Flow pattern identification in oil wells by electromagnetic image logging

Petroleum Science, Aug 2012

Petroleum production logging needs to determine the interpretation models first and flow pattern identification is the foundation, but traditional flow pattern identification methods have some limitations. In this paper, a new method of flow pattern identification in oil wells by electromagnetic image logging is proposed. First, the characteristics of gas-water and oil-water flow patterns in horizontal and vertical wellbores are picked up. Then, the continuous variation of the two phase flow pattern in the vertical and horizontal pipe space is discretized into continuous fluid distribution models in the pipeline section. Second, the electromagnetic flow image measurement responses of all the eight fluid distribution models are simulated and the characteristic vector of each response is analyzed in order to distinguish the fluid distribution models. Third, the time domain changes of the fluid distribution models in the pipeline section are used to identify the flow pattern. Finally, flow simulation experiments using electromagnetic flow image logging are operated and the experimental and simulated data are compared. The results show that the method can be used for flow pattern identification of actual electromagnetic image logging data.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://link.springer.com/content/pdf/10.1007%2Fs12182-012-0213-x.pdf

Flow pattern identification in oil wells by electromagnetic image logging

Pet.Sci. electromagnetic image logging Liu Zaibin 1 Wu Xiling 0 0 College of Geophysics and Information Engineering, China University of Petroleum , Beijing 102249 , China 1 Gas Storage Well and Advanced Technology Development & Application Division, China Special Equipment Inspection and Research Institute (CSEI) , Beijing 100013 , China Petroleum production logging needs to determine the interpretation models first and flow pattern identification is the foundation, but traditional flow pattern identification methods have some limitations. In this paper, a new method of flow pattern identification in oil wells by electromagnetic models are simulated and the characteristic vector of each response is analyzed in order to distinguish the model 1 Introduction In petroleum production logging, in order to calculate the output of crude oil or natural gas in oil or gas wells or observe the oil or gas content change in the strata, important parameters such as oil, gas and water flux, velocity and pressure drop in the pipeline need to be measured online, and these parameters are different in different multiphase flow models. The most widely used multiphase flow model is the model based on flow patterns, so the flow pattern Traditional flow pattern identification method has some limitations, and it is quite difficult to identify two-phase flow patterns accurately using the traditional method, so a new method of flow pattern identification is needed and electromagnetic image logging can play an important role. The two-phase flow patterns depend on the type of fluid and inclination of the conduit. In addition, heat and mass transfer rates, momentum loss, rate of back mixing and pipe vibration all vary greatly with the flow patterns. Therefore, it is necessary to identify the patterns and distinguish their correlation with the flow properties (Mahvash and Ross, 2008) . Flow image logging is a method to probe the interior of multiphase flow through section measurement in a nonlinear way of showing the distribution and movement of the flow in wells by scanning the region detected and applying suitable data processing. It is important for monitoring the dynamic production of vertical, inclined and horizontal wells. Electromagnetic image logging can identify mixed flow patterns based on the electrical property differences between section. Based on previous research on the flow imaging method of electromagnetic measurement in well logging ( Wu et al, 1999 ; 2000; 2008; Wang and Wu, 2009 ; Zhao 2 Flow pattern transformation The most familiar two-phase flows in petroleum production are gas-water flow and oil-water flow. For the churn flow and annular flow, and there are stratified flow and wave flow in horizontal wells ( Hoogendoorn, 1959 ; Oshinowo and Charles, 1974 ; Dulder and Hubbard, 1975; cyclicity, but the stratified model curve has a step change. The signal processing method is used to extract signal characteristics which is relevant with the fluid distribution models, so the models can be separated by the measurement signal characteristics. Through simulation and analysis of different fluid distribution models, the relationship between the measurement signal characteristics and the fluid distribution models can be obtained, that is the curve shape of the measured signal depends on the fluid distribution. For different fluid distribution models, the curve shape is different. If the fluid distribution is centrosymmetric, the curve will show periodic repetition such as the “U”-type of uniform and annular models. For the same kind of distribution model, the numerical value of simulated signal mainly depends on the proportion relationship between the light and heavy phase, i.e. the water holdup. In this sensor model and measurement mode, the greater the water holdup, the smaller the measured value. 4.1 Signal processing Since the shape of the response curve is related to the fluid distribution model, different models can be identified by the characteristics of the response data. The statistical parameters of the simulation data of the eight distribution models are computed by Origin 8.0 software. We computed the mean of the 11 measurement values obtained from the same transmitting electrode, and there are 16 means in all. Then we computed the difference between the maximum and the minimum of the means, and named it as “D(V)”. There are 20 parameters in all which compose the feature vector as shown in Table 2. The fluid distribution models are identified using SPSS Clementine Client 11.1 software which can choose the most suitable parameter. The test results show that any single parameter cannot distinguish all the models, so the parameter combination or substep method is considered. The parameter combination method requires repeated dichotomous classification using multiple parameters to form a rule set or decision tree, and the required parameters process of substep method is step by step, and one model can 6. First, we input all the simulation data. Second, we compute the feature vectors. Third, we identify the stratified model by parameter CoefVar. Last, we identify the eccentric and annular models by parameter D. With this method, the main Simulation data Computed parameter Y N CoefVar>0.467 logging is to identify the combination of fluid distribution models. We use number 1-8 to denote the eight fluid distribution models in Fig. 2, and the basic two-phase flow own ID, as shown in Table 3. sl; 4: sh; 5: al; 6: ah; 7: el; 8: eh. For electromagnetic image logging, the detecting area is the flow section which is perpendicular to the well axis and the measurement data reflect the fluid distribution. After identification, one number can be obtained for each measurement. The measurement through the cross-section is continuous, and a series of numbers (ID) means continuous The disadvantage of this method is multiple solutions. flow patterns and their ID in Table 3 are enough for actual production needs. 5.1 Flow simulation experiment The electromagnetic wave propagation speed is so fast be simulated by the static distribution model experiments. Connecting a network analyzer as both an excitation and measuring device of electromagnetic waves and using a newly developed electromagnetic image logging experimental instrument, the uniform fields include all air and all water and air-water flow with different water holdup experiments are operated. 176 independent full-cycle measurements are recorded in each experiment. Air, tap water and brine are used to simulate natural gas, underground water and strata brine. The water holdup is set to be 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 100% from all air to all Three experiment measurements with the water holdup of 35%, 75% and 85% are used to investigate the precision of and the electromagnetic wave frequency is 3 MHz, the power is 10 dBm. 5.2 Characteristics comparison of the measurement The uniform field in this experiment means singleThe characteristics of uniform distribution experimental measurements can be used to test the accuracy of numerical has periodic characteristics, only 22 all air measurements of emission electrodes 1 and 2 are shown in Fig. 7 for example. For the uniform distribution, the shape of the numerical simulation measurement curve is a continuous regular “U” and the repetitive cycle is 11. The shape of the experimental curve is a continuous irregular “V” and the repetitive cycle is also 11. As shown in Fig. 7, the measurement responses of the numerical simulation and the physical model experiment are value is only related to the electrode location. For a certain excitation electrode, the measured value of the electrode which is furthest away from the excitation electrode is minimum, and the value of the nearest electrode is maximum. 5 . 3 C h a r a c t e r i s t i c s c o m p a r i s o n o f s t r a t i f i e d distribution measurement data Taking the water holdup 50% for example, we analyze data. Fig. 8 shows the contrast between the measured and simulation data of water holdup 50% stratified distribution model. Experimental measurements of all air Simulation measurements of all air The measurement response characteristics of stratified distribution is completely different from that of a uniform field. The measurement data of stratified flow are high and low potentials. In one measurement cycle (11 values), if the receiving electrode is in the brine, the measured voltage amplitude is relatively high but lower than all brine, and if the receiving electrode is in the air, the measured voltage amplitude is relatively low but higher than all air. The measurement value is decided by the fluid around the receiving electrode. From the comparison of the characteristics of the uniform and stratified distribution, the numerical simulation and experimental response data have the same characteristics and change rules. Therefore, the flow pattern identification method based on simulation can be used for the actual electromagnetic image measurement data. In practical application, the relevant characteristic parameters must be amended according to the relationship between simulation and measured data of all gas and all water distribution. yes yes yes yes yes yes yes yes yes yes yes 1 1 3 3 3 3 3 4 4 4 4 4 2 2 yes yes yes yes yes yes yes yes yes yes yes yes Fluid Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine Air/brine 0% 10% 20% 30% 35% 40% 50% 60% 70% 75% 80% 85% 90% 100% The method of flow pattern identification investigated in this paper is validated with 27 groups flow simulation experiment measurement data including uniform and stratified distribution models, as shown in Table 4. Four groups identification results do not accord with the actual models (water holdup 10% and 90% respectively stratified distribution model), and the absolute error is 14.8%. Considering the original error of the experimental data is larger than the simulation data, the results are acceptable. Water holdup 6 Conclusions models, and the flow pattern identification is transformed method, the electromagnetic image logging responses of response characteristics of the fluid distribution models are analyzed. The flow pattern identification is achieved using signal processing. With the network analyzer and developed electromagnetic array sensors, the gas-water two-phase flow in a horizontal pipe is measured in the experimental equipment. The comparison shows that the experimental and simulation data have the same characteristics and the method of flow pattern identification in this paper can be used to process the actual logging data. Barnea D , Shoham O , Taitel Y , et al. Flow pattern transition for gasMultiphase Flow . 1980 . 6 ( 3 ): 217 - 225 Dukler A E and Hubbard M D. A model for gas-liquid slug flow in horizontal and near horizontal tubes . Industrial & Engineering Chemistry Fundamentals . 1975 . 14 ( 4 ): 337 - 347 patterns in vertical and deviated wells . SPE Annual Technical Conference and Exhibition. 5-8 October 1997 , San Antonio, Texas. 102 - 109 (SPE 56108) Hoogendoorn C J. Gas-liquid flow in horizontal pipes . Chemical Engineering Science . 1959 . 9 : 205 - 217 Mahvash A and Ross A. Application of CHMMs to two-phase flow pattern identification . Engineering Application of Artificial Intelligence . 2008 . 21 ( 8 ): 1144 - 1152 Oshinowo T and Charles M E. Vertical two-phase flow Part I: Flow pattern correlations . The Canadian Journal of Chemical Engineering . 1974 . 52 ( 1 ): 25 -35 Journal of Canadian Petroleum Technology . 2000 . 39 ( 6 ): 43 - 55 Spedding P L and Chen J J. A simplified method of determining flow Journal of Multiphase Flow . 1981 . 7 ( 6 ): 729 - 731 Spedding P L and Spence D R. Flow regimes in two-phase gas liquid horizontal pipes . SPE Annual Technical Conference and Exhibition. Denver CO, 6-9 October 1996 . 363 - 375 ( SPE36609 ) Wang X X and Wu X L. Gas-water stratified flow patterns from electromagnetic tomography . Petroleum Science . 2009 . 6 ( 3 ): 254 - 258 Wu X L , Jing Y Q and Wu S Q. Electromagnetic imaging logging method in multiphase pipe flow . Chinese Journal of Geophysics . 1999 . 42 ( 4 ): 557 - 563 (in Chinese) Wu X L , Wang X X , Zhao Y W , et al. Flow imaging method of electromagnetic measurement in well logging . Science in China, Series D: Earth Sciences. 2008 . 38 ( S2 ): 161 - 165 (in Chinese) Development . 2000 . 27 ( 2 ): 79 - 82 (in Chinese) tomography in multiphase flow well logging . Chinese Journal of Geophysics . 2003 . 46 ( 6 ): 870 - 874 (in Chinese) Zhao Y W , Wu X L and Wang X X. Simulation of sensitivity field for electromagnetic tomography in multiphase flow well logging . Chinese Journal of Geophysics . 2007 . 50 ( 3 ): 811 - 816 (in Chinese) (Edited by Hao Jie)


This is a preview of a remote PDF: https://link.springer.com/content/pdf/10.1007%2Fs12182-012-0213-x.pdf

Zaibin Liu, Xiling Wu. Flow pattern identification in oil wells by electromagnetic image logging, Petroleum Science, 2012, 303-309, DOI: 10.1007/s12182-012-0213-x