At the American Water Works Association (AWWA) Conference in Anaheim, CA, in June, I was speaking with an old friend about flow and flow metering.
“Why is it,” he asked, “that I have so much trouble with applications for flowmeters?”
It’s a good question. People keep making the same mistakes, over and over, and the performance of the flowmeters they install suffer for it.
My friend was standing in front of a skid-mounted polymer-feeding system, and I pointed to it and showed him that the flowmeters installed on the skid were, first, low accuracy devices, and second, they’d been installed in locations where there was substantial degradation of the flow profile due to short straight runs and the presence of elbows close to the inlet and outlet of the flowmeters.
This is a very common error, and it is one that is very revealing. What it reveals is that many mechanical designers and process engineers don’t really understand flow. For those of us in the “flow business” this seems quite strange, and a little startling, but if you think about it, it makes a certain sense. In reality, “flow” is a relatively small, and esoteric, discipline for most engineers, even for most control engineers.
One of the errors that cause real headaches in flow metering is making the piping look nice. You know: elbow, close nipple, control valve, short nipple, flowmeter, short nipple, shutoff valve, close nipple, elbow. It looks great, both on the drawing and “as-built,” but it creates a horribly disturbed flow profile, with attendant accuracy and repeatability problems for the flow measurement.
Elbows require 20 or more diameters of straight run after them. You may have to supply many more diameters of straight run if there are two elbows out-of-plane, to control the spiral twist that this piping configuration puts on the flow profile.
Control valves can require at least as many straight run diameters as an elbow between themselves and the flowmeter. Even the close-coupled shutoff valve at the outlet of the flowmeter will provide perturbation in the flow profile, as will the outlet elbow.
Sometimes, it simply is not possible to make a flow-metering run that will work. Some years ago, I was asked to provide a flowmeter for a large water transmission line, in a specific location. The line was 16” diameter, and was mostly buried underground. The location selected was a valve station. The 16” pipe came out of the ground a distance of 8”, and went into a 16”, 90-degree elbow. The elbow was coupled to a v-notch valve by about 12” of 16” diameter pipe. There was another 12” of pipe coming out of the v-notch valve, into another 90-degree elbow, and then the pipe went back into the ground. The engineer asked me to select and install a flowmeter that would accurately measure the flow after the v-notch valve. It took me over an hour to explain why I could not. In that application, there wasn’t a chance of producing an accurate flow measurement. I told him I could give him an output, but that I couldn’t tell him exactly what that output would mean.
Remember my friend at the AWWA show? Not only did his skid have “pretty piping” issues, but he had also used inexpensive flowmeters. There is a well known rule called “The Rule of Any Two” that has bearing here. You can have any two of the following three: fast, cheap or good. If you want fast and cheap, you give up good. If you want good, you have to give up fast or cheap. You can’t have all three.
What this means for flowmeter selection is if you have piping that does not lend itself to good metering runs, you need to compensate by selecting a flowmeter that requires less good piping. Usually, this means that you have to buy a more expensive flowmeter. In an ideal metering run, the installed performance of a paddlewheel flow sensor and a magnetic flowmeter or a swirlmeter may be essentially the same. But as the piping becomes less and less optimum, the performance of the paddlewheel degrades long before the magmeter’s performance or the swirlmeter’s performance does. So, the way to cut down on flowmeter expense is to provide good, long, straight metering runs. Pipe is generally less expensive than flowmeters, so doing this should produce an economical system with good performance.
Another “pretty piping” error is to make the flowmeter either oversized or undersized for the application. If you have a 6” diameter line, and your maximum velocity in the line is about three feet per second, you should probably use a smaller diameter flowmeter than 6”. This means that you need to have sufficient straight run to provide a reducer, the correct number of diameters before and after the flowmeter, and an outlet reducer. Again, this means that the piping has to be asymmetrical, and requires more space. Undersizing the flowmeter does the same things. A typical mistake in cooling tower makeup line design is to use a 2” cold water meter on a 2” line into open discharge. When the makeup line is turned on, the flow through that 2” line is usually greater than 10 feet per second. This will destroy the flowmeter in a matter of months or less.
So, remember “The Rule of Any Two” and keep that piping ugly.
Walt Boyes is a is a principal in Spitzer and Boyes LLC, offering engineering, expert witness, development, marketing, and distribution consulting for manufacturing and automation companies.