Comment

01 March 2005

Batch accuracy

Good, bad, and ugly of Coriolis.

By James Reizner

Coriolis mass flowmeters are the most accurate of the industrial flow measurement technologies and the only flowmeters that claim to measure true liquid mass flow directly.

Coriolis flowmeters see use in continuous and batch processes. In batching processes, they are a less-expensive alternative to load cell weighing systems. As with any other flowmetering technology, Coriolis has its own application "watch outs," including a requirement that the flowmeter be totally and completely full of non-aerated liquid during the measurement.

To be able to properly apply Coriolis mass flowmeters in batch applications, it is important to understand the basic operational concept of the Coriolis meter. It is also important to recognize what the meter is not. Unlike that presented in much of the literature, a Coriolis mass flowmeter does not directly measure mass flow. There is no mass flow sensor internal to a Coriolis meter. Likewise, a Coriolis meter does not measure Coriolis forces; there is no Coriolis force sensor internal to the meter.

Unfortunately, the operation of a Coriolis mass flowmeter is not easy to explain. A simplified way of thinking about the operation of a Coriolis mass flowmeter follows. Fluid flow through the meter generates Coriolis forces that are generally proportional to the mass flow rate, at least for single phase fluids. These Coriolis forces generate a movement within the meter that is generally proportional to the Coriolis forces, vibration of a tube by tube twisting induced by the Coriolis forces. A sensor, such as an electromagnetic detector, measures this movement. The signal from these electromagnetic detectors undergoes processing by sophisticated electronics, which have their own impact on signal accuracy and signal speed of response and latency. There are many references available to those desiring a more theoretical description.

The point of this is to realize that Coriolis mass flowmeters do not directly measure mass flow any more than strain gauge load cells directly measure weight. By understanding the basics of the measurement principle, what the technology can and cannot do, one is apt to be more successful in applying the technology.

Coriolis mass flowmeters generally are less expensive than load cell weigh tanks, especially when considered from a total installed cost perspective. Load cell weigh tanks are complex systems requiring in-depth knowledge in load cell system electronics, piping systems, and structural/civil engineering disciplines. All of this means fairly high engineering costs. In addition, construction/installation costs for load cell systems are high. The structural/civil requirements for load cell systems are stringent.

The structural steel required to properly support a load cell weigh tank can be substantial, and on retrofit jobs, it can be challenging just to get the required strength designed into the structure. Coriolis meters are somewhat expensive themselves, and do require some level of care in piping. Coriolis meters generally do not require any special civil/structural considerations, and overall, the installed cost for Coriolis meters is substantially lower than that for weigh tanks.

Coriolis meters have the ability to handle multiple feeds at the same time, something that weigh tanks cannot do by themselves. For example, a weigh tank may be weighing the direct addition of a dry material via a screw feeder. By adding flowmeters to the system, the system can concurrently measure the addition of multiple liquid feed streams. You can keep tract of the concurrent measurement of the dry material by subtracting the totals of the liquid feeds from the change in the weigh tank reading, resulting in a measurement of the dry material addition itself.

Miniscule measurement

Coriolis meters are complimentary to weigh tanks for accurately adding small weights. Consider the following example: There is a 1,500 kg weigh tank, to which you add 900 kg of product. During the addition of this product, we also want to add 10 grams of a highly concentrated perfume.

Because of the resolution of the load cell system, it would be impossible to accurately weigh the 10 grams of perfume. Enter the complimentary ability of the Coriolis mass flowmeter, and you can accurately and concurrently add the required perfume.

To ensure success, certain application considerations must go into account for Coriolis mass flowmeters. These considerations include the requirement that the tube be completely full of liquid, that the liquid be non-aerated, and that the batch time be commensurate with the capabilities of the meter.

It is a fact that traditional Coriolis mass flowmeters must be completely and totally full of fluid while measuring, or significant errors will occur. Achieving this requirement in industrial piping installations can be challenging, especially in systems where product line cleanouts between batches result in empty pipelines. At startup of systems that are subject to beginning with empty pipes, users will install elaborate systems to ensure Coriolis meters are full before they begin their totalization. In other cases, the meter inaccuracy during its empty/partially full period is accepted, or sometimes not even recognized. At least two of the Coriolis meter manufacturers introduced meters that claim to be able to accurately operate in batching situations when starting from empty.

Key to accuracy This is an example of the type of complex piping system that is sometimes required for accurate batching systems. To ensure accuracy, the automated cutoff valve (E) and the mass flowmeter must be as close to the receiving vessel as practical. Long lengths of piping between the cutoff valve and the vessel can result in a significant quantity of liquid dribbling out of the pipe over time, a problem that will negatively affect batch accuracy and batch speed.

Problem areas

Traditional Coriolis mass flowmeters cannot accurately measure two-phase slug flow. Slug flow is the condition where a slug of liquid is followed by a slug of gas, generally in rapid succession.

Less commonly understood is Coriolis meters have difficulties dealing with aerated liquids.

Errors as large as 58% have been reported, and in one test, 2% to 4% aeration caused meters from eight manufacturers to respond well outside of their stated specifications. At least two of the Coriolis meter manufacturers have products available that they claim have addressed this issue.

Latency will be simply defined here as the time it takes for the flowmeter output to start to respond to a change in flow. Response time indicates how fast a device can respond to periodically changing flow rates. Some in the industry believe Coriolis mass flowmeters are fast-responding devices, but research shows otherwise. Research compared the response time of Coriolis to other flow measurement technologies. Latency and response time are critical parameters for flowmeters used in batching applications, especially in batches of short durations such as bottle filling. If a flowmeter is used to fill a bottle, and that bottle fills in one second, to be accurate that flowmeter better have a very low latency and fast response time. It is generally regarded in industry that the very best currently available Coriolis mass flowmeters have difficulties dealing with batch times shorter than 1/2 second, and many meters have difficulty with batches as long as 20 seconds.

Our experience indicates we could accomplish batch times as short as one or two seconds, depending on the accuracy requirements, but that for any batch time of 10 seconds or less we recommend trial runs to ensure you can meet application requirements.

The topic of Coriolis meter response time is significant enough that the Universities of Oxford and Brunel (both in England) have a joint grant program "to push Coriolis metering technology to its dynamic limit." Several engineers have garnered U.S. patents for a Coriolis meter they say can operate faster than conventional technology.

One area of interest for the dynamic response of Coriolis meters is in their use with pulsating positive displacement pumps. Positive displacement pumps traditionally pump the more viscous liquids often metered with Coriolis mass flowmeters. Many types of positive displacement pumps create substantial flow pulsations, specifically lobe, diaphragm, and piston pumps. In industry, larger pumps with slower but more violent pulsations have proven challenging for Coriolis mass flowmeters to accurately measure, especially when used in short batches. Smaller pumps and faster pumps find their flow pulsations to effectively average out by the relatively slow dynamic response of the Coriolis meters. Patents describe times where, because of this problem, Coriolis mass flowmeters see use as density measuring devices and pump rotation as a volumetric measurement.

Calibration woes

Conceptually, calibrating a strain gauge load cell weigh tank is straightforward. One places traceable test weights on the vessel and records the reading on the scale display. You can either put on or take off weights to generate a hysteresis curve for the system. Anyone who has ever calibrated a 3,000-pound load cell system by carrying 60 50-pound test weights and placing them multiple times on the vessel understands that although straightforward conceptually, in practice it requires a lot of physical effort.

Calibrating Coriolis mass flowmeters for flow rate measurements is a very difficult exercise, specifically because there is no traceable standard for flow rate. You cannot purchase a "pound per hour" anywhere. In general, such calibration of the flow rate for Coriolis mass flowmeters is best left to the manufacturer, who has the sophisticated lab systems required to do this calibration. Fortunately in batching systems, flow rate is not the important variable, rather total quantity of flow is the variable of interest. The most typical method of calibrating Coriolis mass flowmeters for use in batching systems involves batching a fluid over a given length of time into an accurate weigh tank, and then comparing the reading of the weigh scale to that from the totalizer on the Coriolis meter. Various issues arise with this method. First, you must accurately measure time, and hand-held stopwatches are not sufficient for the task. Dribble after valve closure, valve speed of response, valve consistency, weigh scale accuracy, and many other variables enter into this equation.

Coriolis mass flowmeters are a wonderful technology advancement over earlier flowmeters. Their high accuracy, wide rangeability, and mass-flow measurement qualities make them the "meter of choice" for many industrial applications. Coriolis is still a new and developing technology, and as such, has areas for improvement.

Here is a list of improvements that end-users wish to see in future Coriolis mass flowmeters:

Ability to "batch from empty." Since many batch operations require cleanouts between runs, batching from empty (the condition where the pipelines and system are empty or at least less than full of product at the beginning of the batch) is an important area of improvement for Coriolis.

Ability to accurately measure aerated liquids and two-phase flow. Coriolis meters see widespread use to meter viscous fluids, which tend to keep air entrained in them. Improving Coriolis meters so they can handle aerated fluids will greatly enhance the industrial fluid applications for the meters. It can also minimize end-user scrapped batches due to flowmeter errors caused by entrained air.

Faster response time, less latency—ability to do very short batches of 1/2 second or less, even with pumps with pulsating flows. Bottle filling machines are moving to mass flow control using Coriolis meters and away from volumetric fill using magnetic flowmeters or pistons. Coriolis offers many advantages like wide rangeability to handle bottles of varying sizes, high accuracy, and mass flow measurement ability. End-users desire ever-shorter bottle filling times. Pumps that create pulsate flow, positive displacement lobe pumps for example, generate high-speed flow variations that have proven difficult for larger Coriolis meters to accurately meter, especially in short batching applications. Development of faster response time Coriolis meters will help end-users meet this currently unmet need.

Meters easier to calibrate/verify accuracy. Load cell weighing systems are conceptually easy to calibrate, just use traceable calibration weights. Calibration/verification of Coriolis meters is a little more difficult. For continuous flow applications, one issue is a "pound per hour" is not a standard you can purchase. For continuous and batch applications, the metered fluid does have an impact on the calibration of Coriolis meters. So, especially with hazardous fluids, simply calibrating the meter by running water into a scale tank will not adequately calibrate the meter. A simple method for the end-user to verify the meter's performance is per its specifications for the actual metered fluid is important. This verification of metered accuracy is especially important for industries where verification of meter accuracy is required and regulated by a governmental agency such as the Food and Drug Administration (FDA) in the U.S. ISO 9000 and other international standards similarly require verification of flowmeter accuracy, something that is quite difficult to accomplish with today's Coriolis meters. DT

Behind the byline

James Reizner is the section head of corporate engineering at the Procter & Gamble Co. in West Chester, Ohio.