Metering Pumps: The Lowdown on Turndown
Consumers are programmed from an early age to expect that more is better. Why buy a 12 ounce soda when you can get 24 or 32 ounces for a few pennies more? My car has 27 more horsepower than yours! Your pitiful 700 watt audio system doesn’t destroy eardrums nearly as well as my 1,200 watt! So it only makes sense that when we need to buy equipment for a chemical feed process – we once again look for more.
In the world of metering pumps there’s an increasing trend to boast incredible turndown capabilities of 1,000 to 1, 3000 to 1, and even 1,000,000 to 1. On the surface, increasing the turndown ratings to these levels appears to provide a lot more than was previously available. After all, 1000 to 1 is 10 times better than 100 to 1 – right?
Well, not really.
To explore these claims of incredible turndown, there are three very basic questions that need to be answered:
1. What turndown is really required?
2. What does that kind of turndown mean?
3. How can large ratios effectively be measured or verified?
We’ll discuss these three issues, but first let’s define turndown and how it’s related to accuracy, and also look at the background of metering pumps and their turndown capabilities.
Turndown is an easy enough concept - It is expressed as a ratio over which you can adjust the output of a metering pump. If a pump is capable of 10 gallons per hour (GPH) maximum and it has a 10 to 1 turndown ratio, then it can be turned down to 1 gallon per hour. That’s simple. API and the Hydraulic Institute have excellent definitions as well.
But assigning a turndown ratio without it being tied to some form of accuracy is useless. Without noting accuracy, a metering pump’s turndown is basically infinite in that you can go from maximum flow to no flow at all.
Accuracy is another term used loosely. If a pump is noted as having ± 1% over 1000 to 1 turndown, that can mean several things.
For example, assume a pump has a maximum capacity rating of 1000 gallons per hour (GPH). If accuracy is based on the maximum pump rating, that means the pump can vary 1% of 1000 gallons, or 10 gallons, at any capacity setting. Turn the pump down to its minimum capacity of 1 GPH, and flow can vary ± 10 GPH. Table 1 organizes this example to better show why this rating is useless. This table shows that as the pump capacity setting is reduced, the actual resulting accuracy is less and less effective in the process. An actual flow 10 times expected flow doesn’t have the right to be called accurate.
Graph 1 shows the results of a more useful accuracy rating that is based on a deviation from the pump’s current capacity setting, or set point. When the accuracy rating is based on set point, the accuracy expected has much more meaning. It doesn’t matter what capacity the process requires over the pump’s turndown ratio – the pump output will be at the expected flow within a reasonable deviation. In this example the accuracy percentage has been increased from ± 1% to ± 5% so the variance can be seen. If the graph were drawn for the industry standard of ± 1% accuracy, the lines would be on top of each other. This graph shows percent deviation, which is a constant ± 5%.
Graph 2 shows how this relates to the pump’s actual flow rate. In this example the accuracy has been moved back to the industry standard value of ± 1%, so the lines are very close to each other. This graph is more representative of the expected perfomance of a true metering pump. It’s hard to see on the graph, so here are the important flow and performance characteristics:
At maximum capacity setting (100%) where 1000 GPH is expected:
- Maximum allowable actual flow: 1010 GPH
- Minimum allowable actual flow: 990 GPH
At minimum capacity setting (1%) where 1 GPH is expected:
- Maximum allowable actual Flow: 1.01 GPH
- Minimum allowable actual Flow: 0.99 GPH
Graphs 1 and 2 represent steady state accuracy, which more accurately corresponds to the performance of a true metering pump. Steady state accuracy can be defined as a pump’s flow rate variation expressed as a ± percentage of mean delivered flow under fixed system conditions applied over the full turndown ratio of the pump.
Flow repeatability is also a form of accuracy. Also stated as percentage of rated capacity, it describes how much the pump output is allowed to deviate when its capacity setting is changed, and then returned to the original setting. So if the pump is providing 50 GPH at 40% capacity setting and the pump is changed to 80%, when the setting is returned to 40% the pump will still provide a flow near 50 GPH within the flow repeatability accuracy. Typically, metering pumps are rated at ± 3% flow repeatability. Many process engineers order 10 point curves with the pump to show this variation. The 10 point curve shows output at 5 points descending, then the same 5 points ascending, and compares the two.
Manufacturers of many different types of pumps try to attach a variable speed drive to their units and call the resulting product a metering pump. While it’s true that the result will be a pump that varies capacity, the performance and characteristics will not match the industry expectations of a metering pump. Depending on the type of pump attempting this crossover, repeatable performance and resulting flow rate can vary based on a number of factors including:
- Fatigue of elastomers, or hoses, or tubes
- Suction pressure or tank fluid height
- Minor discharge pressure fluctuations
- Hydraulic inefficiencies
Any one of these issues will impact the performance of pump that fails to meet the design criteria of a true metering pump.
Metering pumps, also referred to as dosing, proportioning, or chemical injection pumps, are defined as controlled volume positive displacement pumps with the ability to precisely control process fluid volume by varying effective stroke length. In the past several decades, reciprocating pumps with constant stroke length and variable stroke frequency have also been recognized as metering pumps. In general, the attributes that define a metering pump include:
- Positive displacement - a general category of pumps that differ from centrifugal pumps in that they will displace any process fluid from within the pump head
- Integrated mechanism that enables pump capacity adjustment
- Turndown ratios of 10 to 1 or higher while maintaining steady state accuracy
- High level of steady state accuracy over the pump’s turndown ratio - usually ± 2% or better
- High level of repeatable accuracy – usually ±3%
- Minimal variations in pump flow due to changes in discharge pressure – in other words the change in pump capacity is usually less than 1% for every 100 psi change in discharge pressure
Turndown ratios of 10 to 1 satisfied the vast majority of applications for many years. But several trends influenced the increasing requirement for higher turndowns:
- Chemicals began to be pumped at higher and higher concentrations to avoid larger storage tanks or the use of day tanks
- Process facilities or treatment plants are designed to operate over a wider range. For example, a new water treatment plant may be started at 30% of its full capacity while a region is in the early stages of population growth
- Pumps may be expected to operate with different chemicals within a process, each of which requires its own dosage rate
- Chemical feed calculations based on pilot studies may require a significant change once in the actual process
- Drive technology has advanced over the years and their marketing efforts have been directed at ways to capitalize on the increased capability
Turndown ratios have always been stretched by offering combinations of variable speed along with the pump’s intergral capacity control of effective stroke length. Theoretically, the standard metering pump’s 10 to 1 turndown combined with a variable speed drive’s 10 to 1 turndown yields a total of 100 to 1. In reality, combining the two doesn’t always yield the expected results with required accuracy as the combination starts to introduce hydraulic inefficiencies as both speed and stroke length reach the lowest limits. In the 1980’s, designs were introduced that focused just on variable speed using drives with expanded turndowns and digital technology. These unique drives proved successful in achieving 100 to 1 turndown while maintaining the steady state accuracy required. This success started the industry down the “more is better” pathway. The resulting claims of 1000 to1 all the way up to 1,000,000 to 1 found their way into brochures. That takes us back to the questions:
- What’s required?
- What does high turndown mean?
- How do you measure it?
How Much Turndown Does a Process Need?
Every process engineer needs to answer this question when designing a process. Most industrial processes stay within a small range of pump flows. In these applications, the standard 10 to 1 is usually more than enough range.
Municipal applications tend to benefit more from increased turndown. For example, if coagulant is injected to maintain a required level of turbidity after the filtration process, the amount of coagulant will depend on the quaility of the water source. A large lake with only small variations in turbidity may only need small dosage changes. But a small river or creek that’s subject to large changes based on the storms upstream could require large changes in dosage. Turndowns of 40 to 1 or 50 to1 is a real possiblity. Couple that with the need to change pump flow rate based on plant flow rate, turndown ratios near 100 to1 are realistic.
In actual processes, the need to go beyond 10 to1 is a small percentage of the total of all applications. One reason some engineers like the high turndown ratios is that it can make up for variations or errors in dosage calculations. There’s nothing wrong with recognizing this possiblity, but installing an oversized pump that will always be operating below 1% of its maximum capacity is like using an 18 wheeler to deliver a few music CD’s bought on-line. Efficiency is lost at a number of levels.
What does High Turndown really mean?
When we look at what high turndown really means to the flow of the pump, it’s easy to see the true benefit.
– Start with a pump rated for 100 GPH. At 10 to1 turndown ratio, the pump gives excellent performance between 10 GPH and 100 GPH. The benefit is clear in that 90% of the pump’s maximum capacity is within its usable range.
– If the same pump provides 100 to1 turndown, the range is extended to 99% of its usable range: 100 GPH to 1 GPH. 10 to 1 verses 100 to 1 sounds like 10 times more, but it only increases the usable range by 9% of full capacity.
– Now increase this pump’s turndown to 1,000 to 1 and the real increase in turndown is miniscule. The performance of the pump will now be usable between 100 GPH and 0.1 GPH. The pump now provides an additional accurate flow rating of 0.9 GPH, 0.9 % of full capacity. Again, it sounds like 10 times more than 100 to1, or 100 times more than 10 to 1, but the real result is quite small.
Table 1 provides a way to look at the increase in accurate flow that is achieved by increasing turndown. Observe the amount of extra capacity that is realized by increasing the turndown 10 times that of the row above. Rating a 100 GPH pump with a turndown ratio of 10,000 to 1 only provides 0.09 GPH more range than a 1,000 to 1 rating.
Clearly, wide turndowns do not really increase what the pump can do by all that much. Once the claim of turndown goes beyond 100 to 1, there’s little benefit. In fact, beyond 100 to 1 it raises the question of pump efficiency. The motor and drive is sized to provide 100 gallons per hour. If the pump operates at 0.2 % of its design, efficiency has to suffer. But the other question this raises is just as important.
How do you measure wide turndown ratios?
Flow meters have a useable turndown too, and 1000 to 1 is well beyond the capability of most. Loss in weight measurement of the fluid being pumped (usually a digital scale under the supply tank) is a possiblility, but it’s a rare process that has this hardware installed. Calibration columns require visual observation of the amount of chemical pumped, which introduces human error that causes inaccuracies. At wide turndown ratios, the difference between 10 to 1 and 1000 to 1 may be less than a drop of fluid and there is no chance anyone can see the change in a calibration column, flow meters cannot handle the range, and other potential methods are impractical or too expensive to justify.
So the answer is simple – wild claims of wide turndown ratios cannot be effectively measured or confirmed in the installation. Up to 100 to 1 is easily measured and verified within the steady state accuracy of the pump. 1000 to 1 is more difficult, but can be measured in a controlled setting at the manufacturer’s testing facility. When turndown claims reach 10,000 to 1 or higher, the claim cannot be substantiated. If the rating can’t be measured or proven, that makes it safe to boast exaggerated claims of large turndown ratios. To make a claim that a pump can provide 1,000,000 to 1 turndown, especially without an accuracy rating, is nothing more than marketing fluff without any basis in reality. Companies that make these kinds of claims usually do not manufacture real metering pumps, and resort to meaningless exagerations to project an image of accuracy that isn’t really there.
Can there be an efficent metering pump with 1,000,000 to 1 turndown and a steady state accuracy of ± 2% or better? Let those making the claim prove it, because I seriously doubt it!
Exaggerated claims that cannot be verified are not only misleading, but in the actual installation provide very little benefit to the process. A professional metering pump company rates the turndown of its products based on solid design and verifyable testing methods. In those tests, every pump manufactured can be tested at the turndown ratings while maintaining repeatability and steady state accuracy.
The basic questions that need to be asked related to turndown are:
– What turndown does the process really require?
– If it exceeds 100 to 1, what is the best and most efficient way to address it?
– Will the pump selected provide the turndown needed with the accuracy expected from a true metering pump regardless of pressure fluctuations and the age of the equipment?
If a pump is purchased based on exaggerated turndown that isn’t required by the process, the buyer may not realize some other important issues:
- Most pumps offered with exaggerated claims are primarily plastic housings with significant electronic componentry that, by nature, usually last about 5 years at best. True metering pumps are robust pieces of equipment with 30 year design life.
- Pump designs that try to be metering pumps usually require much more frequent maintenance. For example, peristaltic pumps boast that they can go up to 6 months before needing maintenance. Actual experiences have seen many examples of hose replacement every few weeks. True metering pumps can run for years without maintenance and, in fact, have diaphragms with design life ratings as high as 96,000 hours (almost 11 years running 24/7).
- True metering pumps usually cost less to buy and cost less to maintain. As a taxpayer that is the performance I’d expect at my local municipal treatment plant. Process managers and operators recognize the impact in reducing capital costs and controlling operating expenses.
- Steady state accuracy and repeatability assure accurate dosage and that ties directly to reduced chemical usage while maintaining chemical treatment performance in the process. In potable water, for example, this assures compliant water in the distribution system. Only a true controlled volume metering pump provides the performance required for this level of process integrity.