pressures up/downstream of butterfly valve for reservoir service

[bellmouth]

I am selecting a butterfly valve to be installed on a submerged pipe intake at the bottom of a reservoir, and need to determine the upstream and downstream pressures of the valve so the actuator torque can be determined. The valve is a shutoff valve for the pipe, that when opened allows water to flow to a pump station. The valve is a 48" diameter valve, and the desired water velocity through the pipe when the valve is opened is 15 feet per second. The valve is located at 100 feet below the reservoir surface. If there is a pump at the downstream end of the pipe, do I factor that atmospheric pressure is just on the upstream side of the valve, and include that as part of the upstream pressure? Or assume air pressure on both sides of the valve and neglect it in the calculation, and just use the reservoir height for the upstream pressure?

 

[BigInch]

15 psia + 100 ft * 62.4 pcf / 144 in2 = 58 psia upstream

You should not let your pump suction get below its NPSHr (net positive suction head required) and should be shutting down the pump at that suction head. Say NPSHr equals 20 ft.

Neglect suction velocity head and calculate pressure as,
Min Pump Suction Press = 20 ft * 62.4 pcf /144in2 = 9 psi

Therefore you can have a maximum differential pressure across the valve of 49 psi. Use that differential pressure to calculate your flow through the valve.

"Make everything as simple as possible, but not simpler." - Albert Einstein (1879-1955)
***************

http://virtualpipeline.spaces.live.com/

 

[bellmouth]

Thanks BigInch. You calculate the maximum pressure drop across the valve as 49 psi using NPSHR. Is another way possible to calculate maximum pressure drop across the valve by using velocity pressure and the head loss coefficient K value [K(v^2/2g)] at a valve minimum (cracked) open position? 15 feet/s velocity in the pipe is the desired fluid velocity. The K values approach exponential as the valve closes though so I'm not sure how that would work.

For calculating flow rate, I would use Q=Cv*sqrt(dP), with Cv the flow coefficient of the valve near the closed or at full open position?

Also, the valve manufacturer has requested the maximum downstream pressure for their torque determination. The minimum downstream pressure you established as 9 psi. For maximum downstream pressure, I assume a minimum pressure drop across the valve, should I be using [K(v^2/2g)] for K at full open to determine the minimum pressure drop, and subtract that from the upstream pressure to get the minimum downstream pressure?

 

[BigInch]

Assuming you don't shut down in time and the pump can draw a total vacuum then the absolute max dP is 58 psi.

Use Q = Cv * dp^0.5

So at 15 fps in a 48" d, your target Q = 188 cfs.
You're only going to get 15 fps in the valve and in the pipe when the valve is at full open (assuming the valve is a full bore valve). Your velocity inside the valve will tend to increase as you close, but frictional head loss will increase too, making a net gain in dP. Result is that the velocity will drop and Flow will drop as it enters the downstream pipe. Since Q has dropped to less than 188 cfs, the velocity in the 48" pipe will drop to less than 15 fps.

The valve manufacturer should have requested the maximum upstream pressure and the minimum downstream pressure for his torque calculation. In other words, the highest opening/closing torque is found when the differential pressure across the valve is at a maximum.

On a valve that size, make them do a shop test of the assembled valve and actuator assuring that the torque produced is sufficient to both open and close the valve at the specified maximum differential pressure.

"Make everything as simple as possible, but not simpler." - Albert Einstein (1879-1955)
***************

http://virtualpipeline.spaces.live.com/

 

[bellmouth]

Great thanks! By the way if I use the Q=Cv*dp^0.5 equation with the Q at 188 cfs (or 84,380 gpm), and a valve Cv of 2.0 (from the mfg catalog at full open position) I end up with a dp of 1.78 x 10^9 psi. This doesn't make sense, I need a huge Cv just to get in the range of a 49 to 58 psi drop.

 

[BigInch]

Cv of 2?

You have to be careful with Cv and some mfgrs. Some have their own set of units. I'm already wondering why he wanted the maximum pressure downstream to do torque calcs.

2 GPM/psi^0.5 would be something smaller than a 1/2" valve. You need about 15,000.

"Make everything as simple as possible, but not simpler." - Albert Einstein (1879-1955)
***************

http://virtualpipeline.spaces.live.com/