New Coriolis meter finds hundreds of kilograms of ‘disappearing’ batch product

With the cost of chemicals these days, it’s nice to be able to keep good track of them.

It’s a common problem: the batch of a product has been processed, and is being pumped and metered to the next stage. Typically the flow meter starts empty, suffers the ‘hit’ from the onset of process fluid, meters the bulk of the fluid, and then encounters slugs and bubbles at the end of the batch.

The same basic issue arises in chemical, petrochemical, food and utilities applications, and especially in product transportation: today’s commercial flowmeters perform poorly. Large flow errors arise when the flowmeter is wet and empty, when the flow hits, and with bubbly and slug flow at the end of the batch. No flowmeter technology performs especially well here, but it is ironic that the Coriolis meter, which gives the greatest accuracy when full of fluid, is especially vulnerable to the start and end conditions.

A number of costly workarounds have been developed over the years - such as deploying recirculation loops to ensure the meter is always kept full - but these are often impractical or uneconomic.

Recent trials at Great Lakes Chemicals in Manchester indicate that the new Invensys Foxboro CFT50 digital Coriolis transmitter may be able to resolve the problem. The instrument has demonstrated what engineers say is a ‘remarkable ability’ to meter accurately during the difficult start and end stages of the batch.

Flow metering for batches

Steve James, plant manager of the Manchester site, complained his operations people were having trouble with a batch line that makes a variety of water treatment chemicals for heating systems and desalination.

Roger Marsden, Director of Westmeade Services Limited, who provide technical support for Great Lakes Chemicals, described the problem. ‘There’s nothing unusual about the trial application, or the problems we have seen,’ he said. ‘We take three feeds into the tank, blend and pH balance the batch, and pump it into a storage tank prior to distribution. With some batches we need to filter the product as it leaves, which reduces the flow and causes worse metering problems. We were seeing big inconsistencies between the metered inputs and the output, and it was clear that all the Coriolis meters were being affected by entrained air and fluid hit, even though we’d put in a special piping arrangement as recommended by another Coriolis vendor.’

‘We have other clients with similar problems, and we pride ourselves in supporting the latest technology, to improve the service we can offer,’ said Mr. Marsden. Westmeade heard about the CFT50 and arranged the trial with Foxboro’s Steve Powner. ‘We put the Foxboro flowtube in series with the other vendor’s meter on the outlet, so we could properly compare performance. We don’t recommend any special piping arrangements for air entrainment,’ says Mr. Powner, ‘we just plumbed it straight in.’ A chart recorder was installed to collect data, with a 2 sec update time. Figures 1 and 2 show the start and end of a typical batch. As well as the two mass flow readings, the Foxboro CFT50 density reading is shown to indicate the percentage of air in the fluid.

Prior to the onset of flow, both meters show a zero reading, while the density reading indicates that the meters are ‘wet and empty.’ Once the flow begins (at 18 sec), the CFT50 starts up immediately, whereas the other meter requires some 16 seconds to register the flow.

‘At our flowrates that’s up to 200kg of missing product,’ says Dave Vere, Westmeade’s project engineer. ‘Once the batch is in full flow then the two meters match each other well, but when the last of the product comes through, the other meter often cuts out.’

Figure 2 shows the end of the same batch, some two hours later. Both meters register the drop in flow at 130 minutes, but as the flowtubes drain the other meter stalls. It thus fails to register the final blowthough of product. ‘It depends on the batch how much is missed at the end,’ says Mr. Vere, ‘but when we are filtering it can be as much as a tonne.’ Great Lakes have drawn the obvious conclusions.

‘Operations trust the CFT50, so the other meter will be pulled out until it can be used somewhere else,’ says Mark Wilkinson, I+E Manager at Great Lakes, who funded the trial. ‘We will be recommending the CFT50 to our clients for batching applications where entrained air is an issue,’ adds Roger Marsden. ‘This is a significant extension of Coriolis capability.’ The CFT50 has emerged out of 15-year research partnership between Foxboro and Oxford University. s Engineering Science Department. Dr. Manus Henry, who leads the Coriolis research team, explained why the meter performs so well.

‘There are several key advances. The flowtube drive signal is synthesized digitally, giving good control of the meter in difficult conditions. The dynamic response is rapid: independent trials have shown it can respond to a step change in flow within 25 ms. Finally, we have algorithms for correcting for the effects of two-phase flow. We’ve been hearing of other applications like ethylene oxide with flashing up to 50% gas fraction, and the meter is still doing fine.’

Dr. Henry and his team are working with the Foxboro development group on introducing further improvements. ‘We’ve only started to exploit this transmitter technology, and are pursuing several very promising research lines. This meter is going to keep on getting better.’

For more information, contact Roger Marsden at Westmeade Services, e-mail info@westmeade.co.uk, or Dr. Manus Henry at Oxford University manus.henry@eng.ox.ac.uk