Wet Gas Sampling in the Upstream Area Mr.

Philip A Lawrence Chairman ISO TC193Sc3 (Upstream Natural Gas) Texas USA.
Abstract The measurement of a multiphase flow regime is not simple, recent work by industry has started to focus and concentrate on the measurement of wet gas recognizing that a significant monetary value may be associated due to the hydrocarbon liquid content. Lockhart and Martinelli parameter (XLM) values of up to 0.3 pose extensive issues, in the measurement and sampling of wet gas particularly when using a bulk liquid density to ascertain the effect of a liquid in a D.P. type wet gas meter, all which can provide accuracy issues outside of a meter manufacturers claims. Link that to a lack of methodology to obtain the much needed compositional analysis to input into the PVT system on board these newer wet gas systems and you soon realize there is not much out there to help the meter to perform effectively at its peak performance as a vendor is claiming. A variety of methods have been used including dye tracer, microwave, even bulk separator allocation to try to obtain a representative data set to confirm or help the wet gas meter perform effectively. The metering of a wet gas using single phase devices - etal - TC193 SC#3 WG#2 and ASME MFC19 - prescribes the use of a single phase differential pressure based element or combination thereof, such as a Venturi Meter D.P Cone Meter and Standard Orifice meters or others used in the wet gas stream where applicable. In the majority of cases the D.P. meter will produce an over-reading due to the presence and effect by the liquids , from this knowledge special and sometimes propriety algorithms and methods are used to correct the over reading. To try to correct this Differential Pressure (D.P.) over-reading and to help reduce the measurement uncertainty for the single phase device, it is necessary to adjust the metered rate for the liquid/gas mass ratio. Industry has generally accepted the use of the Murdoch-Chisholm, DeLeeuw or the newer correction method based on Chisholm's algorithm, modified by Steven (ref [1]), with applicability to orifice meters and the newer D.P. cone meters. This paper details the work and techniques that are now available to enable a reasonable representative sample data set to be obtained to allow proper field sample analysis and therefore a calibration, or a fluid composition to be derived and updated into a wet gas meters PVT or flow computer system. High GVF

Due to the nature of the problem (flow homogeneity) many wet gas sampling systems may only be limited to High Gas Volume Fractions (GVF) values, the data presented in this paper is limited to a GVF of around 99% and pressures about 90 Bar G. Data sets and operational evidence for a system called the Mass Phase has been in the field for some years now, whilst operating in this high GVF range of around 98-99 % and around 80 bar these systems are showing some good promise ! The premise that no slugging was expected in the system application is also a factor in its use. The flow regime chart below shows the expected region of operation indicated by high superficial gas velocities - Figure 1.0 (red elliptical highlight)

Figure 1.0 Typical Wet Gas Flow Regime Chart. (Courtesy of ISO TC193Sc3-Wg1) There is no currently available multiphase metering technology that will accurately determine liquid mass-ratio (liquid loading -LL) and composition as a single device the mass phase system is a technology improvement to help end users solve an issue that is on the increase as hydrocarbon gas production systems increase their (XLM) wetness. A variety of methods have been investigated over the years including microwave, dye tracer, dilution, centrifugal separation and gas scrubbers, these can have issues some of which are listed next.

Typical Issues

A) Tracer Dilution Method - inability to trace methanol and there may be a lack of available distance (to allow the tracer to bond to the liquids) and also is not a real time method . B) Microwave Techniques - can only measure water content (and then with a high degree of uncertainty) requiring a process (inc PVT) model for other fluid types C) Gas Scrubber - can have a lack of an available drive mechanism for the volumes of gas required and the inability to discriminate between the methanol and water Typical Gas Composition A typical gas / liquid composition that has been successfully measured by the sampler system in the paper is shown below (see figure 2.0);-

Fig 2.0 Typical Gas and Liquid Composition The fact that the above methods (A,B,C) show some issues in reproducibility, a move for the sampling philosophy to one of a traceable physical sample collection and real time phase fraction determination, would appear to be much better. One method may be accomplished by using a dedicated wet gas / multiphase spot sampling, separation and analysis methodology. An overview of an ideal system methodology is shown in figure 3.0 next -

Laboratory Figure 3.0 Effective System Implementation

A few action steps in the process to achieve a viable result are very necessary if a wet gas system is to provide a consistent, traceable, and useable result ! Spot Sample Philosophy The first action in the system philosophy above is to try to create an even phase distribution and dispersion of the liquids within the gas stream. At the conditions indicted earlier, (98-99%GVF) a mist flow regime should be is expected. There are several technical papers that suggest that an annular liquid flow distribution can be re-homogenized downstream of an expansion. An optimized dispersion quality is expected if the sample collection point is in a position downstream of the wet gas D.P meter. See Figure 4.0 next ;-

Figure 4.0 Wet Gas Spot Sample Point Take off Design How the System Works The basic principal of operation for the mass phase wet gas sampling system is to flow a reasonably representative stream of gas through a mole sieve collection cylinder (Zeolite pack inside) and through a mass flow-meter at process conditions, then allow the downstream outlet gas to vent into a sample collection receptacle under pressure. Provided that the mole sieve can collect the liquids over a defined time base without saturation and dry gas mass flow-meter can record the total mass of gas during this same time then by weighing the mass increase of the mole sieve, the total liquids mass for a given mass of gas can be calculated. The gas meter also records the mass of dry gas thus obtaining a mass fraction ratio. It is considered very important to maintain as far as possible the physical properties of the process stream whilst venting the gas through the system during the spot sample process. If this is achieved then it makes possible a simple control over the flow-rates, velocity etc. during the overall field life. Say 1 liter of wet gas at process conditions has 0.2ml of Condensate, 1.6ml of Water and 0.3ml of Methanol. A typical sample cylinder to take a 1 liter sample of gas weighs 15 Kg and the 2.1ml of liquid collected within would easily be lost on the walls, in the seals or interconnecting tube and fittings.

Even if the cylinder could actively collect the liquids, how would the mass be accurately discriminated ?The accuracy of a mass balance suitable for a 15Kg value usually has a resolution of 0.5 to 1g, which would render any uncertainty in the liquid mass ratio as meaningless. Increasing the overall volume to attain a larger incremental mass is also self defeating (larger cylinder mass) the mass resolution can be increased by separating and collecting fluids from a known mass of gas and taking a separate dry gas sample from this wet gas stream collection by some method. This method requires great forethought and from reviewing the previous flowchart the task of separation and collection may be achieved by using some readily available technology used to dry or scrub gasses. Zeolites Various techniques have been considered to scrub wet gasses including silica gel or benzene traps but ultimately using a mole sieve (liquid trap) made from Zeolite seems to be the best and offers a robust solution. Zeolite Technology has already been in common use for scrubbing gasses in other industries for many years now. The Russian federation for example (Gazprom) use Zeolite technology for the purposes of the drying and scrubbing of sales gas supplied to many commercial users so we know from history that this method is a robust method. (It is documented in national standards in that region also) A good answer for the issues is to collect the liquids and gas (wet gas) in a control cylinder whilst allowing the gas to pass through this cylinder via a drying agent and then exit to a mass flow meter, the subsequent dry gas mass being measured per unit time whilst , the mass rate of the liquid is also measured in the cylinder using a timed weighing system which can also act as a sample receptacle. Uncertainty There are various things that have to be taken care of in wet gas sampling particularly in regards to factors that affect the overall uncertainty. (Ignoring for the time being, the spot sample take off-system design.) The uncertainties in the sampling methodology can be broken into a few different areas. Determination of the mass ratio is dependent on the ability of the mole sieve pack (Zeolites) to collect and retain the liquids. The recorded measurement resolution of the mole sieve mass change and the gas mass flow-meter uncertainty. Determination of the composition depends on the ability to recover and analyze the liquids collected properly.

Liquid & Gas Mass Ratio Determination The liquid collection cylinder comprises of a cylinder packed with the Zeolites which are crystalline, highly porous, materials with a three dimensional structure having pores of known predictable sizes.

The pores have an affinity to sieve out polar molecules that fit within their structure due to a capillary action and surface tension effects. They are manufactured in a range of pore sizes, typically 3, 4, 5 and 8 Angstrom although there is a variation in the way the manufacturers express the sizing. (Zeolite13X has an 8 Angstrom porosity).See Figure 5.0.below.

Figure 5.0 Zeolite - Typical Structure & Construction Generally product molecules larger than the structure of the Zeolite will not be trapped, a typical example of some sizes are shown below in figure 6.0.

Figure 6.0 Typical molecule sizes for some commonly measured fluids

Given a choice of molecules to adsorb, there is also a known hierarchy with reference to our particular process interest. The hierarchy is such that the most polar molecules are adsorbed first. In the case of the previous data case and in this white paper these are;- Water, Methanol, & Ethanol The hydrocarbons take a specific order according to the structured order of the alkenes (The less the saturation the more strongly liquid is adsorbed ) followed by the alkanes. With each of these series the higher molecular weight member will be adsorbed first. i.e. C8's before C1's. There will be some cross over between the two series, i.e. Ethane before Propane. Zeolites will adsorb about 20% at best by mass and during the process generate heat (Exothermic reaction), the adsorption process is reversible using a variety of techniques. Also Adsorption is clearly not instant so consideration has to be given to the velocity through the column (residence time), the contact area (bead size) and the mass of liquid required to provide a reasonable uncertainty. The issue is also to be determined was a reasonable time for the sample process during which a dry gas sample could be taken , this is a compromise based upon the expectation for the process and our initial choice was made trying to attain a mole sieve total mass below 2 kg. This will provide an internal volume of about 250ml for this set of flow parameters and regimes typical cylinder deign shown below.

Figure 6.0 - Typical Collection Cylinder (Mole Sieve Pack fits inside) The mass flow meter was of a Coriolis type, it must be noted that these types of meter must be installed with enough pressure to facilitate a useable gas density and thus derive a gas flow rate. This means that the pressure and the gas flow-rate must be above a certain level to facilitate a density sufficient to cause a change in the Coriolis effect detectable by the meter. If the flowing process parameters are not sufficient to meet minimum requirements then a serious measurement error may occur. The fact that liquid content (2 phase) can affect the measurement performance of a dry gas Coriolis meter should not be an issue for this application since the gas is very dry at the cylinder exit due to the Zeolite packs drying ability.

Coriolis Meter Typical gas flow rate performance accuracy claims from one leading manufacturer that provide Coriolis meters are as follows +/- 0.35% of Mass Flow reading with a 0.2% repeatability value on dry gas. See figure 7.0 - Typical performance of a 2 inch Dry Gas Coriolis Meter.

Figure 7.0 - Typical Dry Gas Coriolis Meter Performance (U-Shaped Design) The entrained liquid in gas effect on a Single Phase Gas Coriolis Meter has been demonstrated by the data release from the wet gas JIP at CEESI some years ago (see Figure 8.0). Generally a 2 phase fluid effect on a Gas Coriolis meters accuracy & performance is large and changes with pressure, this is why the hydrocarbon gas when metered in this kind of application must be as dry as possible or meter performance will give a poor result.

Figure 8.0 - Entrained Liquid Effect on a Single Phase Dry Gas Coriolis Meter.

PID Generally mass phase systems can be manufactured and designed to provide samples from each of two flow streams. A PID is shown next regarding the system design as seen in Figure 9.0 next.

Figure 9.0 Wet Gas Sampling System-PID Initial Testing of a Mass Phase System In the original system design for the prototype of the mass phase system during the test phase the ability of the mole sieve (Zeolites) to extract liquids from the gas stream at Isokinetic flow velocities (as defined by the sample inlet configuration) was reviewed and analyzed. This testing was completed by using compressed air and water. Liquid fraction recovery in the range 85 95% was experienced using factory supplied (out of the box) mole sieve material. At the start it proved difficult to get a stable baseline for the liquid content of the air stream despite the use of chillers and cooling coils. However the tests served to prove that the basic principle was valid and moved the design along to a finished product. Also during the first tests it became apparent that some basic performance for the sieve could be ascertained by locating the peak temperature of the exothermic reaction. At the highest flowrates this zone would spread and peak within 7 cm of the outlet whereas at low flow-rates, it would be evident closer to the inlet. (et-al Jiskoot & Payne 2003) and Figure 10.0 next.

Figure 10.0 Exothermic Zone seen in the Collection Cylinder Performance Enhancement It has been determined from testing that the diameter of the mole sieve be doubled (volume quadrupled) and that the column be packed with a mixture of 4 and 8 angstrom material at 4mm diameter this was to further enhance the performance It was also discovered that this could still be achieved within a fully loaded weight of below 5 Kg and show a measurable value on a laboratory grade mass balance system to a resolution of better than 0.1 gm ! With this design of mass phase system it can trap the available liquids in this high GVF flow regime pressure (hydrocarbons and other) values in excess of 95% (under optimum conditions 100%, under the worst test of the volume sieve 80%) of the liquids and also measure the total mass to a resolution better than 0.1gm in a nominal 30-50gm liquid sample (0.20% 0.33%). Couple this together with the performance of the dry gas Coriolis meter performance as stated previously and a viable stable wet gas sampling system is possible! After describing the system in detail above overall a typical production wet gas sampling system should comprises a take-off probe and a bypass loop in an enclosure containing a mass flow meter, two molecular sieves and two sample collection receivers. The take-off probe is mounted in the line at a suitable position (using the principles mentioned earlier) and has multiple inlet points to help profile the pipeline and flow regime properly.

The number and size of the inlets are designed for each application dependant on the pressure and multiphase wet gas flow regimes present. The flow loop is designed to maintain sample representivity through the by-pass loop and to have a minimal dead volume.
Whats Next Analysis! The next phase in the application of a wet gas sampler is how to is to manage transport and extract the sample of the gas and liquids without compromising the products. This can be daunting for single phase sample collection at the best of times, so the same criteria of cleanliness and maintaining traceability as used in normal sample methodology has to be transferred into this special multiphase sampling area with some thought and common sense.

Jiskoot and Payne NSFMW-2003 postulated that the liquids might be extracted / removed by vacuum extraction from the Zeolite cylinder pack however they quickly determined that NO liquids could be extracted under vacuum at room temperature! Jiskoot and Payne also at that time conducted a limited number of tests by heating the mole sieve to about 100 deg C also showing little improvement. Finally a simple test was performed to determine if the application of temperature would remove water from the mole sieve material, which it did so a new methodology and procedure was produced for heating and extracting the liquids which in the final form is shown next. Figures-11a & 11b.

Significant work has been undertaken regarding the laboratory processes for both extraction and preconditioning of the mole sieves (Zeolite packs) and further tuning of this together with a system for the correct selection of Zeolites which is required, currently recovery rates in the order of between 88 95% are being been achieved Figure 11b Figure 11a

Mole Sieve Heater as used in the extraction process (Figure 12.0)

Figure 12.0 Mole Sieve Liquid Extraction Heater One primary objective for the system was to ascertain the liquid mass ratios in a wet-gas stream and the recovery uncertainty was not thought to be really relevant to the primary compensation of the metering result although management of a wet gas stream uncertainty and fluid composition is becoming more and more important in the industry today Consider ISO TC193 SC3Wg5 Wet Gas Sampling, a new ISO technical report that will be developed over the coming year showing the importance of this subject and developing techniques and methods to help in this exacting subject. Conclusions Multiphase & Wet Gas Metering plus wet gas sampling all fit together in a specific niche area, with the advent of the new issuance of multiphase standards and technical reports by ASME(MFC19) ISO TC193 Sc3 and API - CPMA (draft Chapter 20.3) the importance of this subject is gathering pace. The development of wet gas and multiphase sampler systems is a daunting challenge due to many factors some major points of which are 1) Mixing of the multiphase components - Representative sample collection issues 2) Collection of the components - Changing properties during the collection process 3) Design of the collection probe - Correct Representation of the flow stream regime * ( also linked to #1 *) The use of partial separation by the mole sieve methodology seems to be a viable method and has been in use for some years however proper selection of the flow regime and flow characteristics is very key in using the system type mentioned in the paper.

The propriety nature of many of these wet gas and multiphase systems can also be an issue for the standards writing committees, this is why many of the multiphase and wet gas standards documents are technical reports and not standards. The sampling system and analysis process discussed in this paper has been developed to allow the determination of the liquid content (LL) and composition of a wet gas stream to an improved accuracy than previously seen, it is at the leading edge of the art and science of this multiphase niche area ! The mass phase process should be suitable for adaption to many applications for testing of wet gas systems provided it is selected over the correct GVFs, flow and pressure regimes , it lends itself also to being used as a portable device which could have a real positive impact in the hydrocarbon industry world - wide for calibrating a wet gas meters PVT inputs , or checking a wet gas streams liquid content as a physical wet gas meter itself.

References
1) Mark A. Jiskoot. & Ken Payne- A New Method to Determine the Liquid Content of a Wet Gas Stream and Provide a Sample of the Liquid Phase for Compositional Analysis. 21st International NSFMW Meeting Tonsberg Norway 2003. 2)

Wet Gas JIP Data Release - CEESI 2007 Estes Park Colorado USA

3) ISO-TR.26762-ISO TC193Sc3Wg1 - Allocation Measurement of Natural Gas and Condensate in the Upstream Area. 2008 4) ASME MFC19 -Wet Gas Flow-Metering Guideline - 2008 5) Lawrence, Philip. A - Wet Gas Measurement Class 1320 International School of Hydrocarbon Measurement (ISHM) Oklahoma May 2008 6) Lawrence, Philip. A - Wet Gas Measurement Class 1320 International School of Hydrocarbon Measurement (ISHM) Oklahoma May 2010.

7) ISO TC193 Sc#3 Wg2 - Draft Wet Gas Technical Report 2011- reviewed ISO I.S.O. Nanjing Plenary Meeting China. 8) ISO TC193 Sc#3 Wg5 - Wet Gas Sampling Resolution Nanjing China - 2011