DRUM LEVEL CONTROL

G

Thread Starter

GENE

Hi,
Can anyone give me an example of a good high pressure boiler drum level control, 3 element, level, steam flow, and feedwater flow. Boiler also has unmetered continuous blow down. I am using Bailey SLC'S.
Thanks for the help
GENE
 
V
Can you provide the following info:

Drum pressure
Steam flow at full load
Number of feedpumps available
Type of feedwater flow control (feedwater control valve, feedpump speed ?)
Types of measurement available (levels, FW flow, steam flow)

Then we can start doing something

Vince
 
P
Generally speaking cascade control with two PI controllers should do the trick (we've used this on several boilers up to 80bar working pressure). The primary controller matches the flows (you obviously need mass flow measurements to match - steam flow must be corrected for pressure & temperature either by your instrument or PLC): the steam flow is the setpoint, the feedwater flow is the actual value. The secondary controller compensates for the blowdown and other
minor disturbances (drifts) by influencing the setpoint of the primary controller (usually limited to +-30% or so) according to the actual level reading and level setpoint. This controller should be relatively slow, because you don't want it to react to "fictitious" level changes caused by sudden changes in steam demand. Instead you want it to correct longterm changes, like
increased blowdown... IMHO the key part is to get the steam mass flow measurement right. You can add some additional "fuzzy" features like opening the feedwater valve 100% when level is very low... (we have this but it was not activated once...)

Regards,
Peter Kosin

Process control engineer
INEA d.o.o.
http://www.inea.si
 
I should hope not, Peter - opening the feedwater valve 100% will add large amounts of cold water and cause the level to fall even further ...... (Commonly known as "squashing the drum" in this part of the world.)

Bruce Durdle
[email protected]

Peter Kosin wrote:
You can add some additional "fuzzy" features like opening the feedwater valve
100% when level is very low... (we have this but it was not activated once...)
 
A key factor in Water Level and Boiler Feed Water control is the size of the steam and mud drums. If too small they can thwart attempts to
effectively control drum level. Years ago I participated in a study conducted to mathematically relate drum size to drum level transients caused by sudden increase or decrease in steam flow.

The former results in the phenomena referred to as "drum swell", while the latter causes "reverse drum swell" or in Bruce's vernacular "squashing the drum". Both are caused by the sudden change in the water/vapor ratio resident in the boiler's riser steam tubes and the thermal input.

The mathematical thermal transfer study was conducted. It was quite revealing. If you are interested in a copy I will look for it, but it
may take some time.

Phil Corso, PE
Trip-A-Larm Corp
 
To those requesting a copy:

Please send a self-addressed envelope (large) with postage to cover about a 1/2-pound (250-gram) document. If you want it sooner, send UPS,
FEDEX, or DHL account number, and I will ship it as "Collect / 3rd Party Billing".

Phil Corso, PE
Trip-A-Larm Corp

>A key factor in Water Level and Boiler Feed Water control is the
>size of the steam and mud drums. If too small they can thwart
>attempts to effectively control drum level. Years ago I participated
>in a study conducted to mathematically relate drum size to drum
>level transients caused by sudden increase or decrease in steam
>flow.
>
>The former results in the phenomena referred to as "drum swell",
>while the latter causes "reverse drum swell" or in Bruce's
> vernacular "squashing the drum". Both are caused by the sudden
>change in the water/vapor ratio resident in the boiler's riser steam
>tubes and the thermal input.
>
>The mathematical thermal transfer study was conducted. It was
>quite revealing. If you are interested in a copy I will look for it, but
>it may take some time.
 
Steam drum level control:
Three element system apply to boilers where greater capacities are accompanied by high pressures and temperatures, and where the drum and other water capacitiy sections of the boiler are relatively small.
A change in steam demand will create: considerable swell on increasing load and vice versa, shrinkage of the drum level on decreasing load.
Swell caused by an increase in steam mas flow rate
within the water due to an instantaneous reduction of pressure in the drum results in an increase in drum-level controller to call for a reduction in feedwater valve opening.
This directly opposes the needs of the boiler.
To counteract this effect, the three element
system employs the ratio of steam flow and feedwater flow as the source of principal action. Readjustment of the drum-level controller provides
compensation for the steam flow water-impulse, should the actual variation in drum level remain.
Steam flow is ratioed to feedwater via proportional action (two input controller).This output signal is one of two inputs to the feedwater controller, its second input is the output of the level controller. The feedwater controller is P+I.
In the past, these controlling actions were achieved via controllers designed for boiler control.
An important point to remember is the calibration of the level XTR. It must be done based on design operation condition, because the large decease in water SG as temperature increases.
For starting from cold, one may equipe the drum level with several XTR's calibrated at various operating temperature or correct for the level reading. This is no problem anaway because the procedure for cold start calls for manual reading of the level through glass gage. Your Bailey system will do the job providing you reproduce the original analog tuning parameters, pertaining to your boiler.
You say high pressure, what operating pressure ?
Is it constant?
Near critical point becomes delicate.
We do not enter in that today.
 
For a 2520 psig, 565 MW generating unit with full load steam flow of 3875 klbh and 30" drum diameter controlling at 50% (15") -

Density compensated drum is PV (process variable) of PI controller with the constant SP (setpoint) of 50%.

The output of this PI controller is summed with the % steam flow FF (feedforward) signal and limited to 0-100% of feedwater capacity. This becomes the SP for the feedwater flow PI controller which uses measured feedwater flow as the PV.

The result is then the SP signal for the final control element. In our case the final control elements are positioners for variable speed hydraulic transmissions between the turbines and the boiler feedwater pumps.

Drum level measurement is inverted dP (0-30" H2O) with the density compensation function derived from drum pressure. We employ mutiple mesurements of each with median selects for reliability.

The feedwater and steam flow must be ranged identically. Feedwater flow with temperatures above 300 degrees F should be temperature compensated.

Steam flow is determined by HP turbine control stage pressure.

Resulting control is +/- 1" of SP. Under extreme transient conditions we'll see +/- 4". For us these results are adequate and we don't attempt to deal with drum swell/shrink as distinct issues.

Most of the devils are in the details of constructing tracking, anti-windup, and limit conditions.
 
how do you apply the density compensation from pressure, i need to get the specifics of the curves and the formulas applied.

thanks,
 
please try following:

Make one PID controller
apply (Drum Level + feed water flow - steam flow) to it's input as process value and it's manipulated value should regulate the full load valve.
make the provision that when you put this controller on auto it's Set point should be equals to present process value.
Ask plant operator to maintain the desired level manually by looking in to hydrastep etc. and when he achived little stability the controller is to put on auto.
this will work very good.
 
A key factor in Water Level and Boiler Feed Water control is the size of the steam and mud drums. If too small they can thwart attempts to
effectively control drum level. Years ago I participated in a study conducted to mathematically relate drum size to drum level transients caused by sudden increase or decrease in steam flow.

The former results in the phenomena referred to as "drum swell", while the latter causes "reverse drum swell" or in Bruce's vernacular "squashing the drum". Both are caused by the sudden change in the water/vapor ratio resident in the boiler's riser steam tubes and the thermal input.

The mathematical thermal transfer study was conducted. It was quite revealing. If you are interested in a copy I will look for it, but it
may take some time.

Phil Corso, PE
Trip-A-Larm Corp
 
Density compensated drum is PV (process variable) of PI controller with the constant SP (setpoint) of 50%.

The output of this PI controller is summed with the % steam flow FF (feedforward) signal and limited to 0-100% of feedwater capacity. This becomes the SP for the feedwater flow PI controller which uses measured feedwater flow as the PV.

The result is then the SP signal for the final control element. In our case the final control elements are positioners for variable speed hydraulic transmissions between the turbines and the boiler feedwater pumps.

Drum level measurement is inverted dP (0-30" H2O) with the density compensation function derived from drum pressure. We employ mutiple mesurements of each with median selects for reliability.

The feedwater and steam flow must be ranged identically. Feedwater flow with temperatures above 300 degrees F should be temperature compensated.

Steam flow is determined by HP turbine control stage pressure.

Resulting control is +/- 1" of SP. Under extreme transient conditions we'll see +/- 4". For us these results are adequate and we don't attempt to deal with drum swell/shrink as distinct issues.

Most of the devils are in the details of constructing tracking, anti-windup, and limit conditions.
 
Top