Burgeoning demand for flowmeter calibration in China
China's economic transformation has resulted in a dramatic increase in the demand for industrial equipment. Bryan Franklin, flow products manager at ABB in the UK, explains how the company is investing in new facilities in China to meet the demand for accurately calibrated flowmeters.
The first of the rigs, built at the company's factory in Stonehouse, Gloucestershire, is due to become operational in early 2005 and will be used to calibrate meters sized 15-50mm, with a flow volume of 20L/s. The second and third rigs, which can calibrate flowmeters of between 65-150mm and 200-400mm respectively, will also come on line during 2005.
Although designed and built to calibrate primarily electromagnetic and vortex meters, all of the rigs can actually be used to calibrate almost any type of flowmeter. Built to the same ISO standards as ABB's existing flow calibration rigs worldwide, the rigs incorporate a 'carousel' design, which enables up to three flowmeters to be calibrated simultaneously.
The calibration process itself is fully automatic, with specially developed software used to run the process and collect the calibration data.
Why is calibration accuracy important?
Calibrating a flowmeter before it is used is the only way of ensuring that it will deliver the level of accuracy expected by users and promised by manufacturers. Most manufacturers use a process known as 'wet calibration' to test flowmeters under good flow conditions on a flow rig. This process is critical in ensuring the flowmeter delivers the best possible accuracy in everyday use.
The resulting data is then used by the manufacturer to show that the flowmeter has been calibrated for accurate measurement within the limits of its specification.
In some applications, flowmeters are used to give an indication of the rate at which a liquid or gas is moving through a pipeline and high accuracy is not a crucial factor. However, for processes such as in the food processing or chemical industries, flowmeter accuracy is often the deciding factor between optimum product quality and, potentially costly, wasted product.
How do the rigs work?
The lack of any single standard imposing uniform conditions for flowmeter calibration means there is no single agreed method for how a rig should be built or tested.
For this reason, two main calibration methods have emerged, static and dynamic. For a flowmeter to be properly calibrated, the conditions under which calibration takes place ideally need to be as close to real-life conditions as possible. It is well known that if a step change occurs in a 'stable' flow regime, then it can take up to five minutes for the flow to stabilise again.
Most flow rigs cannot achieve a step change, however, needing a finite time for valves in the line to fully open and close, for example. Therefore, for a true simulation to take place, any testing needs to take place once a stable condition has been restored to ensure any sources of potential uncertainty have been minimised as far as possible.
In a static test, measurements take place when nothing is changing. The rigs built for ABB's Chinese factory incorporate a loop with a weigh tank as a reference device. Water is circulated around the loop, with its flow rate being measured by a flowmeter installed in the line. Circulation continues until the flow is stable and at the desired rate.
When the flow is stable, the water is diverted as quickly as possible into the weigh tank. As soon as water starts to flow into the weigh tank, a timer is started and the flow is measured into the tank. When the tank is full, the flow is diverted back to the recirculating mode, away from the weigh tank, and the timer is stopped.
By utilizing the weight of water collected, compensating for density and the time taken to collect the water, a flowrate can be accurately calculated. This flowrate is then compared against that of the flowmeter.
In a dynamic test, no time is allowed for water levels to settle. In this method, a tank is first filled with water. When the tank is full, a valve is opened and the flow starts.
As the water level drops, the flowrate is computed by measuring the time taken to pass level switches mounted in the side of the tank. When the level has dropped to a certain switch, the test is started and then stopped when the level passes another switch lower down in the tank.
The most significant problem with this method is that the level and volume are not the same. The level of water will vary according to the shape of the tank itself. Moreover, the shape of the water surface will change depending on the depth of water remaining in the tank and the flowrate of the water. As the test takes place, the water surface will probably be choppy, making it difficult to know the exact level of water in the tank.
The effect of the valve opening and the consequent changes in the water level will also add to the uncertainty during the test. As the water level falls, there will be less pressure to push it out, decreasing the flow rate. To maintain a constant flow rate, the valve therefore has to adjust continuously.
Taken together, all of these factors introduce too many unknowns into the equation for a truly accurate flow calibration test to take place.
In common with all of ABB's flow calibration rigs, worldwide, the rigs being installed in China incorporate both a weigh scale and a master meter as reference devices. The main advantage of this method is that these devices provide a fundamental value against which to compare flowmeter measurement performance. By calibrating the master meter against the weigh scale, it is also possible for the rig operators to ensure the continued accuracy of the rig.
How to ensure your flowmeter is properly calibrated
Having your flowmeter calibrated properly is vital to ensuring accurate performance. With no single standard imposing uniform conditions for flowmeter calibration, some flowmeter manufacturers make claims about their products' accuracy that are just not true and could cost you dear.
To be sure that your flowmeter has been calibrated as accurately as possible, you should ask manufacturers the following questions (see next page):
Does the calibration figure include flow rig uncertainty?
Flow rig uncertainty takes into account all the factors that could affect the meter's flow measurement accuracy during calibration. Sources of uncertainty can include variations in testing times, valve response, the effects of obstacles in the line or variations in the properties of the fluid being measured.
In reality, manufacturers tend to base their calibration uncertainty figures on the uncertainty of their reference device. These devices provide a known value against which to compare measurement performance and can include methods such as a weigh tank, volume tank or meter prover. Any deviations from this value are then used to calculate the uncertainty value.
However, the resulting uncertainty value is the uncertainty of the primary device only, not the uncertainty of the whole calibration process. A reputable manufacturer will use this figure as the basis for their uncertainty calculations, the less scrupulous might claim that this is the calibration uncertainty itself.
How was the uncertainty value calculated?
Various standards exist which define procedures for measuring and expressing uncertainty. For example, the 'Uncertainty and Confidence in Measurement,' (Document M3003) published by the UK Accreditation Service (UKAS) sets out procedures for accredited laboratories for estimating uncertainty during calibration.
This publication defines in detail the complex calculation necessary to arrive at the real uncertainty of a calibration facility. If properly adhered to, it should enable comparisons of calibration laboratories' uncertainties.
However, even this publication does not account for factors such as the dynamics of the rig or its operating procedures. The situation is further complicated by the response of flow to a change in conditions.
Another common area of uncertainty is the duration of a calibration test. ISO standards 9368 and 4185 set out guidelines for uncertainty in the measurement of time during calibration, with a recommended period of at least 30 seconds. Most flow calibration experts will agree that a calibration time of at least 60 seconds is required. Despite this, some manufacturers use times that are only 25% of this figure, introducing still more uncertainty into the calibration.
For these reasons, amongst others, users need to be extra vigilant when comparing uncertainty claims from manufacturers or indeed calibration laboratories. Those who claim that their rig is as accurate as their primary device are simply misleading their customers.
Contact: ABB Ltd
Tel: +44 1480 475321