Pressure Drop across an Orifice Plate

[sports2012]

Hi All,

I have a question regarding non recoverable pressure drop through an orifice plate. Below is my scenario but with simplified (made up) numbers. Any help would be much appreciated.


I have a centrifugal pump, with a minimum continuous flow rate of 10 gpm. I need to design a re-circulation line that will open to protect the pump for when the main process line is shut off. This will ensure that at least 10 gpm of fluid is always being pumped. My main process line uses 6" piping, and my re-circulation line uses 3" piping. My pump discharge pressure is 100 psi while operating at the 10 gpm. The tank the bypass line feeds into has a back pressure of 10 psi on the bypass line. I can design any orifice plate I want, to generate whatever pressure drop is needed to create false head against the pump (so the pump works at the desired 10 gpm).

My question is, what pressure drop do I want out of my orifice plate? Should it be 90 psi? Because I am going from a 6" main line pipe to a 3" pipe in the re-circulation line, the fluid velocity will increase. My pressure should therefore decrease on the inlet side of the orifice. If this is correct, and we assume 3" pipe has an initial 80 psi just before the orifice plate, and 10 psi after the plate, well now my pressure drop across the orifice is only 70 psi.


So which is it? Should the CV of the orifice be designed for a 90 psi pressure drop, or a 70 psi pressure drop?

I think my confusion goes back to not understanding the difference between recoverable and non-recoverable pressure drop.

 

[BigInch]

100-10=90
some of that 90 psi pressure drop may be recovered after the orifice. You may want to put a globe valve in there between orifice plate and tank to manually adjust pressure into the tank if necessary.

you must get smarter than the software you're using.

 

[GHartmann]

You can do a more accurate calculation of the drop in the 6" line and into the 3" line. However the pressure drop from 10 GPM of flow in a 3" line is going to be pretty small. You are not likely to have 10 psi of drop from the pump to the orifice.

I would therefore use a pressure drop of 90 psi. Then the equations you use for the orifice size should take into account what part is not recoverable.

Even if you have sized the orifice too big, creating a flow greater than 10 GPM (wherever the hydraulics balance out) and protecting the pump. But you still might want the globe valve to maintain your max flow from the pump to its destination (i.e. 10 GPM of minimum flow plus your required flow). Under this case the pump head will decrease your dP available will be less than 90 psi and the flow through the kick back line will be less than 10 GPM which is OK. (Unless you have a shutoff valve on the kickback line).

 

[Compositepro]

Why would you not use a back pressure control valve rather than an orifice plate? This will allow for recirculation when the pump is deadheaded but otherwise stay closed, thus not wasting energy continuously though an orifice.

 

[katmar]

The velocity for 10 gpm is a 3" pipe is less than 0.5 ft/s. Your pressure drop due to friction and velocity changes (Bernoulli) will be negligible. Use the full 90 psi to size the plate, but as stated by GHartmann make sure that you use the correct equation to calculate the overall pressure drop and not the vena contracta pressure drop that the instrumentation people would use.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[GHartmann]

The reason you might NOT want to use a back pressure valve for controlling the minimum flow is that most pump curves are pretty flat towards their shutoff head at the minimum flow requirements. Thus you can have a broad area of flow from the pump that satisfies the pressure or head requirements, but not the flow requirements. You would need to have a really good idea of your system resistance curve and where it might be relative to the opening of the kick back line.

The other reason might simply be cost of the installation.

There are available automatic recirculation valves which operate on the principle of pressure (against internal spring ARC) which are used commonly for boiler feedwater pumps.

 

[ione]

I agree with comments above about differential pressure to be used for sizing the orifice plate. In order to get such a differntial pressure (90 psid) with an orifice plate fitted in a 3" pipe and for a flow rate of 10 usgpm, you need a definitely small bore (small beta ratio) so the pressure recovery after vena contracta should be something negligible.

 

[sports2012]

Thank you all for the replies. What would an acceptable beta ratio be if I want negligible pressure recovery?

So with 10 gpm, the pressure drop due to the line going from 6" of flow to 3" of flow is negligible. If we say the minimum continuous flow rate was something higher like 100 gpm, I assume I would need to check the fluid velocity in the 3" line to see what kind of decrease in pressure I will experience when going from a 6" pipe to a 3" pipe. This will give me a new orifice inlet pressure, which I could then use to find my new delta P.


And for those wondering, the orifice plate is being used to save money over a backpressure control valve.

 

[LittleInch]

I am no expert in restriction orifices, but you need to be aware of other issues such as vaporisation of the fluid at the critical point in the restriction, which for your example could restrict the pressure drop across one orifice to around 70psi, thus you might need 2 in series or a further variable restriction like a globe valve. If you have vapourisation the collapse of th bubbles you'll make a lot of noise and not have much of an orifice left after a few days... Also be aware of the velocity through the orifice. Once you get past 10 m/sec you can get either quite a lot of noise and some significant erosion hence your RO won't be working the same after a relatively short time unless you make it out of something very strong. At low actual diameters of the orifice itself (anything less than 5mm or so), the risk of plugging with grit / debris increases. So sometimes it's the practical side that causes issues, not the theoretical ones.

I can't see your concern over sudden contraction being a issue at those sorts of relative numbers. If though you were only looking for a pressure drop of say 5 psi, then it might become more relevant - the devil is in the detail.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Duwe6]

"If you have vapourisation the collapse of the bubbles you'll make a lot of noise . . "
And that cavitation will cut through the pump impeller, and possibly the actual casing, in short order. Continuous cavitation is amazingly destructive.

 

[e43u8]

To respect definition of terms it's better such orifice to be titled "Restriction Orifice (RO)" rather than "Orifice Plate" which is to be used in flow measuring...

In the returning back lines such as minimum flow for pumps or anti surge lines for compressors the differential pressure across the orifice or control valve is to be considered per presuures in main discharge line and in destination, either suction line or suction drum, at the moment of returning back activation for full flow because the minor flow restriction due to upstream/downstream lines, if any, would be compensated by the overdesign factor has already been considered in the peak of such minimum flow estimation; then 100-10=90 bar is reasonable differential pressure drop for your case...

As ione mentioned, you need a RO with definitely small beta ratio so the pressure recovery after vena contracta should be something negligible.

 

[sports2012]

Thank you e43u8. What would a small beta ratio be considered? Is anything less than 1 acceptable?

 

[Latexman]

"What would a small beta ratio be considered? Is anything less than 1 acceptable?"

(p1-p4)/(p1-p2)=1-[Β][sup]2[/sup]

p1=upstream pressure tap
p2=downstream pressure tap
p4=fully recovered downstream pressure
[Β]=diameter ratio

[Β][ ] % Pressure Loss
0.1[ ] 99
0.3 [ ] 91
0.5[ ] 75
0.75 [ ] 44
0.9[ ] 19
0.99 [ ] 2

I would say [&Beta;] < 0.45, which gives about 80% pressure loss, but for a specific case one must really run through the calcs to obtain the required permanent pressure loss.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[sports2012]

Thank you Latexman.

If I go with a multi-hole orifice, instead of a single hole orifice, how would this affect the beta ratio? Would I need to convert the equivalent cross sectional area of the multi-hole orifice to a single hole orifice to determine the beta ratio?

 

[Latexman]

Use "Search" between Forum and FAQs in the upper left of webpage and you can find many relevant posts, like this one:

thread378-205170


Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[katmar]

I don't think that a multi-hole orifice plate is a good idea here if you are concerned with plugging the holes with debris. If a single hole is used the orifice bore will be in the region of 1/4", and if more holes are used they will be smaller and more likely to plug. LittleInch suggested 2 or more orifice plates in series, and this would result in each orifice bore being a bit bigger than the case for a single plate.

But even if you use 3 plates in series and allow 30 psi per plate the orifice bores only increase to about 5/16" - hardly any more resistant to plugging. The answer is probably to install a strainer before the orifice with a mesh size of 1/16" or less.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Latexman]

The OP did not mention debris or plugging, LittleInch did. I assumed it was to reduce noise. Which is it, sports2012? Why multi-hole?

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[sports2012]

Yea, there is no debris. The water being pumped should be clean. It is to help reduce noise under the specified limit.

 

[LittleInch]

Latexman - are you sure about those numbers? I agree with katmar and for the OPs (made up) figures of 10 gpm with a 90 psi pressure drop in a 3" pipe I get about a 6mm hole in a 70mmID pipe - beta ratio of 0.08...

velocity about 21m/sec, no idea if it flashes but is a possibility.

I was just pointing out that there are practical issues, potential plugging being one of them and if it happens then your pump isn't doing 10GPM any more it's doing about 2.... I didn't do the calcs, but for that flow rate I agree multi orifices are unlikely to make it any better.


My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Latexman]

It's for approximating concentric circular orifice with vena-contracta taps (Perry's 5th page 5-16). They look good to me, but if you see something wrong, please show us.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.