12 replies to this topic

[Logan@adm]

Hi everyone,

 

I guess I just had a couple things that i'm having trouble wrapping my brain around. For a centrifugal pump, I understand that by throttling the discharge valve, I am artificially imposing an additional component of friction head upon the system and decreasing the flow rate, so on it's characteristic curve the pumps operating point should shift to the left.

 

However, as the flow rate decreases due to the throttled valve, the velocity will also decrease which will lead to a decrease in friction head. If the velocity decreases enough to compromise and overcome the additional friction imposed by the throttled valve, won't the total friction head actually decrease for the lower flow rate? I do understand that it depends on your system conditions, but is this scenario where the friction head actually decreases when the valve is throttled at all possible?

 

My second issue is say that a pump is designed for a certain system, but when the pump is installed it is found out that the friction head was overestimated, so the system has less frictional resistance than what was anticipated. How will the pump adapt to the new operating conditions? Will its operating point shift to the right, creating more flow through the pump? Because if that was the case wouldn't the velocity of the fluid just increase, therefore increasing the friction head? Or will the pressure at the end of my discharge line just be greater, so the increased pressure head compensates for the lower friction head? Any help on the subject is greatly appreciated

[fallah]

Hi,

      

About the centrifugal pump think dynamic...and consider the system curve as an indicator of any change of resistance to flow...then considering just a decrease in friction loss due to velocity reduction might lead to a confusion...

 

Throttling in discharge line will shift the system curve toward left side due to dynamic behavior of the centrifugal pumps even if reduction in friction loss will almost compensate the pressure drop across throttling point; e.g. in nearly flat section of the pump curve...

 

And yes, if friction loss was overestimated operating point will shift to the right and create more flow through the pump than expected and it is natural because the system curve due to actual lower resistance to flow than expected, will shift to the right side and its intersection with pump curve as well...

[Logan@adm]

Fallah,

 

Thank you for the reply. Regarding the second topic, when there is less resistance I understand that there will be an increase in flow through the pump. But this increase in flow will result in an increase in velocity. So won't the flow resistance increase as well due to the increase in velocity? I guess I am just having a little trouble understanding how it shifts my system head curve to the right. In a general sense of things, I understand that with the decreased resistance it will shift to the right therefore increasing the flow rate and decreasing the system head. But this increase in flow rate will result in more velocity head, which directly leads to more friction head as well. So where does the decrease in the system total dynamic head occur then? Will it be in the discharge pressure?

[Logan@adm]

Or will the increase in velocity head and friction head due to the increased flow rate be less than the original decrease in flow resistance, resulting in a net decrease of system head?

[Logan@adm]

Also, digging deeper into my first issue, can there ever be a case where you throttle your discharge valve, but your friction head will decrease? I understand that by throttling the valve, there is an extra pressure drop within your discharge line, but the throttled discharge valve results in a decrease in velocity. Therefore your skin friction along the length of your pipe will dramatically decrease.

 

Hypothetically speaking if you throttle your discharge valve to half open, the k frictional loss value will increase, which will increase the pressure drop across the valve. I guess what I am asking is that can the resulting decrease in skin friction head due to the decrease in velocity ever be greater than the increase in friction head due to the throttled valve? Say if it was a gate valve and had a low k value? Which would then result in a net decrease in friction head? Because if that were the case, then the system head curve would not shift to the left, even though your flow rate is decreasing. Then the system head curve would not intersect the pump curve. I guess I am just having a little bit of trouble understanding this. Any help is greatly appreciated.

[Ajay S. Satpute]

Dear Logan,

 

Why not take some numbers (flow rate, pressure, pipe length, elevations, pump curve  etc.) and simulate the pumping system in hysys or excel sheet? I'm sure it would be a great fun when you solve this mystery yourself using your own calculations.

 

Thanks and regards.

 

Ajay S. Satpute

[fallah]

Or will the increase in velocity head and friction head due to the increased flow rate be less than the original decrease in flow resistance, resulting in a net decrease of system head?

 

Hi,

 

Yes, the increase in friction loss due to velocity increase would be less than the reduction of pressure drop due to throttling valve opening; resulting in a net decrease of pump's differential head...

[fallah]

Also, digging deeper into my first issue, can there ever be a case where you throttle your discharge valve, but your friction head will decrease? I understand that by throttling the valve, there is an extra pressure drop within your discharge line, but the throttled discharge valve results in a decrease in velocity. Therefore your skin friction along the length of your pipe will dramatically decrease.

 

Hypothetically speaking if you throttle your discharge valve to half open, the k frictional loss value will increase, which will increase the pressure drop across the valve. I guess what I am asking is that can the resulting decrease in skin friction head due to the decrease in velocity ever be greater than the increase in friction head due to the throttled valve? Say if it was a gate valve and had a low k value? Which would then result in a net decrease in friction head? Because if that were the case, then the system head curve would not shift to the left, even though your flow rate is decreasing. Then the system head curve would not intersect the pump curve. I guess I am just having a little bit of trouble understanding this. Any help is greatly appreciated.

 

Hi,

 

Let's go ahead via pumping system curve analysis after throttling as follows:

 

Please focus on the system curve equation and draw/imagine this curve for a pumping system. Then, keeping constant all other system parameters, a little bit throttle the discharge valve and start to draw/imagine the new system curve after throttling. Starting from zero flow on H axis of the H-Q system if after throttling you would want to have the same flow rate for a point on the original system curve, you must have higher differential head than before means the new point would be at top of the original system curve. On the other hand, if you would want to have the same differential head for that point on the original system curve, you must have lower velocity (flow rate) than before means the new point would be on the left side of the original system curve. Then throttling the discharge valve, in any case, will result in the new system curve intersects the pump curve at the left side of the original system curve.

[Logan@adm]

Ok thanks for the input, i appreciate it. But in the case where the valve is throttled, where does the increase in system head come from then? I understand that the increase would normally come from the increased resistance due to the throttled valve, but what if the decrease in skin friction head due to the decreased velocity becomes larger than the increased resistance due to the throttled valve? This would result in a net decrease in friction head. So where does the increase in system head occur? Would the pressure at the end of my discharge pressure just increase? Leading to an increase in system head? 

[paulhorth]

Logan,

Imagine that your pump had no discharge piping at all downstream of the valve - that it simply discharges to atmpsphere, like your bath tap. Then all the head in the pump is taken up across the valve - you can see this I hope? As you open the valve, the flow increases and the DP across the valve decreases but is still equal to the head of the pump. The pump discharge pressure goes down the curve as the valve is opened, whie the pressure downstream stays at atmospheric.

Now, imagine adding a meter or so of piping on the discharge. Now, for the same flow, you have to open the valve a tiny bit more, and the DP across the valve is a tiny bit less due to the friction loss in the 1 m of piping. The pump discharge pressure stays the same, for the same flow.

All OK thus far?

Now, increase further the length of discharge piping, in small steps. At each step, you have to open the valve a little more to get the same flow. The DP across the valve becomes smaller each time you lengthen the piping but the pump discharge pressure stays the same.

With sufficent steps you will end up with the actual length of piping that you require. Now, for the same flow, the valve is open to its required opening for the system, and the pump DP is shared across the valve and piping ( and any static head). End of problem, with no difficult point to understand.

 

Of course if the valve is too small, it will be fully open without reaching the required flow. If it is too large, it will be only cracked open, which makes controlling the flow difficult. A good rule of thumb says the valve should be sized to be about 70% open at the normal flowrate with the assumed system losses. If, when built, the system losses are different, then the valve opening will change until a new pressure profile is reached.

I don't see what is problematic with this concept.

 

Paul

[Logan@adm]

Ok thanks for the comments guys, i appreciate it

[Logan@adm]

Hey guys I guess I do just have one more question, which I think I know the answer to but just want to double check. Say I have a centrifugal pump that is pumping water to be tied into a cooling water return line. Will the pump create enough flow so that the pressure at the end of my pump discharge pipe line will be approximately equal to the pressure of the returning cooling water? Or will it produce a smaller amount of flow at a higher pressure at the end of my discharge pipe that will be in correlation with the pump motor's hp? The reason that I ask is I currently have a 3 hp motor on a pump. From the pump curve the total dynamic head (tdh) is 115 ft at about 67 gpm. From my friction and static head, my calculations indicate that at that tdh/gpm ratio, the pressure at the end of my discharge line will be approximately 33 psia. However, the estimated cooling water return line pressure at the tie-in point is about 20 psia. So will the pump produce more flow therefore lowering the pressure to about the pressure of the cooling water at the end of the discharge line? Which could potentially overload my motor? Or will it act in accordance to the motor capacity and produce the higher pressure at the lower flow rate? This is all assuming that there is not any throttling occurring within the discharge line

[Steve Hall]

Logan,

 

Due to conservation of energy, the pressure at the junction is a singular value. That means that whatever streams come together at the junction their pressures are the same. When you have a circulating flow - your cooling water return line - then suddenly introduce a new flow into it - your questionable line - then the flow rate of BOTH streams might adjust to achieve the required balanced pressure. I say both streams assuming your cooling water return flow is induced by a centrifugal pump, or even gravity flow. The combined flow rate from the two sources will develop increased pressure drop from the junction point to the terminus of the line.

 

Now, your question is about the pump you've described in this thread. It MUST and WILL ride its curve. And so will the pump in the cooling water return line. To solve the problem you have to perform trial-and-error calculations, working backward from the terminus, to find the flow rates that create the correct pressure profiles and also intersect the pump curves.

 

I've uploaded some relevant pages from my book that might help - see http://www.cheresour...-pressure-drop/ - although your specific situation isn't described perhaps the explanations of pump and system curves will help with your understanding.

 

Steve