12 replies to this topic

[justcrazy85]

Dear all,

I am a young chemical engineer at the beggining of my carrier. I need to design a volumetirc pumping system, but unfortunately all my academic studies about pumps were focused on centrifugal ones, so my background on this topis is poor.
I have 5 questions:

1) are positive displacement pumps a kind of volumetric pumps? which is the main difference between positive displacement pumps, metering pumps, and piston pumps?

2) In case of centrifugal pumps, discharge pressure changes by liquid according to its specific gravity, but the discharge head (in meters) is always the same for any liquid, once you have defined the pumping system. Is it true even in case of volumetric pumps (metering, Positive displacemente) ?

3) which is cavitation effect on volumetric pumps? I guess that pump will work anyway, but discharge pressure shall be lower in case of cavitation. Moreover you could have pump failures due to vibrations as with centrifugal. Is it true?

4) I heard that in case of incipient cavitation a centrifugal pump will absorbe more power. Can you explain me the reason? is it true even in case of volumetric pumps?

5) My volumetric pump has to elaborate a flammable liquid, so it is very important to avoid cavitation foir safety reasons. I was suggested to install a device to measure the power absorbed by pump,so that if cavitation is occurring then pump will absorbe more energy and the system will automatically shut down the pump. Do you think this system is efficient?


Thanks for your help and sorry for my long message

Regards and Wishes

Antonio

[ankur2061]

Hi,

Going into the detailed theory of positive displacement pumps on the forum would be difficult. My suggestion would be to read through the basic principles of PD pumps. Also read through the principles of cavitation for PD pumps. The links are provided below for the free resources on the internet:

http://www.engineeri...umps-d_414.html

http://www.pumpworld..._pump_basic.htm

http://en.wikipedia.org/wiki/Pump

http://www.pumpschoo... vs centrif.pdf

http://www.google.co...lient=firefox-a

Also refer the articles on cavitation at:

http://www.cheresour...ation-and-npsh/

http://www.cheresour...ding-cavitation

A book reference is provided which would be very useful and is considered a treatise on pumps (centrifugal and PD)

"Pump Handbook" by J Karassik

Hope this helps.

Regards,
Ankur.

[justcrazy85]

Thanks a lot for your help Ankur... i WILL analyze later that material...i had just a quick look but it seems that there is no any reference to power absorbtion in cavitation status..
Is it true that a centrifugal pump absorbe more power in cavitation conditions? is it true as well as for PD pumps?
as last point....what do you think about question number 5 in my previous message?

Thanks again

Regards
Antonio

[ankur2061]

The brake power or absorbed power for any pump is calculated as

P = Q*H*ρ*g / (3600*η*1000)

where:

P = Absorbed power or brake power, kW
Q = Pump Capacity, m3/h
H = Pump head, m
ρ = liquid density kg/m3
g = gravitational acceleration, 9.81 m/s2
η = efficiency expressed as a fraction

The above formula is applicable for both centrifugal and PD pumps.

Cavitation causes the efficiency to drop sharply as well as decrease in pump capacity and the pump head. The efficiency drop is sharper than the capacity and head drop thereby causing the absorbed power to increase marginally. However this increase in absorbed power may not be very noticeable.

The most noticeable aspect during a cavitation phenomena is the noise and vibration of the pump.

Vibration monitoring using vibration sensors would be a better idea to shutdown the pump on high vibrations rather than using a pump trip based on absorbed power.

Regards,
Ankur.

Edited by ankur2061, 26 December 2011 - 11:58 AM.

[S.AHMAD]

justcrazy

1. The danger for flammable liquid is the fact that during cavitation for most liquid, the pump suction pressure is under vacuum. Air can get through any leak and thus when flammable vapor is compressed in the pump, the heat may be enough to initiate ignition.

2. Using power to detect cavitation is not effective method. In my opinion, it is better to measure the suction pressure. Cavitation occurs when the suction pressure is lower than the vapor pressure of the liquid at the operating temperature.

[kkala]

1. I have heard the cavitation noise of a pump in the fertilizer unit (1978), but at that time power consumption of pump motors was not monitored. Drop of efficiency must be strong during cavitation, if the pump motor is to to increase power (probably to one third, when head falls to half). Can you bring examples or references of power increase during cavitation? Engineers working in Operations can see it, if power consumption of each motor is registered. http://nitw.academia...ntrifugal_Pumps may contain data, hard to be understood (by me).
2. Irrespectively of the above, cavitation detection through power consumption measurement has not been met. Cavitation could be detected by sensing vibrations http://www.incosys.c...avitation-e.pdf, or acoustic signals http://www.imperiali...tionsensor.pdf.
3. Nevertheless we apply low level alarm and low level cut off (stopping pump motor) on the suction tank / vessel, to prevent pump from cavitation. Other more precise methods (e.g. of para 2) have not been observed in local refineries.

Edited by kkala, 01 January 2012 - 04:59 PM.

[kkala]

....1. The danger for flammable liquid is the fact that during cavitation for most liquid, the pump suction pressure is under vacuum. Air can get through any leak and thus when flammable vapor is compressed in the pump, the heat may be enough to initiate ignition.

I have also heard that cavitation of centrifugal pumps tends to destroy mechanical seal, thus creating leak out. This can cause fire in pumps transferring flammable liquids. In local refineries the pump is tripped down just when the liquid in suction reaches a level low enough for possible cavitation (post No 6).

Edited by kkala, 02 January 2012 - 06:39 AM.

[justcrazy85]

First of all, thanks to all of you for your support and useful hints...

I absolutely agree with you that power absorbion is not an effective method. I saw this application in a chemical application similar to mine. As i saw it I had several doubts because i never saw or heard it before. Nevertheless I asked for your support due to my short experience...and an advice is always useful I guess :-)

Kkala, I already proposed a system based on low leavel detection but (as you know) unfortunately the system is not currently fail-safe. So, I was wonderind and evalauting which option is more effective: to make the alarm loop fail-safe or to install a further barrier (as vibration detector for example). I know you can not provide me this information..it depends on the specific system, but i reallyappreciate all data and information you provided.
Just as further point, a collague of mine talked me about a further detection system: some wet detectors to install in the pump...if there is no liquid (in case of cavitation) then pump power should be cut off. Actually, i have again doubts about that: a certain amount of liquid product can be in the pump body even if cavitation is happening (mainly at the beginning during flashing) and before wet sensors are activated you could already have a flammable atmosphere.
Do you agree with my point?

Thanks for your support

Antonio

[justcrazy85]

Kkala....just to be more clear.....when i say "more effective" , it should be read as "cost-effective"

Thanks again

Antonio

[latexman]

On PD pumps with pulsing flow, you have to include "acceleration losses" into the "Net Positive Suction Head" (NPSH) equation. Make note of these new terms when you read through the principles of cavitation for PD pumps.

[kkala]

I agree to non suitability of wet detectors, since pump can transfer liquid in cavitation, even though the head is reduced. Trace of air bubbles (before they collapse) could be better, but still doubtful seeing that web search does not reveal such detectors. Vibrations or sound detectors seem to be the means applied in practice according to web literature, but I have not seen any such instrumentation before. Only LLA and LLLS stopping the pump.
The operation of LLA and LLLS seems safe enough in the majority of cases. The destruction due to cavitation occurs because none is near the pump for immediate action. The LLA warns, so that somebody can go there. If not heard, stop will occur after a while through the LLLS. It is quite improbable that none of the two operates (when needed).
More sophisticated systems (e.g. voting 2 out of 3 LLLS) have not been seen as protection against pump cavitation.
One can think that LLA / LLLS stop the pump prematurely, based on max temperature and more volatile of handled liquids (e.g. naphtha can have varying vapor pressure). But LLA just notifies the operator for preventive action. He can deactivate the trip switch, close a discharge valve partially, and let the pump operate under supervision for more time, if this is the intent.

Edited by kkala, 04 January 2012 - 04:01 PM.

[justcrazy85]

kkala

i absolutely agree with you from a control point of view, but in my case the issue is different for two reasons:

• low level and low low level signals in my case are not fail safe: I mean that even if the control logic is proper, the electrical connection/wiring is not fail safe. If circuit breaks down, system does not work and no any alarm signal is generated. And it is a big issue. This is the reason why i am trying to understand ifit is more cost effective to make my low level signal fail safe or to install a second barrier...but your previous advises are very good. Thanks
• You said that it is difficult none of two alarms operates. Actually it is only partially true. It would be true in case of independent alarms/sensors/circuits, but independent systems are usually used for Hih/High High level detection. In very few cases you can found independent system for low/low low level detection. And it is not absolutely my case :-)

Above all, thanks for your support. Information you provided have been very useful

Thank a lot

Antonio

[kkala]

Thanks for the information to know the case.
Speaking of local refineries, LLA would be connected to normal control system, while LLLS to the independent ESD (emergency shut down) system. These two systems should have even separate sensors, orifices, etc. But deviations from the last rule are spotted in rear cases. Of course reliability is limited if both LLA and LLLS are connected to one system.
I understand that emergency power for instrumentation is not available in your case.