2017 E-zine Articles

Extreme Measurement: Chlorine

By Walt Boyes
Many measurements, and especially level and flow applications, can be looked at as a continuum from “too easy to bother measuring” to “too hard to do.” Most flowmeters can work well on the left-hand side, the “easy” side. As you go further to the right on the continuum line, fewer flowmeters are capable of handling the applications. For example, almost anything can measure water. Many flowmeters can measure air.
If you start talking about applications that become more difficult than clean water or dry air, the number of devices and measurement technologies available to you drops off sharply.
Chlorine gas is a byproduct of the manufacture of caustic soda, and is cheap and available. It is also one of the finest bleaching agents and, to date, the best large-scale disinfectant ever found. It is used in many industries from paper to water and wastewater. It is relatively inexpensive, too. Measuring the flow of liquefied chlorine gas is one of the most difficult applications there is, and measuring the flow of gaseous chlorine is almost as difficult.
Measuring Chlorine

Chlorine is approximately 2.5 times heavier than air, has an expansion ratio of about 450:1 and supports the combustion of iron and steel at low temperatures (approximately 450 degrees F). It is also hygroscopic, and when moist, it immediately produces several acidic byproducts, such as hydrochloric and hypochlorous acids. Industrial grade chlorine is hardly ever found without at least 2 to 3 percent impurities, such as ferric chloride, ferric sulfate and ferric oxide, as well as several varieties of waxes.
More next month.
From Flow Control (November 2001)

Think Again: Alternatives to Installing Flowmeters to Specified Straight Run

By David W Spitzer
Previous articles suggested that many flowmeters are affected by the velocity profile of the fluid upstream and/or downstream of the flowmeter. One approach to developing a good velocity profile is to install flowmeters with sufficient straight run as specified by the flowmeter manufacturer for that particular flowmeter.
This sounds good (on paper) but there are many installation where there is not enough space to install the specified straight run. Modifying the design to accommodate the specified straight run in these installations is often expensive or impossible. Nonetheless, there are some practical actions that can be taken.
For example, a flowmeter requiring 15 diameters of upstream straight run and 5 diameters downstream of downstream straight run must be installed in a piping run that is 15 diameters long. One could install the flowmeter with 10 diameters upstream and 5 diameters downstream. However, most flowmeter designs are affected more by their upstream straight run than by their downstream straight run. Therefore, it might be prudent to reduce the downstream straight as much as practical so as to increase the upstream straight run — perhaps with 13 to 13.5 diameters upstream and 1.5 to 2 diameters downstream in this example.
Sometimes more straight run is available such as when the above flowmeter can be installed in a piping run that is 30 diameters long. The flowmeter would look nice and meet the manufacturer specifications if it were installed with 15 diameters upstream and 15 diameters downstream. However locating the flowmeter 20-25 diameters upstream and 5-10 diameters downstream might be more pragmatic given that piping configurations in the field rarely match the piping configurations tested by the manufacturer.
These applications illustrate ways to partially mitigate the effects of a poor situation (insufficient straight run) and how to make the best of a good situation (more than sufficient straight run) respectfully. Similar opportunities are rampant in industry — if you understand what is happening and know where to look.
This article originally appeared in Flow Control magazine (January 2016) at www.flowcontrolnetwork.com.

Calculating the Incremental Cost of Electricity

By David W Spitzer
What is the approximate incremental cost of electricity for a plant that purchases electricity at US$ 0.06 per kilowatt-hour?
A. US$ 300 per year per kilowatt-hour
B. US$ 400 per year per kilowatt-hour
C. US$ 500 per year per kilowatt-hour
D. US$ 600 per year per kilowatt-hour
Commentary
Knowing the incremental cost of electricity is important because it can be used to quickly estimate the annual energy savings associated with a process change and/or energy conservation project such as the installation of a variable speed drive. Conversely, it can also be used to quickly estimate energy costs associated with actions that increase energy consumption.
Increasing or decreasing the electrical load by one kilowatt for the entire year can be estimated as follows.
(US$ 0.06 per kilowatt-hour) * (8760 hours per year) = US$ 526 per year
Answer C is closest to the calculated amount.

Additional Complicating Factors
This discussion sheds light on the importance of knowing incremental utility costs. However the actual calculations for electricity in a typical plant will tend to be much more complicated because they often involve time-of-day billing, demand charges, hatchet/ratchet clauses, and power factor penalties described in various rate structures.
Similarly, it is also important to know the incremental costs of other applicable utilities in the plant such as natural gas, coal, waste fuels, industrial gases, cooling water, chilled water, compressed air, and steam.
This article originally appeared in Flow Control magazine (January 2016) at www.flowcontrolnetwork.com.

Extreme Measurement: Chlorine Properties

By Walt Boyes
While chlorine gas is usually stored at low pressure (approximately 60 psig), its high expansion ratio makes it dangerous to handle. I personally witnessed a 150-lb upright cylinder of chlorine gas turn into an unguided missile when it was dropped accidentally off the tailgate of a delivery truck onto its valve, which promptly broke off, venting gas through what became immediately a rocket nozzle. In addition, chlorine gas re-liquefies with a temperature change of about 1 degree F, unless it has been heated.
In fact, most people who use it don’t even try to measure flow. Inventory is all done by weight on load cells. Gas feeders for water and wastewater use inaccurate and inexpensive glass tube rotameters in the vacuum line, while automatic metering valves have always been the bane of chlorinator manufacturers’ and users’ existences. The small orifices are where all those impurities I mentioned wind up, after all.
Chlorine is also very hard on materials. Few metals are inert to chlorine. Tantalum is one. Sterling silver is also immune. There are some specialty alloy steels, like the Hastelloy Cs, which have very high resistance to chlorine. Even plastics are degraded by chlorine. ABS, PVC, CPVC, TFE Teflon, all are more or less resistant to chlorine but will become brittle and decayed after only a relatively short time, perhaps a couple of years.
More next month. 
From Flow Control (November 2001)

Coriolis Mass Flowmeter Zero Stability

By David W Spitzer
The first edition of The Consumer Guide to Magnetic Flowmeters (co-authored with Walt Boyes) was published in 2002. The tabulation of quantifiable differences in performance between the various magnetic flowmeters available at the time presented useful information for both users and manufacturers.
Initial research for The Consumer Guide to Coriolis Mass Flowmeters (co-authored with Walt Boyes) performed shortly thereafter was another story. When the spreadsheet tabulating the performance of every model and size Coriolis mass flowmeter available at the time was about 50 percent populated, I noticed that the performance of almost all of the Coriolis mass flowmeters was 0.1 percent of flow rate for liquids. My concern was that my time had been wasted because most Coriolis mass flowmeters had the same performance and that there would be no market for the Guide. I buried these thoughts (after a short panic attack) and completed the tabulation.
My initial thoughts were unfounded — even though the performance of most Coriolis mass flowmeters was 0.1 percent of rate for liquids — because the zero stability for each model and size was different. The zero stability can be thought of as how well the zero of a Coriolis mass flowmeter can be calibrated. This (typically) automated procedure is initiated with the flowmeter full of liquid under zero flow conditions. The error associated with zero stability is a fixed value (such as 0.05 kg/min) that affects the flow measurement by that amount throughout its operating range. Therefore zero stability has a greater effect on performance at lower flow rates when performance is expressed as a percentage of the actual flow rate.
After the zero stabilities for each size and model were incorporated into the finalized table, the performance some of the best flowmeters was as much as ten (10) times superior to the worst performers in a number of models and sizes. All Coriolis mass flowmeters were NOT created equal.
This article originally appeared in Flow Control magazine (February 2016) at www.flowcontrolnetwork.com.

Flowmeter Performance

By David W Spitzer
A Coriolis mass flowmeter with an accuracy of 0.1 percent of flow rate, an upper range limit of 300 kg/min, and a zero stability of 0.01 percent of upper range limit is calibrated to operate between 0 and 100 kg/min. What is the approximate performance of the flowmeter at 30 percent of its full scale flow?
A. 0.1 percent
B. 0.1 percent of actual flow rate
C. 0.2 percent
D. 0.2 percent of actual flow rate
E. 0.3 percent
F. 0.3 percent of actual flow rate
Commentary
Answers A, C and E are incomplete performance statements and can be deemed inappropriate.
The zero stability of the flowmeter is 0.01 percent of 300 kg/min, or 0.03 kg/min — independent of flow rate. The flowmeter is operating at 30 percent of its 100 kg/min full scale, or 30 kg/min. The overall flowmeter performance is the sum of accuracy and zero stability — in like units.
One way to calculate the overall performance is to add the performance specification expressed as a percentage of flow rate to the zero stability similarly expressed as a percentage of flow rate. In this application, the performance specification is 0.1 percent of flow rate and the zero stability is 100*0.03/30, or 0.1 percent of flow rate. Therefore, flowmeter performance is (0.1+0.1), or 0.2 percent of flow rate. Answer D is correct.
Additional Complicating Factors
The performance of this Coriolis mass flowmeter varies with flow rate. For example, operating the flowmeter at 15 kg/min will increase the zero stability to 0.2 percent of flow rate — degrading the overall performance to 0.3 percent of flow rate. This may not appear to be much of a concern, but operating the flowmeter at much lower flow rates can significantly degrade flowmeter performance. In many gas applications (not considered here), flowmeter performance can be well over 1 percent of flow rate.
This article originally appeared in Flow Control magazine (February 2016) at www.flowcontrolnetwork.com.

Extreme Measurement: Coriolis Option for Chlorine

By Walt Boyes
A colleague of mine discussed a chlorine application, using Coriolis meters, on an Internet discussion group. He had noted that were “several hundred meters installed” in this application. Since Coriolis mass flowmeters are generally quite expensive, I was extremely interested in knowing what application was so critical that they required hundreds of meters.
Typically, these meters replaced rotameters and guesswork. In the aluminum industry, for example, it is necessary to bubble chlorine gas through the melt to remove the impurities in the molten metal. Chlorine is released to the atmosphere. Eventually, this led to the EPA requiring that all chlorine users get a better handle on usage in process, so they could more accurately report air releases.
The application isn’t perfect, even though the Coriolis meters work quite well in it. The meters are not inexpensive. They are made of Hastelloy C, and the gas is presumed to be clean and dry. Obviously, when it isn’t, or when the gas reliquifies, or they get a liquid slug, maintenance of the Coriolis meter follows. Particulate deposition in the measurement tube is also a maintenance issue.
But they can meet the performance requirements of this EPA driven application: better than 1 percent of rate, at very low flow rates. The minimum flow in this application was 1 SCFH.
More next month.
From Flow Control (November 2001)

Differential Pressure Transmitter Temperature and Pressure Effects

By David W Spitzer
Many flow applications (including in the oil and gas industry) are implemented using differential pressure flowmeters such as orifice plate and Venturi primary flowmeter elements. Differential pressure flow transmitters are used to convert the differential pressure signal at the primary to an electronic signal. Most flowmeter standards exist tend to focus on the primary flow element but tend to de-emphasize the differential pressure transmitter.
Shortly after starting a new job about 30 years ago, a salesman paid a visit to tell me about the differential pressure transmitters that he was selling. My immediate response was, “What is the temperature coefficient?” to which he replied, “Thank you for saving me 20 minutes.”
Specifications for differential pressure transmitters have improved dramatically over the last 30 years but the underlying concerns regarding their performance remain the same. Notwithstanding that the performance of almost all differential pressure transmitters is expressed as a percentage of set span or calibrated span (not a percentage of the measured differential pressure or the measured flow rate) there are other factors that can affect the performance of differential pressure flowmeters.
In particular, differential pressure transmitters are affected by variations in ambient temperature and process pressure. Manufacturers publish specifications that can be used to quantify these effects in actual applications. These specifications have improved over the past 30 years but temperature and pressure effects can still significantly affect differential pressure flowmeter performance and can exceed the transmitter accuracy by a factor of 3 or more in some applications. These effects become more evident when performance is expressed as a percentage of flow rate — especially at low flow rates when the produced differential pressure can be relatively low.
It is still important to consider temperature and pressure effects when calculating overall differential pressure flowmeter performance — even though they have been progressively reduced by improved technology and calibration techniques.
This article originally appeared in Flow Control magazine (March 2016) at www.flowcontrolnetwork.com.

Effects of Ambient Temperature Fluctuations

By David W Spitzer
A differential pressure transmitter with an upper range limit of 500 mbar and a set span of 0-250 mbar (approximately 0-100 inches) is installed outdoors. The transmitter has an ambient temperature effect specification of 0.001% URL/degC plus 0.002% set span/degC. Approximately how is the flow accuracy of the transmitter affected by ambient temperature fluctuations of ± 20 degC?
A. 0.04 percent of flow rate
B. 0.08 percent of flow rate
C. 0.1 mbar
D. 0.2 mbar
E. None of the above
Commentary
Ambient temperature fluctuations can affect the differential pressure measurement by 0.01*0.001*500 plus 0.01*0.002*250, or 0.01 mbar/degC. This corresponds to a 0.2 mbar potential error for the ± 20 degC temperature fluctuation. Answer D might appear to be correct but it is an absolute (fixed) differential pressure error and not a flow error.
The 0.2 mbar potential error represents 0.08 percent of the set span of 250 mbar and would introduce approximately 0.04 percent of flow rate error when the flow is at 100 percent. The flow error at other flow rates will be different so Answer A and Answer B are not correct.
Answer E as correct.
Additional Complicating Factors
Not only will the flow error be different at different flow rates but the differential pressure transmitter can also be significantly affected by the static pressure at which it is measuring. In many applications, both the temperature and pressure effects should be considered.
This article originally appeared in Flow Control magazine (March 2016) at www.flowcontrolnetwork.com.

Extreme Measurement: Other Options for Chlorine

By Walt Boyes
Other meters, other than rotameters, have been used successfully to measure chlorine, including target meters and vortex-shedding flowmeters. Target meters have been used for decades in sizes up to at least 6″ diameter to measure chlorine gas. The target is made of CPVC, the arm is made of Hastelloy C and the meter body can either be plastic or Hastelloy. All Hastelloy vortex meters have also been used in 2″ and 3″ sizes, and probably larger. Diaphragm-style gas meters have been designed for this service, but none is currently in manufacture. Liquified chlorine gas can also be measured using transit-time ultrasonic flowmeters on a Hastelloy or Tantalum spool section, and there is even a clamp-on transit-time meter for gases, including chlorine.
Even on the far right-hand side of the difficulty continuum, flowmeter manufacturers are producing more choices for engineers and users.
From Flow Control (November 2001)

Getting the Gas Out

By David W Spitzer
Current events seem to have (temporarily) altered the tone of television interviews from cordial/congenial to confrontational. Whereas the person being interviewed used to answer the interviewer’s questions, the person being interviewed now seems to deliver their preplanned message no matter what the question. More than one interviewer has stopped the verbal tirade and actually lectured the person being interviewed (sometimes in no uncertain terms) as to the inappropriateness of not answering the question.
Most vendors are generally knowledgeable of their products. If there are questions beyond that person’s knowledge and expertise, most have no problem contacting the proper person in their organization for clarification. Yet, the tone of recent interviews reminds me of a particular vendor presentation.
Some years ago, an officer of a company came to chat. I politely listened to his presentation about his flowmeters. When he finished, I wanted to know how gas is removed from the flowmeter so that the flowmeter can remain full of liquid at all times. After all, special features with bells and whistles are nice, but if the flowmeter does not remain full of liquid, the quality of measurement can suffer.
What was the vendor’s response to my question?
A. The flowmeter is not self-filling, so the user has to run high flow through the flowmeter to remove the gas before zeroing the flowmeter.
B. The flowmeter is not self-filling, so the user has to determine when gas accumulates in the flowmeter during operation and run high flow through the flowmeter to remove the gas.
C. The flowmeter is not self-filling, so the user has to determine when the piping in the area of the flowmeter becomes partially empty and run high flow through the flowmeter to remove the gas.
D. Once the gas is removed from the flowmeter, the flowmeter should operate accurately.
E. The flowmeter has many virtues, including…
Commentary
Answers A, B and C describe the realities associated with the installation of this flowmeter. For proper operation, the flowmeter must remain full of liquid. Any accumulation of gas in the flowmeter provides the potential for degradation of the measurement. Collectively, Answers A, B and C describe actions that the user should take to maintain accurate measurement. In many applications, these procedures can be impractical, unfeasible and inconvenient. Answer D is a true statement, but it does not address the issue of reintroducing gas into the flowmeter. The response given to my question was Answer E. It really did not matter what the question was — the answer was already predetermined.
Additional Complicating Factors
Some suppliers provide a procedure in their instruction manual to remove gas from their flowmeter. Other suppliers address the issue by stating that gas must be removed from the flowmeter for proper operation. The user is advised to examine the instruction manual before purchasing.
From Flow Control (July 2002)

Evaluating Specifications

By David W Spitzer
Speaking with the many suppliers of flowmeters and other instrumentation while researching “The Consumer Guide…” book series led to the interesting observation that almost all suppliers claim to sell the best equipment. Therefore, finding the best equipment can be a challenge. Unless these suppliers sell equipment that performs identically (which the research shows is not the case), equipment from these suppliers cannot all be the best. In contrast, a few suppliers will admit that their equipment is not the best, but is adequate for many applications. It is also interesting to note that the price of the best equipment is often competitive with the price of average and below average equipment.
Which of the following are logical approaches to determining the quality of an instrument?
A. Evaluate the supplier’s verbal statements.
B. Obtain and test instruments from different manufacturers. Compare the test results.
C. Compare equipment specifications.
D. Determine the operating parameters for the application. Calculate performance at the operating flow rates.
Commentary
There is some value to all of the answers. In Answer A, the salesperson’s statements should provide a broad filter that may be used to determine the potential usefulness of the instrument in the application. This should be done with the understanding that there are financial incentives for each supplier to claim to supply superior equipment, so their statements may be somewhat optimistic.
Obtaining and testing equipment of different manufacture (Answer B) is perhaps the best approach for determining the quality of an instrument. It gives the user hands-on experience with the instrument and independent test data. However, with some instruments (including flowmeters) this can be so costly and labor intensive that only the largest companies can afford it.
Comparing instrument specifications (Answer C) relies on supplier claims that describe instrument performance. Understand that accuracy statements can be manipulated because some suppliers’ specifications push the limits while other suppliers are more conservative. While this approach may not be able to discern whether an instrument ranks 43rd or 44th out of 60 instruments (who cares anyway?), the information gathered can generally find the 3-5 instruments that should be seriously considered for purchase.
Answer D could be considered a subset of Answer C and it also relies on supplier claims. The performance of some instruments varies with full scale flow rate. For example, the accuracy of a magnetic flowmeter model may be identical for a wide range of flow rates in many sizes. In contrast, the accuracy of a given size and model Coriolis flowmeter varies with flow rate. Supplier specifications can be used to calculate the performance of these flowmeters at various operating flow rates. The results can be compared to determine which flowmeter would perform better in this application.
Additional Complicating Factors
Instrument specifications are usually complete, but at times can be complicated. In addition, they are often expressed in a manner that makes the instrument appear to perform better than other instruments by emphasizing the portion of the specification that puts their instrument in a positive light. Caveats and constraints are often located in the fine print. The user is advised to be sure that performance calculations include all portions of the specifications.
From Flow Control (August 2002)

Just Calibrate It!

By Walt Boyes
Repeatedly, I have seen applications in which there is a pipe, followed by a flowmeter, followed by a pipe. There are multiple things wrong with this picture. But it is entirely understandable how things got to that state.
What’s wrong with it? For one thing, there are no valves. That means there is no way to shut off the flow. So? Well, suppose this flowmeter is in a 36″ line that serves drinking water and fire flow to a town of 20,000? What do you do if the flowmeter breaks? What you have to do is to shut off the water at the nearest valve, remove the flowmeter and replace it, either with a new flowmeter, or a spool piece, so that the original meter can be repaired. Then, if you have elected to repair the original, you have to repeat the process when the flowmeter is repaired and ready to be re-installed. In the meantime, the town goes without water for several hours, maybe even a day or two.
Of course, nobody would let the town go dry. So what happens? The flowmeter sits unrepaired until some agency demands that it be fixed.
Good design requires that any flowmeter be installed with shut-off valves and a bypass capable of the entire flow through the meter. Often, this simply isn’t done. It takes up space, costs money and the many other excuses for not doing it that I have heard.
That’s the second thing wrong with the picture. There is no bypass. Unless the line you are working with is capable of being shut down completely for a considerable amount of time, it should have a bypass.
More next month.
From Flow Control (October 2001)

Detecting Leaks in Water Distribution Systems

By David W Spitzer
Measuring the flow through water distribution systems is important due to the large economic value of the large quantities of water that are transported from one location to another. In many applications, these measurements are used to monitor flows but, on occasion, they are used for custody transfer and billing purposes.
Water is typically distributed in large pipes that can range from (say) 12-inches to upwards of 120 inches in size depending on the location. Full-bore flowmeters are often installed where accuracy is important even though much more economical insertion flowmeters can provide reasonable accuracy when properly applied and installed.
Insertion flowmeters measure flow at one location in the pipe and infer the total flow making it subject to additional errors caused by improper sensor location and velocity profile effects. However they are an economical alternative to a full-bore flowmeters when flow monitoring — especially for temporary flow measurements as may be required to analyze operational and distribution problems.
Some time ago, I was asked to analyze the operation of a relatively large water distribution system with full bore flowmeters where closure of the water balance approximately doubled to almost 5 percent. In this location, the water district was concerned that a poor water balance could result in large undetected water loses that could continue for long periods of time. This water district provides water to almost 2 million people so the economic impact of water leaks could be substantial.
Problems were found relating to the large incoming water flowmeter, some distribution flowmeters, and the operation of the majority of the downstream flowmeter stations. The water balanced to well within one percent after the flow measurement issues were corrected. Modifying the flow metering station measurement and control strategies resulted in the system operating “smoother” with less ongoing maintenance.
This work took a few onsite visits plus analysis afterwards and helped provide water to a couple million people more efficiently and with confidence that leaks can be detected in a timely manner.
This article originally appeared in Flow Control magazine (May 2016) at www.flowcontrolnetwork.com.

Venturi Flowmeter System Accuracy

By David W Spitzer
A Venturi flowmeter system in a water distribution system that is properly designed and installed is calibrated for 0-250 mm WC and typically operates between 10 and 40 of full scale. Will the Venturi measure accurately if the differential pressure transmitter can measure the differential pressure accurately between 10 to 100 percent of full scale?
Commentary
This is somewhat of a trick question because the information presented is not complete — even though it may appear to be complete. Note that the first sentence clearly addresses the entire system — Venturi and transmitter. The second sentence is a bit dubious because the question only refers to the Venturi and the question would be incomplete because no information is provided to evaluate the hydraulic operation of the Venturi.
If the question is whether the Venturi flowmeter system will operate accurately, we should calculate the differential pressures at 10 and 40 percent of full scale — which are 2.5 and 40 mbar respectively. The transmitter will operate accurately above 10 percent of its full scale, or 25 mbar. Therefore, without examining the operation of the Venturi primary flow element, the Venturi flowmeter system will not measure accurately over most of its anticipated operating flow rates.
Additional Complicating Factors
The Venturi flowmeter system may not operate accurately over most of its anticipated operating flow range but it will operate nonetheless with reduced accuracy that might be acceptable for the application. You might say (especially in difficult application) that sometimes poor performance is not that poor.
This article originally appeared in Flow Control magazine (May 2016) at www.flowcontrolnetwork.com.

Just Calibrate It! (Checking and Testing)

By Walt Boyes
Flowmeters should not be treated like they are just pieces of pipe. They are precision instruments that need to be cleaned and calibrated on a regular basis.
Mechanical flowmeters, such as propeller and turbine types, must be calibrated at least once annually. Mechanical parts wear, and wearing causes error in the flowmeter’s readings. Orifice plates, Venturi tubes, Pitot tubes and V-cones need to be checked for wear and calibrated at similar intervals. Electronic flowmeters don’t usually wear out, but they need to be regularly checked for other reasons as well. Depending on the service, electronic flowmeters, such as magnetic flowmeters and ultrasonic flowmeters, should be checked and tested at least once a year.
Checking the electronics of a flowmeter is easy. Whether the metering element is a turbine, or an orifice plate or a magnetic primary, simulating a signal from the actual metering element can test all of the secondary electronics. This is what many people think of when they talk about “calibrating” a flowmeter. But this isn’t testing the flowmeter itself.
To test a flowmeter itself, you have to be able to test the primary measuring element: the turbine spool, the orifice plate, the V-cone, the magmeter primary, the pipe the ultrasonic meter is clamped to and so forth.
That’s why the bypass is so necessary. So you can take the metering element out, and take it to a calibration stand or prover, or to a flow lab.
More next month.
From Flow Control (October 2001)

The Fine Art of Variable Speed Drive Operation

By David W Spitzer
Applying variable speed drives is a combination of art and science. Many applications are “academic” in the sense that a control strategy utilizing a control valve is replaced with a control strategy utilizing a variable speed drive. In other words, instead of installing a pump or fan operating at full speed and using a control valve to throttle to the desired flow, the control valve is eliminated by operating the pump or fan motor with a variable speed drive to meet demand. This generally reduces operating costs by only generating the hydraulic energy necessary to meet system demand. This strategy reduces energy and is a “good thing” on many levels.
However “academic” applications typically yield energy savings of 20 to 30 percent. Simple payback is often approximately 3 years. This is nothing to sneeze at — especially when large motors are involved.
The art behind variable speed drive applications is to make process changes in conjunction with “academic” changes. This is not always possible, but when it is, the savings can be significantly higher. How significant? Some years ago, I installed a variable speed drive on the plant air compressor that was operated by the largest motor in the plant. Without getting into the details, the project actually increased electrical energy consumption while decreasing fuel costs much more. The project had a six (6) month simple payback. In a matter of days, the project politics went from, “You want to do WHAT?” to, “How fast can you get it done?”
Projects such as this are typically conceived by users at the operating companies. Manufacturers and vendors are knowledgeable of their variable speed drive equipment and the “academic” applications in which it can be utilized. However it is the end-user and appropriate consultants who are in the best positions to apply some art and find the gems.
This article originally appeared in Flow Control magazine (June 2016) at www.flowcontrolnetwork.com.

A Formula for Energy Conservation

By David W Spitzer
Approximately how much does it cost to operate a 10 horsepower motor continuously for a year?
A. USD 3000 / year
B. USD 5000 / year
C. USD 7000 / year
D. USD 9000 / year
Commentary
Every professional working in a plant should be actively involved in the profitability of the plant and, as such, should know how much money will be saved by simply turning off a piece of equipment that is simply not needed at the time. With that in mind, any plant professional should be able to answer this question in 10 seconds or less — even if it is not within their area of expertise.
Let’s assume that a motor is 90 percent efficient and operates for 8760 hours per year. Further, the incremental cost of electricity is USD 0.07 per kilowatt-hour. Using these assumptions (they could be different in your location), the energy that a fully-loaded one horsepower motor that operates continuously will consume is approximately (0.746 kW/hp / 0.90) * (USD 0.07 / kWh) * 8760 hours/year, or USD 508 / year.
Therefore a fully-loaded 10 horsepower motor would cost approximately USD 5080 per year to operate continuously. Answer B is the best answer using the above assumptions. Another answer might better apply in your location.
Additional Complicating Factors
Plants are typically subject to time of day billing, demand charges, hatchet clauses, ratchet clauses and other tariffs that can make this calculation quite complex. You might consider having one person perform this calculation for your plant and subsequently publicize the results throughout the plant to increase awareness of how energy conservation can dramatically affect the bottom line.
This article originally appeared in Flow Control magazine (June 2016) at www.flowcontrolnetwork.com.

Just Calibrate It! (Flow Lab and Calibration Truck)

By Walt Boyes
At a flow lab, they run a known quantity of fluid (almost always water) through the metering element, and compare the reading of the meter with the actual throughput of the lab. The difference, if any, is the inaccuracy of the meter. Adjusting the meter’s calibration factor (sometimes known as “k-factor”) to account for this inaccuracy is the process of calibrating the meter.
So, what do you do if you are on the receiving end of those numerous applications where the flowmeter can’t be dismounted because there is no bypass and you can’t shut the flow off for a day or so?
The Calibration Truck
You can do any of several things. If you have the ability to divert all of the flow through the meter for a very short time, you can use a “calibration truck.” This method works well for water utilities with interties that can’t be shut down. You get a construction water truck, and weigh it empty to get tare. Then you wait until about 3 a.m. and shut the intertie off. What water usage there will be can be satisfied (hopefully) by stored water in the served grid. Then you fill the truck, read the total on the meter and turn the intertie back on. After you get a weight on the full truck, you can compare it to the flowmeter’s total, and adjust the flowmeter accordingly. It isn’t as accurate as doing it in a lab, but for annual calibration of a large bore intertie meter, it is the most accurate way to do it.
More next month.
From Flow Control (October 2001)

Velocity Profile Instability

By David W Spitzer
Many things can go wrong and cause a flowmeter to measure improperly. Typical errors cause the flowmeter to measure incorrectly by a certain percentage of flow rate, a fixed flow rate, or by a certain amount of signal. The problem can usually be fixed or corrected after the problem is identified and analyzed. This process typically requires specialized knowledge and is not necessarily simple or straightforward.
However some problems are more difficult to include quantifying and correcting for the effects of velocity profile on the flow measurement. The performance of many flowmeters can be adversely affected by a velocity profile that is not sufficiently developed before entering the flowmeter. In other words, measurements can be affected by varying velocity profiles because insufficiently-developed velocity profiles can vary with flow rate and (at times) composition to affect the measurement in different ways at different flow rates (and sometimes) at different times. Identifying this problem is often difficult — especially in larger pipe sizes where velocity profile distortion can propagate well beyond published straight run requirements.
This instability problem is often difficult to find and diagnose because the actual and measured flow rates often cannot be compared. It sometimes takes a “sixth sense” to even suspect that this type of instability is even occurring — let alone how to correct for it.
Even flowmeter suppliers and users who know their instruments and processes respectively are at a distinct disadvantage when it comes to diagnosing this type of instability so it helps when both work together. Fortunately this type of instability does not happen often but you should be aware that it can bite you — hard.
This article originally appeared in Flow Control magazine (July 2016) at www.flowcontrolnetwork.com.

Flowmeter Instability

By David W Spitzer
Which of the following could make a liquid flowmeter measurement unstable?
A. Loose flowmeter fitting
B. Loose flowmeter wiring
C. Air in impulse line
D. Swirling flow profile
E. Flowmeter located between two reducers
F. Defective piping support
G. Oscillating flow
Commentary
If you thought that all of these issues could cause a flowmeter measurement to become unstable… you would be almost correct. Depending upon the process and flow technology employed, any of these issues could cause the flow measurement to become unstable except (perhaps) Answer G — oscillating flow. In this case, it is the process that is unstable — not the flowmeter.
Additional Complicating Factors
The complicating factor associated with addressing flowmeter instability is that there tens (if not hundreds) of potential sources of flowmeter instability to consider. The first order of business should be to determine if the instability is a process problem or a flowmeter problem. Stabilizing the process by “fixing” the flowmeter is analogous to an ostrich putting its head in the sand and is clearly counterproductive but (unfortunately) can be politically expedient. Hiring an outside consultant to diagnose the problems should be considered when plant personnel cannot locate and remedy the root cause of the problem.
This article originally appeared in Flow Control magazine (July 2016) at www.flowcontrolnetwork.com.

Just Calibrate It! (Clamp-on Flowmeter)

By Walt Boyes
Another way to “check” the calibration is to use a clamp-on ultrasonic flowmeter. Now, remember that in order to use one flowmeter as a secondary standard to calibrate another, in situ, you need a flowmeter that is at least one order of magnitude more accurate than the one you are calibrating. That will usually not be the case with an ultrasonic meter, even though transit-time technologies are quite accurate. So you cannot actually use a clamp-on meter to “calibrate.” What you can do is to use it to check the reading of another flowmeter. If the clamp-on meter’s reading differs greatly (more than 3 to 5 percent of reading, for example) from the meter under test, this indicates that there is something wrong. The meter under test should be removed and tested properly. Of course, what people do is to adjust the test meter so that it reads what the clamp-on tells them it “should” and they call it “calibration.”
So, if you are a design engineer, put those valves and bypasses in. If you are an owner or an owner’s rep, insist on bypasses and valves around flowmeters. If you are a contractor, point it out to the owner and the engineer. If you are a maintenance engineer or foreman, keep bugging your superiors about putting in bypasses around selected flowmeters at the next plant shutdown. Remember, you can’t say it is calibrated for ISO 9001 purposes, or for FDA purposes unless you can take it out and calibrate it!
From Flow Control (October 2001)

Flowmeters Downstream of Pumps

By David W Spitzer
Pumping liquid from one location in the process to one other location in the process is a common hydraulic configuration encountered in processing systems. Not to be confused with pumping to multiple locations, this configuration is typically comprised of an originating vessel or tank, pump, flowmeter, destination vessel or tank, and their interconnecting piping.
Any of these components could potentially cause problems. For example, the vessel or tank sizes could be incorrect, the pump could exhibit cavitation, an improper flowmeter could be selected, and/or the interconnecting piping could be too big or too small. However even when these items are correct, there is a distinct tendency to install the flowmeter near the pump — often with insufficient straight run which can compromise flowmeter performance.
Straight run is one method that can be used to generate a good velocity profile entering the flowmeter. Straight run essentially provides space in which velocity profile distortion present in the liquid can dissipate. Standards and manufacturer recommendations detail straight run requirements for various flowmeter technologies for various upstream configurations. The problem is that this information typically focuses on piping configurations that might not include an upstream pump or the specific type of pump that is actually installed.
Pumps impart hydraulic energy into the fluid and can distort the velocity profile with strong eddies that take a long straight run to dissipate — especially in large pipes. There is an understandable general preference to install the pump on a solid foundation on the ground so as to be mechanically stable and provide ease of maintenance. There is also a tendency to locate the flowmeter immediately above the pump — especially when an instrumentation professional is not available for consultation. The installation may look nice… but locating the flowmeter one or two decks above the pump with more than sufficient straight run is advisable to reduce the probability that velocity profile distortion entering the flowmeter can adversely affect flowmeter accuracy.
This article originally appeared in Flow Control magazine (August 2016) at www.flowcontrolnetwork.com.

Required Pump Pressure

By David W Spitzer
Approximately how much pressure head must a pump located at grade produce to pump water to a tank located 20 meters above grade?
A. 1 bar
B. 2 bar
C. 3 bar
D. 4 bar
Commentary
This may look like a straightforward question… but looks can be deceiving. This question is grossly incomplete because it does not include whether the top, bottom, top of its influent piping, or some other reference is 10 meters above grade or if the tank is pressurized. In addition, flow rates and pipe sizes that affect friction losses are not presented.
That said 10 meters of water column is equivalent to approximately one atmosphere which is approximately 1 bar so in this application approximately 2 bar of pressure will be needed to begin to start flow (if friction losses are neglected). Answer A and Answer B can be eliminated.
It is reasonable to conclude that more the 2 bar of pressure will be required. How much more is dependent upon the details of the application.
Additional Complicating Factors
Additional factors can add to the pressure required. For example, what if there is a flowmeter and flow control valve in the piping system? What if the piping takes a tortuous route to the tank? Pump sizing can be a multi-discipline activity that can quickly get complicated.
This article originally appeared in Flow Control magazine (August 2016) at www.flowcontrolnetwork.com.

Pumps and Flowmeters: Working Hand-in-Hand

By Walt Boyes
There are many types of pumping applications, and few other process parameters are tied so closely to pumping as flow measurement. But because the transferring of fluids is not a cut and paste operation, you can’t just say, put a pump here, followed by a flowmeter.
Some types of pumps and pumping applications work extremely well with flowmeters, and then there are types of pumping applications that simply do not. In some cases, you have to accept a less-than-perfect result, and in some cases, you can’t find a flowmeter that will work well, or even at all.
Sometimes this is a function of the application, sometimes it is a function of the choice of pump and sometimes it is a function of the expected cost of the system, or the required payback. For example, it is certainly possible to measure the output of a small, diaphragm-operated chemical metering pump using a Coriolis mass flowmeter. But since Coriolis meters cost several thousand dollars each, this would be an over-specification for the level of accuracy required for this application.
More next month.
From Flow Control (September 2001)

Flowmeters for Dirty Liquids

By David W Spitzer
Although every application is different, there can be considerable similarity between many applications. Interestingly instrumentation (and other) professionals with experience can often see commonality in applications in diverse industries. Some years ago, our plant used a technology to introduce chemical feeds that was similar to that used to bottom-blow oxygen to refine steel. On a side note… the controls were entirely different for these processes.
Dirty liquid flow in general and sludge flow in particular is a case in point. Although their properties may be somewhat different (one is a subset of the other), both contain solids suspended in a liquid. As such, there are certain common items that can be considered when selecting an instrument to measure the flow of these streams. It is often helpful to consider attributes of the flowmeter that should be avoided while understanding that special flowmeter designs may be available in the avoided technologies to address these applications.
In general, flowmeters with impulse tubing, moving parts, and surfaces subject to wear are generally not advisable in these applications so differential pressure, positive displacement, thermal, turbine, and variable area technologies (among others) are usually avoided. Flow technologies that are obstruction-less and less subject to wear tend to be more suitable such as magnetic and ultrasonic flowmeters. Depending on the application, Coriolis mass flowmeters may or may not be subject to wear and plugging.
Every application should be considered as different. However they may have commonality that can eliminate technologies and help focus on the few technologies that promise to be more applicable.
This article originally appeared in Flow Control magazine (September 2016) at www.flowcontrolnetwork.com.

Tank Level Control

By David W Spitzer
A tank at the outlet of a process upstream is used as the feed tank for a downstream process. The flow into the tank depends on the operation of the upstream process while its outlet flow is set by the operator. The tank typically overflows about once a month and the pump has run dry on occasion. What control strategies should be considered to better control the tank level and avoid these undesirable operating conditions?
A. Manual
B. Feedback
C. Cascade
D. Feedforward
E. Advanced
Commentary
Let’s investigate the existing operation before considering modifications. The information given suggests that the inlet flow is not controlled per se but that it is affected by the upstream process and hence, reflective of the stability of the upstream process. It appears that the operator periodically adjusts the outlet flow (to the downstream process) to keep the tank level within reasonable limits. Overflowing the tank and running the pump dry likely occur when the operator does not or cannot properly adjust the outlet flow rate. For all practical purpose, the tank level control is essentially manual.
Feedback control of the tank level is desirable to allow the operator to monitor the tank level instead of being its controller. The outlet flow control loop apparently already exists to provide a stable feed flow to the downstream process. Installing a feedback level control loop (Answer B) where the tank level manipulated the existing flow control valve would degrade the now-stable flow to the downstream process. Therefore cascade tank level control (Answer C) should be considered.
Additional Complicating Factors
Consideration should also be given to allowing the level to vary somewhat around its setpoint in order to reduce changes to the outlet flow. After ensuring that the upstream process operates in a stable manner, feedforward control (Answer D) should be considered — especially when the downstream process operates well when its inlet flow is allowed to vary.
This article originally appeared in Flow Control magazine (September 2016) at www.flowcontrolnetwork.com.

Pumps and Flowmeters: Basic Pump Types

By Walt Boyes
We can divide pumps into several basic types for our purposes. First, are the turbine and centrifugal pumps, whose output is consistent, without major pulsations. Nearly any type of flowmeter can be used with these pumps, so long as they are pumping clean fluids. That’s why water utilities still use basic turbine flowmeters, whose design has changed little over 50 years. It is why you can use paddlewheel flowmeters and various kinds of insertion meters.
When centrifugal pumps with recessed impellers are used for fluids containing solids, such as raw sewage, flowmeters with rotating elements are not used. Instead, metering devices, ranging from flow nozzles and orifice plates, to magnetic flowmeters and Doppler ultrasonic flowmeters, are used. Basically, it costs more to meter the pumping of fluids with entrained solids than it does the pumping of clean fluids. Both the pump and the flowmeter have to be made of the appropriate materials, when abrasives or corrosives are being pumped. Sometimes this can make the flowmeter incredibly expensive.
Other types of pumps cause pulsation in their output flows. Diaphragm pumps of various different kinds and sizes, piston pumps, progressive cavity pumps and vacuum pumps all cause the flow to pulse. Pulsating flow is hard to measure with any sort of volumetric flow device. Positive displacement flowmeters are generally fairly expensive, so they are only used in situations where the cost/benefit of doing so outweighs the initial expense and maintenance expenses required to keep the flowmeter operating.
More next month.
From Flow Control (September 2001)

Iron Ore Level Measurement

By David W Spitzer
Sometimes you can look good if you are armed with the right knowledge at the right time in the right place. One of my projects at US Steel involved the instrumentation at an iron ore beneficiation plant in the late 1970’s where iron ore was crushed, concentrated and agglomerated into pellets for use in blast furnaces. The level of iron ore in various bins and hoppers was measured throughout the plant where the size of the material ranged from boulders to rocks to gravel to stones to powder — depending on the location in the process.
The choice of level technology that could be applied to the concentrator feed bins was limited because the bins were tall (30 meters) and the iron ore was abrasive and dusty. Contact level transmitters were out of the question so the plant used ultrasonic level technology for these measurements. If radar level technology was available at the time it would have been very expensive.
There were a number of manufacturers that made ultrasonic level transmitter systems at the time but few had transducers that were able to measure the required distance. The plant had previously experimented and settled on the one transducer design that used the same downward-pointing sensor surface to both send and receive ultrasonic energy. While not perfect, these instruments could be operated and maintained by plant personnel. The same transmitters were purchased for new bins that were installed in the expansion.
Sometime later, my supervisor called me into his office and told me that the ultrasonic level transmitter system used to measure hot coke (and other materials) was not measuring properly in another plant. The system was purchased and installed as part of a larger mechanical project that did not receive an instrumentation review. Simply put, I was told be at the plant at 8:00 am on Monday morning to assess the system and determine what could to be done.
More next month!
This article originally appeared in Flow Control magazine (October 2016) at www.flowcontrolnetwork.com.

Vortex Shedding Flowmeter Size

By David W Spitzer
A 3-inch stainless steel vortex shedding flowmeter is sized to measure a maximum water flow of 100 gpm full scale. Can this flowmeter measure the desired 10 gpm minimum flow required for startup?
Commentary
Vortex shedding flowmeters have four common constraints — materials of construction, available size, Reynolds number and velocity. The availability of the 3-inch size and its stainless steel construction are not issues in this application.
The Reynolds number associated with a water flow of 100 gpm in a 3-inch pipe is (3160*100*1) / (1*3), or approximately 100,000. Vortex shedding flowmeters operate above Reynolds numbers of 10,000 to 40,000 depending upon manufacturer and size. Details should be considered, but most 3-inch vortex shedders will not operate properly at 10 gpm with a Reynolds number of approximately 10,000.
In a 3-inch pipe, 100 gpm flows at approximately 4.4 feet per second so the velocity at 10 gpm is approximately 0.44 feet per second. This relatively low velocity will cause a problem because vortex shedding typically ceases below approximately 1 foot per second in liquids.
In summary, the Reynolds number constraint may or may not be violated in this particular 3-inch vortex shedding flowmeter. However, this flowmeter will not measure the desired a flow of 10 gpm because it will turn off and cease to operate below approximately 20 to 25 gpm.
Additional Complicating Factors
Using a 2-inch vortex shedding flowmeter would increase the velocities at 100 and 10 gpm to 9.74 and 0.97 feet per second respectively. This will increase Reynolds number and should increase the velocity sufficiently to allow the 2-inch flowmeter to measure a flow of 10 gpm.
This article originally appeared in Flow Control magazine (October 2016) at www.flowcontrolnetwork.com.

Pumps and Flowmeters: Pulsation Dampers and Alternatives

By Walt Boyes
Sometimes, pulsation dampeners of various kinds are used to permit measurement in cases of extreme pulsation, like diaphragm metering pumps, bellows pumps, piston pumps and similar devices. Pulsation dampeners and suction stabilizers are used to make metering possible. A pulsation dampener is a cylinder with a bladder that permits some of the liquid to enter on each “pressure stroke” of the pump. Compressed gas outside of the bladder recompresses the bladder, forcing the fluid out when the pump is “backstroking.” This dampens the pulsations in the fluid, and permits a velocity flowmeter to operate properly.
But what do you do if you can’t measure the flow easily, or reasonably inexpensively?
Here’s where there are a number of interesting alternatives to flowmeters to produce some of the same results. One of the major reasons to measure flow is to determine if the pump is running. Another is to make sure that the pump is running properly, not just turning. There are a couple of ways to do that without a flowmeter.
One of the most common ways is to take a contact closure from the auxiliary contacts on the pump motor starter. This is typically taken to some means of alarm or a PLC or a SCADA or DCS input. The problem with this is that it only indicates that the motor starter’s contacts were pulled in, not that the pump started. So, this method is often used in combination with a pressure sensor. A pressure surge occurs when the pump is started, and a combination of the motor starter’s aux contact and the contact from the pressure sensor indicate that the pump did start.
More next month.
From Flow Control (September 2001)

Hot Coke Level Measurement

By David W Spitzer
As a result of my involvement with ultrasonic level measurement systems used bins in our iron ore beneficiation plant (described in my previous article), I was asked to visit another plant that used an ultrasonic level measurement system to measure the level of hot coke in tall bins.
On Monday morning, I entered the plant and was escorted to the top of the bins. One look into a manhole in the hot coke bin revealed the magnitude of the problem — swirling clouds of hot coke dust obstructed viewing beyond (say) 0.5 meters into the bin. We then went to the shop to examine a transducer that had already been removed from service. I took one glance at the transducer and said, “that won’t work” to which the plant technician replied, “we know that” followed by a gesture to escort me back to the gate.
Seeing his reaction, I asked if the plant had an ultrasonic level transmitter made by the same manufacturer whose equipment was installed in the iron ore beneficiation plant (described in my previous article). Fortunately there was one such transmitter in the plant that we could see in operation. After seeing the unit and describing why the new ultrasonic sensor had a much better chance of working, the plant technician agreed to try one instrument in the most difficult service — hot coke.
Making a long story short, one week later, the plant installed the ultrasonic transmitter on the hot coke bin and contracted a factory service technician to install and start it up. A week later after installation, the plant called my supervisor (by chance I was in his office) and asked when the rest of the system would be installed. The new level transmitter worked so well that although having to be coaxed into installing just one transmitter, the plant was now blaming us for not having installed level transmitters on all of the bins.
As I suggested at the top of my previous article — sometimes you can look good if you are armed with the right knowledge at the right time in the right place.
This article originally appeared in Flow Control magazine (November 2016) at www.flowcontrolnetwork.com.

Common Flowmeter Performance Statement

By David W Spitzer
What is the measure by which flowmeter performance should be compared?
A. Percent of flow rate
B. Percent of full scale
C. Percent of calibrated span
D. Percent of set span
E. Percent of meter capacity
Commentary
All of these performance specifications (and more) are found in practice. Explanations of these terms and examples of their use can be found in our various Consumer Guides (spitzerandboyes.com/products) and in my book Industrial Flow Measurement (ISA Press). In theory, any of these measures could be used for comparison but some are more cumbersome (difficult) to use than others.
The specification used to describe the performance of many flowmeter technologies is a percentage of flow rate (Answer A) within its minimum and maximum flow rates. This form of specification typically results from the absence of a zero adjustment in the fundamental flowmeter measurement. A flowmeter with a percent of rate statement can usually be readily compared with other such flowmeters because the same straightforward specification applies throughout the flow range.
However the error associated flowmeters expressed as a percent of full scale (Answer B), calibrated span (Answer C), set span (Answer D) and meter capacity (Answer E) is typically a constant measurement error (such as ±1 lpm) throughout the flow range. Comparison can be complicated because the measurement error is dependent upon a somewhat arbitrary selection of full scale, calibrated span, set span and/or meter capacity.
Fortunately, percent of full scale, calibrated span, set span and meter capacity measurement errors can be converted into percent of flow rate errors at each flow rate of interest and compared. Answer A is correct.
Additional Complicating Factors
Complete flowmeter performance specifications typically include other influences such as analog output accuracy, voltage effects, ambient temperature effects, process pressure effects, and the like that can significantly complicate flowmeter performance calculations.
This article originally appeared in Flow Control magazine (November 2016) at www.flowcontrolnetwork.com.

Pumps and Flowmeters: More Alternatives

By Walt Boyes
In simple applications, in clean fluids, in normal temperature and pressure applications and in applications where there are no corrosives or abrasives, flow switches are often used.
Poppet-type flow switches are often used with metering pumps. There must be enough fluid per stroke to move the poppet far enough that it trips a switch. One switch trip counts as one pulse from the pump. Some of these devices are almost accurate enough to measure flow.
Another way to do this is with a current sensor or current switch. When a pump’s motor starts, there is a current surge that can be sensed by a switch, usually an induction sensor. Even a variable frequency drive’s motor current can be sensed. Some current sensors permit you to sense jams downstream (high current because the pump is deadheading) or loss of suction (low current because the pump isn’t working very hard).
When you have a pumping application, give some thought to the things you want to know. You may need to measure flow, but you also may not need to.
From Flow Control (September 2001)

Replacing a Starter with a Variable Speed Drive

By David W Spitzer
In that blur that seems to pass for e-mail these days, I seem to recall someone asking whether replacing a starter with a variable speed drive to operate a centrifugal pump was in order. The implication is that the existing motor/pump was being turned on and off to meet demand.
Should the starter be replaced with a variable speed drive to operate the motor/pump? Why or why not?
A. Yes, because it will allow the pump to be started and stopped slowly, and will reduce wear and lower maintenance costs.
B. Yes, because the pump will be operating at a lower speed and generate a lower head and lower forces within the pump, and hence will reduce wear and lower maintenance costs.
C. Yes, because it will put the process under better control.
D. No, because it will cause more energy to be expended.
E. No, because energy savings will not justify the modification.
F. Yes, because it will save energy, and the energy savings will justify the modification.
Commentary
Without further clarification, any of the above answers could be correct, and in certain instances, more than one answer might apply. In any case, Answer A is correct because the variable speed drive will start and stop the motor/pump more gently than the starter. Soft starts and stops exert less stress on the pump/motor, so one would expect the equipment to last longer and cost less to maintain. However, higher motor temperature due to harmonics coupled with reduced fan cooling of the motor might mitigate some of the benefit.
If the motor/pump is operated at reduced speed, the pump will produce a lower head than that of the motor/pump operating at full speed. The lower head produced will reduce forces within the pump, so one would expect the equipment to last longer and cost less to maintain. Answer B is correct when the motor/pump is not operated at or above full speed.
If the process can be made continuous by installing a variable speed drive and appropriate controls, then Answer C will likely also be correct. This is because in many processes, the elimination of cycling improves the ability to control the process in a steady manner. Note that the economic benefit of this improvement can be nonexistent, such as the case where a pump is turned on and off to maintain a minimum level in a feed tank.
If the motor/pump continues to turn on and off and operates its variable speed drive at full speed to produce the same hydraulic output using the same mechanical input, electrical energy consumption will increase due to the losses associated with the variable speed drive. This strategy may obviate the need to modify the controls, but it can be a detriment because the expenditure to replace the starter with a variable speed drive results in increased operating costs. In this case, Answer D would be correct.
Further, if the motor/pump operates continuously at a speed that is reduced slightly from full speed, the energy savings may not offset the losses of the variable speed drive, so no energy savings are realized. As the operating speed is reduced further, the energy savings may offset the variable speed drive losses and produce energy savings. However, these savings may not be sufficient to justify the cost of replacing the starter with a variable speed drive, so Answer E would apply.
As speed is reduced further yet, the energy savings may dwarf the variable speed drive losses and result in significant energy savings such that Answer F applies and the modification can be justified.
Additional Complicating Factors
All processes are different, so there may be overriding factors that necessitate the installation of a variable speed drive. In one application, the installation of a variable speed drive and a larger pump were dictated because the fluid contained crystals were handled more “gently” at lower motor/pump speeds.
From Flow Control (September 2002)

Premium versus Standard Motors

By David W Spitzer
Premium motors that are more efficient than standard motors have been available for years. Consider the decision as to whether to use a standard motor or a premium motor in a centrifugal pumping application. Simply stated — evaluate how long it will take to save enough energy to recover the cost differential between a premium motor and a standard motor.
Which of the following are needed to decide whether to select a premium motor or a standard motor?
A. Cost of each motor
B. Efficiency of each motor
C. Nameplate speed of each motor
D. Pump efficiency
Commentary
The cost of the standard motor is subtracted from the cost of the premium motor to calculate the additional investment needed to purchase the premium motor. The additional investment should be justified by the energy cost savings generated when operating the premium motor. Therefore, the cost of each motor (Answer A) is needed.
If the two motors have identical efficiencies, there would be no energy savings and the additional investment would not be recovered. If the premium motor reduced energy consumption as compared to a standard motor by a few percent to produce the same brake horsepower, energy savings would be realized. The amount of energy savings would be used to calculate how long it would take to recover the additional investment associated with installing the premium motor. Therefore, the efficiency of each motor (Answer B) is needed.
The motor nameplate speed is the actual speed of the motor during full speed operation at full mechanical load. This will be less than synchronous speed due to slip. Premium motors typically exhibit less slip than standard motors, so the nameplate speed of a premium motor is typically higher than that of a standard motor. Because the premium motor runs faster than the standard motor, the pump will operate faster and require more mechanical energy to operate at the higher speed. In the case of a centrifugal pump, an increase in nameplate speed from 3450 to 3500 revolutions per minute (rpm) will increase brake horsepower requirements by the cube of the speed. In this example, the increased motor/pump speed increases brake horsepower requirements (and energy costs) to (3500/3450)3 times the requirements at 3450 rpm, or approximately 4.4 percent higher. Therefore, the nameplate speed of each motor (Answer C) is needed.
Depending on the availability of certain pump curves, the efficiency of the pump may be required to calculate the additional operating costs associated with the speed increase (Answer D).
In summary, energy savings attributable to the installation of premium motors can be offset by the increased energy consumption associated with the increased mechanical loading caused by the higher motor/pump speed.
Additional Complicating Factors
If the motor is operated by a variable speed drive, the nameplate speed of each motor (Answer C) and the pump efficiency (Answer D) are not needed because the motor will typically operate at reduced speed to meet demand.
From Flow Control (October 2002)