Q: have a question about orifice flowmeters: Why do we put it before the control valve (CV) and not after it (upstream of the CV, not downstream)? This seems to be the case in the refineries around 98% of the time. I need to know if there are standards that require designers to put the orifice in this position or could we put them after the CV if there is enough length of pipe? If there is enough straight run, does it matter? Thanks for your help.
Sadeem Al-Baghdadi / firstname.lastname@example.org
A: If you must, you can place a valve before an orifice flowmeter (for example, after a subheader on-off ball or gate shutoff valve). But there are many reasons to prefer upstream placement.
The flow path is tortuous through a control valve, causing the Reynolds number (Re) to vary. Therefore, the fluid will need to pass through a long straight line before stabilizing. The throttling also causes variation of the pressure of the flowing fluid, which changes the orifice coefficient. Also, in liquid applications, if the vena contracta pressure drops below the flash point of the liquid, flashing occurs that makes the measurement meaningless.
We want the velocity profile (the flow regime) at the orifice plate to be stable and the same as it was during testing of the plate. The orifice coefficients are determined using long, straight pipe runs during testing, and therefore, if we want to keep the orifice coefficient correct, we should reproduce the tesing conditions, both in terms of straight runs and in terms of Re. The Re must exceed 20,000 even at minimum flow to guarantee turbulent conditions.
Control valves are usually one size smaller than the pipe and therefore, the expander after the CV adds to the straight run requirement. By the way, over the years I did not find much value in flow straighteners.
The straight run recommendations vary widely. When the orifice-to-pipe-diameter ratio β is 0.8, I use 50 for reducers and throttling valves, and 40 for elbows in two planes, but Spink, for example, recomends 60 to 100. If all you need is repeatability and are not concerned with accuracy (the measurement is not for accounting purposes), the length of the straight runs can be less than in Figure 1. Otherwise the longer the run, the better the accuracy. With flange taps, the downstream straight line requirements (6-8) are unaffected by the upsream disturbances, but if pipe taps are used, they require twice that length (12-16). Flange taps are used above 2-in. pipe size and corner taps are used below.
There are really no such things as standards. There are recommendations, and they tend to disagree. Probably the best is ISO5167, Parts 1 and 2, which gives similar advice to the recommendation in my handbook (Figure 1).
The American standards AGA and API tend to be a little more conservative than ASTM. Of the standards of other nations, the German and the French are probably the most conservative. I also found over the years that vendors’ recomendations can be highly unreliable.
Naturally, there is more to installation than straight runs, and there is more to orifices than installation. For a full treatment of selection, sizing, etc. see Chapter 2.6 in my handbook.
Béla Lipták / email@example.com
A: If you don’t have access to a copy of the relevant volume of Liptak’s Instrument Engineers’ Handbook, "Process Measurement and Analysis," see the international standard ISO5167 Part 1 and 2 . Note that this is not necessarily conservative.
The API standard (API MPMS 14.3 ) allows the use of 19-tube flow straighteners ahead of the orifice, but these have been demonstrated to be a cause of error rather than an improvment.
Ian H. Gibson, Process, control and safety engineering consultant / firstname.lastname@example.org
A: Yes, indeed, you can make a case to place the orifice downstream of the control valve for liquid service where there are sufficient straight-run lengths. However, this is more complicated for gas service, where you will have a variable pressure/density downstream of the CV. You can use a multivariable DP transmitter to compensate.
Straight run rules are meant to be applied in the majority of cases without doing analysis (rules for everyone). One can always break the rules if you know the basis and what further analysis is required.
In the case of orifice placement, it is not just about straight run lengths.
Simon Lucchini, CFSE, MIEAust CPEng (Australia), Chief controls specialist, Fluor / Simon.Lucchini@Fluor.com
A: Yes, you can put the orifice plate downstream of the control valve, provided you have the upstream and downstream lengths available to give you a stable flow regime, where the turbulence caused by the tortuous flow thorough the control valve stabilizes into stable flow. The ASME standard assumes a stable flow, and flow regime in a range of Reynolds number, on which the flow coefficients are based. Therefore, for flow computation to be valid, you have to be in that flow regime.
You can send me the sketches/iso-meteric drawings and I can review it for you, at no cost.
Romel S. Bhullar, P.E. ISA Fellow , Senior technical fellow/director, control systems, Fluor / Romel.Bhullar@Fluor.com
A: The orifice coefficient research is done with long, straight, run of pipe because this is the only arrangement that can be duplicated. Anything less than that is likely to have a nonstandard flow pattern, and it will affect that coefficient. See the standards or a good flow handbook to see what is required for good accuracy. A partly open control valve will certainly create a complicated flow pattern.
Cullen Langford P.E., Member, ASME Main Flow Committee / CullenL@aol.com
A: The reason you should not place an orifice plate downstream of a control valve is simple: The pressure downstream of a control valve will vary as the valve opens and closes. If the orifice plate is upstream of a control valve, the pressure is always constant, which allows for correct measurement.
So, as indicated before, no matter how many downstream straight runs you have, the pressure will constantly be modified if the control valve is upstream, causing measurement errors.
Alex (Alejandro) Varga / email@example.com
A: I suggest that you read the ISO5167 Part 1 and 2 standards, which may answer most of your questions about how to install the orifice flow meters. Another good reference is the 2013 Chemical Processing article, "Think straight about orifice plates.”
Regarding the second part of your question, control valves are generally one size smaller than the pipe size. To minimize pressure losses, pipes are restored to their normal size downstream of the control vale. If you install the orifice plate downstream of the control valve, then the reduced pipe size will remain in place for ~ 50 pipe diameters. That is a long run of reduced pipe size, which will increase your pressure loss in pipe. This pressure loss has to be compensated using hydraulic analysis. So why would you spend extra money, when you can achieve the same effect by placing the flowmeter upstream of the valve?
Raj Binney / firstname.lastname@example.org
A: Measurement of liquid flow with an orifice flowmeter depends on the liquid remaining totally as a liquid and not as a vapor. At the vena contracta following the sharp-edged orifice, the pressure drop will be a maximum (minimum pressure.) We carefully evaluate the temperature and pressure of the flowing liquid stream and size the orifice so the pressure at that vena contracta will not be close to the liquid boiling point. If boiling occurs, we call that flashing; therefore Bernoulli’s law no longer applies and the relationship between pressure drop and volumetric flow rate becomes unknown—you can’t use orifice flowmeters to measure the flow of flashing liquids. By locating the orifice plate/flowmeter before the CV, any pressures close to the vapor pressure are shifted downstream of the control valve, rather than downstream of the orifice.
Richard H. Caro, CEO, CMC Associates / RCaro@CMC.us
This column is moderated by Béla Lipták, automation and safety consultant and editor of the Instrument and Automation Engineers’ Handbook (IAEH). If you have an automation-related question for this column, write to email@example.com.