Orifice plate high outlet velocity 2

[mjpetrag]

I am sizing a bypass line for a pump where I have already calculated the required square edged orifice dimensions for the flow and pressure drop I need. However, the outlet velocity of the orifice is extremely high. I need about 30 GPM of flow through a 3/4" SCH 40 pipe with a pressure drop of about 50 psig discharging into an atmospheric pressure vessel. The orifice diameter is 0.45". The inlet velocity is about 17.5 ft/s and the outlet velocity is around 60 ft/s right out the orifice center. This high velocity is about 2 ft away from a 90 bend back into the vessel.

Is there any way to size the orifice by reducing the velocity to safe operating limits mitigating against erosion and hammering effects? I am also worried about the deformation of the orifice plate due to these velocities.

Do I even need to worry about these effects? This is my first time doing the calculation so please understand if the question sounds naive.

-Mike

 

[Hydromechdude]

Would it be possible to add a second restrictor in series with the first one? This would reduce the pressure drops through the individual restrictors.

What is the system pressure held at?

 

[mjpetrag]

System pressure is the pump discharge pressure which hovers around 50-60 psig. The bypass line off the pump discharge goes into an atmospheric pressure tank.

-Mike

 

[ione]

Erosion could become a serious issue here.

When fluid passes through an orifice, pressure drops and velocity increases, and these two phenomena go together (Bernouilli principle).
In your first post the value of 60 ft/s is in the vena contracta, that is in the throat section. Beyond this section the fluid expands again and its velocity decreases. Using a 1” pipe and for the required flow (30 gpm) velocity is about 12 ft/s (3.7 m/s), which is still a bit high. With a 2” pipes velocity drops to about 3 ft/s. Note it takes the so called “development length” for the fluid to become fully developed. For turbulent flow (and this your case) it has been suggested a length of about 70 pipe’s diameter is required (Perry, A. E. and Abell, C. J., Scaling laws for pipe-flow turbulence, Journal of Fluid Mechanics, 67, 1978, 257–271).

Final note: using a 3/4” pipe, upstream the orifice plate, the fluid velocity (for a flow of 30 gpm) is kind of 22 ft/s, that definitely is a too higher value.

 

[mjpetrag]

Thanks Ione. Just to confirm I'm doing this correctly, I will increase the pipe to 1" SCH 40 with a .4" diameter orifice.

dP across orifice = 53 psig
Discharge coefficient = 0.59 from the Crane book
23 GPM across orifice
Inlet velocity = 8.6 ft/s
Outlet velocity = 59 ft/s
Permanent dP loss = 83%
The length of the pipe after the orifice will be at least 70" then.

What happens to the 17% pressure that gets recovered downstream if it is going into an atmospheric vessel?

-Mike

 

[ione]

First of all a dp is never a relative pressure so no psig but just psi (or equivalent unit).

I have checked this separately (water at 20°C) and have found results comparable to yours:

1. V2 = 55 ft/s (outlet velocity)
2. Q = 27.3 gpm (flow)
3. C = 0.6 (coefficient of discharge)

Your question puzzled me a bit: in effect you do not have any pressure gain or extra pressure. This is not the way things go.
When sizing a pump, one has to compare the performance curve of the pump (pressure vs flow) with the pipeline curve. The working point of the pump is given by the intersection of the pump curve and the pipeline curve. So your pump will give a certain flow for a certain pressure drop determined by the intersection of the two curves afore mentioned. So for a given pump changing the pressure losses (that is changing the pipeline configuration) will determine different flows.
Please consider that the fluid has to face other pressure losses to reach your open vessel (both distributed pressure losses due to friction over the pipe length and minor losses).

 

[Latexman]

The 17% that is "recovered" downstream of the orifice will be lost due to friction of the pipe and fittings between that point (downstream of the orifice) and into the tank.

Good luck,
Latexman

 

[mjpetrag]

The pump is oversized and the process limits how much flow can go downstream to the main line. So the pump is throttled back too far. I looked at the pump curve for where it should operate, found that pressure and flow, and used the extra flow needed to go to the bypass line. The pump discharge head would drop for the higher flow and that was how the pressure was found for the upstream orifice conditions. Please correct me if I'm wrong.

-Mike

 

[Latexman]

Thats about right, but will the flow to the main line decrease due to the lower head, or is it flow controlled? The rigorous answer is iterative if the main line is not flow controlled. Depending on the characteristics of the pump curve and system curve (bypass included) the difference in your approach and a rigorous approach may or may not be significant. Plotting the system curve onto the pump curve will show you.

Good luck,
Latexman

 

[ione]

It seems Latexman and I are both tuned on the same wavelength.
When dealing with oversized pumps that work at constant speed, bypass is one of the methods available to regulate the discharge capacity and so change the flow in the main line and consequently the pressure drop and consequently the working point of the pump. Do not treat the by-pass line as a separate entity (if you know what I mean). Consider the overall system curve without splitting apart the bypass and check with your pump curve the working point. This is at least the way I would approach the problem.

 

[davefitz]

Other solutions are possible.

For example, relocate the orifice plat at a flange directly on the vessel nozzle, so the high velcoity is to a large reervoir instead of a pipe elbow.

Alternately, consider the use of a long, small diameter "capilary tube" , SS about 5/8" nominal and XXX long in lieu of the orifice, with final discharge either to a "erosion tee" or into the large reservoir.

 

[mjpetrag]

Got ya. The pump is throttled by a valve controlled by the level of the atmospheric tank feeding the pump. We'll most likely have to play around with the control valve settings to get the right flow but the discharge conditions stay relatively constant.

-Mike

 

[btrueblood]

Dave's idea is a good one. You can also adjust (increase) the drop in a capillary tube by adding a few bends.