Bent Orifice Plates 3

[engineeringguy]

Is there a correlation that can determine the error in an orifice plate reading after it has warped? For example, if I use an orifice plate to measure steam, then syrup (or heavy crude) and the viscosity of the syrup warps the plate, is there a method to determine that the flow rate is within some sort of window? (E.g. +/- 50%)



"If experience was always the only factor, how would we get to the moon?"

http://www.piping-designer.com

 

[JLSeagull]

Whether viscous crude or steam, high pressure differential pressure or some kind of slug may have warped the plate. I don't know of any compensation for a warped plate. I suspect that the accuracy is still much better than 50% at full scale.

 

[vpl]

Is there any way that you can put a portable ultrasonic meter on the pipe and determine experimentally the correlation you need?

Patricia Lougheed

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[zdas04]

Change the damn plate. The assumptions that underlie the AGA-3 calculations are so sensitive to plate configuration that a backwards plate will give you upwards of 26% downward bias, a nick in the leading edge can be 10% down. A bowed plate?

There are too many variables in a distorted plate (how bowed is it?, is the distortion symmetrical about both the vertical and horizontal axes?, did the event that damaged it also nick the leading edge?, etc.) that I've never seen an estimate of the amount of bias it creates.

If you need the bias to create an adjustment, then you need a point of reference. The reference point can be the sum of the upstream meters or the sum of the downstream meters, or a portable prover, or a strap on Ultrasonic.

David

 

[engineeringguy]

I know the problem is a bowed plate (actually several hundred of them). There is an inherent problem with the original design that I'm also charged with fixing. I'm not trying to compensate for the flow rate. That's next to impossible, especially what zdas04 mentioned. I'm trying to solve a larger problem.

The plate controls a steam flow control valve for injection. Despite having pressures where they need to be, the valve isn't opening enough to get the desired flow rate through. Either the valve isn't sized properly or the its not being controlled properly. After review, the valve is sized properly. Next step is to replace with a known good plate and try to replicate the problem.



"If experience was always the only factor, how would we get to the moon?"

http://www.piping-designer.com

 

[zdas04]

I'm not sure I'm completely understanding your applicatiion. In similar kinds of control scenarios I look at the delta between what I want and what I'm getting and then nudge the valve in the direction that it needs to go, wait a minute and check it again. In this scenario the meter is telling what direction and how much to nudge the valve so the absolute value of the flow rate tends to not be as critical as some applications (if you meet your target and downstream processes are starved then bump up your target).

It kind of sounds like you have a table in your algorithm that says for 15% open I will get "x" flow and for 40% open I will get "y" flow. If that is the case, then good luck. That is more reliance on the lineararity of a valve than I've ever been able to accept.

David

 

[dcasto]

CIG use to have a nice table of common orifice plate issues and an estimate of the errors


Guidelines For Measurement Errors Caused By Buckled Orifice Plates

http://findarticles.com/p/articles/mi_m3251/is_7_235/ai_n27985270/

 

[hacksaw]

let me get this straight, you have several hundred bent plates in service, or several hundred plates?

what dimensional atandard for the plates are you using, and what range of delta-P?

 

[engineeringguy]

Yep. Maybe several hundred is a bit of an exaggeration but a couple hundred isn't. They meter their steam on every cyclic well and selected orifice plates as the technology to do it. I've decided to take a more pragmatic approach and manually control the valve to 100% and measure the output flow rate on a troubled orifice, then change the orifice, repeat.

The end result isn't something to the accuracy of a University study or development of a set of equations but being able to go to the client and say (with reasonable certainty) "Yep, the problem's here; now here's how we're going to fix it" Thanks for the help





"If experience was always the only factor, how would we get to the moon?"

http://www.piping-designer.com

 

[hacksaw]

you need thicker plates in geothermal or enhanced recovery wells, you also need to consider eccentric plates or use drain/bleeds in you plates.

this is not a service for the traditional 1/8" plate

 

[JLSeagull]

I agree with thicker plates. The standard orifice flow element tables apply with pressure differential less than 1000 inches water column. You may want to contact a few orifice plate peddlers and select one that will provide deflection and thickness calculations in addition to the bore to accommodate your maximum pressure differentials.