Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
Results
Abstract Knowledge of liquid holdup and multiphase flow regimes in pipes is very important in pressure drop calculations and design of surface facilities. To predict the holdup, one has to define the flow regime at a given pressure, superficial gas and liquid velocities, temperature, and other flow characteristics. Several flow maps, regression and mechanistic models are available in literature to predict holdup and flow regime. However, some of these models do not predict the liquid holdup correctly. This paper presents two Artificial Neural Networks (ANN) models to identify the flow regime and calculate the liquid holdup in horizontal multiphase flow. These models are developed using experimental data - 199 data points - and utilizing three-layer back propagation neural networks. Superficial gas and liquid velocities, pressure, temperature and fluid properties are used as inputs to the network. The output of the first network is the flow regime, while the other network predicts the liquid holdup. One-half of the data was used to train the ANN models, one quarter to cross-validate the relationships established during the training process and the remaining quarter to test the models and evaluate their accuracy. The results show that the developed models provide better predictions and higher accuracy than the empirical correlations developed specifically for these data groups. The developed flow regime model predicts the flow regime correctly for more than 97% of the data points. The liquid holdup model outperforms the published models in terms of the lowest absolute average percent error (9.407), the lowest standard deviation (8.544) and the highest correlation coefficient (0.9896). Introduction Multiphase flow is defined as the concurrent flow of two or more phases, liquid, solid or gas, where the motion influences the interface between the phases. The flow regime or flow pattern is a qualitative description of the phase distribution in the pipe. There are three types of flow regimes in horizontal gas-liquid flow, namely, segregated, intermittent, and distributive flows. Segregated flow is further classified into stratified smooth, stratified wavy, and annular flow regimes. The intermittent flow regimes are slug flow and plug (elongated bubble) flow. Distributive flow regimes include bubble, and mist flows. Other investigators classified flow regimes in horizontal gas-liquid flow as: bubble flow in which gas bubbles tend to float at the top in the liquid; stratified flow in which the liquid flows along the bottom of the pipe and the gas flows on top; intermittent or slug flow in which large frothy slugs of liquid alternate with large gas pockets, and annular flow in which a liquid ring is attached to the pipe wall with gas blowing through. Usually, the layer at the bottom is very much thicker than the one at the top. The flow regimes in horizontal gas-liquid flow are illustrated in Fig. 1. Flow regime maps are used to predict flow regimes in horizontal gas-liquid flow. These maps are plots of superficial liquid velocity versus superficial gas velocity. One of the first maps used in the oil industry is that of Baker (1953). Ten years later, it was modified by Scott. Beggs and Brill, and Mandhane presented other flow pattern maps, which were constructed, based experimental data. Taitle and Dukler developed a theoretical model for the flow regime transitions in horizontal gas-liquid flow. Lately, other studies have been carried out for the prediction of specific transitions. Separate models have been developed for stratified flow, slug flow, annular flow, and dispersed bubble flow. An example of these flow pattern maps is given in Fig. 2.
- Research Report > New Finding (0.66)
- Research Report > Experimental Study (0.48)
Abstract Traditionally, transient testing is used to determine important reservoir properties, such as reservoir permeability, formation damage (skin) and reservoir boundaries. The data during transient testing are acquired by placing a pressure gauge downhole close to the reservoir. This requires a wireline unit and a vessel for offshore wells. Also, it requires the well to be shut in after producing for an extended time at a stable rate. Hence, It is difficult to do frequent transient testing for wells experiencing a drop in their productivity/injectivity due to logistics & economic constraints. In addition to the economic considerations in running such a test, the process is time consuming in acquiring the data and analyzing the results. This paper presents a new approach that uses historical surface performance data at surface, acquired during production mode, to estimate reservoir properties. This approach is based on establishing the trend of the performance of a well from production data. From that data, a regression analysis is used to obtain a pseudo productivity index that is function of rate. That parameter is corrected to the actual productivity index of the well using the limit function at shut-in condition. Newton-Raphson approach is then used to estimate the average reservoir permeability and total skin using from the actual productivity index based on pseudo-steady state equation. Classical methods can be then used to calculate the breakdown of the total skin. The application of this approach in the field revealed that it could give satisfactorily results that compare favorably with transient test results. It introduces a new and quick measure of evaluating well performance and completion efficiency. It can also show the variation of the IPR of a well, due to increased gas saturation, through the production life as previously presented by Muskat. This approach uses the available abundant data that is intended to serve other objective with computer applications to achieve the objective of transient testing. It can be utilized in cases where it is difficult to conduct transient testing. Introduction It's important to characterize formations in order to make rational decisions on reservoir management. Very often, it takes to know whether a positive skin exists to decide on stimulation job. Thus, in most cases, the primary objective of running pressure transient testing is to determine if a positive skin exists to remove the damage. That is especially true for wells after workover or in developed and mature fields where the permeability is predetermined or estimated from porosity transforms using well logs. As with a producing well, injection wells may also build "skin" over their operating life. This skin may result from the accumulation of foreign matter screened or filtered from the injected fluid. The PI/II of a well is conventionally measured, during transient testing, by placing a pressure gauge downhole to measure the pressure drawdown across the perforations that make the rate measured. The complexity of estimating the PI of the well from surface lies in the difficulty of estimating the difference of flowing pressure at surface and bottomhole. That can be estimated through the pressure traverse calculations. The difference in shut-in and flowing pressure, though in flowing wells at surface and bottomhole is not the same especially at high flow rtes. That is attributed to frictional pressure losses and the change of fluid gradient due to the increase of gas influx into the well. The density is involved in evaluating the total energy changes due to potential energy and kinetic energy changes.
- North America > United States (0.28)
- Asia > Middle East (0.28)
Abstract Production logs (PL) are often run to diagnose unwanted water production before seeking remedial actions. The diagnosis, involving accurate estimation of slip velocity during two-phase oil-water flow, is a precursor to taking successful corrective measures. This study entails finding the most reliable holdup or slip-velocity model by examining 78 production logs, run in Kuwait's Greater Burgan field. We observed that most models provided the desired level of accuracy when the downhole results were compared with those measured at surface. The notion of direct use of flow-pattern maps in PL interpretation is introduced in this work. Introduction As an oil field matures, a producer experiences coproduction of oil and water because of aquifer encroachment and/or water injection. Consequently, wellbore oil-water flow is encountered in many producing provinces throughout the world. Controlling water production is one of the major objectives of reservoir management. However, diagnosis of water flow in a layered completion must precede any remedial actions. Production logging allows identification of layers contributing most of the water. For accurate PL interpretation, a reliable holdup model is required. Therefore, any downhole remedial actions for water control are critically dependent upon the accuracy of the model used in PL analysis. Empirical models were proposed for water holdup or slip velocity calculations. Other investigators probed flow patterns. Flow-pattern-based models for holdup have also been advanced to explain the oil-water flow behavior. Yet another group of studies sought response of various PL sensors in large-diameter pipes, mimicking wellbores. Despite these theoretical and/or laboratory investigations, lack of field verification has always clouded a model's applicability. Therefore, in this study, we sought to find the most reliable holdup or slip-velocity model by examining 78 production logs, run in Kuwait's Greater Burgan field. Both in-situ and surface rate measurements formed the backbone of this investigation. These measurements show a wide range of water cut, 1 to 75%. We compared the widely used Choquette model with those of Hasan-Kabir, Flores et al., Nicolas-Witterholt, and Zuber-Findlay. These models are intended for vertical wellbores, the exception being the Hasan-Kabir model. The Zuber-Findlay model, originally developed for gas-oil flow, is used for comparative purposes only. Note that only the Flores et al. and Hasan-Kabir models are based upon flow-pattern maps. We also sought to make direct use of a flow pattern map to aid interpretation of production logs. This integration effort revealed elements that often go unexplored in many cases. Perhaps more important, one can learn about the mechanics of two-phase flow and the tools attempting to capture them. Comparison of Holdup Models Performance of Holdup Models. Various models are available for computing holdup in oil/water flow. We selected a few to compare and contrast their relative strengths. Also included is the Zuber-Findlay model to provide the benchmark for gas/liquid flow, rather than as a tool for analyzing the flow problem at hand. All the models are presented in the Appendix for easy reference. Production logs run in 78 flowing wells, with surface water cut ranging from 1 to 75%, provided the data in this comparative study. In all cases, water holdup comparison is made of that measured downhole, above all perforations, with that computed from a model. Volumetric rates measured at surface were used for holdup calculations. Model results are sensitive to both in-situ and surface measurements, however. We explore this sensitivity question in the Error Analysis section in detail.
- Asia > Middle East > Kuwait > Ahmadi Governorate > Arabian Basin > Widyan Basin > Greater Burgan Field > Wara Formation (0.99)
- Asia > Middle East > Kuwait > Ahmadi Governorate > Arabian Basin > Widyan Basin > Greater Burgan Field > Ratawi Formation (0.99)
- Asia > Middle East > Kuwait > Ahmadi Governorate > Arabian Basin > Widyan Basin > Greater Burgan Field > Mauddud Formation (0.99)
- (3 more...)
Abstract Production logs (PL) are traditionally interpreted by collecting depth-indexed information recorded by four major sensors; that is, pressure, temperature, density, and flow velocity. Although foot-by-foot data are gathered during flowing and shut-in passes, flow interpretation is normally done over selected discrete depth intervals, known as stations. This study explores data analysis on a continuous depth basis rather than that done at discrete stations. In so doing, we obtain unbiased in-situ flowmeter calibration, leading to improved understanding of wellbore flow behavior and production log interpretation. Introduction Production logs, when used correctly, have the potential for providing an array of valuable information throughout the life of a well. This array of information includes well performance evaluation, workover and completion evaluation, and reservoir performance monitoring. Conventional production logging as practiced today traces its origin to the introduction of temperature logs in 1937 and, subsequently, to the use of other sensors in 1965. Over the past three decades, continuous enhancement of sensor resolution paved the way for improved downhole diagnosis. One cornerstone to the early success lay in the ability to run multiple sensors in one run, although surface transmission of only one sensor response at a time was permitted in 1971. About a decade later, simultaneous transmission of signals from all sensors to the surface became possible. Advances in data gathering techniques and instrumentation continue to this day. At the same time, our understanding of mechanics of gas-liquid and liquid-liquid flows in conduits of various orientations has also matured over the years. Despite these advances, integration of mechanistic models with sensor responses has lagged behind. In other words, a holdup model may be used, but without an explicit usage of the flow-pattern map associated with it. We also observe that the traditional approach of station-based analysis, dealing with a few data points, do not often explain the anomalous flow behavior. Stated differently, one wonders whether certain sensor response is induced by the complex flow behavior, poor sensor resolution, logging speed, or a combination thereof. In this study, a rigorous depth-indexed analysis of flowmeter response is introduced. In the companion paper, we address how flow-pattern-based mechanistic models can be integrated with log responses to gain superior understanding of flow mechanics, leading to improved interpretation. Although the illustrative field examples involve oil-water flow from Kuwait's Greater Burgan field, the principle is intended for all flow situations. Data Analysis Conventional Approach. The ideal spinner response owing to fluid influx at different tool velocities is illustrated in the synthetic example A in Fig. 1. Between the perforation sets, the spinner speed remains constant. However, the spinner revolutions increase where fluid enters into or exits the wellbore. For each pass at a constant cable speed, a depth-dependent spinner profile or spinner response will develop. Ideally, the spinner response will be parallel to each other. From these data, one can interpret and quantify the amount of fluid produced, injected, or cross-flowed within the wellbore. Using the conventional approach, for each of the four depths A, B, C, and D as depicted in Fig. 1, a linear regression of the data yields a slope and an x-axis intercept. Fluid density, viscosity and the spinner bearing friction affect the slope and position of the line in the calibration plot.12 Nonetheless, theoretically speaking, the slope of the lines will be parallel to one another for constant fluid properties, as illustrated in Fig. 2 between the 100% and 0% flow depths. In this case, depth "D" represents the no-flow interval below all perforations or static well conditions, while depth "A" represents the maximum flow above all fluid entry from the formation.
- Asia > Middle East > Kuwait > Ahmadi Governorate > Arabian Basin > Widyan Basin > Greater Burgan Field > Wara Formation (0.99)
- Asia > Middle East > Kuwait > Ahmadi Governorate > Arabian Basin > Widyan Basin > Greater Burgan Field > Ratawi Formation (0.99)
- Asia > Middle East > Kuwait > Ahmadi Governorate > Arabian Basin > Widyan Basin > Greater Burgan Field > Mauddud Formation (0.99)
- (13 more...)