Method and Device for Multiphase Fraction Metering Based on High Pressure Xe Filled Ionization Chamber

This invention relates to metering devices, more specifically, to the measurement of parameters of liquid and gaseous media and can be used for the control of fluid flow parameters, more specifically, for the condition monitoring of oil and gas field by controlling the composition of the multiphase mixture delivered from the well.

Prior Art

In the oil industry, probably, the simplest method of multiphase fraction metering is the radioactive method based on the attenuation measurement at several gamma ray energies. The main idea of the method is the comparison of the count rate of gamma rays passing through the multiphase mixture (production fluid) with the count rate of gamma rays passing through the empty pipe, and pure oil, water and gas.

Hereinafter the term 'fraction meter' will mean the part of the device that measures the volume fractions of various phases, whereas the 'multiphase flowmeter' will mean the whole device that measures the volume flowrates of all phases.

The International Patents PCT/EP 94/01320, EP 0236623, and WO 9742493 describe the method of multiphase mixture fraction metering based on the attenuation measurement at two or more energy levels. More than two energy levels are not needed for the characterization of three phase fractions, but they can be used for the determination of a forth component (for example, mechanical impurities), or for the measurement of the sulfur content in oil and salt content in water.

US Patents 6,097,786 and US 7,075,062 describe the main principles of multiphase fraction measurements with X-ray tubes instead of chemical sources.

The methods of multiphase fraction metering described above are applied in flowmeters manufactured by Schlumberger Vx, Haimo MPFM and Roxar MPFM (it should be mentioned that Roxar flowmeter actually uses only one energy level to determine three phase fractions; it also performs electromagnetic measurements of multiphase mixture properties). The part responsible for the fraction metering uses properly collimated radioactive source, i.e. a tube with multiphase mixture and a gamma ray detector. The gamma rays emitted by the source pass through the multiphase mixture and are incident upon the detector. The detector usually comprises a NaI scintillation crystal, a photomultiplier, and appropriate logic.

The NaI scintillation crystal has an energy resolution of about 6-10 and is preferable in comparison with other detectors since it is very widespread and can operate under severe conditions (up to 120 ⁰ C). The disadvantage of the scintillation detectors (not only equipped with NaI, but also many others) is their low energy resolution. To gather all the gamma rays from the same energy level, one should use a very wide energy window (for example, for 30 keV gamma rays the preferable energy window is 20-40 keV). In this case the high energy gamma rays hitting the detector can be scattered inside the crystal, leaving only part of their energy. Then they will be detected as low energy gamma rays. Moreover, to avoid very strong statistical fluctuations (Poisson noise) the total source activity can not be too low (at least several hundred MBq for a Ba 133 source). The gamma rays coming to the detector at the same time (up to the scintillation time) will be detected as gamma ray of double energy (accurate to the energy resolution). These challenges lead to the systematic errors in the detector response and must be corrected.

In comparison with scintillation detectors, semiconductor detectors have a significantly higher energy resolution, 0.1-0.5%. This means that energy

windows can be chosen very narrow, thus reducing the systematic error. Such detectors can resolve very close energy levels; the operator may assign several energy windows which is unattainable for a scintillation detector system. The more energy windows are used, the more detailed information about multicomponent fluid in the tube can be obtained. The patent EP 0696354 describes semiconductor detector used for multiphase fraction metering. However, semiconductor detectors are more costly than scintillation detector systems and require low cryogenic temperatures for operation. Moreover, only the best semiconductor detectors allow measuring high gamma ray fluxes (10 ⁵ counts/sec) that are required for reliable multiphase measurement.

The gamma-detection unit installed in the throat of the Venturi pipe is described in US Patent 7,105,805 (Schlumberger Technology Corporation). The operation of this unity is based on the attenuation of gamma-rays emitted by Ba- 133 (the main energy peaks are 32 keV, 81 keV and 356keV) that are used for density and composition measurement of multiphase mixtures. According to the invention, the detector units comprise a filter for selectively detecting the photons it receives at a first energy level and at a second energy level, said first and second levels being predetermined amongst the energy levels of radiation from the source. The filter comprises a scintillator crystal whose dimensions are such that said crystal mainly detects gamma rays that are emitted at said first and second energy levels. This design allows provides for better spectral characteristics at the first energy peak (32 keV) and improves the measurement accuracy. However, the peak width remains large enough and affects the accuracy of the result.

The suggested invention is to use a Xe-filled high pressure ionization chamber for gamma ray detection in the standard multiphase flowmeter. The gas-filled detector holds an intermediate position between scintillation detectors and semiconductor detectors concerning the energy resolution. For example, the commercially available ionization chambers have energy resolution of around 2-

3% at the Cs 137 peak (662 keV) [S. E Ulin, V. V. Dmitrenko et al. Instruments and Experimental Techniques, vol.. 37, JSTs. 2, part 1, p. 142-145, (1994); A. Bolotnikov and B. Ramsey. Nuclear Instruments and Methods in Physics Research, A 396, ρ.360-370 (1997); G. J. Mahler et al., IEEE Transactions on Nuclear Science, vol. 45, No. 3, p. 1029-1033 (1998)], and the maximum level of detectable gamma ray fluxes is comparable to that practiced for scintillation systems. The Xe filled ionization detectors can operate almost under any conditions (temperature above 200 ⁰ C and high pressures) and their performance is independent of temperature (if temperature-stable logic is used for signal processing). There is no need for a photomultiplier, and therefore the design is robust enough for field application. Finally, the cost of these detectors is less than for semiconductor detectors.

Summary of Invention

The present invention relates to the methods of multiphase mixture fraction meter, more specifically, to the development of instrumentation for the detection of gamma-radiation by the high pressure xenon-filled ionization chambers. The use of such detector allows obtaining good gamma ray energy resolution, detection of gamma rays at more energy levels, finally increasing the accuracy of the results. As is known to inventors, this use of gas filled detectors for multiphase fraction metering is novel.

The object of this invention is to provide a new multiphase mixture fraction determination method for mixtures delivered from the well and a device for the implementation of this method.

Said object can be achieved by providing the multiphase mixture composition characterization method suggested herein comprising the steps of passing gamma rays through a multiphase mixture flow and detection of the transmitted gamma radiation with further composition characterization of the

multiphase flow from the energies detected, wherein said detection is carried out using a high pressure Xe filled ionization chamber, further wherein said ionization detector comprises a gamma rays transparent window, a cathode, a wire anode and a shielding grid. To obtain a sufficiently high detection efficiency one should provide a Xe pressure in the detector of 30-50 atm.

When implementing the method the volume fractions of oil, water and gas are preferably determined using preliminarily measured absorption coefficients for a multiphase mixture at least at two energy levels, wherein the absorption at one of these levels is due to photoeffect and that at the second level is due to Compton scattering, and the volume fractions of the three phases are calculated using the following set of equations:

μfa _0ll λ-μr ^r a _waler +μra _gas = μ _x μfa _oύ +μT ^er a _wmer +/*T<V = λ x where ^μ > is the X phase attenuation coefficient, ^aχ is the volume fraction of said phase in the mixture and / = 1, 2 is the energy window number.

However, the fractions of oil, water and gas can be determined using more than two energy levels by finding the volume fractions of oil, water and gas at which the following sum takes on its minimum value:

∑N,(ju:' ¹ a _oύ +//,— «« _* , +μr oc _sas -μj = min

I=I , where ' is the count rate in the z-th energy window and n is the number of energy levels.

The method can be implemented using a multiphase mixture fraction characterization device based on an ionization chamber filled with high pressure Xe and further comprising a gamma ray source and a Xe filled detector comprising a gamma ray transparent window, a cathode, a wire anode and a shielding grid. Preferably, said device comprises a collimated gamma ray source. The gamma ray source can be either chemical source or an X-ray tube. The detector window is preferably made from a material with a low absorption

at 30-80 keV. Said gamma ray source should preferably provide at least two energy levels.

Detailed Description of the Invention

Gas filled detectors are primarily of interest for their high energy resolution achieving 2% at 662 keV in some counterparts, whereas the theoretical value can be even higher (0.5% without electronic noise). The energy resolution of a gas filled detector can be described as follows:

M . ₂ .35 ^ ^{+ Fε E} E E where E is the gamma quantum energy, δ£ is the energy resolution (absolute value), a is the electronic noise related energy determination error, ε is the Xe atom ionization energy (22 keV) and F is the Fanot factor equal to 0.13 for Xe.

The gas filled detectors (depending on voltage) can operate in two basic modes: ionization and proportional. For composition detection, the best mode is the ionization one which requires a higher Xe pressure in the chamber, but comparatively lower electric field. The optimum pressure for high energy resolution (one of the key parameters of the multiphase fraction meter) was determined in [S. E Ulin, V. V. Dmitrenko et al. Instruments and Experimental Techniques, vol.. 37, JNa. 2, part 1, p. 142-145, (1994); A. Bolotnikov and B. Ramsey. Nuclear Instruments and Methods in Physics Research, A 396, p.360- 370 (1997)] and is 40-50 atm under standard conditions (Xe density within 0.6 g/cm ³ ). This level of xenon density provides for a reasonably high detection efficiency at expected gamma ray energies, from 30 to 1000 keV. The actual level of detection efficiency depends on the detector size and may be 10-15% at 662 keV in some Xe detector counterparts [S. E Ulin, V. V. Dmitrenko et al. Instruments and Experimental Techniques, vol,. 37, N2. 2, part 1, p. 142-145,

(1994); G. J. Mahler et al., IEEE Transactions on Nuclear Science, vol. 45, No. 3, p. 1029-1033 (1998)]; the typical detector length is about 20-40 cm. The typical layout of the multiphase fraction meter is shown in Figure 1. Note that the layout is exemplary only, and the actual geometry may vary. The diameter of the detector is not of great importance for the collimated beam, ranging from 5 mm to 50 mm.

Collimated gamma rays from radioactive source 1 (chemical source or X- ray tube) pass through the tubing with multiphase liquid 2 and enter the detector. The Xe filled detector comprises inlet window 3 made from a high- strength material (typical internal pressure inside the gas detector is 30-50 bar). The window material has a low attenuation coefficient at the source energy (for example fiberglass, fibercarbon, kevlar, beryllium etc.). The detector further comprises cathode 4, anode wire 5 (it can be a cylinder) and special shielding grid (Frisch grid) 6.

The operation principle is the same as for any ionization chamber. The gamma rays interacting with the gaseous matter (by the photoelectric effect or Compton scattering in the energy region which of interest) generates photoelectrons that drift along the electric field between the cathode and the anode wire. The electric field should not be very high so the chamber remains in the ionization mode (up to 3-5 kV depending on gas pressure and hydrogen and other gas impurities in xenon; anyway the field should be insufficient to generate an electron avalanche). The magnitude of the signal taken from the anode wire is proportional to the energy lost by single gamma ray quantum. The shielding grid allows measuring the electron component of the total signal thus increasing the performance of the gas filled detector and making the signal independent from the point of the gamma rays/gas molecules interaction.

One can also provide a device based on a gas filled detector operating in proportional mode with a moderate amplification coefficient. To this end one should provide a higher electric field sufficient for the development of electron

avalanche thus increasing the overall collected charge. The advantage of proportional mode is that it provides for a higher signal-to-noise ratio (total charge or photoelectron current). This reduces the requirements to the electronics or provides somewhat higher energy resolution. The disadvantage of proportional mode is that a very stable electric field should be provided. Indeed, the amplification coefficient is an exponent function of the field, and small field fluctuations cause dramatic changes in the coefficient, bearing the threat of impairing the energy resolution. Furthermore, it is a serious technical problem to provide a high electric field as the field strength required for proportional mode is a linear function of gas pressure in the detector. The fundamental design of a gas filled detector operating in proportional mode is similar to that of an ionization chamber.

When a Ba- 133 source is used, the following energy windows are available for the high-accuracy parameters of the multiphase fraction meter:

Wl : 29-38 keV (peak at 32 keV)

W2: 76-86 keV (peak at 81 keV)

W3: 267-284 keV (peak at 276 keV)

W4: 295-312 keV (peak at 303 keV)

W5: 346-365 keV (peak at 356 keV)

W6: 374-394 keV (peak at 384 keV).

Note that the number of energy levels can be greater, for example, seven in this case if the logic keeps the noise level low enough to resolve the 31 and 35 keV peaks in the Ba 133 spectrum. Measuring the attenuation coefficients at five windows gives us seven variables (such as the volume fractions of oil, water, and gas, salt content in water, sulfur content in oil, sand concentration etc.) if the testing time is sufficient (at least few minutes at several hundred MBq source activity) and if the tested fluid is sufficiently stable during measurements. Other application (more important) is to use the attenuation data in additional energy windows for gaining a higher accuracy of oil, water, and

gas volume fraction determination (using the standard root-mean-square technique).

As noted above, the detector performance can be changed to suit the requirements (electric field strength, shielding grid installation and parameters, gas pressure etc.). The energy resolution achievable is significantly better in that case than for the scintillation detectors. The detection efficiency per unit volume of a gas filled detector is lower than that of the scintillation detectors at gamma ray energies of about 200-400 keV, but this drawback is not critical as the high signal speed allows using a higher activity source. Furthermore, the same detection efficiency can be achieved as in the scintillation detector if the gas filled detector has a larger size. For on-surface applications of the testing tool, a 10-30 cm increase in size is not critical. Moreover, since the most important information can be obtained from the low energy gamma rays (32 keV peak for Ba 133) that are sensitive to the fraction composition, and high energy gamma rays mainly provide information on density (their attenuation is proportional to the mixture density and is less sensitive to the fraction composition), high statistics for all of the energy windows is not obligatory. For example, a Xe filled (50 bar) detector with a detection length of 20 cm is suitable for the reliable detection of 81 keV gamma rays. As to higher energy gamma rays (windows W3-W6), their statistics can be useful for the r.m.s calculation of oil, water, and gas fraction corrections.

Unlike standard scintillation detectors, the signal from the Compton scattered gamma rays with 276, 303, 356, and 384 keV energies will not generate strong noise in the Wl and W2 windows since their width is small enough. The existing background (systematic error) will affect the attenuation coefficient (it depends on material thickness) and can be eliminated by a simple correction of the count rates in the Wl and W2 windows to the count rates for other windows using empirical coefficients. These coefficients can be

determined by attenuation measurements of known thickness samples (liquid or solid).

The oil, water and gas volume fractions can be found using the following equation:

∑ N, {μf a _oύ + μ:° ^ler a _waler + _/ , ^«« a _gm - μ, ) ² = min ι=l

X where **• is the X-phase attenuation coefficient, ^aχ is the volume fraction and ' is the count rate at the z-th window.

Example

By way of example we consider a device for measuring the phase composition of a multiphase mixture that comprises an ionization chamber as the gamma detector. The detector is 30 cm long and 15 cm in diameter and has an anode (a 3 mm diameter cylinder) in the center. The Frisch grid is installed at 4 mm from the anode surface. The cathode-grid voltage is 2 kV, the anode-grid voltage is 6 kV and the anode to grid distance is 5 mm. The detector has an analog to digital converter and a computer converting the anode current into gamma quantum energy. The energy resolution of the device is 2% at Cs 137 (662 keV). With a Ba 133 detector, the following levels can be detected: 32 keV (31+35 keV), 81 keV, 276 keV, 303 keV, 356 keV and 384 keV.

The Xe density is 0.5 g/cm ³ (the 2O ⁰ C pressure is about 50 bar). The detection efficiency under the above conditions is about 50% for 356 keV gamma quanta. The inlet window is made from PEEK (polyether-ether-ketone), thickness 3 mm, and the device is installed in the tube in a way that the Ba 133 source is at one side of the tube and the gas filled detector is at the other. The metallic walls of the tube have openings in which boron carbide windows are installed (boron carbide has a relatively low absorption coefficient in the 30 keV

gamma quantum energy region). The tube diameter in the device installation area is 5 cm.

Before measuring gamma quantum absorption coefficients of water, oil and gas, the count rates are calibrated for an empty tube and then in sequence for a water, oil and gas filled tube. The latter measurement is made for the gas separated from the multiphase mixture and injected into the tube under high pressure. If this calibration is impossible, the gas absorption coefficient can be calculated if the gas composition is known at least approximately. Then the data are corrected in a way to subtract the contributions from higher energy gamma quanta scattered in the crystal and detected as low energy gamma quanta. The gamma quantum absorption coefficients of water, oil and gas are calculated as

— = In — -, where d is the tube diameter (gamma quantum path length) and

Px Px ^d Nx'

⁰ is the empty tube gamma quantum count rate measured in an air filled tube and then corrected for vacuum based on the air absorption coefficients of gamma quanta with different energies.

Gas volume fraction and water content measurements in the multiphase mixture are carried out by measuring the mixture absorption coefficient and comparing it with the absorption coefficients obtained for each of the component media as described above. The volume fractions of water, oil and gas and possibly other parameters (mechanical impurities, water salt concentration etc.) are calculated from the absorption coefficients obtained as above. This multiphase flowmeter allows measuring water content accurate to 1- 1.5% with a gas volume fraction of up to 95%. At higher gas volume fractions water content measurement becomes a difficult task, but it is possible to estimate the gas volume fraction or, which is more valuable, the liquid volume fraction in the multiphase mixture.