 450 F and 430 psi conditions, is it OK to use the density of the steam at these conditions, or do I have to take into account the water density?

From steam tables, at above conditions: density=51.55 lb/ft^3


Posted by Gerrit on 26 June, 2012 - 7:40 am
PV, the density you list must be liquid (or water), not vapour (steam).

My metric steam table gives in similar conditions (pressure 30 bar and saturation temperature 233.858 degC) the following specific densities:

Liquid: 0.00121670 m^3/kg
Vapour: 0.0666641 m^3/kg

Please convert these to imperial units with your favourite unit-conversion tool, if needed.

As you can see, the ratio of liquid density to vapour density is approximately 50. Therefore, if you are operating in the usual saturated steam conditions (and not for some reason with liquid only), you do need to take into account both liquid and vapour densities or specific volumes.

Posted by David Todd on 26 June, 2012 - 10:32 am
The most favoured method of coping with shrink and swell is to ignore it, by letting the forward loop carry the load, while maintaining loose settings on the level controller. Drum level controllers customarily require a proportional band near 100% ( Gain about 1 ) and several minutes of integral time.

Posted by PV on 2 July, 2012 - 9:32 am
Thanks Gerrit

Your use of this site is subject to the terms and conditions set forth under Legal Notices and the Privacy Policy. Please read those terms and conditions carefully. Subject to the rights expressly reserved to others under Legal Notices, the content of this site and the compilation thereof is © 1999-2014 Nerds in Control, LLC. All rights reserved.

Users of this site are benefiting from open source technologies, including PHP, MySQL and Apache. Be happy.

Human beings were created by water to transport it uphill.
Advertise here
View free setup and multi-vendor EtherCAT demo videos online
Servo, steppers, analog, digital & web HMI - Fully Integrated!
Servo, stepping motor control, analog & web HMI in one system!
Time to incorporate data handling, web HMI and motion in one system! is the largest Automation community on the web. Learn how to advertise here now...
our advertisers
Help keep our servers running...
Patronize our advertisers!
Visit our Post Archive