Advances in Coriolis Flow Metering Technology Help Keep Profit Flowing by Resolving Toughest Measurement Challenges

by Wade Mattar

Flow Specialist, Invensys/Foxboro

http://iom.invensys.com/UK/Pages/Foxboro_MandI.aspx  or http://iom.invensys.com

 

Coriolis technology offers unprecedented accuracy and reliability in measuring material flow, and is often hailed as among the most superior flow measurement technologies. But conventional Coriolis meters have had one significant limitation: they haven’t performed well in measuring two-phase flow conditions, flow which involves a combination of gas and liquid mass. Two-phase flow can cause process interruptions, measurement inaccuracies, and poor product quality, which can quickly sap the profitability of a production run. This paper addresses the application of Coriolis technology, explains the challenges of measuring two-phase flow, and describes technology from Invensys that enables two-phase flow measurement in even the most demanding applications.

 

How Coriolis measurement works

 

With worldwide revenues expected to grow from $412 million in 2000 to $671 million in 2005, Coriolis meters are among the fastest-growing flow measurement technologies.1 These meters measure flow by analyzing changes in Coriolis force of a flowing substance. Coriolis force is generated in a mass which is moving within a rotating frame of reference. That rotation produces an angular, outward acceleration, which is factored with linear velocity to define the Coriolis force. With a fluid mass, the Coriolis force is proportional to the mass flowrate of that fluid.

 

 Because industrial process lines do not typically feature rotating pipes, exploiting the Coriolis principle for flow measurement requires oscillating a section of pipe. This produces the Coriolis force, which can be sensed and analyzed to determine the rate of flow. To use Coriolis force for measurement, a Coriolis meter has two main components: an oscillating flowtube equipped with sensors and drivers, and an electronic transmitter that controls the oscillations, analyzes the results, and transmits the information.

 

Reliable Coriolis measurement depends on consistent, reliable oscillation, which is determined by the following four factors: the density of the liquid, the balance of the tubes, the dampening caused by the flow stream itself, and the physical isolation of the tubes from the environment. Compromising even one of these factors will degrade Coriolis meter performance. Yet two-phase flow compromises every one of them. Thus applications involving negligible amounts of entrained gas — even as little as 2 percent volume — have been poor candidates for Coriolis measurement. This has been particularly troubling for those running batch operations, where reliable, highly accurate flow measurement can confer considerable bottom-line advantage, but where it is also necessary to begin with an empty or partially filled flowtube.

 

Making matters worse, entrained air may not emerge as the culprit until after a frazzled process engineer has invested many hours trying to figure out why he can’t get the results he needs. One senior engineer at a leading international producer of specialty chemicals believes that process engineers spend more time trying to resolve problems with Coriolis mass flowmeters than with all other technologies put together, and that the inability to measure two-phase flow accurately is the most common source of their woe. Our own analysis shows that up to 92 percent of all Coriolis measurement problems are due to entrained air or gas, yet in the vast majority of cases two-phase flow is not even recognized as the problem.

 

Where accurate flow measurement matters

 

New methods of production cost accounting are beginning to reveal just how bad the problem really is. Where conventional methods of monitoring profit do not provide managers with enough granularity to identify and capture hidden profit opportunities, new “profit-per-minute2” analyses can show quite dramatically how continuous inability to measure two-phase flow compromises product quality and causes costly production downtime.

 

Coriolis technology is highly accurate in single-phase flow, with perhaps a ± 0.02 percent error level. Two-phase flow boosts the error rate to 20 percent or higher. At low levels of two-phase flow, conventional Coriolis meters tend to under-predict actual flow; at higher levels, they tend to over-predict.3 Following are some of the profitability drains that inaccurate flow measurement causes:

·       Lost production: In flow-intensive operations such as most pharmaceutical and specialty chemical operations, thousands of dollars’ worth of lost production can pass undetected in minutes.

·       Inaccurate pricing: In custody transfer applications, where measured amount transferred defines payment price, faulty measurements raise financial havoc on either end of the transfer.

·       Compromised quality: In batch operations, for example, where quality is very much a function of recipe consistency, inaccurate flow measurement can diminish the quality of the output and sap profitability through waste, rework, and customer dissatisfaction.

·       Excess downtime:  When traditional Coriolis meters encounter entrained air, they render inaccurate measurements — and if the condition persists, will shut down. Whatever material is in production at the time of shutdown is lost, along with valuable production time. Rework and restart add additional costs as well.

 

Living with the enemy

 

Until now, process engineers by and large have accepted the problem of two-phase flow measurement as something they “just have to live with.”  Often not even connecting their difficulties to measurement of two-phase flow, they attribute unaccountable material loss or overage to process irregularities (such as erratic cure times) and process interruptions, and struggle to make their numbers despite such limitations.

 

Those who do identify two-phase flow as the culprit often create costly workarounds. One example involves installation of venting devices upstream in the process. This proves ineffective, however, because it does not capture 100 percent of the air, and also exposes the process to the atmosphere, which itself may compromise product quality. In pharmaceutical, food and beverage, and other sanitary applications that require a closed-loop system, installing venting devices is not an option at all.

 

Another workaround involves removing air from the process liquid. This gets quite complicated because air can enter from a wide variety of possible points, including inbound transit of raw materials, leaks in pump suctions, tank agitators that whip air into the product, and tank infeeds with long-distance delivery systems. Eliminating or modifying so many sources would be costly. Removal of entrained air through deaeration technology is also expensive, and in some cases impractical.

 

Some companies have actually addressed the problem through marketing, by promoting entrained air as a product feature. This is especially possible in beauty and bath, food, and dairy products, where presence of bubbles in a clear tube, for example, becomes part of the packaging.

 

To catch a thief

 

To determine the extent of the problem and find a cost-effective solution, Invensys commissioned a survey of process engineers. This revealed that entrained air was indeed a major problem in the industry, and that what customers really wanted was a cost-effective meter that could provide accurate measurement despite the presence of air. Invensys partnered with researchers at Oxford University in England to develop digital technology for accurate measurement of flow, even when air was entrained in the flowtube.

 

Working closely with the Oxford researchers, engineers at the Foxboro Measurements and Instruments division of Invensys developed a transmitter that applied the Oxford measurement principles via a dual digital processing system.4,5 In the resulting patented product, called the Foxboro CFT50 digital Coriolis flow transmitter, one processor controls the flowtube using digital circuitry. The drive waveform that initiates and maintains flowtube oscillation is synthesized digitally at the correct amplitude, frequency, and phase. Because it is digital, the drive can be adapted to meet changing flow conditions rapidly and precisely. A second processor digitizes the flow measurement data signals from the Coriolis meter for precise fluid measurements.

 

Because the processor can compensate for physical conditions, the CFT50 can be used in any fluid metering application, including traditionally difficult-to-handle materials such as slurries, non-homogenous fluids, and problematic fluids that foam or flash. When used with a 3-A approved flowtube, the CFT50 flowmeter is also suitable for sanitary applications. The CFT50 is the only available Coriolis mass flowmeter that can perform in batch applications starting with empty flowtube conditions. Even with the flowtube empty, the CFT50 responds ten times faster than traditional Coriolis transmitters, which reduces startup time while increasing production throughput and profitability. The following examples demonstrate how the system is used to produce accurate flow measurement in three challenging applications: measuring ethylene oxide flow, batch processing, and unloading of trucks and railcars.

 

Ethylene oxide
 

Ethylene oxide (EO) is commonly used as either a raw material or sanitation fluid, and presents many metering challenges. It is, for example, generally used at or close to its boiling point, and also often pushed with nitrogen rather than pumped. So when EO flows, it has a tendency to boil, break out dissolved nitrogen, or both, resulting in two-phase flow. Nonetheless, many users have tried to measure EO flow with Coriolis and have obtained accurate measurements for about 80 percent of their process, until the EO flashes or breaks out nitrogen.

 

During these periods, the gas void fraction (GVF) — the percent by volume of gas in the liquid — can be 50 percent or greater for brief intervals. Even when the short-lived two-phase event ends, the traditional Coriolis meter may take tens of seconds to recover and give accurate measurements. This of course makes the measurement unreliable and incapable of being the sole measurement in the process. So at best, some technique other than flow control is needed to regulate the process.

 

The CFT50 overcomes this problem. Figure 1 shows typical measurements from the CFT50 over a 20-hour period in an industrial EO application.  Note that, although the GVF gets higher than 50 percent quite often, the Foxboro Coriolis meter continues to provide a useful mass flow measurement.

 

Figure 1: Ethylene Oxide Application

 

This particular user had tried to employ traditional Coriolis meters in this very location to control EO batching, but without success. Since using the new Coriolis flowmeter, he is now able to trust the mass meter and is using it as his primary measurement. Note also in Figure 1 the frequency of troubling levels of GVF during this 20-hour period. Only 49 percent of the time is the flow without some level of entrained gas. For 20 percent of a typical day’s runtime of 20 hours, there is in excess of 5 percent GVF — often actually above 10 percent GVF. So for at least 20 percent of the time, traditional Coriolis flowmeters would have serious problems with this flow measurement.

 

Batching from empty

 

Batching processes must either start or finish with an empty or partially full flowtube. This is a common occurrence in sanitary applications, where the entire wetted area has to be cleaned or sanitized either daily or between batches. It is generally not possible to keep a meter and associated pipework full, at least for the first batch of the day. So many users have accept that the first batch of the day will be somewhat “off,” and may even have to be reprocessed.

 

Another issue in batch processing is that many users continue to draw from emptying supply tanks until they are practically dry. This introduces large amounts of gas when the supply tank is nearing the bottom. As a result of such issues, Coriolis meters have serious measurement problems at both the start and the end of the batch. Figures 2 and 3 compare a CFT50 with traditional Coriolis meter performance in a batching application that starts partially full and finishes partially empty. 6  The application, at Great Lakes Chemicals in Manchester, UK, had been using a traditional Coriolis meter, which had proved unsatisfactory. An opportunity arose to install a CFT50-based flowmeter in series with the traditional meter, so that real-time performance comparisons proved possible.

 

Figure 2: Start of Batch

 

Figure 2 shows the startup. Values have been normalized so that density and mass flow can be shown on the same graph. Densities have been normalized to the maximum value of 1156 kg/m3 and flows have been normalized to the maximum value of 14.4 tonne/hour.

 

At the very beginning of the batch, the meter is partially full, as signified by the indicated normalized density of about 0.1. During this period both the new and the traditional Coriolis meters are indicating zero flow, as they should. After about 18 seconds, as indicated by the increase in density, flow begins. Although the meters are not full until about 30 seconds into the batch, there is flow that needs to be metered immediately. Notice in Figure 3 how the traditional Coriolis meter is unable to deal with the partially full condition. This is manifest not only by its missing the process fluid passing through; the meter produces a substantially negative indication, which will obviously affect the overall batch total.  Even after the meters have reached a full condition, at about 30 seconds the traditional Coriolis meter exhibits an additional 5-second lag before indicating full flow. Meanwhile the CFT50 not only indicates full flow immediately when full, it also gives a flow indication while partially full. So roughly 15 seconds’ worth of flow at the beginning of the batch is measured by the new technology that would otherwise have been lost. Over a series of short batches, this lost measurement can produce substantial losses.

 

Figure 3: End of Batch

 

Similarly, Figure 3 depicts the end of the same batch. At this stage, the desire is to pump in as much material as possible to avoid any waste or additional manual labor needed to utlilize the bottom of the supply tank. Yet as soon as the density reading indicates a partially filled condition, the traditional Coriolis meter shuts down. It stays offline for several minutes, entirely missing the final blow-through

of product, where the flow is never entirely free of air. By contrast, the CFT50 keeps on operating, providing an accurate indication of the true mass flowrate. This allows the user to drain the supply tank while still in automatic control, preventing any waste or additional labor costs.

 

Unloading railcars and tank trucks

The end-of-batch problem depicted in the previous section is also encountered when unloading a railcar or tank truck. The requirement is to empty out the tank completely, which invariably introduces air as the level approaches bottom. This is exacerbated by the fact that in most cases unloading is done at as high a flowrate as possible to speed up the process. This high flowrate tends to suck air into the flowmeter. Where a conventional Coriolis meter would shut down in this situation, the CFT50 Coriolis meter will continue to provide a useful flow measurement, enabling faster, more complete unloading of tank trucks and railcars.

 

Is two-phase flow a problem in your process?

 

The above cases are but a small sampling of the many ways in which the benefits of Coriolis accuracy can be attained in areas that have been traditionally out of reach. Every day we are seeing new applications wherein Coriolis flowmeters are being successfully used to measure processes involving chocolate, whipped butter, ice cream, tomato paste, paint, and even for substances such as ground beef!  There is no reason the digital technology in the CFT50 cannot be applied in these situations as well.

 

So take a look at your flow measurement challenges. Are you simply writing off lost materials without knowing exactly what is causing them? Are you experiencing downtime that could be better spent producing profit? Are you investing in complex, costly workarounds to remove air or gas from your processes? If you answer yes to any of these questions, or just feel that better flow measurement would improve your process in any way — entrained gas may be the culprit, and implementing a CFT50 flow meter may be the solution to the problem you never even knew you had.

 

Parts of this article were presented by the author at the IEE Seminar on Advanced Coriolis Mass Flow Metering, Oxford University, United Kingdom. Those parts are copyright The Institute of Electrical Engineers, London, 2003. Used with permission.

 

References:

 

1 Yoder J., 2002. The World Market for Coriolis Flowmeters. Market study conducted by Flow Research and Ducker Worldwide.

 

2 Maxager Technology, Inc.,1999. White Paper: Rethinking Profit Analysis: Profit per Unit vs. Profit per Minute.

 

3 Reizner J., 2003.  IEE Seminar on Advanced Coriolis Mass Flow Metering, Oxford University, Coriolis – The Almost Perfect Flowmeter.

 

4  Henry et al., Oct. 30, 2001. United States Patent No.:

US 6,311,136 B1. Digital Flowmeter.

 

5  Henry et al., Jan. 14, 2003. United States Patent No.:

US 6,507,791 B2. Digital Flowmeter.

 

6 Control Engineering Europe, June/July, 2003. New Coriolis meter finds hundreds of kilograms of ‘disappearing’ batch product

 

 

 

This article was written by Wade Mattar, Flow Specialist at Invensys/Foxboro and is re-published here with the permission of Invensys Foxboro.  Foxboro is part of the Invensys Production Management Division and provides world class Foxboro information technology, automation, and process solutions to a wide range of manufacturing applications for the cement, chemical, metals & mining, oil & gas, pulp & paper, power, pharmaceutical and specialty chemicals. Besides Foxboro, the division includes APV, Baan, SimSci-Esscor, Triconex and Wonderware.

The Invensys Production Management Division of Invensys plc works closely with customers to provide solutions that help them maximize return on investments and optimize performance across their supply chain. Invensys has performance improvement expertise and technology encompassing the entire value chain -- from the production line to executive offices; from customer relationship and procurement management to distribution logistics. With more than 50,000 worldwide installations, we serve the oil, gas, and chemicals sectors; power generation; food, beverage and personal healthcare; fine chemicals/pharmaceuticals; pulp and paper; mining and cement; and discrete and hybrid manufacturing sectors. The division headquarters are in Foxboro, Massachusetts USA.

For more information on Invensys Foxboro, please visit their website at www.foxboro.com/m&i or www.invensys.com.