26 replies to this topic

[Field Engineer]

Oil is pumped to loading terminal (tanker) via export pumps from 10m (height) oil Tanks (capacity of oil tanks 8 tons)

Now it is requested to install control valve at the pumps discharge line going to terminal

Dear colleagues!!! could you please advise what kind of valve to install to control the pump suction pressure,

with the decreasing of oil level at the tanks pumps suction pressure also decreases and trip pressure is 0.035 bars at the suction. Should it be butterfly, globe or ball valve, which will be appropriate ? Thanks

 

 

 

[Field Engineer]

pump discharge pressure is normally 8-9.2 bars and line size is 16" (manual gate valve was used to control the suction pressure for 3 years)

[kabtik]

Are you saying that the height of the liquid provides so much pressure at the suction of the pump when the tank is full so you have to control the suction pressure by throttling the suction and then gradually opening the suction valve as the level climbs down? Please clarify

[thorium90]

As kabtik has suggested, it does seem odd. Furthermore, 0.035bar is not even high, (although you might have a very large tank of different material so maybe)... You also mentioned it is decreasing to the trip value of 0.035bar which implies 0.035 is a Low pressure trip am I right? If you throttle down the suction valve, the suction pressure will go even lower isnt it? Therefore you should have a gate valve for the suction and leave the suction valve fully open to give the pump as much NPSH as it wants. The discharge valve can be a globe valve to control the pressure and flow you want

Edited by thorium90, 22 February 2013 - 09:43 AM.

[ankur2061]

Field Engineer,

 

Having a control valve on the pump suction is something very unusual. At least I have not encountered it.

 

In a batch transfer operation from a tank what I have seen are the following type of controls (trip).

 

1. Low Level in the suction source (tank) trips the pump.

 

2. A pressure switch (transmitter) in the pump suction line trips the pump at the trip set-point. This trip is somewhat tricky and I have seen spurious trips due to this type of pump protection control.

 

The above two controls are most commonly used to prevent the transfer pump running dry without liquid.

 

Can you further explain the logic behind throttling the suction flow using a control valve?

 

Regards,

Ankur.

[kkala]

Referring to post no 4, indeed no closing of suction valve is recommended for a centrifugal pump. Can Field Engineer clarify the intended scheme? Post no 1 mentions control valve at discharge and suction line. Better to provide a sketch, indicating minimum flow return line (if any).

Edited by kkala, 22 February 2013 - 10:02 AM.

[Field Engineer]

I am very sorry for poor explanation.

Please advise what kind of control valve to install  ON THE DOWNSTREAM OF the PUMP, so that it will control the pump suction pressure (this means, when the suction pressure decreases, control valve will gradually close to provide more suction pressure),
with the decreasing of oil level at the tanks, pumps suction pressure
also decreases and trip pressure is 0.035 bars at the suction. Should it be butterfly, globe or ball valve, and what size ? Line size is 16"

Pump discharge pressure during operation is normally 8-9.2 bars.(manual gate valve was used on the DOWNSTREAM (instead of control valve) to control the suction pressure for 3 years)

Thanks

[thorium90]

Ahh, I get it. You are afraid as the pump pumps down the tank, the pump will suck out the tank until the low suction pressure trip.

 

The simplest way is to have a control valve at the discharge. The setpoint is decided upon by the operator who takes a look at the suction pressure before inputting the discharge pressure setpoint.

 

The second way automates it a little. Its output will be from a calculation block taking into account both the discharge and the suction pressure. For example, the discharge pressure can be set at 8bar and the control valve will open to achieve the pressure downstream of the valve. However, when the suction pressure approaches the setpoint for the suction, say 0.05bar, the output from this block will increase which is subtracted from the total output. Therefore the output to the control valve will remain the same. You would need abit of experience to get such a controller design to work as planned since both are pressure controllers otherwise you might end with one always taking precedence.

 

The third method does not control the suction until a low alarm is reached. The output to the valve will also be on a block that has both the suction and discharge pressure going to it. The block acts as a low selector. During normal circumstances, the discharge pressure will control the output to the valve. The suction pressure will be on tracking mode and follow the output from the discharge controller. When the suction pressure alarm is activated, the controller will turn the discharge pressure controller to manual and use the suction pressure controller for control which will then output to the control valve based on a preset suction pressure setpoint and throttle down the valve to increase the suction pressure.

Edited by thorium90, 23 February 2013 - 12:12 AM.

[Field Engineer]

Thank you sir

We are planning to install control valve at the pump discharge and the next question is what type ?

My option is : Butterfly valve, because more liquid can pass through the valve and cost effective, globe valve will cause the pressure loss due to contorted path through the valve and costs pretty much, ball valve will cause some challenges with seats.

What is your opinion gentlemen ? Thanks a lot

[kkala]

1.16" discharge line does not seem to comply with tank capacity of 8 ton (post no 1), can you please clarify? At present much bigger suction tank capacity will be supposed.
2. Seeing that NPSHa represents absolute pressure at pump suction (minus vapor pressure), there should be no NPSH issue for the pump. Mentioned 0.035 barg represents a tank level of 42 cm from pump suction center line, if liquid specific gravity (sg) is 0.85. It seems better to trip the pump through a level switch to avoid air ingression, than try to "control" suction pressure.   Fuel pumps for ship loading here apply this principle, and they have only manual valve at discharge.
3. Probably you want to lower flow rate before trip (as tank level is approaching the value to cause trip) just because lower flows permit lower tank levels without air ingression.  In this (sophisticated) case a level control valve would be installed at pump discharge. Is the intent so? This would be done manually here (think if control valve should be failed close or open, also it will have a ΔP even quite open that could reduce max flow).
4. At any case question concerns recommended type of valve. Local Process Dept would first think of a butterfly valve, on the ground that globe valves available in local market are up to 8" size (this with delay, after order); and ball valves are actually on/off (not for control), available at smaller sizes. Since tight sealing is not necessary for the intended service, the butterfly valve need not be of a sophisticated type (http://en.wikipedia....Butterfly_valve).
However searching the web you can find globe valves of higher size (up to 24"), <http://www.dpv-uk.co...lobeextract.pdf>, Perry mentions globe valves of up to 16" (class 150 Lb); besides control valves are usually of lower size compared to connected pipe (16"). Probably information to local Process Dept (not responsible for valve selection) is not precise, or globe valve of proper size can be found in another market at a reasonable price. Opinion from somebody knowing these matters would be welcomed. Flow reduction with globe valve (quite) open had better be estimated in advance.
 
Note: This post was prepared and sent before seeing post no 9 by Field Engineer

Edited by kkala, 23 February 2013 - 07:47 AM.

[Field Engineer]

Thank you very much for priceless information, I am looking forward to suggest globe valve and working on sizing, I already sent an e-mail to manufacturer to provide Cv values. On the other hand thinking about butterfly valve, could it be safe, reasonable and handle the control ?

If you have any Cv values for 16" and 12" glove valves, please let me know, I can understand that it varies depending on manufacturers valve sizes, but anyway. I want to calculate approximately. Its a big pleasure to deal with professionals   

Thanks a lot

[kkala]

1. No experience in globe valves of size > 8" (post no 10). Web indicates availability of 12" - 14" - 16" globe control valves (e.g. Valtek), but their Cvs cannot be found. C R Branan's "Pocket Guide to Chemical engineering" (Gulf 1999), Control valve design - Table 1-6 - liquid flow, indicates  Cv=12*d² as a rough estimate  in the absence of specific data, where d=valve diameter (inches). Nevertheless no limit of d size is mentioned. The coefficient is 32 (average) for butterfly valves, instead of 12 for globe valves.
Note: Suppose you install a 12" globe, understood Cv = 12*12²=1728 GPM(psi)^--0.5; ΔP due to flow restriction (16"/12") and enlargement is not included. Coefficient is 11.5 for single-seat top guided globe.
2. Not having experience on butterfly control valves in liquid service,  they are used so and should be also considered. Probably pricing will be the decisive factor. An expert could clarify it now, apart from technical matters. I do not know whether a butterfly valve of lower size (e.g. 14" or 12") would be recommended.
"Behaviour" of two manual butterfy valves was once investigated on a branched gas flow (1980); control was satisfactory. No problem was heard after the implementation.
A book from Instrument Society of America was used to estimate ΔP of butterfy valve in function of its opening angle.

Edited by kkala, 24 February 2013 - 06:56 AM.

[Field Engineer]

Gentlemen appreciate the effort applied by everyone.

Mr KKala, thanks a lot

[Raj Mehta]

Well, I agree with KKala. but I found PDF attached below, saying that butterfly valve is available in sizes from 2" to 48" and globe valves upto 8". Experts clarity is required here. Globe valve certainly provides the best throttling & controlling service but at the cost of high pressure drop, when compared to butterfly valve of same size. Butterfly is cheap and has many benefits. Following PDF explains it. Thanks. Regards, Raj Mehta.

Attached Files

•  valve selection essentials.pdf   164.84KB   274 downloads

[kkala]

Yes, butterfly valves are available in large sizes. Investigation of two butterfly valves mentioned in post no 12 concerned 20" or 24" valve size on a same size gas pipe.

A globe control valve installed on 16" pipe would usually be of lower size (probably 14", 12", etc, depending on requirements). Similarly I wondered (in post no 12) whether a butterfly control valve could be of lower than 16" size too, without some drawback. Is it common practice?

Such size reduction has not been seen in manual globe or butterfly valves.

No doubt a lower size butterfly valve will have higher ΔP, for same flow and valve opening.

[Raj Mehta]

Yes Kkala,

 

I have seen butterfly valves of sizes lower than 16" and that too preferred over globe valve. I have seen that ball valves are preferred for smaller line sizes whereas, Butterfly valves for higher line sizes.

 

Globe provides better controlling feature and is used for critical services where proper & efficient flow control is desired, else, butterfly is used (I suspect, cost might be a reason for this.)

 

Thanks.

[gegio1960]

Thanking Kostas and Raj for their answers giving parts of the solution, I'd like to see a post from a professional expert in valve selection.

Maybe there are more than one among the forum people.

Regards and smiles 

[saeed]

Dear Field Engineer,

 

Referring to your question regarding CV values in Globe control valves sizes 12" and 16" i would like to mention CV values of SAMSON Globe control valve size 12" (Type 3241) which are 190 - 290 - 420 - 735 - 1150 and finally 1730, these values are just for this specific type from Samson AG. products and of course in case you use another valve type from another manufacturer the values could vary but for sure they can not propose you much larger CV values for this valve size (i mean for 12" Globe valve the range of available CVs are this and no manufacturer can claim that they have GLOBE valve, 12" with for example CV value 3000)     and for Size 16" (Type 3254) CV values are 420 - 735 - 1150 - 1730 - 2300 and finally 2900.

 

Regarding selecting the type of valve, as the other friends explained before, Globe valve is best choice when you are thinking about controlability of valve (but there are also many situations that we prefer or have to use butterfly valves, for example while requiring very low DP which results very high required CV value or ... ) but price-wise, butterfly valves are more economic (also you have to note that we can not always use BF valves instead of Globe valves in control applications, for instance when we require very low CV values such as 2 or 0.2 or...    or while dealing with very high DP in comparison to inlet pressure and ... ).      but i want to ask you to consider another type of valve for your application too, this type is called "Rotary plug valve" or "Rotary Globe valve" (The advantage of this type of valves is that in one specific size they have CV values even up to two times more than maximum CV value of one Globe valve in the same size, so with smaller valve size you can get same cv value (of course if you are in need of high cv value in your case which i'm not sure about it)) 

Finally for brochures of Samson globe valves types 3241 and 3254 you can search the web and you can find them easily.

[kkala]

Thanks, saeed, for the useful information. Samson's table 1.1 of <http://www.rencor.co...n_Cv_Values.pdf> gives following data for globe valves.

Cv=1730 for 12" size (12")

Cv=1150 for 12" size (10")

Cv=735 for 8" - 10" - 12" sizes (8")

Cv=420 for 8" - 10" - 12" sizes (6")

figure in parenthesis is the valve size that  would give practically same Cv as per Cv=12*d_² by Branan (see post no 12).

It seems that a valve of lower size than 12" (even 6") can have outlets (flanges) of 12" to fit 12" pipe. In this case pipe reducers would not be used. Valve is sized according to required Cv. A globe valve of 12" flanges can have Cv from 1730 to 420, or even lower.

Edited by kkala, 09 March 2013 - 11:26 AM.

[saeed]

Dear KKala,

 

You are absolutely right about ( Cv=12*d_² by Branan ), it is a good rule of thumb formula for calculating maximum possible CV value of Globe valve in each size.

 

In the meantime, While selecting the control valve, we should consider:

1- velocity in valve outlet        

2- required CV value        

3- if there is Cavitation, Flashing (in liquids)   or Chocking (in Gases),    

 

with taking into account all above mentioned parameters, we can select the best  possible control valve size.   but also we have to note that valve size selection depends on pipe size as well, for instance, we can not use 6" valve while pipe size is 16".     The common standard in OIL and Gas industry is that we have to select a size for valve at least one size more than half of the pipe size.   i mean that for example if the pipe size is 16", so half of it is 8" and so minimum possible valve size would be 10" which is one size bigger than 8".

[kkala]

Thanks, saeed, for the  information on control valves, not known to me because local instrument Dept deals with it. But this is useful for a process engineer to know. I can also understand why instrument Dept does not say "size of this control valve is 16", when it is installed on a 16" line. Cv rather than pipe diameter indicates valve size.

[saeed]

Dear Kkala,

Yes, you are right. it is not mandatory to use valve size equal to pipe size when you have the flexibility of using smaller valve size.

I'm chemical engineer but i used to work with valves rather than in process department. so my process abilities is negligible and instead i just know something about the valves and their accessories.

[kkala]

Every activity has its merits and is useful when it is not felt as a fatigue-duty.
A relevant topic found: <http://www.cheresour...-pressure-drop/>

[Ajay S. Satpute]

Some more useful info. on valve CV values from vendor.

 

Regards.

 

Ajay

Attached Files

•  Flowserve Cv tables.pdf   1.18MB   222 downloads

[jr31]

I belive that you need to look at minimum required pump flow which should be available from pump vendor & size the valve for this flow as a normal flow for sizing the control valve. Possibly this should require a smaller size of valve. This minimum recirculation flow I feel should be first level of protection. A tank low level trip (or sunction Pressure)  of pump would be additional protection. We have,  for a 8" discharge line  a 3" size  recirculation Globe valve (Cv-app 110)  & this arrangement is working fine.  

The post indicates use of a Manual Gate Valve for 3 years. The gate valves are not best for control applications. So possibly a butterfly valve which has lesser controlability than globe valve could be considered espcially the prssure drop across the valve should be around 9 Bar. This would more economical than a Globe valve.