In the current issue of INTECH, I wrote an article based on my after-dinner talk, “An Hour with Dr Flowmeter.” Today I got fanmail…it was an excellent letter.
Here’s my reply, sandwiched between his comments.
Walt, I’m an engineer whose specified and worked with flowmeters a lot over the decades. I read your article on flowmeters in Intech and have a few comments. A lot of what I’ll tell is info reported to me by others; some is from my own experience.
First, you rate price as a more important factor than I do. Especially these days, people are stretched thin so I regard reliability more highly than price, although it depends on circumstances. It can be hard for techs to find time to look into and fix problems, and sometimes the result is process downtime, too. Second, the cost to install – or reinstall a failed flowmeter – should also be a consideration. The need for accuracy and repeatability will depend on circumstances.
WB> you aren’t wrong. My main point was that accuracy and repeatability depend on circumstances, exactly as you said. Whether you take price as the important first cut, or try to come up with some way to use reliability as a first cut, that’s hard to say. Price is easy. Coming up with a value for reliability isn’t.
Here are some of the flowmeter problems I’ve heard about over the years. One, a Dietrich annubar (essentially an averaging pitot tube) was installed across a boiler forced draft fan air duct. This didn’t work out because the duct expanded after it was installed. Had it been supported in a flexible manner on one end, it might’ve worked.
WB> that’s a good one. I’ll have to remember that. I built a custom insertion mag for a 10 foot wide rectangular channel once. Because it could not be supported at the outboard end, it waved up and down with an appreciable effect on the velocity measurement. I also built a custom insertion turbine meter for a penstock once, and it lasted about 30 seconds. We put it in, and the shaft bent the turbine up to the inside wall of the pipe. The velocity through the penstock was higher than expected (grin). I don’t much like insertion mags.
Second, an Instrument Maintenance supervisor once told me that an ultrasonic flowmeter used to measure flow in an open-channel flowmeter mounted outside didn’t work because wind would affect it too much. Third, another Instrument Maintenance supervisor described a situation where a magnetic flowmeter worked fine initially, but over time it read higher and higher because grease (this was in a sewage treatment plant) would harden on the liner and progressively restrict its opening.)
WB> I’ve had both happen to me. I replaced the ultrasonic open channel meter with a pulse radar meter and it didn’t actually work much better. I went to a guided wave radar transmitter and it worked fine. The magmeter is a huge problem. This happens all the time and people don’t realize it. When it happened to me, the treatment plant was downstream from a huge commercial laundry that did sheets, tablecloths, etc. The 12” magmeter restricted to just over 2” ID in about 48 hours and was just practically filled with papier mache.
Open channel flowmeters are not very accurate in general because a small change in measured signal often results in quite a variation in flow on one end, but on the other end, sometimes not much flow change at all. Proper installation of these meters is crucial.
WB> Proper installation of the primary flow element is crucial. You can have a perfect installation of the level transmitter, and have the Parshall Flume (for example) be installed high on one end. Or not installed with its flume floor straight. Or other nightmarish things contractors do.
Personally I like magmeters a lot if they can be used. I don’t have much experience with coriolis flow meters. I worked on one app about 1991 where one supplier could undercut MicroMotion on price in an app. We decided to go with MicroMotion because they were the pioneer in coriolis meters, and we wanted be as confident as possible. I’ve read coriolis meters have problems when air is entrained in fluids.
WB>I like magmeters, too. I’d use them wherever I could. I am grateful that the pricing has come down so much. When I started playing with them, they were $1000 an inch diameter, above 4”, and $2000 an inch diameter below 4”. Now look at how cheap they are. Coriolis meters have gotten cheaper, better, more accurate and more reliable. But I’d still use a magmeter unless I really wanted mass.
Doppler (ultrasonic and transit-time) meters often work great 98% of the time but in many apps they don’t work all the time. If they could be made more reliable they’d be great for more apps because they are not that costly. The 98% statement applies to other instruments besides flow meters. It’s often hard to find something that reliably ALL the time, but it’s easy to find instruments that work well MOST of the time.
WB> I’ve come to the conclusion that doppler ultrasonics aren’t worth the trouble, but transit-time ultrasonics can be very good and can be quite reliable. Multipath transit time flow meters are very accurate, and very reliable. I’m starting to see transit-time ultrasonics that can even be used in light slurries.
Anyway, I’m having a slow afternoon.
WB> Everybody deserves a slow afternoon once in a while. (grin)
Technical Services And Strategic Consulting For Technology Companies
