There are dozens of horror stories. Flowmeters have been installed backwards or sideways or immediately between two elbows because that was the only place the contractor could make it fit. Flowmeters have even been installed upside down.

Often, you don’t get to see the flowmeters until after they’ve been designed, purchased and installed. If you can get involved at the design stage, you can then make sure that some of the more difficult problems are minimized. You can make sure that the flowmeter specified will work for the service. You can make sure that the drawings show the flowmeter installed correctly and that notes the things the contractor must not do.

If you get involved in the installation, you can check the ongoing installation of the flowmeters against the design drawings and spot where the contractor has deviated from the drawings — either in procuring a different flowmeter or in installing the flowmeter where it was convenient for the contractor, instead of where the plans showed it to go.

Real problems can be caused by improper selection of the flowmeter. Many cooling towers are relatively small, and have a two-inch inlet. So, many times, a two-inch cold-water meter is selected, with a pulse contact head, as the makeup water flowmeter. Most cooling tower makeup lines are into open discharge, and the control valve is usually situated before the flowmeter in the pipeline. You should use a two-inch magnetic flowmeter — or a two-inch turbine flowmeter — both of which will easily handle a 300 gallons per minute surge, while a two-inch cold-water meter will not handle more than about 200 gallons per minute. But the cost differential between a pulse-equipped cold-water meter and a magnetic flowmeter or a turbine flowmeter is large. So, what can be done at the start-up stage to solve this problem?

• Move the flowmeter to the back of the control valve so it is always full of water.
• Replace the snap-opening solenoid valve with a slow-to-open, slow-to-close valve.
• Put a flow control valve before the flowmeter that limits the flow to under the rated maximum flow rate of the meter.

Another typical design issue is the use of flowmeters where there is not adequate straight run, either upstream or downstream from the meter. Designers tend to try to use the same flowmeter throughout a project. There are valid reasons for this, among which are ease in installation of common designs, ease in maintenance, lower spare parts requirements and ease in operation when all the flowmeters are the same. The problem, however, is that while it is possible to use all the same flowmeters, the performance of these flowmeters may vary widely depending on where they are installed.

After the design phase, it often isn’ft possible to change the piping configuration. After all, if there isn’ft enough space, there isn’ft enough space. In tight quarters, you should use flowmeters like magnetic flowmeters, V-cones, swirlmeters, fluidic meters and positive displacement meters because they don’ft require long distances of straight pipe run.

A startup nightmare is the flowmeter that is correctly specified and appears to be installed at the correct location, but which does not appear to work properly. For example, in all types of insertion flowmeters, the flowmeter might be installed sideways to the flow instead of in line with it. This is somewhat difficult to see, depending on the model of flowmeter, and causes erratic readings. Other examples of installation problems are improper wiring of analog outputs, no power connection, pipe not full of fluid and similar conditions to watch for.

Often, you have to live without even a decent solution to the problem. In wastewater treatment plants, often the Parshall flume is cast incorrectly into the concrete. About the only thing you can do, other than to shut the plant down completely, is to dye test or do velocimetric testing to determine an actual calibration curve for the flume and then reprogram the device. This holds true for closed pipe flowmeters as well. Typically, flowmeters in process plants are used for control rather than metering and, therefore, repeatability of the signal is much more important than absolute accuracy. So, if you can get a repeatable signal you can live with, you can commission the flowmeter and make the system work.

Calibration in-situ is usually impractical or impossible. Test and measurement instrumentation companies make calibration devices. But you cannot use them to actually calibrate the flowmeter. The only way to calibrate a flowmeter is to pass a known quantity of fluid through the meter and determine the deviation of the flowmeter’fs reading from the actual value of fluid flow. Unless you can run the output of the flowmeter to a weigh tank — or to a tank truck and weigh the truck — you are stuck with comparing rather than calibrating.

You can use a “loop calibrator” to check the output of the flowmeter at arbitrary values. You can use a primary element simulator. These devices allow you to simulate the output of the transducer to the transmitter, but not to test the operation of the transducers themselves or calibrate them.

And you can use a flow comparison device. The most popular of these is the clamp-on portable transit time ultrasonic flowmeter. These devices are quite accurate, but not accurate enough to be used as a calibrator. You can use these flowmeters for a critical comparison. If the reading on the clamp-on portable transit time meter reads close to the reading of the installed flowmeter, the installed flowmeter is likely in good calibration. If they do not closely agree (within 5-10 percent of reading), something is likely wrong.

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.