It's just a PHASE

The National Engineering Laboratory's campus, near the Scottish New Town of East Kilbride, contains many strange sights. Entire aircraft wings and the feet of oil rigs suspended in mid-air, being shaken by machines that simulate thunderstorms and earthquakes; huge bags, suspended from spidery towers, that are filled with water until they burst. Its newest addition looks less dramatic but is just as impressive: a cavernous space, dominated by giant bright red tanks and rows of brightly-coloured piping, punctuated by sensors and meters. But these pipes are helping to solve on of the oil and gas industry's biggest problems - how do you measure a flow containing oil and gas?

The building of the multiphase flow laboratory and the associated multiphase advisory centre (MAC) began in 1991, triggered by the discovery of `marginal' oilfields in the North Sea, where the oil pumped out of the wells is accompanied by large volumes of seawater. `There was a consultation exercise with the Department of Trade and Industry over the funding,' says Flow Centre manager Brian Millington. `We argued that we needed state-of-the-art equipment - and we got it.' The engineers' wish-list cost a total of £7million, of which £5million went on equipment.

The MAC is the home of the UK national standard for multiphase flow, which is used to calibrate meters and other instruments. This `standard flow' is supported by the National Measurement Standards Policy Unit, part of the DTI.

Apart from maintaining the national standard, most of the MAC's work is concerned with researching `generic' issues of flow measurement - chiefly the reliable measurement of multiphase flow - and with anticipating what its client industries will need in the next decade.

The MAC's research programme is coordinated by a working group of 12 people, including representatives from the utilities, such as the water companies; and firms like British Gas and ICI. Each programme runs for three years. The current programme, involving subjects including ultrasonic metering, has two years left to run, says Millington, and the working group is now beginning to review possible projects for its successor.The centre is thriving, says Millington, but latest trends in multiphase research may begin to encroach on its territory. `There's always a question mark over large-scale testing, because of its cost,' he comments, `and there's more and more emphasis being placed on computational methods. We're still grappling with what that means for NEL.' However, computer packages that can model the complexities of multiphase flow are very much in their infancy. `There are a few codes available, but their results are always very dubious.'

Slugs, bubbles and strata

Almost 99 per cent of the MAC's work is for the oil and gas sector, but most of the process sector experiences multiphase flow of some sort, particularly the food, cosmetics and pharmaceuticals sectors (which have to deal with suspensions and emulsions, both multiphase mixtures) and the brewing and distilling industry. NELis now pursuing a policy of expanding its business into these sectors, and Millington hopes that the process industries will soon begin to catch on to the possibilities of working with NEL. Its links with strong chemical engineering universities, especially Strathclyde and Imperial College (see May PE), may help with this, he believes.

Millington's hopes may be helped by the variety of multiphase work at NEL. The MAC itself concentrates on mixtures of liquid phases, but many natural gas wells produce liquid condensates along with gas. The handling and measurement of these flows is the province of the head of gas flow, Paul Johnson, in the NEL's wet gas flow laboratory.

The lab is based around a recirculating loop system which can pump gases at 70 bar pressure through a four-inch pipe, into which various meters, such as venturi, ultrasonic and orifice plate, can be inserted. An injection system, resembling a steel porcupine, fires liquid at 130 bar into the gas flow. Altering the pressure of gas and liquid can create the range of flow conditions inside the pipe which would be found in a real condensate-laden gas stream, Johnson explains.

The flow can be stratified (with liquid running along the bottom of the pipe); laminar (where the liquid forms a film along the pipe's inner surface); or in two transitional phases, known as `bubble' or `slug' flow. The characteristics depend on the liquid-to-gas proportion, the flow rate, and on the physical properties of the fluids in the pipe, such as viscosity and surface tension. This is where the current NEL set-up falls down. At the moment, the wet gas rig uses air and water, but these are very different fluids from natural gas and condensate.

To solve these problems, says Johnson, NEL is spending `several million pounds' on a new liquid injection system and gas/liquid separation equipment; a wider pipeline, and a new working fluid closer to natural gas an its condensates.

The ideal fluid would be the `rich' natural gas itself - a mixture of the six simplest alkanes, with traces of carbon dioxide and nitrogen, `but the safety aspects of using natural gas in a residential area are horrendous, even if we used an inert gas like nitrogen as the gaseous phase,' says Johnson. Instead, NEL has opted for a hydrocarbon-based product made by Exxon, a kerosene substitute with properties similar to octane, as the liquid phase, with nitrogen as the gas. The Exxon product has a high flash-point temperature, and should model the behaviour of hydrocarbon condensates without posing the risk of explosions, even if it should leak.

Riddle of the sands

Liquid and vapour aren't the only phases that oil and gas recovery involves. The third phase is sand. And it's obstructive, abrasive and destructive.

Inevitably, when drilling under the seabed, sand gets into the pipelines. Every component suffers, but the worst affected are valves, whose moving parts can be quickly destroyed by the particles.

NEL's `dirty flow' rig subjects manufacturers' valves to rigorous tests to see whether they can withstand the depredations of sand. The director of the valve-test centre, John Peters, explains that, until recently, there was no recognised quantitative method for assessing a valve's performance in sandy service. NEL's task was to design a testing regime that would be recognised internationally by both valve manufacturers and operators.

The first stage was to design a testing rig that could produce wear identical to that found in the field (see left). The rig is in fact very simple, with a pump circulating a working fluid through a closed loop of piping, which can be configured to hold up to 15 valves under test, with pumping pressures up to 15 barg. The test fluid is water, which can be clean, or can contain either 1 per cent or 2 per cent sand. The valves are generally opened and closed 1000 or more times, and the valve seat is tested for leaks at least three times during the test. The torque or force needed at the valve stem to open and close the valve are also monitored constantly.

Companies supplying valves for test tell NEL how they expect their products to perform - the valve-seat leakage and operating stem torque or force after a certain number of cycles, although the test always runs for `significantly' longer than stipulated. To pass the test, the valves have to perform within these parameters, and survive an inspection of their component parts. Valves which pass the test are then recognised by NEL - and most of the oil companies - as suitable for sand.

John Peters sums up the valve test rig neatly. `It may look a bit Heath Robinson, but believe me, it's state-of the art. In this industry, it has to be.'