Deaerator level control

[ananya2011]

Hi all,

I am Bala Krishna Godavarthi, i am an DCS operator in QAFCO. We have a strange problem with newly commissioned Deaerator,

i.e level control at higher loads is becoming difficult, controller is not responding properly, even though it was tried to

tune several times. I suspect a problem with feed water temperature as feed water is beng supplied at 45 deg c, as there are

no preheaters at the feed inlet and steam consumption for Deaeration is higher, i suspect swelling or shrinking in the level

when fresh water is fed at low temperature, but this is not the case at lower loads, but this comes into picture while

increasing loads and at higher loads. What can be done with level measurement? is the level can be biased with pressure just

like the case in boilers??

 

[TD2K]

When you say the controller is not responding correctly, what do you mean by that?

Have you checked that the valve is responding correctly in the field to a changing set point? Is there a local level indication, is it showing the same level as your instruments when this is happening?

 

[steamdog]

The deaerator should be sized for your maximum load, so that the temperature in the storage part of the DA should not change with a higher flow rate and the shrinking of the water level should not be caused by colder water. You have not outlined your control method for the level so that it is difficult to comment on this.

 

[hacksaw]

Dearators can present special difficulties at high loads where make-up water can create a process upset that the level control cannot handle. You need to get a process/ mechanical engineer familiar with the material and energy balance of the dearator involved. Typically piping/equipment changes are required to stabilize control.

Dearator pressure control at all loads is critical and you may need to preheat your makeup water.

 

[ananya2011]

Hi all,
Thanks for the replies, may be the problem was not quite elaborated.Let me try to explain,i previously worked in a process plant where deaerator is operating at 125 deg C and feed is entering at 110 deg C, so steam required for mechanical deaeration was quite less, so pressure change in deaerator during process upsets was not much significant. But in this case
as water is entering at 45deg C and operating temperature was at 125 deg C steam consumption is higher,for example to treat
200t/hr of BFW we are injecting around 35t.hr of steam to keep the deaerator pressure constant, the Deaerator operating pressure is at 160 kpa to keep the outlet DO <5 ppb. So with change in water flow during process upset this pressure is varying which makes me think that it causes either swelling or shrinking in the level.we have a master controller on the level cascaded with a flow control, master control is level and slave is on flow control.I think level indication is working correctly as all the LT's including trip transmitters are varying in the same manner.

 

[thorium1]

If you dont mind, is it possible to post a trend? It would be useful to see the following trends:
Deaerator pressure and level
Flow rate of water to the deaerator; the control valve, SP, PV, OP
Flow rate of steam to the deaerator; the control valve, SP, PV, OP (I assume the steam is directly sparged to the deaerator)

I believe the problem lies in the fact the feed water is much colder than the temperature in the deaerator. When there is an upset in the level, say the level drops, the control valve dumps more cold water in, lowering the temperature and pressure in the deaerator and causing a reduction in boiling but increasing the level. The lowered pressure causes the control valve to increase the steam to the deaerator to increase the pressure. The lag time between the the cold water getting boiled to saturation causes the steam valve to over compensate causing the water to boil more and raising the pressure. The control valve controlling the feed water encounters more pressure in the deaerator and therefore for the same OP, less water goes in the deaerator. This causes the level to drop again. This cycle starts to continue and depending on the tuning of the two valves maybe keep on going and a dynamically coupled loop. Analysing the trends can quite easily confirm this.

 

[thorium1]

This phenomenon is principally created by the cooling effect of the cold water and exacerbated by the lag times for the temperature to rise and for the water to flow through the packings in the deaerator to reach the bottom and be registered as level. Typically preheated water will almost not experience this. The effect is also compounded by the dynamic coupling of the controllers for the steam to the deaerator and the water to the deaerator. For example, when the controller for the feedwater tries to control the level at the SP by adding more water, it cools the deaerator and acts a disturbance for the controller using the steam to control the deaerator pressure. When the steam controller adds more steam to maintain pressure SP, it acts as a disturbance to the level controller.

 

[kkala49]

Two deaerators in a local refinery are fed with water of 10-91 oC (but often 60-90 oC, due to preheating with waste streams). Controls are LCV for incoming water and PCV for steam.
Though not close to boiler Operations, I have never heard of above problem, probably because useful volume of each deaerator surge drum (just below spray-tray tower) has 20 min hold up time at maximum outlet rate of water (102 oC). This may offer adequate inertia.

 

[thorium1]

yes kkala, I believe you are right, this problem is unlikely to happen if the water is preheated or the deaerator is large enough or there is enough process noise that it sort off gets masked out. Also, I note that this problem wouldnt be observed if the upsets are small enough. The instability is set off only after a large enough disturbance. During steady operations, one never experience the problem.

 

[thorium1]

one simple way to solve would be to nerf the steam PCV tuning parameters. The reason is that deaerator levels would have trips associated with them while deaerator pressure is not as critical. The other way is to just place the controllers on manual until the fluctuations die out.