Handling the heavier crudes

With heavy oil today representing around 70% of the world’s confirmed oil reserves, extraction of the heavier crudes is growing rapidly and there is an increasing need for flow measurement of high viscosity hydrocarbons. NEL's Chris Mills reviews the technological challenges ahead.

Innovative methods of recovery, transportation and measurement are essential to fully exploit the world's heavy oil (high-viscosity‡) reserves and to bring the benefits of a diverse energy supply. Real-time monitoring, for example, is seen as crucial for efficient production control and for overall recovery optimisation, while more accurate measurement of the viscous flow streams is vital for allocation and custody transfer purposes.

In terms of accurately measuring heavy oil, a variety of issues have been identified and reported with respect to some of the measurement technologies currently being applied to viscous fluids[1]. These include the higher viscous friction of the fluid being metered; the increased pressure losses incurred across internal bends and restrictions; the possibility of extreme or varying velocity profiles; and the increased susceptibility of viscous liquids to entrain secondary components such as solids or gas. It is therefore reasonable to predict that different metering devices will be affected by these phenomena in different ways.

The Coriolis mass flow meter and the ultrasonic flow meter (USM) are both relatively modern flow measurement technologies and continue to capture increasing market share as operators gain confidence in their performance. But do they offer an alternative approach to the problem of metering high viscosity liquids? A claim often made by manufacturers of Coriolis flow meters is that they can be calibrated in water and that the calibration will transfer to high viscosity service with no requirement for characterisation or modification. However, research carried out on Coriolis flow meters at the UK National Standards Oil Flow Measurement Facility at NEL revealed that at high fluid viscosities (100-300cSt) there was a notable under-reading of the mass flow rate when using the original water calibration[1]. The accompanying graphs show the performance of a commercially available 4in Coriolis flow meter when applied to a low viscosity fluid and high viscosity fluid respectively.

‡ For the purpose of this article, ‘high’ viscosity in relation to hydrocarbon liquids is taken as a kinematic viscosity > 100 cSt.

It can be seen from the low viscosity data (Figure 1) that the results fall within the manufacturer’s stated performance for the entire test data. However, the high viscosity data (Figure 2) displays a growing under-reading at low flow for the majority of the high viscosity test points. To understand this phenomenon it is worthwhile plotting the entire data against pipe Reynolds number (Figure 3).

The test data now appears to display a notable trend with decreasing pipe Reynolds number. This deterioration suggests that there might be a shift in the calibration factor at low Reynolds numbers. In highly viscous fluids, it is possible to attain low Reynolds numbers with a moderate flow velocity relative to the fluid properties. Thus the effects observed cannot solely be attributed to low fluid velocity.

Research carried out by Tschabold, Kumar & Anklin[2] suggests that the shift in calibration factor is caused by oscillatory shear forces within the flow tubes of the Coriolis device. It is believed that the Reynolds number effect shown here is caused by the interaction between oscillating Coriolis forces and oscillating shear forces. The oscillating shear forces are found to be far more dominant than the oscillating Coriolis forces in low Reynolds number flow than in high Reynolds number flow. However, it is possible for manufacturers to apply an online compensation for this effect using on board signal processing within the Coriolis device.

Another possible issue with Coriolis flow meters is that there can be a substantial pressure drop across the device at high fluid velocities. This is due to the size and design of the measuring tubes within the device. Generally, Coriolis meters have two smaller bore twin measurement tubes that are curved or arced in shape. This smaller size, coupled with the curved design, can cause additional pressure drop across the measurement system. However, recent advancements in Coriolis meter designs mean that new models are now available with a straight tube design. As a result, the pressure drop across these models is generally much lower than previous designs.

Figure 1: Coriolis mass flow error – low viscosity.

Figure 2: Coriolis mass flow error – high viscosity.

Figure 3: Coriolis mass flow error – pipe Reynolds number.

Coriolis meters also offer significant flow meter diagnostic capabilities. Whilst currently not as advanced as ultrasonic flow meter diagnostics, there exists the potential for using the diagnostics to indicate flow characteristics and even potential shifts in parameters between calibrations. Some Coriolis manufacturer’s also state that the energy required to oscillate the measuring tubes could potentially be used to indicate the presence of a second phase, ie gas in liquid flow or liquid in gas flow.

In high viscosity applications the principle advantage offered by USMs is that they are non-intrusive and therefore no additional pressure drop is experienced across the device. On the face of it, this reduces the effect of frictional losses on pumping rates and minimises the chance for cavitation or flashing at higher flow rate.

However, it must be remembered that the device typically requires at least 10 diameters of straight pipework upstream of the device and five diameters downstream. Manufacturers normally specify that a flow conditioner is installed just before the 10 diameters upstream. If this 15 diameters of additional pipework and the flow conditioner is referred to as a ‘metering section’, then the overall pressure drop across the ‘metering section’ can be significant and becomes comparable with other measurement technologies.

USMs are also highly susceptible to flow profile changes. They require a fully developed flow profile which for laminar flow can mean that over 100 diameters of upstream straight lengths of pipe are required[3]. Of course, in most industrial applications this is not normally achievable due to space constraints. Bends, valves, expansions, contractions, misaligned pipework all have an effect on the flow profile and often cause asymmetry and swirl. This means that as the USM assumes a fully developed profile, it can under or over-read the actual flow. While flow conditioners can be used to remove the asymmetry and swirl, these have an associated pressure drop and hence cancel out the main advantage of USMs.

The USMs themselves may provide an answer to this problem in the use of the secondary information that they can provide. Not only do they provide flowrate, they also offer a large amount of diagnostic information that relates to path velocities, ultrasonic signal strength and speed of sound. This diagnostic information could possibly provide the end user with estimations of any asymmetry or swirl within the pipe and help to diagnose any non-ideal conditions present in the flow.

Another issue affecting USMs used in measurement of high viscosity oils is the potential for temperature gradients within the pipe, especially in areas where pipelines experience extremely cold temperatures. At low Reynolds numbers the flow profile is peaked and parabolic in shape. There is also very little heat transfer between layers of oil, which can result in temperature gradients forming along the cross sectional area of the pipe and in the transducer ports. Distinct temperature layers, each with different speed of sound, then refract the ultrasonic signal, causing minute changes in propagation angle, which in turn has a significant effect on measurement accuracy.

Chris Mills is a technical consultant at NEL, an international provider of specialist technical consultancy, research, development, testing, measurement and programme management services to the oil & gas industries as well as government. Part of the TUV SUD Group, NEL is a global centre of excellence for flow measurement and fluid flow systems and is the custodian of the UK's National Flow Measurement Standards.

Calibration of USMs is another important factor to consider. In order to ensure maximum accuracy they should be calibrated in ‘as close to operating conditions as possible’. In the past this has been difficult, due partly to the scarcity of suitable test facilities capable of providing viscous flow in combination with accurate and traceable reference instrumentation. However, NEL has recently upgraded its viscous oil flow facility to calibrate meters up to 1500cSt. This is an important step in not only calibrating meters at their operating viscosity, but also allowing important research and development work to be carried out.

To help lower the uncertainty in the measurement of heavy oils, NEL, supported by the UK National Measurement System, is currently investigating the performance of established flow metering technologies, including Coriolis and USMs, across a range of kinematic viscosities (200-1500cSt).

With support from key members from industry, this research should provide solutions to the difficult challenges currently presented by high viscosity flow measurement. OE