Best practice for flow measurement

D

Thread Starter

DAVIDRAJ

I have a question. like for certain flow measurement what would be the best practice? like Control Valve should be first and then Flow meter or reverse?

clarify
 
A
There is no matter to used control valve or flow meter because both are in series loop. But mostly people used control valve first.
 
For water applications, I usually place the control valve downstream of the flowmeter, because in that location, the valve will not be able to disturb the flow profile before the water hits the meter.
 
> There is no matter to used control valve or flow meter because both are in series loop. But mostly people used control valve first. <

I would disagree.

It is usual to expect some pressure drop across a flowmeter and the usual recommendation is to ensure sufficient back pressure to prevent cavitation or vaporisation/gas breakout in the meter.

Control valves usually generate significant pressure drop and are best downstream of the meter as part of the back pressure contribution.

They can also be a source of gas bubbles or cavitation downstream because they are often the cause of a higher pressure drop than the meter. SO if upstream of the meter they would tend to create gas bubbles or cavitation which would flow into the meter. Some meters require a particular flow profile and valves are a source of turbulence.
 
W
I'm sorry, but I can't let this go.

You are unfortunately dead wrong.

Hydraulically, putting a control valve upstream of a flowmeter can decrease the accuracy and repeatability and linearity of the flowmeter up to 40-50% of reading, and sometimes make it unusable.

For example, a fully open butterfly valve spawns Von Karman vortices that propagate downstream for as many as 50 diameters or more.

If you are going to locate a flow meter downstream of a control valve (or any valve that produces an obstruction in the flow) you should either put a flow conditioning device upstream of the flowmeter or use a flowmeter with a built in flow conditioner (V-cone, swirlmeter, PD meter, etc.) Even magnetic flow meters are susceptible to this error.

Put the damn valve downstream of the flowmeter (at least 10 diameters) or you'll have more trouble than you want.

Walt Boyes
Editor in Chief
Control and ControlGlobal.com
555 W. Pierce Rd Suite 301
Itasca, IL 60143

[email protected]
www.controlglobal.com
630-467-1300
 
R

Rohit Chandak

Installing Control Valve downstream of the flow meter not only helps good flow measurement by avoiding velocity profile disturbances but also helps doing a good control loop application. As control valve takes signal from flow meter to operate. More is better but at least have 5D downstream of the flow meter & upstream of the control valve.

I fully subscribe to what Walt has suggested. There are other metering technology which are proprietary DP Flow Measurement like ACCELABAR, V-CONE which could be a good choice on such metering system.
 
> I have a question. like for certain flow measurement what would be the best
> practice? like Control Valve should be first and then Flow meter or reverse?

i doesn't matter either you put your control valve donwstream or upstream, you just need to satisfy the straight-run requirement of your flow element
 
P

peter walker

Walt

Some advice please.
It is common practice to estimate flow rate from control valve position where a dedicated flow meter is not available.

I need to estimate the uncertainty of flowrates derived from this process.

I consider there are four parts to this uncertainty :-
1) the uncertainty in the stated CV of the valve, and its characteristic.

2) the uncertainty of the valves position (i.e % lift)

3) the uncertainty in the pressure / temperature measurements (i.e the secondary instrumentation)

4) the uncertainty of the process data, density viscosity etc.

could you comment on the 1) and 2) above please.

For 1). What would be a typical manufacturing uncertainty of say a simple globe valve with a contoured plug, or a cage type trim. how do they perform against catalog data when tested in a flow lab?

For 2). What uncertainty would you expect from a well maintained valve and positioner assembly (I am inclined to use +/- 1%)I may be optimistic here as I appreciate valves in the real world can be much worse than this, but suggest that +/- 1% should be achievable. If memory serves me well I thing I have seen specs of +/- 0.25%, but this seems very optimistic!

For 3) and 4) I can estimate these myself.

The applications would be gas flows to a flare system, so P2 is atmospheric or very near, and could be assumed constant.

Many thanks

Peter
 
Top