Gas-water stratified flow patterns from electromagnetic tomography

254Pet.Sci.(2009)6.254-258DOI 10.1007/s12182-009-0040-xGas-water stratified fow patterns fromelectromagnetic tomographyWang Xiaoxing and Wu Xiling*School of Resources and Information Technology, China University of Petroleum, Beijing 102249, ChinaAbstract: Because of gravitational differentiation of multi-phase fluids, gas-water flow is usuallystratified in highly inclined or horizontal gas wells. By using electrode arrays to scan flowing fluids,electromagnetic tomography can identify the flow patterns of mixed fluid from the different electricalproperties of gas and water. The responses for different gas-water interface locations were calculated andthen physical measurements were undertaken. We compared the results of the numerical simulation withthe experimental data, and found that the response characteristics were consistent in the circumstances ofuniform physical fields and stratified flows. By analyzing the signal characteristics, it is found that, withthe change of the interface location, the response curves showed“steps" whose position and width weredecided by the location of fluid interface. The measurement accuracy of this method depended on thevertical distance between adjacent electrodes. The results showed that computer simulation can simulatethe measurement of the electromagnetic tomography accurately, so the physical experiment can bereplaced.Key words: Electromagnetic tomography, stratified flow, gas-water, numerical simulation, physicalexperiment1 Introductionscanning measurement.In the recent years, flow imaging techniques, as a non-Because of gravitational difentiation of multi-phase linear measuring method, have made rapid progress withfluids, gas-water flow is usually stratified in highly inclined or the development of measurement science and computerhorizontal gas wells. Conventional flow measuring techniques techniques (Peyton et al, 1999; Bolton et al, 2007; Tan et al,are linear methods that average the local values and cannot 2007; Yang and Liu, 2000; Li and Huang, 200; Bolton andgive accurate descriptions of the fluid distribution in the wellPrimrose, 2005). For the special circumstance of multi-phase(Wu, 2004).flow in the well, Wu et al (1999; 2000; 2008) introduced flowIn order to obtain stisfactory results, the Schlumberger imaging techniques to the well logging field and proposedCompany developed a new production logging instrument foran electromagnetic tomography method for the multi-phasehorizontal wells- Flow Scanner, which has four miniature flow according to the dfferences of electrical propertiesspinners, five electrical probes, and five optical probes among oil, gas, and water (Zhao and Wu, 2003; Zhao et al,designed to measure the velocity profile of the borehole fluid, 2007). Based on the previous work, we used electromagneticand the fraction of water and gas present in an interval oftomography for the gas-water stratified flow, and studied thepipe (the local water and gas holdups). The electrical probes charateristics and changing rules of the measuring responsesmeasure fluid impedance to distinguish hydrocarbons andby numerical simulation and physical experiments.water and the optical probes GHOST (Gas Hold OpticalSensor Tool) are used to identify oil and gas from their 2 Measuring modelrefractive indices (Wang et al, 2007). The CAT (CapacitanceArray Tool) developed by the Sondex Company has an arrayThe electromagnetic tomography sensor is designed asof 12 capacitance sensors placed on the backup arm of theshown in Fig. 1 and Fig. 2. A composite structure of theMAC (Multi-Arm Caliper) device, which can distinguishfirst and the third layers are the shielding electrode layers,oil, gas, and water according to their dielectric constantsaround the probes (Ni and Zheng, 2004) The MCFM (Muli- and the middle one is the main electrode layer. SixteenCapacitance Flow Meter) developed by the Baker Atlaselectrodes are located around the circle of the inner wall inCompany works on a similar principle. It is clear that theeach layer. In order to reduce the influence of the steel casingpresent tools just measure the local fluid without a wholeon the measuring中国煤化工re intalld on acorrosion-resistaDuring thefYHC N M H Gan electrode in*Corresponding author. email: wuxiling@sina.comthe main electrode layer to emit signal, and two focusingReceived August 18, 2008electrodes and two shielding electrodes close to it on thePet.Sci.(2009)6:254-258255-V.(eVφ)+∠v.(σVφ)=0 (φ∈ D){φ=1(T :φ∈source electrode)(1)Shielding electrode layerφ= 0(T2 :φ∈earthed electrode)意Main electrode layerwhere, D is the area to be solved; F is the boundary of thetransmitting and focusing electrodes; F2 is the boundary ofthe grounding electrodes.Fig. 1 Schematic diagram of clectrode arrays3 Simulation of the measuring responsesAccording to Fig. 2, a sine signal with the voltage of1 V and the frequency of 3 MHz is applied to the transmittingGrounded electrodeand focusing electrodes. The real and imaginary parts ofi 2 Focusing electrodeMeasuring electrodethe complex number result can be obtained by FEM (Finite .1 Transmitting electrodeElement Method). Then, the amplitude and phase can beGrounded electrode 10; Focusing electrodecalculated as follows:13 14a(r)2 =Q,(r)2 +Q}(r)2(2)Fig. 2 Section of the main electrode layer and measuringmodetan(r)=2(rQ,(r)(3)left and right, above and below emitted the same signals. where, a() is amplitude,中(r) is phase; Q(r) and 9() are realThen, the other electrodes in the main electrode layer were part and imaginary part, respectively.switched to be measuring electrodes in turn, and received theFig.3 shows the potential distribution when the firstresponses to the fluid medium in diferent directions. The four electrode is the main transmitting electrode and the fifthelectrodes close to the measuring electrode are all grounded. electrode is the measuring electrode. The gas-water (blue isBecause of 11 measuring data for one transmitting electrode，the saline water of 1 Q2m, white is air) interface lies betweenthe number of measuring data in a period is 16x(16- -5)= 176. the feenth and the siteenth electrodes.Zhao (2002) proved that the electric potential near theIt is clear that the distribution of media affects themain eletrodes is uniform in the Z-axis by analyzing the potential dstribution strongly. In the air region, equipotentialdistribution of the 3~D electromagnetic field. Therefore, 3~D lines are denser which means the potential changes faster. Thesimulation can be transformed to 2-D, which can simplify the values of the real part are greater than those of the imaginaryproblem. When the frequency of the electromagnetic wave is part. The potential distributions corresponding to otherlower than 3 MHz, it can be proved that the electromagnetic combinations of measuring elctrodes are similar.field is a near field in the range of pipeline diameter and isAs the changes of the amplitude are more regular, wea time-varying field under quasi-static conditions (Shen，ianalyze the response curves of the amplitude. Fig. 42006; Zou, 2002). An excitation signal with a common shows a period of the response curve (176 measured values)harmonic voltage is emitted. Ignoring the coupling of of the model in Fig. 3. The main transmitting eletrodeelectric and magnetic fields, the definite problem of the 2-D rotates from the thiteenth electrode counterclockwise. Forelectromagnetic field can be described as follows according convenient comparison, the response curves of uniform fieldsto Maxwell's equations:of air and saline water are displayed together.0.RV3 1.050.C0.950.00.0230.020.850.6520.550.0140.45 :0.011-0.02 .0.35-0.020.250.005 .16 -0.04-0.05中国煤化工15-0.061213 1-0.0 -0.04 -0.02 0 0.02 0.04 0.06MHC NMHG4 0.06x, mm入，111Fig. 3 Potential distribution of real and imaginary parts of gas- water sratified flow256Pet.Sei.(2009)6:254-258.35-- Uniform saline waterUniform air30-0.15 t20VUVUUUUUUUUUUUU0.1005-.00-1 22 3344 55 6677 .8899110121 132143 154165 176Measuring sequenceFig. 4 Simulated response curves of gas-water stratfied flow (five electrodes in the water)Fig.4 indicates that because of the same medium saline water. Fig. 5 shows the comparison of the experimentaldistribution of the uniform field, the response curves of the and simulated data. Because of the repetition of the curve16 electrodes have the same“U”shape according to the characteristics with one- phase fluid, only 22 measured valuesdifferent distances between the transmitting electrode and when the first and the second electrodes were the transmittingmeasuring electrode. The lowest point corresponds to the electrodes are presented as an example. Fig. 6 shows thevalue of the measuring electrode opposite to the transmitting comparison of the simulated data and experimental data of theone. Moreover, the values of the saline water are greater than gas-water stratified flow model in Fig. 3. Good consistenciesthose of the air.were shown in Fig. 5 and Fig. 6, which proved the accuracyFor the measurement of a single transmitting electrode of the numerical simulation.(11 measured values), when the medium in the pipeline isstratified, wherever the transmitting and focusing electrodes- Simulation results of saline waterare, if the measuring electrode is in the air, the measured-o- Experimental resuts of salie watervalues are always between those of the two uniform fields.However, if the transmitting and focusing electrodes areSimulation resuts of airin the air and the measuring electrode is in the water, the- Experimental results of airmeasured values are almost zero, which makes the responsecurve show a“step”, whose width depends on the number ofthe measuring electrodes in the water. Many“steps” in the0.15response curves of a single transmitting electrode composethe“step”in the curve of a period (176 measured values). Thewhole width depends on the number of the electrodes in the言0.05air of all 16 electrodes.4 Physical modeling246810121416182022Air and saline water (resistivity is 1 Q2:m) were selectedfor the static physical model experiment. The measurementFig. 5 Comparison of the experimental and simulated data with a single-system consisted of a measurement sensor, control circuit,phase fluidsimple flow simulator, and an HP8753C network analyzerwhich was used to create and measure the electromagneticfields. As the network analyzer provides a constant power, 5 Analysis of signal characteristicsthe attenuation values of power are measured. The voltageamplitude can be calculated from the attenuation values ofBecause the computer simulation can simulate thepower.measurement of the electromagnetic tomography accurately,the physical experiment can be replaced. Fig. 7 shows theresponse curves when the gas-water interface is near theUpu/Roupuutop of the pipeline, taking the interface between the firstF=101g pumu =10lgupn(4) electrode and the second electrode for example. When theFimput0 input/ Riputtransmitting and focusing electrodes are in the air and themeasuring electr中国煤化工ep”still appears.The differences fYHc N M H Ghe sratfid flowwhere, F is the attenuation values of power.in Fig. 4 are that un umul. ui cMIU vaiuus are more whichAt first, the experiment was performed with either air or means the width of the“step” is larger because of morePet.Sei.(2009)6:254-25857. - Simulation results0.16- Experimental results0.140.12 .0.100.06 一0.04 -0.020.00Measuring sequenceFig. 6 Comparison of the experimental and simulated data from gas- water sratified flow.57Uniform airGas-water sratified flow)4-vvvvvwwvvvvvvv0.0一11 22 33 44 556677 8899110 121132 143 154 165 176Fig. 7 Simulated response curves of gas-water strtified flow (nine electrodes in the water)electrodes in the water. The width of the whole“step”in a decided by the places where the transmitting, focusing andperiod of measurement (176 values) decreases because fewermeasuring electrodes are.electrodes are in the air.2) The measurement accuracy, namely the discernibleIn order to study the measurement accuracy, the response height of the stratified flow, is the vertical distance betweencurves for different locations of interface were simulated. The two adjacent electrodes above and below according to theresults showed that if the gas-water interface was between curve characteristics.the same electrodes but at different places, the shapes of the3) Electromagnetic tomography can identify gas-watercurves were exactly the same. With the increase of the water stratified flow and provides a new method of flow profilecontent, the measured values became greater. Therefore，measurement. Computer simulation can simulate theaccording to the curve characteristics, the measurement measurement of the electromagnetic tomography accurately,accuracy, namely the discernible height of the stratified flow, so the physical experiment can be replaced.is the vertical distance between two adjacent electrodes aboveand below.ReferencesWhen the equipment was placed in the well, the relativeBolton G T and Primrose K M. An overview of electrical tomographicposition of the equipment and the gas-water interface wasmeasurements in pharmaceutical and related application areas. AAPSunknown, and the interface may be slanting. In this situation,PharmSciTech. 2005. 6(2): E137-E143the charateristics of the response curves still met the above- Bolton G T, Benett M, Wang M, et al. Development of an etrialmentioned rules, which were also proved by numericaltomographic system for operation in a remote, acidic and radioactivesimulation.environment. Chemical Engineering Journal. 2007. 130(2-3): 165-6 ConclusionsLiHQandHuang.中国煤化工。=gy and Application.Hangzhou: ZhejMHC N M H Gin Chinese)1) With change of the intrface location, the response Ni G J and Zheng x xtApplcaton ot a new model multiple phasecurves showed “steps” whose position and width wereholdup logging instrument in horizontal well- Capacitance array258Pet.Sci(2009)6.254-258multiple phase holdup logging instrument (CAT). Well Testing. 2004.electromagnetic measurement in well logging. Science in China,13(4): 86-89 (in Chinese)Series D: Earth Sciences. 2008. 38(S2): 161-165Peyton AJ, Beck M S, Borges A R, et al. Development of electromagnetic WuX L, Zhao L and Liu D J. A fundamental study on electromagnetictomography (EMT) for industrial applications. In: Proceedings of thewave imaging logging in multiphase flow. Petroleum ExplorationWorld Congress on Industrial Process Tomography. Buxton: Greaterand Development. 2000. 27(2): 79-82 (in Chinese)Manchester. 1999. 306-312Yang W Q and Liu S. Role of tomography in gas/solids flowShen X N. Electromagnetic Fields and Elctromagnetic Waves. Beiing:measurement. Flow Measurement and Instrumentation. 2000. 11(3):Science Press. 2006. 78-81 (in Chinese)237-244Tan C, Dong F and Wu M M. Identification of gas/liquid two-phase flow Zhao L. Study on flow imaging logging using electromagnetic wave.regime through ERT-based measurement and feature extraction.h.D. Thesis. Beijing: China University of Petroleum. 2002 (inFlow Measurement and Instrumentation. 2007. 18(5-6): 255-261Wang G F, Wu L and Liu T. Flow Scanner: A new flow tool for Zhao L and Wu X L. Calculation of sensitivity field for electromagneticproduction logging in horizontal wells. Petroleum Instruments.tomography in multiphase flow well logging. Chinese Journal of2007. 21(2): 33-36 (in Chinese)Geophysics. 2003. 46(6): 870-874 (in Chinese)Wu X L. Petroleum Production Logging Principles. Beijing: Higher Zhao Y W, Wu X L and Wang X X. Simulation of sensitivity fieldEducation Press. 2004. 213-236 (in Chinese)for electromagnetic tomography in multiphase flow well logging.WuX L, Jing Y Q and Wu S Q. Electromagnetic imaging loggingChinese Journal of Geophysics.2007. 50(3): 811-816 (in Chinese)method in multiphase pipe flow. Chinese Journal of Geophysics. Zou L. Modeling of electrical resistance tomography. Master Degree1999. 42(4): 557-563 (in Chinese)Thesis. Beijing: Tsinghua University. 2002 (in Chinese)WuX L, Wang X x, Zhao Y W, et al. Flow imaging method of(Edited by Hao Jie)中国煤化工YHCNMH G