Available online at www. sciencedrect.comNUCLEAR。0.0 ScienceDirectSCIENCEANDTECHNIQUESNuclear Science and Techniques, Vol.18, No.2 (207) 73- 76Gamma-ray attenuation technique for measuring void fraction inhorizontal gas-liquid two-phase flowLI Zhibiao* WU Yingxiang LI Donghui(Instiute of Mechanics, the Chinese Academry of Sciences, Beijing100080, China)Abstract The measurement of void fraction is of importance to the oil industry and chemical industry. In this article,the principle and mathematical method of determining the void fraction of horizontal gas- liquid flow by using a sin-gle- energy γr-ray system is described. The Y-ray source is the radioactive isotope of 24IAm with r-ay energy of 59.5keV. The time-averaged value of the void fraction in a 50.0-mm id. transparent horizontal pipeline is measured undervarious combinations of the liquid flow and gas flow. It is found that increasing the gas flow rate at a fixed liquid flowrate would increase the void fraction. Test data are compared with the predictions of the correlations and a goodagreement is found. The result shows that the designed Y-ray system can be used for measuring the void fraction in ahorizontal gas- liquid two-phase flow with high accuracy.Keywords r-rays, Single-energy, Void fraction, Two-phase flowCLC number TP3911 Introductionracy. Daidzic et al.u have measured void fraction us-ing the magnetic resonance imaging (MRI) system.The imaging and measurement of multiphaseThe shortcoming of this approach is that the liquidflows have received much attention in recent years,should not be conductible. Kendoush and Sarkis'2]largely driven by the need in the oil industry to meas-have measured the void fraction using X-ray absorp-ure the mass flow rates of oil, water, and gas in thetion. Yang et al.3] have measured the void fractionproduction pipelines. The ability to see the inside of anusing the impedance method. Wojtan et al.4I haveobject and make quantitative measurements of the en-measured the dynamic void fraction in stratified typesclosed materials and structures has a wide range ofof flow. Harms et al.5 have proposed a void fractionapplications. However, the measurement over a widemodel for annular flow in a horizontal pipe.range of flow regimes and the ability to measure massRadiation technique is being considered as a bet-flow rates of each component with high accuracy re-ter option to get the details of a multiphase flow struc-quire a detailed knowledge of the hydrodynamics ofture, as information about phase distribution can bemultiphase flows, especially the phase fraction on theobtained destructively. Stahl et al.6l have discussed thecross-section of a pipeline for each component, as theaccuracy of void fraction measurements by sin-fast changing phase fractions directly control the mul-gle-beam gamma densitometry. Takenaka et al.7] havetiphase flow behavior and flow rates, and also the ba-measured the void fraction by neutron radiography.sic information to reconstruct the flow pattern images.The motivation for the study presented in this articleTherefore, in recent years many researchers have de-was to examine how γ-rays could be used in a particu-voted their attention to determining phase fractionslar fiel中國煤化工is, the imaging andand to improving phase fraction measurement accu-measuJ1ase flows in pipe-YHCNMHGSupported by National Natural Science Foundaion of China (No.10572143)*E-mail: zhibiaoli@ 163.comReceived date: 2007-01-0474NUCLEAR SCIENCE AND TECHNIQUESVol.18lines. Specifically, the possible role of Y-ray techniquesrepresents the intensity of the Y-ray beam before itin an application of oil industry that requires figuringpasses through the bottom wall. The attenuation of airout the phase fraction, by accurately measuring flowis so small that it can be neglected. Hence, Eq.(3) canrates of liquid and gas in the oil production pipesbe written asshould be examined.The radiation technique considered in this studyu.L=-ln(份)(4)is associated with a single-energy Y-ray system. Liquid[L=d-L,in the test section attenuate y-radiation without depos-iting significant amounts of energy. As the gas phaseBy solving Eq.(4), the thickness of air in the testhas lttle attenuating power for γ radiation, the at-section can be obtained. So, the void fraction can betenuation of gas can be neglected. Therefore, by de-written astecting the attenuation of the γY-radiation beam, the2(d-4)V4(d-4)phase fraction in the flow channel can be measured. Innarccos(1-4)-πd2以<號this article, the aim is to apply the single-energy γ-raya=radiation technique to determine the cross-section void.2124√4(d-4)fraction in a horizontal gas- _liquid two-phase flow.二arcos(-+d(5)2 Measurement principleAttenuation of ry-rays through a sample of thick-3 Experimental systemness L is given by3.1 Experimental setup=-μL(1)In this study, the Y-ray source is the radioactiveisotope 24Am, which emits y-ray with energies of 59.5where μ is the mean linear attenuation coefficient ofkeV. The radioactive isotope is assembled andthe sample, Io represents the incident (upon the sample)shielded in a thick lead pot to prevent the harmfulintensity of the Y-ray flux, I is the outgoing intensityemission of 241Am Y-rays. The radiation activity of thefrom the sample and L is the sample thickness. Whenisotope is 3.7 GBq. A collimated single y-ray beam (inthe Y-ray passes through different materials, the rele-20 mm diameter) radiates from the bottom of thesource pot and can be turned on/off by a mechanicalvant relations becomeswitch, to ensure operation safety (see Fig.1).In=-4-4L-L(2)where subscripts 1, 2, ... denote the materials withdifferent linear attenuation coefficients.According to Eq.(2), when the Y-ray passesthrough the object composed of water and air, the in-tensity attenuation yields:μL +μL =-In|Fig.1 Experimental setup.The scinillation detector is made of NaI crystal.lL +L=dA中國煤化工with a size of 40 mmwhere subscripts W and a denote the material of water(heitom IYHconnected with a pho-and air, respectively, d is the pipe diameter.CNMHGreterofthedetectorisIn this study, Io represents the intensity of the55 mm and the length is 220 mm. In addition, a colli-γ-ray beam after it passes through the upper wall, and Imationholewithasizeof50mmx30mmx150mmNo.2LI Zhibiao et al: Gamma-ray atenuation technique for measuring void fraction in horizontal gas liquid two phase flow5(length X width x height) is mounted on top of the de-for the prediction of the homogeneous model, andtector (see Fig.1).MVF is the measured void fraction. From Fig.2, it canThe nuclear instrument presented in this study isbe seen that void fraction increases with the increasingdesigned as a muli channel type. The system is oper-of mixture velocity. MVF and HVF have the sameated in count mode. It is made up of high voltagetendency, but at high mixture velocity MVF is lesspower supply, amplifier, shaping amplifier, and pro-than HVF. The reason for this fact can be explained asgrammable data acquisition system.follows. According to the assumption of the homoge-The experiment was carried out in the Mutiphaseneous model, gas and liquid have the same velocity inFlow Laboratory of the Institute of Mechanics, Chi-the pipe flow. But the real situation is, when the super-nese Academy of Sciences (see Fig.1). The length official velocity of gas is increased, actual velocity ofthe test section is 10 m and the plexiglas pipe innergas is larger than actual velocity of liquid, so the slipdiameter is 50 mm. The superficial velocities of airvelocity between gas and liquid causes MVF to be lessand water are 0.075~1.248 m/s and 0.539~1.617 m/s,than HVF at high mixture velocity. The superficialrespectively.velocities of the liquid in Fig.2 (a) and (b) are 0.8985When the pipe is empty, from Eq.(1),mas' and 1.6173 m/s, respectively. The maximumrelative error between MVF and HVF is less than= -24.4(6)10%.0.6where 1I2 and 1 represent the intensity of the γ-ray0.5-0- MVFbeam after and before it passes through the test section,0.4 trespectively, p denotes the pipe. So, the value of I andlocan be obtained:),3 I10 =I,/exp(44)(7)021= I2exp(4)0.1 tUs=0.8985m+s"From Eqs. (7), (4), and (5), the void fraction can0.0be calculated.3 1.8 2.0 2.2Mixture velocity Uu1 m+s--3.2 Static calibration0.450.40 -0- HVFb)o口Table 1 is the static calibration data, which showsMVFthat the system can be used to measure the void frac-2820.30tion accurately in static conditions.0.25 tTable 1 Static calibrationActual voidMeasured voidRelative error/ %0.15fraction0.100.104080.104630.53324Ua=1.6173m-s5-'0.266210.26308-1.176890.527720.52708-0.121661.0.73430.72922-0.69222Mixture velocity Uu 1 m-s-'0.90990.9075-0.26409Fig.2 Comparison of the test data with the homogencousmodel.Results and discussion4.2 Comparison with other correlations4.1 Experimental data中國煤化工cal correlation:The relation between void fraction and mixtureYHCNMHG(8)velocity are shown in Fig.2. X- axis denotes mixture“1+(0.2+1.2/(Usc/Us))velocity and Y-axis denotes void fraction. HVF stands76NUCLEAR SCIENCE AND TECHNIQUESVol.18This correlation has been widely used for the5 Conclusionscalculation of void fraction in horizontal gas- liquidThe measurement of the void fraction is of con-two-phase flows. Modified Armand equation wassiderable importance to the oil and chemical industries.proposed based on the test data:The principle and mathematical method of determin-1+ (0.2+ 0.98574/(Us/Ux)(9)ing the void fraction of a horizontal gas- -liquid flow,by using a single energy Y-ray system is described.The time-averaged value of a void fraction in aThe result is shown in Fig.3 (a). The X-axis de-50.0-mm i.d. transparent horizontal pipeline wasnotes the ratio of the superficial velocity of gas to thesuperficial velocity of liquid, and the Y-axis denotesmeasured. The measurements were made at variouscombinations of the liquid flow and gas flow. It wasthe void fraction. The relative error is less than 8%.found that increasing the gas flow rate at a fixed liquidflow rate would increase the void fraction. Test dataModifed armand oquation0.6-were compared with the predictions of the following0.5-correlations:.4(1) The maximum relative error between test data andAmand equationhomogeneous model prediction was less than 10%.(2) Test data had good agreement with the Armand0.2-equation.口Measurdata](3) The maximum relative error between the test data(aand Franca correlation prediction9 was 5%..0;223.0Ugc1UaAcknowledgments0.The authors acknowledge Wang Keren, MaFranca 19920.5____ +5%Naiqing, Yuan Maozhu and Ma Runhai for their col-laboration and suggestions in the phase fraction de-0.4termination.g 0.3References1Daidzic N E, Schmidt E, Hasan M M et al. Nuclear Engi-(bneering and Design, 2005, 10: 1163-1178.2 Kendoush A A, Sarkis Z A. Experimental Thermal and121.2.0Mixture velocity 1ms-'Fluid Science, 2002, 25: 615-621.Fig.3 Comparison of the test data with correlation of Armand3 Yang H C, Kim D K, Kim MH. Flow Measurement andand Franca.Instrumentation, 2003, 14: 151-160.Franca et al.9 measured the void fraction using a4. Wojtan L, Ursenbacher T, Thome J R. Experimentalquick valve in the horizontal pipe, and proposed anThermal and Fluid Science, 2005, 29: 383-392.empirical correlation of the void fraction:5HarmsTM,LiDQ,GrolEA,etal.IntJHeatandMassUsG_Transfer, 2003, 46: 4051-4057.(10)0.98Um +0.16 .6 Stahl P, von Rohr P R. Experimental Thermal and FuidThe comparison between the test data and theScience, 2004, 28: 533-544predicted value is shown in Fig.3 (b). The maximum7 Takenaka N. Asano H. Experimental Thermal and Fluidrelative error is 5%. The comparison of the test data中國煤化工with the correlations of Armand and Franca shows thatCNMH Gth Inst 1946. 1: 1623.the Y-ray system designed here can be used for meas-ACKE 1rans. NO. 820.uring the void fraction in horizontal gas- -liquid) Franca F, Lahey R T Jt. Int J Multiphase Flow, 1992, 6:two-phase flow with high accuracy.787-801.