7 replies to this topic

[marthin_was]

Hi all, I Need you suggestion.
I have high pressure pump (330 psig,4000 gpm), running paralelly, it is designed to ship water through pipeline to reservoir (atm condition),from hydraulic simulation, it takes only 180 psig to deliver the water, but the pipe line is old and thin, the problem is, is there any way to reduce disch pump pressure (to avoid replacing/repairing too many pipes) without trim the pump impeler? Is it good options to use PCV at common discharge line and install common minimum flow return line to dispose extra flow?

[Art Montemayor]

Marthin:

Before you receive my suggestions, you are going to receive my criticism. You are mis-calculating and probably misapplying pump hydraulic information and design.

First, you are using the wrong impeller for the pump you have selected IF you only want to transport the water – and do it with a minimum of discharge pressure. I believe you are reading the pump curves completely wrong and this I attribute to a lack of understanding what they represent.

A pump does not generate pressure automatically with the flow it transports. A pump responds to the resistance forces that are imposed to that flow which it is trying to transport. Therefore, you don’t have a “high pressure pump (330 psig, 4,000 gpm)”. What you have is a pump capable of withstanding high pressure (330 psig). If the same pump meets little or no resistance in transporting 4,000 gpm, its discharge pressure will be NIL (or very little).

You indicate on your pump curves where the pump will operate; this is only true if you (or the system) can impose the corresponding total developed head (TDH) indicated on the ordinate of the curve. Since you say you only want to “transport” the water to a reservoir at atmospheric conditions, then the obvious horse-sensical thing to do is to allow the pump to only overcome the systems resistance (you say that this is 180 psig – 415 feet of water). If your calculations are correct, then the pump can operate with an 11” diameter impeller and yield what appears to be 3,500 gpm at 415 feet of TDH. This is the lowest pressure that you can impose on the discharge piping.

I strongly recommend you (as I do all engineers with pump applications) to generate a SYSTEM CURVE that overlays your pump PERFORMANCE CURVE. With this system curve you will see the pressure generated by the system’s resistance as the flow rate is increased. Where the system curve crosses the corresponding impeller’s performance curve is where your pump will operate. You can raise the TDH on the pump if you want to (I don’t see any reason for this) by throttling the discharge line at the point where the water discharges into the reservoir. This will increase the pressure on the transport line of course. That’s why I would run the system wide open – there should be no throttling on the discharge line and the only pressure built up should be the system resistance imparted to the water – your estimated 180 psig (415 feet of head).

I would not trim the existing impeller, but would rather buy a new, 11” model and install that one instead.

I don’t understand your question involving PVC and “common minimum flow return line”.

Your attachment of the pump performance curves was a smart move on your part. It is the perfect way to furnish basic data and to explain what you are trying to do. It enabled me to quickly spot what you are assuming wrongly.

I hope this helps you out.

[marthin_was]

Thanks art for your critics, since I need to learn from experience engineer like you, let me explain more detail...

the original design of water shipping are like this :

Existing Shipping pump -------------> atm Reservoir + Booster pump ----------> field
'old' 30" pipeline pipeline
Our current condition, the booster pump is not available since last 5 years, so the water is shipped directly from shipping pump (bypassing booster pumps) to the field that requires high pressure since it takes long distances. Now we want to activate again the booster pump and repair the old pipeline, but without eliminating flexibililty to ship the water directly from shipping pump. So I aggree with your suggestion not to trim the impeller. The PCV purpose is to throtling the pump pressure, if the pressure is reduced then the flow will increase so I need to return back the flow back to tank.CMIIW

Let me clear with your explanation, did you mean if we have pump that capable to deliver 4000 gpm at 760 ft head, and I have system (pipelines+etc), that give back pressure 180 psig, the pressure reading at pump discharge only 180 psig? is this happen only if we full open the discharge valve?

Thanks again for your explanation, I'm waiting for your next repply..

Regards,
Marthin

[Nirav]

Hello marthin,

I suggest you to read the following article. This is the best article on this subject.

http://www.driedger...._cp/CE1_CP.html

You said that....

The PCV purpose is to throttling the pump pressure, if the pressure is reduced then the flow will increase so I need to return back the flow back to tank.CMIIW

This is only partially correct (in case if you try to find flow just by considering 180 psig as the required pump head). As Art has rightly said, you must know your "system resistance". Pump can not produce pressure if there is no resistance to flow. For example, if you remove discharge line totally and just put 1 feet of discharge pipe open to atmosphere, pump discharge pressure can NOT reach to 330 psig. Because there's very very little resistance in this case.

So, why I said you are partially correct above is due to following.
When you put PCV, you are adding resistance to flow. Which means, now, your system resistance would be 180 psi (pressure drop in pipe line) + pressure drop in PCV.

And if you decide to have pressure drop in PCV = 150 psi, your current pump will work fine with the "same original flow of 4000 gpm". You do not need any return line in this case.
Because, now, system resistance = 180 psi + 150 psi = 330 psi.
This means, it will not give any extra flow and still meet your "ultimate" requirement of 180 psi resistance in pipeline.

This is, in brief, how you design your control valves in pump discharge systems. If you really want to put some kind of PCV in your system, you need to do detailed hydraulics and proper selection of PCV.

Thank you,

[marthin_was]

Nirav, thanks for your input,
Did you mean, when PCV is throtling to reduce pressure (in this case the valve must be reduce the opening), it would not reduce the water flow, and move the operating point of the pump at pump curve?

Thanks,
Marthin

[Nirav]

Did you mean, when PCV is throtling to reduce pressure (in this case the valve must be reduce the opening), it would not reduce the water flow, and move the operating point of the pump at pump curve?

Marthin,

Just one question. Did you read the article I wanted you to read?
If not, I would again recommend you to read it. It will give answer of your question.

See Fig.1-3 on that article. You basically move your "system curve" (called "modified system curve") upwards when you throttle the valve in discharge line. OR move your "pump curve" downwards to meet system requirement. Both are same phenomena. Ultimately, you ask your pump & control valve to provide pressure required to kill the resistance in path.

Just to reply to your specific query of reduction in flow when you throttle the valve. In normal condition, it will reduce the flow if there is no change in upstream pressure of valve.
REMEMBER : Flow is always dependent on the "flow area" and "pressure drop" across valve opening. This means, if you can provide higher pressure in upstream, you can get same flow even if your area is reduced (throttled), considering same downstream pressure.

Suppose, there is no valve in discharge. Means you have resistance in discharge as 180 psi for 4000 gpm flow. BUT based on your pump curve, you find that it must have 6000 gpm flow WHEN pump pressure is 180 psig. This can be considered as "full opening" position of PCV (in other words, there's no valve as i said initially).

Now, when you throttle the valve, you will start reducing the flow and stop when you reach 4000 gpm. You move upward on pump curve. In this way, you have also increased system resistance also by means of "pressure drop" in PCV (because valve is throttled). Now, your system resistance is NOT 180 psi. But it is = 180 psi + dP of PCV. So, dP of PCV would be as below (these are brief calculations).
>> Find discharge pressure of pump at 4000 gpm from pump curve.
>> Deduct dP of pipe line (resistance of pipe line) from pump disch.pressure. In your case, it is 180 psi.
>> Remaining is dP of PCV (resistance of PCV). In your case, it is 330-180=150 psi

Earlier, Art said that, system curve must cross at pump curve at required point. Your required point is at 4000 gpm on pump curve. Therefore, you have to increase your system resistance from 180 psi to the value at which it will cross at the required point on pump curve. You are doing it by providing PCV. You are increasing resistance so that it will cross at your required point.

I hope you understood above.

Again, remember that pump never produces pressure without resistance in discharge. Your pump is capable of producing 330 psig pressure if there is resistance. So, you are providing additional resistance ( in terms of PCV) to get the flow of 4000 gpm.

If you don't want PCV, you must change pump which will give 180 psig pressure at 4000 gpm. Probably, 11" pump impeller as suggested earlier by Art.

I guess it is clear in your mind now.

[marthin_was]

Ok Nirav,
Thanks for your explanation, it help me a lot for understanding what the article means, to be honest, I got confused when looking modified pump curve that move clockwise... ^___^

[marthin_was]

Dear All,

If I have 3 pumps running parallely, is it good engineering practice to install PCV at common header pressure instead of install PCV at each discharge pump and For large diameter pipe (30"), is it ok to use butterfly PCV? Since as far as I know globe valve for that size not available in industrial market..

REgards,
Marthin