relatively thick orifice

[dbachovchin]

I want to calculate pressure drop across a small (0.038 in), non-sharp-edged relatively thick (0.060 in) orifice. Flow is not choked, but Reynold number is high. There are big pipes on both sides.

Using Perry, discharge coefficient is 0.61. This corresponds to 2.7 velocity heads.

If, instead, I treat it as a sudden contraction and a sudden expansion, then I have 1.5 velocity heads, which corresponds to a discharge coefficient of 0.83.

Which is correct? Is the orifice so thick that the flow attaches? Does this result in a higher discharge coefficient? Is it generally true that if a small pipe is used as an orifice then the flow will go up for a given pressure drop?

 

[hacksaw]

you need to clarify your comment about the edge profile such as its radius of curvature.

 

[dbachovchin]

It is sharp edged. The hole is just a drilled hole through a sheet of metal.

 

[hacksaw]

if you are using the device as a flow restriction then you would normally assume a worse case discharge coeff ~0.63 to calc. the thick plate over all pressure loss case.

Perry's figures are for flow measuring devices for specific tap locations. When you take the sudden contraction / expansion case you are estimating overall pressure pressure loss.

if you are using this orifice as a flow meter, then you need to decide where your pressure taps are located so you can determine the discharge coefficient from available correlations. The ASME has meter correlations for small bore orifices.

The data you need is generally available in handbook form and in some cases on the web, but you'll need to determine the specifics of your orifice configuration: location of pressure measurement reference planes, the i.d. of the adjacent pipe, definition of your application as a flow estimate for a given pressure drop and fluid, its use as a flow restriction, or its as a flow measuring device, etc.

 

[dbachovchin]

To be concise ---
How do I determine the discharge coefficient for a thick (L/d =2 ), non-critical flow orifice?

 

[hacksaw]

The short answer is that your discharge coefficient will be a function of the location of your pressure measurements.

At issue is the amount of stagnation and pressure recovery at the pressure measurment locations.

At the risk of stating the obvious, this dependence is clearly visible in the coefficient graphs contained in Perry's Handbook.

There are at least a half-dozen standard pressure tap locations (close-up,flange,radius,pipe,vena-contracta, and over all pressure loss). So your question has at least six answers.

If you have a configuration that cannot be traced to actual flow testing, i.e. standard configurations and hydraulic similarity, then handbook data is the wrong place to start.

You may have to perform you own coefficient measurements. That is a subject beyond the capability of this forum.

Good luck,

 

[dbachovchin]

The pressure drop depends on both the position of the taps and the thickness of the orifice. I know how to account for pressure tap position, such as by using charts in Perry. But all of these are for thin (sharp-edged)orifices. What is the effect of orifice thickness? Where are there charts for thick orifices?

 

[BobPE]

Whay would you think that is a thick orifice. A thick orifice is called a tobe and there are formulas for tubes. You orifice is not a tube.

BobPE

 

[hacksaw]

Once you get to away from the configurations that have been subjected to careful experiments all you can do is estimate it.

The most common assumption os to use Cd~0.60-0.63 or so or consider using a tube.

You can perform a CFD calculation in lieu of an estimate or lab test.

With small orifices, even if you do characterize its discharge coefficient, with time the edge will lose its sharpness and you have to anticipate what affect this will have.

There are mass flow controllers for lab gases etc. You'll need one to calibrate the orifice you are designing.

Perry's figures are not for small diameter plates nor are they particularly useful for detailed calculations.

Good luck,