"Ask the Experts" is moderated by Béla Lipták (http://belaliptakpe.com/), process control consultant and editor of the Instrument Engineer's Handbook (IEH). Now that the 4th edition has been published, I am recruiting contributors to work on the 5th edition. If you can participate, or if you have questions for our team of experts, please write to me: firstname.lastname@example.org.
Q: We have traditionally specified a vortex flowmeter for a particular service because high rangeability and low pressure loss are required. We have had no problems reported by either the vendors (generally Yokagawa) or the plants up to now. However, on the project we are now working, the vendor (Invensys) has come back and said their meter will not work in this application.
The process conditions are
- Line size: 10"
- Fluid: EB gas
- Max flow: 22,000 lb/hr
- Inlet pressure: -11.5 psig
- Temperature: 200 ºF
- MW: 99
- Viscosity: 0.008 cP
- Z: 1
- CP/CV ratio: 1.08
Is there any reason why some vendors' meters would work in this application while others would not? Or is a vortex a poor selection, and we have just been lucky up to now?
On a related note, in Table 2.1b of your Handbook, you show vortex meters as having a high pressure loss, while we have generally found them to have a low pressure loss, at least in the services we use them.
Shaw Energy & Chemicals Group
A: Let us start with the sizing of this meter by converting your mass flow of 20,000 lb/hr into actual cubic foot per minute (ACFM). The specific weight (SW) of one ACF is SW = (99 Mw x 26.2 PSIA)/(10.73 x 660 ˚R) = 0.366 lb/ft3.
Twenty-thousand lb/hr is 333.3 lb/min; therefore, your flow is 910.1 ACFM (333.3/0.366). Keep in mind that this flow is correct only under the conditions you have stated, but if the pressure drops or the temperature rises, this volumetric flow will further be reduced.
Now, if you take a look at Figure 1 (which is Fig. 2.30i in Volume 1 of my handbook), the detectable flow range of a 10-in. vortex flowmeter is about 600 to 8000 ACFM. In selecting the size of all flowmeters, we want to get the normal flow into the center of the range. Therefore, your meter is oversized, as average flow is near the minimum detectable. In other words a 6-in. or 8-in. meter should have been used. For a vortex meter to work, you must have a minimum Reynolds Number of 10,000 (Coanda units go lower).
You say that there is a conflict between the statement in the Instrument Engineer's Handbook that the pressure drop requirement of vortex flowmeters is high, while your experience is that it is low. This oversizing is the reason. The pressure drop produced by a vortex flowmeter is about 2 velocity heads (v2/2g), and because the pressure drop rises with the square of velocity, therefore the ΔP of an 8-in. meter would be 2.5 times, and for a 6-in. meter it would be 7.5 times the ΔP you are experiencing today in your 10-in. flowmeter.
Your main problem is over-sizing. The reason why one manufacturer's meter works under your conditions and the other's does not is because the minimum flows of the different manufacturers' vary. The reason why your supplier's representative did not know why your installation failed is probably because he/she is better at arranging business lunches than at sizing vortex meters. (These days this is often the case.)
By the way, if the size of your meter is less than line size (and it usually is), you need 10 to 15 pipe diameters of straight upstream pipe run in order for the vortices to properly evolve. This is about the same length as required for an orifice installation with a beta ratio of 0.7. In order to keep the velocity profile symmetrical, when the meter is installed downstream of a control valve, some manufacturers recommend a minimum length of 30 diameters of upstream straight run.
A: I experienced similar problems in a couple of my applications years ago. YEW worked well and other manufacturers' transmitters failed to perform. The design of the bluff body and sensor type and location besides the signal strength is the only explanation. This is usually the case when vendors could not explain the cause of failure. Also, fluid pressure and temperature variations seem to have some contributing effect.
Q: I want to use vortex flowmeters for bidirectional flow measurement. To do that, I will place two vortex meters in the same pipeline, one in the normal and the other in the reverse direction. Because these two meters will be installed to read in the opposite direction, what will be the limitations for this?
Is it sufficient to provide 5D pipe runs downstream length for each meter?
A: One of the disadvantages is high cost.
Another is that because you need a Reynolds Number of 10,000 to develop before each of the meters (in their own directions) will become reliable, you will have a "blind spot" under their minimum flows (Figure 1).
A third disadvantage is that you will probably double the pressure drop in both directions (Figure 1).
Lastly, the straight pipe run distance required between the two meters (which distance is the downstream run to both meters) is rather unpredictable, and much depends on the meters' designs. But, in any case I would use about 10D to about 15D to be on the safe side. As to the upstream runs required, I would use more than 5D. As to the upstream straight runs, if there are no obstructions, I would use 10D to 15D, and if there is a control valve in either direction, I would use 25D to 30D.
The bottom line is that this is not a very good idea, and you should check if inherently bidirectional single meters might not meet your requirements.
A: The real answer is that using two vortex-shedding flow meters for bidirectional flow is an improper application. The vortices shed by both bluff bodies propagate really far beyond the pipe, and may affect the other meter's readings.
In an application such as this, where I believe high accuracy is wanted, I would use a transit-time ultrasonic flowmeter—they are generically bidirectional and can indicate which direction the flow is going. In many HVAC applications, paddlewheel or insertion turbine flowmeters are used, not necessarily because they are inexpensive, but when fitted with a quadrature sensor assembly, they indicate bidirectional flow very well.
If I absolutely had to use a fluidic flowmeter, such as a vortex-shedding meter, I would use two vortex precession meters instead. This is the Swirlmeter made by ABB and by a couple of companies in China. Because they have built-in flow conditioning, they can be close-coupled and work fine. I just don't see the reason to use them in this case.
A: You will also need to be sure that the meters are suitably spaced apart, so that the vortices from one bluff body will not be affecting the measurement (bluff body/shedder bar) for the second flowmeter. Why not consider a multi-path ultrasonic or magnetic flowmeter, both of which better support bidirectional flow?