TY - JOUR
AU - Neupane,, Bhupati
AB - Abstract An accurate inversion of original reservoir resistivity is an important problem for waterflood development in oilfields in the middle-late development period. This paper describes the theoretical model of original resistivity recovery for a conglomerate reservoir established by petrophysical models, based on the stratigraphic model of reservoir vertical invasion of the conglomerate reservoir of an oilfield. Likewise two influencing factors of the resistivity change with a water-flooded reservoir were analyzed. The first one is the clay volume decrease due to an injected water wash argillaceous particle and the reservoir resistivity changes are influenced by it, and the other is to inject water to displace crude oil in the pore space leading to the increase of the water-bearing volume. Moreover the conductive ions of the injected water and the original formation water exchange and balance because of their salinity difference, and the reservoir resistivity changes are also influenced by them. Through the analysis of the above influential factors based on the fine identification of conglomerate lithologies the inversion models of three variables, including changes in the amount of clay, the resistivity of the irreducible water and the increase of the water bearing volume, were established by core analysis data, production performance and well logging curves information, and accurately recovered the original reservoir resistivity of the conglomerate. The original oil saturation of the reservoir was calculated according to multiple linear regression models. Finally, the produced index is defined as the difference of the original oil saturation and current oil saturation to the original oil saturation ratio, and it eliminates the effects of conglomerate lithologies and heterogeneity for the quantitative evaluation of flooded layers by the use of the principle of relative value. Compared with traditional flooding sensitive parameters which are oil saturation and water production rate, the interpretation accuracy of the production index can achieve 82%, provide technical support for the development programs determination and the well adjustment pattern in the second development of the oilfield. inversion method, conglomerate reservoir, original resistivity, petrophysical model, reservoir resistivity, original oil saturation ratio, well adjustment pattern 1. Introduction Due to the special sedimentary environment, the conglomerate reservoir has features such as complex and multi-changing lithology and serious heterogeneity, etc. These features can influence the evaluation precision of water-flooded layers. For example, the complex and multi-changing lithology can lead to the following conditions in which the rock skeleton seriously interferes with the well logging information of pore fluid, and the results of the saturation evaluation are wrong. In addition, the serious heterogeneity can lead to great changes in the calculation results of the water production rate for the same formation, and the water-flooded level cannot be accurately evaluated. Thus, in order to reduce the influence of the above two factors, accurately calculating the original oil saturation of the reservoir, and based on the same lithology condition, evaluating the flood level by using the relative value of the reduced bulk of crude oil, has become the key to improving the evaluation accuracy of the water flood layer in a conglomerate reservoir. Presently, there are mainly two ways to calculate the original oil saturation. One is a regional experience formula based on the physical property changing patterns of the water-flooded layer. This method uses the change of shale content and porosity after a reservoir waterflood to estimate the original oil saturation, and the established formula has obtained the certain effect evaluation (Geng et al2005, Okotie and Ikporo 2014). But this method has a lack of theoretical reasoning and is limited in terms of applicability. The other is using capillary pressure curves to calculate the original oil saturation. Heping Pan makes use of core analysis data to establish the transformational relationship between laboratory capillary pressure and formation capillary pressure, then changes the formation pressure into reservoir altitude and finds the formation of the original oil saturation according to the capillary pressure plate (Pan et al2000). But this method needs a large quantity of experimental data and can be accurate only in the condition of a special reservoir (Hamilton et al1998). On the other hand, most kinds of resistivity affecting factors and inversion models are systematically researched (Mclachlan et al1987, Argaud et al1989, Walid and Khaled 2014, Pardo and Torres-Verdin 2014), but the original resistivity recovery of a water-flooded reservoir lacks a theoretical model and rigid theoretical reasoning, and different kinds of resistivity affecting factors and mechanisms should be systematically used to reach a conclusion. Therefore, based on impact factor analysis of reservoir resistivity change, establishing a kind of original oil saturation recovery method which has a strong theory and wide applicability is the basis of the quantitative evaluation of the water flooded layer for complex oil and gas reservoirs. 2. Electrical model establishment of a water-flooded reservoir After water-flooding, the erosion of clay caused by the flood water and the ion exchange balance between the flood water and the original formation water seems to be affected by the conduction of the formation, and further changes the measured value of resistivity (Jakosky and Hopper 1937, Hill and Milburn 1956). The characteristics of the conglomerate reservoir in the Karamay oilfield include being based near the source, a multi-stream and rapid changing depositional environment leading to severe heterogeneity, and argilla mainly existing in the form of disperse clay. Thus, the petrophysical volume model of both water-flooded and unflooded reservoirs is established according to the longitudinal invasion pattern in the process of drilling (figure 1). Figure 1 indicates that the unflooded reservoir is kept in the original condition of the reservoir, so pore fluid property is decided by bound water and crude oil. The conductivity of the uninvaded formation is compounded with dispersed clay, a conglomerate framework and bound water with a parallel composition model. The material balance formula (equation (1)) explained that the effect on the pore structure and the pore fluid property is caused by the flood water in the water-flooded formation (Givens 1986). For the water-flooded reservoir, the conductivity of the original formation is compounded with dispersed clay, a conglomerate framework, bound water and flood water within a parallel composition model. The material balance formula is shown in equation (2) (Clavier et al1977): Vsh+Vma+ϕ=1ϕwi+ϕh=ϕ1 where Vsh, Vma—represent the content of the clay and conglomerate framework respectively (decimal); Figure 1. Open in new tabDownload slide Petrophysical volume model of reservoir with longitudinal invasion (a) unflooded reservoir's longitudinal invasion model (b) flooded reservoir's longitudinal invasion model. Figure 1. Open in new tabDownload slide Petrophysical volume model of reservoir with longitudinal invasion (a) unflooded reservoir's longitudinal invasion model (b) flooded reservoir's longitudinal invasion model. ϕwi, ϕh, ϕ—represent bound water porosity, oil-gas porosity and effective porosity (decimal). Vsh′+Vma+ϕ′=1ϕwi+ϕh′+ϕw=ϕ′2 where Vsh′ ⁠, Vma—represent the content of the dispersed clay and conglomerate framework after the flood formation respectively (decimal); ϕwi, ϕh′ ⁠, ϕw, ϕ′—represent bound water porosity, postflood oil-gas porosity, injected water porosity and postflood effective porosity (decimal). The research on the conductivity model of shaly sands goes through a development process from a single conductivity model to a multi-medium common conductive model. At present the conductive mechanism and conductivity model establishment have been researched thoroughly by experts, with many innovations achieved in theory and technology. In 1942, Archie studied the resistivity law of pure sandstone with 100% water saturation, and his research results became the basis of all conductivity models from then on (Archie 1942). In the following decades, with the understanding of the conductive mechanism of shaly sands becoming deeper and the evaluation precision on oil saturation of a complex reservoir improving, the following experts such as Waxman, Clavier and Silva et al have established, respectively, three kinds of dual-electric layer resistivity models which are W-S, D-W and S-B (Waxman and Smits 1968, Clavier et al1984, Silva and Bassioni 1985), and also have been used widely in well logging interpretation. In recent years, based on Givens's model and Argaud's parallel conduction theory (Givens 1986, Argaud 1989), Crane has extended Archie's law which has general meaning according to experimental observation and a variety of discussions (Crane 1990), so that the conductive model has generalization. In conclusion, the theoretical basis of all conductivity models is the same, namely the final conductivity result is contributed by the parallel conduction of every conductivity part in the reservoir according to different shale types and pore structures. For example, Crane's theory considers that the total conductivity of rock comes from the parallel results of four conductivity parts which are, respectively, large pores, micro pores, conductive minerals and rough surfaces, etc (Crane 1990). According to the petrophysical volume model in the condition of both water-flooded and unflooded reservoirs and the theoretical basis of reservoir parallel conductance (Waxman and Smits 1968), taking the empirical equation into account, the conductivity models of the undisturbed formation of water-flooded and unflooded reservoirs are respectively established, and equation (3) is the resistivity model of the unflooded undisturbed formation (Argaud 1989, Crane 1990): 1Rt=VshRsh+VmaRma+ϕwiRwi3 where Vsh, Vma, ϕwi—represent the content of dispersed clay, conglomerate framework and the bound water porosity respectively (decimal); Rt, Rsh, Rma, Rwi—represent the resistivity of the undisturbed formation, dispersed clay, conglomerate framework and the bound water (Ω · m). Equation (4) is the resistivity model of the water-flooded undisturbed formation, which is supposed to be different from the unflooded formation in two ways (Waxman and Smits 1968, Argaud 1989, Crane 1990). Firstly, because of the injection water washing, the content of the dispersed clay can be changed. Secondly, the oil and gas in the pore volume is driven by the flood water which causes the changes in the fluid property. Both factors affect the change of the resisitivity of formation. 1Rt′=Vsh′Rsh+VmaRma+ϕwiRwi+ϕwRw4 After the reservoir is flooded, because of the different salinity between the undisturbed formation water and the bound water, the ions can exchange and balance which further affects the water filled porosity and formation water salinity, and finally changes the resistivity contribution value of the pore fluid property for the whole reservoir resistivity (Amidu and Dunbar 2008, Chanh et al2011). The formation resistivity is decided by the resistivity of bound water in the undisturbed formation and the resistivity of the injection water during the oil field development, together with the condition that the water saturation value is composed of the bound water saturation and the injection water saturation. Therefore, the resistivity of the formation water can be used to replace the common contributions of reservoir resistivity which are the bound water and injection water (Tan et al2010b). As for equation (4): 1Rt′=Vsh′Rsh+VmaRma+ϕwsRws5 where Vsh′ ⁠, Vma, ϕws—represents the content of the dispersed clay, conglomerate framework and formation water porosity after the flood formation respectively (decimal); Rt′ ⁠, Rsh, Rma, Rws—represent the resistivity of the flooded undisturbed formation, dispersed clay, conglomerate framework and formation water (Ω · m). The following equation is acquired by subtracting the flooded formation resistivity model with the unflooded formation resistivity model: 1Rt−1Rt′=Vsh−Vsh′Rsh+ϕwiRwi−ϕwsRws6 According to equation (6), we can see that the flooded formation resistivity is decided by two factors. One is the decreased content of the dispersed clay caused by the wash of injection water, which diminishes the contribution of argillaceous attached conduction to the whole reservoir conductivity. As a result, the resistivity represents an exponential increase with the decreased content of the clay. On the other hand, when injection water enters the reservoir and replaces the oil and gas in porosity, then the water saturation increase and formation conductivity ability seems to improve. The resistivity curve represents the decreased trend. Besides, because of the salinity difference between the injection water and formation water the ions can exchange and balance. As a result, the desalination process is the main cause for the increase of the resistivity curve. During the flood development in the reservoir, both factors commonly exist and affect the changing of the reservoir resistivity, which formulates the resistivity as an irregular ‘U’ trend. Therefore, for the recovery of flood formation resistivity, based on the current resistivity and together with the geologic logging data, petrophysics and dynamic production etc, evaluating the results of two influential factors and calculating reservoir resistivity accurately finally provides the technical support for the evaluation of high water flooding formation. 3. Analysis and inversion of resistivity models 3.1. Lithology identification of a conglomerate reservoir The conglomerate reservoir in the Karamay oilfield belongs to a positive cycle piedmont fluvial facies. The depositional features, which include being based near the source, multi water systems and a multivariant depositional environment, cause the complexity and multi-changeability of reservoir lithologies and oiliness (figure 2). To improve the accuracy of inversion, the recovery of the original resistivity must be based on the subdivision of petrophysics. In consideration of the complex lithology of a conglomerate reservoir, the traditional mathematical statistics method cannot accurately calibrate the nonlinearity mapping relationship between the logging data and conglomerate lithologies. Instead, the data mining method, which has the ability of self-organization, self-learning, reasoning and a high-accuracy nonlinear modeling, can well resolve this problem (Tan et al2010a). Data mining is a process of extracting abstract, tacit, undiscovered but useful information and a knowledge process model from vast, partial, noisy, fuzzy and random data (Usama et al1996). Its tasks are classified into description and prediction. The description is making a summary through the deduction of the potential relationship in the data. The prediction is making forecasts through the deduction of the current data; the modeling methods for forecasting come from machine learning, artificial intelligence, pattern identification, statistics and other research areas (Witten and Frank 2005, Han and Kamber 2006). The main modeling methods include a neural network, support vector machines, a Bayesian network, and decision tree methods etc. Figure 2. Open in new tabDownload slide The change of lithology and oil-bearing properties of a conglomerate reservoir. (1) Gray mudstone, medium; (2) Shale with conglomeratic sandstone, medium, oil patch; (3) Conglomeratic gritstone, medium, oil patch; (4) Conglomeratic gritstone, medium, oil immersed; (5) Conglomeratic gritstone, loose, oil rich; (6) Glutenite, medium, oil patch; (7) Glutenite, loose, oil patch; (8) Conglomerate, tight, oil trace; (9) Conglomerate, medium, oil patch; (10) Conglomerate, medium, oil immersed. Figure 2. Open in new tabDownload slide The change of lithology and oil-bearing properties of a conglomerate reservoir. (1) Gray mudstone, medium; (2) Shale with conglomeratic sandstone, medium, oil patch; (3) Conglomeratic gritstone, medium, oil patch; (4) Conglomeratic gritstone, medium, oil immersed; (5) Conglomeratic gritstone, loose, oil rich; (6) Glutenite, medium, oil patch; (7) Glutenite, loose, oil patch; (8) Conglomerate, tight, oil trace; (9) Conglomerate, medium, oil patch; (10) Conglomerate, medium, oil immersed. The identification model establishment of conglomerate lithology must consider two factors by using data mining methods. On the one hand, the modeling process of the mining algorithm must be the ‘white box’ model which can know how the grader works as well as the relative importance of different kinds of lithological parameters (Han and Kamber 2006), in order to deeply analyse sensitive parameters and the classification process of the lithological model to give some sound guidance in geological and geophysical research for the next step work. On the other hand, the selected model performance must be better, that is the highest repeated cross-validation accuracy on data volume. By analysis of the lithology identification accuracy of several algorithms (figure 3), the accuracy of the decision method C5.0 is the highest, which can reach 96.97%. Based on the above analysis the decision method C5.0 is selected as the data mining algorithm of conglomerate lithology identification. The decision tree is a tree structure which is similar to a flow chart. It starts to test the data samples from the root node and the data samples are divided into different subsets according to different results, and each subset of data samples forms a minor node. Each branch represents a test output, and each leaf node represents a category or category distribution. The basic algorithm of the decision tree is a greedy algorithm, which builds a decision tree by recursion from the top to the bottom and mainly includes two processes of construction and classification. The data mining process is a continuous cycle of optimization, and data mining of the decision tree includes six steps such as data preprocessing, data selection, data transformation, data mining, model evaluation and knowledge discovery (Tan et al2010a). Figure 3. Open in new tabDownload slide The repeated cross-validation accuracy of different data mining algorithms. Figure 3. Open in new tabDownload slide The repeated cross-validation accuracy of different data mining algorithms. The sealing core wells are selected as the lithology identification data of the conglomerate reservoir to arrange and process. Firstly, the lithological profile of sealing core wells are established according to the theory of sequence stratigraphy, then different log response characteristic values are read to establish the correspondence relationship between conglomerate lithology and logging data. The lithologies acquired by data mining can be classified into eight types which are: conglomerate, sandy conglomerate, glutenite, conglomeratic gritstone, fine sandstone, conglomeratic mudstone, silty mudstone and mudstone. The total seven parameters, which are the undisturbed formation resistivity, gamma ray, spontaneous potential (SP), caliper, neutron porosity, acoustic travel time and compensated density logging, are selected as a mining field to indicate lithology. The data mining result of the conglomerate lithology is shown in figure 4 based on the decision tree. According to the figure, the decision tree method firstly selects three sensitive parameters from seven mining fields, and then makes use of the tree structure to recognize lithology from the top to the bottom. Each branch of the tree represents a kind of lithology recognition rule and the branch point represents the attribute parameter and numerical intervals of the recognition rule. The decision tree model shows clearly the parameter combination and detail process of the lithology recognition, with a comprehensive identification accuracy rate of above 90%. Figure 4. Open in new tabDownload slide Identification model of conglomerate lithology based on the decision tree method. Figure 4. Open in new tabDownload slide Identification model of conglomerate lithology based on the decision tree method. The main oil-layer lithology of the conglomerate reservoir in the Karamay oilfield includes conglomeratic gritstone, glutenite and conglomerate, thus the analysis and calculation of different factors during the original reservoir resistivity inversion are based on the subdivision of three kinds of lithology, by which to eliminate the influence of lithology and improve the accuracy of inversion. 3.2. Analysis and inversion of shale factors The injected water washing to mud is mainly affected by two factors such as the injected water volume and reservoir pore structure. To research the rule of the clay changing amount with increasing injected water, samples which have both a different lithology and the same lithology but different pore structures are used to make displacement experiments in the laboratory. Figure 5 shows the changing trend curves, representing the changing amount of clay with water saturation increasing for the conglomerate reservoir with the lithology of conglomeratic gritstone, glutenite and conglomerate. As shown in the figure, the changing amount of clay in three kinds of lithology all increase with the increase of injected water. Compared with the low flood level condition, the trend slowed significantly while the injected water reached middle-high level although the relationship between the two factors is also positive. The main principle is that the trend presents a two-stage change feature which has a breakpoint. In addition, three kinds of lithologies have this changing trend. Figure 5. Open in new tabDownload slide Crossplot of clay changing amount and water saturation for samples with different lithologies. Figure 5. Open in new tabDownload slide Crossplot of clay changing amount and water saturation for samples with different lithologies. Figure 6 is the conglomeratic gritstone's changing trend curves which represent the clay changing amount of different pore structure samples with the water saturation increasing, and the reservoir quality index is selected to scale pore structure. From the figure we can see that the clay changing amount of different pore structure samples all increase with an increase in injected water. Because the samples with good pore structure have features such as low heterogeneity and uniform clay distribution, the clay of these samples is washed strongly by injected water and the clay changing amount is higher than the samples with poor pore structure for the same water injection. Therefore, the total trend of the clay changing amount of four samples also presents a two-stage change feature and has a breakpoint value when the samples reach a middle-high water flood level. Figure 6. Open in new tabDownload slide Crossplot of clay changing amount and water saturation for samples with different pore structures. Figure 6. Open in new tabDownload slide Crossplot of clay changing amount and water saturation for samples with different pore structures. From the above research results, we found that the petrophysical implication of the breakpoint value is the decreasing amount of clay when the injected water forms a main flow network in the reservoir pore space. The pore and throat in the rock formed a whole flow system. Because of the influences of capillary pull and aeolotropy the fluid in some pore paths does not flow, where in some pore paths the flow is restricted and in some others it flows in advance. Clay distributes all around these pore spaces. During the water driven process, the clay distributed around the advanced flowing pore spaces is first washed. With the injected water amount increasing, the changing amount of clay represents a single tendency. When the driven water reaches a certain extent (middle-high flooded level), water in the main flowing network becomes a continuous phase and further forms a high transmissibility path. Then, with the injected water amount increasing, the washing affect on the clay is weakened. Thus, the changing amount of clay shows the slower trend. From the above analysis, the changing models of the clay volume make use of water saturation and reservoir quality index to carry out a multiple fit based on different lithologies, and the changing trend of different flooded periods must be taken into account. Equations (7)and (8) are conglomeratic gritstone's calculation formulas of the clay changing amount in a low-middle flooded and middle-high flooded period, and the sample number is 113. Other lithological parameters and factors in a different flooded period are shown in table 1. ΔVsh=Vsh−Vsh′=0.1131*Sw+0.000173*K/ϕ′−0.0309R2=0.94937 ΔVsh=Vsh−Vsh′=0.032*Sw+0.000245*K/ϕ′+0.016R2=0.91278 where ΔVsh—clay decreasing content between an unflooded and flooded reservoir (decimal); Sw—water saturation of a flooded reservoir (decimal); ϕ′—effective porosity of a flooded reservoir (decimal); Table 1. Model parameters of clay changing content for different lithologies. Model . Clay changing content model . Low flooded period . Middle-high flooded period . Parameters . Sw . K/ϕ′ . Constant . Sw . K/ϕ′ . Constant . Conglomeratic gritstone 0.113 1 0.000 173 -  0.030 9 0.032 0.000 245 0.016 Glutenite 0.109 5 0.000 164 -  0.034 2 0.031 7 0.000 236 0.014 8 Conglomerate 0.102 7 0.000 153 -  0.040 2 0.029 5 0.000 214 0.012 7 Model . Clay changing content model . Low flooded period . Middle-high flooded period . Parameters . Sw . K/ϕ′ . Constant . Sw . K/ϕ′ . Constant . Conglomeratic gritstone 0.113 1 0.000 173 -  0.030 9 0.032 0.000 245 0.016 Glutenite 0.109 5 0.000 164 -  0.034 2 0.031 7 0.000 236 0.014 8 Conglomerate 0.102 7 0.000 153 -  0.040 2 0.029 5 0.000 214 0.012 7 Open in new tab Table 1. Model parameters of clay changing content for different lithologies. Model . Clay changing content model . Low flooded period . Middle-high flooded period . Parameters . Sw . K/ϕ′ . Constant . Sw . K/ϕ′ . Constant . Conglomeratic gritstone 0.113 1 0.000 173 -  0.030 9 0.032 0.000 245 0.016 Glutenite 0.109 5 0.000 164 -  0.034 2 0.031 7 0.000 236 0.014 8 Conglomerate 0.102 7 0.000 153 -  0.040 2 0.029 5 0.000 214 0.012 7 Model . Clay changing content model . Low flooded period . Middle-high flooded period . Parameters . Sw . K/ϕ′ . Constant . Sw . K/ϕ′ . Constant . Conglomeratic gritstone 0.113 1 0.000 173 -  0.030 9 0.032 0.000 245 0.016 Glutenite 0.109 5 0.000 164 -  0.034 2 0.031 7 0.000 236 0.014 8 Conglomerate 0.102 7 0.000 153 -  0.040 2 0.029 5 0.000 214 0.012 7 Open in new tab K—permeability of flooded reservoir, ×10-3μm2. The conductivity of clay mineral is mainly controlled by the cation exchange and absorption (Winsaure and McCardell 1956, Hamdi and Srasra 2013). Because the surface of a clay particle charges negative electricity, when it meets with water solution the clay particle absorbs the cation in it to keep ionic equilibrium (Cosenza et al2015). As a result, the structure of the double layer including the adsorbed layer and diffused layer are formed on the clay particle surface. Discovered through research, the conductivity of clay is mainly decided by the diffused layer, since the ions in the adsorbed layer cannot move and do not have a contribution to conductivity. Thus, the cation exchange capacity (CEC) is usually used to quantitatively scale the conductivity of clay mineral. To research the relationship between clay resistivity (Rsh) and CEC, the clay mineral samples are selected and their CEC value is also measured by using the method of luminosity analysis to establish the relationship between clay resistivity and the CEC of the conglomerate reservoir in the Karamay oilfield (figure 7). From the figure we can see that the CEC variation range of the clay mineral in this research area is between 28 ~ 75mmol/100 g. With the CEC improving, the clay resistivity decreases, and there is a good correlation between these two factors. Therefore, the CEC of clay mineral can be used to calculate its resistivity. Equation (9) is established on the basis of 36 samples: Rsh=97.153*e−0.0613*CECR2=0.94139 where Rsh—clay resistivity (Ω · m); CEC (mmol/100 g). Figure 7. Open in new tabDownload slide Crossplot of clay resistivity and cation exchange capacity (CEC). Figure 7. Open in new tabDownload slide Crossplot of clay resistivity and cation exchange capacity (CEC). The above analysis indicates that the clay resistivity is relevant to its CEC which is fundamentally decided by the type and content of the clay minerals (Vladimir et al2007). In order to calculate CEC accurately, the type of clay minerals should be analyzed first in the Karamay oilfield. The conglomerate reservoir in the Karamay oilfield belongs to a set normal cycle deposition of piedmont alluvial fan. Some aluminum silicate minerals in the land source, such as feldspar and glimmer, become a great quantity of glimmerton when they degrade and weather in a warm and wet geologic climate. Augite, hornblende and blackmica in detrital particles become montmorillonite when suffering from half way chemical erosion. However, the montmorillonite is not seen in the research area, and the main reason is that the clay minerals change into autogenetic clay minerals in the later stage of catagenesis of oil and gas reservoir because of the difference of the formation water property (Naidja et al1995, Annunziata et al1999, Tan et al2009). The montmorillonite belong to a kind of swelling clay mineral and its forming condition is pH = 6 ~ 8 range. With an increase in the formation alkalinity, the montmorillonite become dehydrated and change into illite (pH = 7 ~ 8) or chlorite (pH = 8 ~ 9) which are non-expansion clay minerals. Because the main water type in the region of interest is NaHCO3, where the average pH value is 7.8, the montmorillonite formed with the land source detrital material change into illite or an illite–montmorillonite mixture layer under an alkaline environment. On the other hand, the aluminum silicate minerals are leached by an acidic medium to form kaolinite. However, the research region belongs to an alkalinity environment, which made a little amount of kaolinite that could only be seen in some specific wells. From all the above analysis, the main types of clay minerals in this area are illite and an illite–montmorillonite mixture layer together with a little of kaolinite. On the basis of ascertaining the types of clay minerals in the region of interest, the relevant calculation models of the amount of different clay minerals need to be established. According to the laboratory analysis data combined with the logging response features of clay minerals, the basic features of different kinds of clay minerals in the Karamay conglomerate oilfield are concluded and are shown in table 2. From the table we can find that there is a big difference in terms of gamma, density, neutron and acoustic travel time value among three kinds of clay minerals. Thus, taking the above four logging curves of a conglomerate reservoir, the calculation models of different clay mineral amounts are established according to the multiple factor linear regression method. This method is the key point with which to establish prediction models through the correlation analysis between two or more than two independent variables and one dependent variable, and process data samples to make a decision and prediction. When the relationship between the independent variable and dependent variable is linear, it is named as the multiple factor linear regression method. At present, the ordinary least squares are generally selected to find the best function by use of the quadratic sum to minimize the error. In this paper, the well-known international statistical and analysis software, SPSS, is used to analyse multiple factors of mineral content calculation and establish corresponding calculation models. Table 2. Characteristic parameters and content calculation models of different clay minerals in the Karamay oilfield. Type of clay mineral . GR/API . DEN/ (g cm-3) . AC/ (μs m-1) . CNL/ % . CEC/ (mmol/100 g) . Calculation model of clay mineral content . Vsh . ϕS/ϕN . ϕS/ϕD . ϕN/ϕD . Constant . I  +  M mixture layer 170 ~ 220 2.48 283.85 27 60 ~ 80 (72) VI + M=1–VI–VK Illite 250 ~ 300 2.84 172.41 22 10 ~ 30 (23) 0.210 4 2.348 2 -  1.753 8 1.247 2 -  0.976 3 Kaolinite 90 ~ 130 2.61 217.39 38 3 ~ 15 (8.5) 0.084 6 1.296 4 -  1.583 9 1.183 6 -  0.639 2 Type of clay mineral . GR/API . DEN/ (g cm-3) . AC/ (μs m-1) . CNL/ % . CEC/ (mmol/100 g) . Calculation model of clay mineral content . Vsh . ϕS/ϕN . ϕS/ϕD . ϕN/ϕD . Constant . I  +  M mixture layer 170 ~ 220 2.48 283.85 27 60 ~ 80 (72) VI + M=1–VI–VK Illite 250 ~ 300 2.84 172.41 22 10 ~ 30 (23) 0.210 4 2.348 2 -  1.753 8 1.247 2 -  0.976 3 Kaolinite 90 ~ 130 2.61 217.39 38 3 ~ 15 (8.5) 0.084 6 1.296 4 -  1.583 9 1.183 6 -  0.639 2 Open in new tab Table 2. Characteristic parameters and content calculation models of different clay minerals in the Karamay oilfield. Type of clay mineral . GR/API . DEN/ (g cm-3) . AC/ (μs m-1) . CNL/ % . CEC/ (mmol/100 g) . Calculation model of clay mineral content . Vsh . ϕS/ϕN . ϕS/ϕD . ϕN/ϕD . Constant . I  +  M mixture layer 170 ~ 220 2.48 283.85 27 60 ~ 80 (72) VI + M=1–VI–VK Illite 250 ~ 300 2.84 172.41 22 10 ~ 30 (23) 0.210 4 2.348 2 -  1.753 8 1.247 2 -  0.976 3 Kaolinite 90 ~ 130 2.61 217.39 38 3 ~ 15 (8.5) 0.084 6 1.296 4 -  1.583 9 1.183 6 -  0.639 2 Type of clay mineral . GR/API . DEN/ (g cm-3) . AC/ (μs m-1) . CNL/ % . CEC/ (mmol/100 g) . Calculation model of clay mineral content . Vsh . ϕS/ϕN . ϕS/ϕD . ϕN/ϕD . Constant . I  +  M mixture layer 170 ~ 220 2.48 283.85 27 60 ~ 80 (72) VI + M=1–VI–VK Illite 250 ~ 300 2.84 172.41 22 10 ~ 30 (23) 0.210 4 2.348 2 -  1.753 8 1.247 2 -  0.976 3 Kaolinite 90 ~ 130 2.61 217.39 38 3 ~ 15 (8.5) 0.084 6 1.296 4 -  1.583 9 1.183 6 -  0.639 2 Open in new tab In order to reduce the inaccuracy caused by data property differences in the linear regression process, firstly the original data are normalized to acquire four factors such as neutrion porosity (ϕN), acoustic porosity (ϕS), density porosity (ϕD) and shale content (Vsh). Then three kinds of porosity values are handled dimen-sionlessly to construct three variable parameters such as ϕS / ϕD, ϕS / ϕN and ϕN / ϕD. Finally, the calculation models of the clay mineral amount are established and combined with the shale amount. Equation (10) is the formula of the illite content and the sample number is 136. The calculation parameters and factors of the illite–montmorillontie mixture layer and kaolinite amount appear in table 2. VI=0.2104*Vsh+2.3482*ϕS/ϕN−1.7538*ϕS/ϕD+1.2472*ϕN/ϕD−0.9763R2=0.874610 On the basis of the accurate calculation of the clay mineral amount, the petrophysical volume model is used to calculate the total CEC of the dispersed clay, as shown in fomula (11). Based on the results of the research above, the change amount of the clay minerals of the flooded reservoir and the clay resistivity can be calculated by abundant logging data. Therefore, the contribution of the clay factors on formation resistivity can be recovered effectively. CEC=CECI*VI+CECK*VK+CECI+M*VI+M11 where CEC, CECI, CECK, CECI+M—represent the total clay CEC, illite CEC, kaolinite CEC and illite–montmorillonite mixture layer CEC, respectively (mmol/100 g); VI, VK, VI+M—represent the volume of illite, kaolinite and illite–montmorillonite mixture layer respectively (decimal). 3.3. Analysis and inversion of bound water factors The bound water saturation in the original reservoir is decided by lithology, porosity, permeability, median grain diameter and pore structure (Asquith 1990, Scott and Dani 2005). Thus, in order to obain the accurate original resistivity inversion of a flooded reservoir, the above various factors should be researched, and then calculated models with high accuracy and wide applicability should be established. Due to the special sediment environment the complicated lithology is the biggest feature for the conglomerate reservoir, and different lithologies have big differences in terms of physical property, seepage property and pore structure. The aeolotropism of three lithologies which are conglomeratic gritstone, glutenite and conglomerate gradually enhance and lead to bound water saturation increasing. Thus, it is very necessary to establish bound water saturation models in a conglomerate reservoir on the basis of lithology division. For two factors of porosity and permeability, we can see that the bound water saturation gradually decreases with the values of porosity and permeability increasing according to the sealing core well data, and it is linearly relevant to porosity and is index relevant to permeability. The influence on bound water saturation caused by the pore structure is decided by the peak value of the micro pore throat. The higher the peak value, the higher the bound water saturation. The influence on bound water saturation caused by a medium grain diameter displays mainly two aspects. Firstly, the size of the grain diameter affects the size of the throat; when the size of the grain diameter becomes smaller the capillary pressure can become higher and the bound water saturation also becomes higher. Secondly, the size of the grain diameter decides the contact area of the rock and water. In the condition of the same volume, the little size of the grain diameter is the bigger contact area and the membrane bound water also is more. Because the reservoir quality index can not only represent the physical features of the reservoir, but also can scale pore structure, the bound water saturation models of the conglomerate reservoir can be established by using a reservoir quality index and medium grain diameter based on lithology division. Equation (12) is the calculation formula of bound water saturation for conglomeratic gritstone. The sample number is 87 and the other lithology models are shown in table 3. Swi=−0.00306*K/ϕ′−0.02094*Md+0.2462R2=0.90812 where Swi—bound water saturation (decimal); Md—medium grain diameter (mm); ϕ′—effective porosity of flooded reservoir (decimal); K—permeability of flooded reservoir (×10-3μm2). Table 3. Saturation model parameters of different lithologies. Model . Bound water saturation model . Water saturation model . Parameters . K/ϕ′ . Md . Constant . lgRt′ . lgϕ′ . ΔSP . Constant . Conglomeratic gritstone -  0.003 06 -  0.020 94 0.246 2 -  0.197 5 -  0.767 4 0.028 3 2.983 6 Glutenite -  0.002 84 -  0.018 47 0.228 3 -  0.199 8 -  0.786 2 0.132 8 2.998 6 Conglomerate -  0.002 72 -  0.017 59 0.263 7 -  0.221 9 -  0.199 2 0.028 7 2.251 6 Model . Bound water saturation model . Water saturation model . Parameters . K/ϕ′ . Md . Constant . lgRt′ . lgϕ′ . ΔSP . Constant . Conglomeratic gritstone -  0.003 06 -  0.020 94 0.246 2 -  0.197 5 -  0.767 4 0.028 3 2.983 6 Glutenite -  0.002 84 -  0.018 47 0.228 3 -  0.199 8 -  0.786 2 0.132 8 2.998 6 Conglomerate -  0.002 72 -  0.017 59 0.263 7 -  0.221 9 -  0.199 2 0.028 7 2.251 6 Open in new tab Table 3. Saturation model parameters of different lithologies. Model . Bound water saturation model . Water saturation model . Parameters . K/ϕ′ . Md . Constant . lgRt′ . lgϕ′ . ΔSP . Constant . Conglomeratic gritstone -  0.003 06 -  0.020 94 0.246 2 -  0.197 5 -  0.767 4 0.028 3 2.983 6 Glutenite -  0.002 84 -  0.018 47 0.228 3 -  0.199 8 -  0.786 2 0.132 8 2.998 6 Conglomerate -  0.002 72 -  0.017 59 0.263 7 -  0.221 9 -  0.199 2 0.028 7 2.251 6 Model . Bound water saturation model . Water saturation model . Parameters . K/ϕ′ . Md . Constant . lgRt′ . lgϕ′ . ΔSP . Constant . Conglomeratic gritstone -  0.003 06 -  0.020 94 0.246 2 -  0.197 5 -  0.767 4 0.028 3 2.983 6 Glutenite -  0.002 84 -  0.018 47 0.228 3 -  0.199 8 -  0.786 2 0.132 8 2.998 6 Conglomerate -  0.002 72 -  0.017 59 0.263 7 -  0.221 9 -  0.199 2 0.028 7 2.251 6 Open in new tab The formation of water salinity and temperature are the two main factors which affect formation resistivity. According to the water experiment and formation temperature data in the region of interest, the crossplot of the formation water resistivity and salinity under different reservoir temperatures is established (figure 8). From the figure we can observe that the formation water resistivity gradually decreases with its salinity increasing in the condition of the same formation temperature, and under the same formation salinity the formation water resistivity also displays the decreased trend with formation temperature increasing. Thus, the bound water resistivity (Rwi) can be determined by using figure 8 according to historic salinity data in this area. The bound water salinity of the conglomerate reservoir in the Karamay oilfield is 6158 ppm in its original reservoir condition, and combined with the geothermal gradient (3.2 °C/100 m) the resistivity of the bound water can be calculated in different depths. From all the above, the porosity models of bound water (ϕwi = Swi*ϕ′) can be established on the basis of lithology division and the resistivity of bound water can also be acquired by the crossplot. As a result, the resistivity contribution of bound water to a flooded reservoir can be recovered effectively. Figure 8. Open in new tabDownload slide Crossplot of formation water resistivity and salinity under different reservoir temperatures. (Note: the red points are the oil saturation value of corrected core analysis and the blue curves are the value of the model calculation.) Figure 8. Open in new tabDownload slide Crossplot of formation water resistivity and salinity under different reservoir temperatures. (Note: the red points are the oil saturation value of corrected core analysis and the blue curves are the value of the model calculation.) 3.4. Analysis and inversion of formation water factors The injected water displaces crude oil in reservoir pores when entering the reservoir which makes the water saturation increase. On the other hand, the difference of salinity between the injected water and formation water can change the resistivity of the original formation water. The above two factors have the most important influence on the resistivity of the flooded reservoir. Thus, the prerequisite of inversing the original reservoir resistivity accurately is to establish a kind of saturation calculation model which has a good applicability to a conglomerate reservoir. Through analyzing the accuracy of the current various saturation models (Archie formula, correction Archie formula, Simandoux formula, three-water conductive formula and multiple linear regression formula, etc) based on core data and logging curves, finally the multiple linear regression model is selected as the water saturation calculation formula in a conglomerate reservoir (figure 9). This model is based on the deformation of the Archie formula, which uses the resistivity of the original formation, effective porosity and the normalizing value of SP to regress the water saturation formula. Nearly all the water saturation models deal with the problem of formation water resistivity calculation. However, the factors affecting the resistivity of the formation water are various and complicated, which make accurate calculation difficult. Then the multiple linear regression model avoids this problem, and it uses the continuous normalizing value of SP to replace the resistivity of formation water in order to improve the model accuracy. Equation (13) is the water saturation calculation formula of conglomeratic gritstone. The sample number is 125 and other lithology models are shown in table 3. lgSw=−0.1985lg Rt′−0.7674 lgϕ′+0.0283 ΔSP+2.9836R2=0.84713 where Sw—water saturation (%); Rt′—undisturbed resistivity of flooded reservoir (Ω · m); ϕ′—effective porosity of flooded reservoir (%); ΔSP—normalized value of SP. Figure 9. Open in new tabDownload slide The comparison chart of the calculation results for different saturation models. Figure 9. Open in new tabDownload slide The comparison chart of the calculation results for different saturation models. There are mainly two factors which influence the formation water salinity. One is the property of the injected water, and the other is the dissolving effect on dissoluble ions in the rock skeleton caused by the formation water. When the formations are flooded, both factors exist at the same time and affect the formation water salinity. The trend figure of the formation water salinity which changes along with the development time is established according to the formation water analysis data over the years (figure 10). The overall trend presents a ‘w’ form and it can be divided into four stages: A, B, C and D. Stage A is the desalination stage, in which the formation water salinity gradually decreases with the injection of fresh water in the beginning of the development. Stage B is the reinjection of the formation water and dissolution process. When the formation water salinity equals the injected water salinity the formation water salinity no longer decreases. In this case, the salinity of the formation water increases caused by the dissolution of soluble ions in the rock skeleton and the reinjection of the formation water. Stage C is the stage that occured during the time of the 1990s, in which the Karamay oilfield came into the key stage of adjustment and many developing wells were put into production. As a result, the desalination effect takes the predominant position and the salinity of the formation water decreased again. Stage D is the produced water reinjection stage, in which the salinity increases. Formation water resistivity of different years can be calculated accurately according to figure 8. Finally, the porosity models of the formation water can be established on the basis of lithology division (ϕws = Sw*ϕ′) and the formation water resistivity can be calculated accurately. Therefore, the resistivity contribution to the flooded layers caused by the changing of the formation water can be effectively recovered. Figure 10. Open in new tabDownload slide Changeable trend of salinity with production years. Figure 10. Open in new tabDownload slide Changeable trend of salinity with production years. 4. Calculation and application of original oil saturation 4.1. Calculation of original oil saturation Because of the special depositional environment the conglomerate reservoir has some special features such as complicated lithology and severe heterogeneity. The two traditional flooded formation evaluation factors, which are water saturation and water production rate, have a bad applicability in the quantitative interpretation of the conglomerate reservoir, and it is impossible to reach the accuracy in the real application. Therefore, the sensitive parameter construction of the water flooded layer for the conglomerate reservoir must eliminate the effects of the complicated lithology and severe heterogeneity. Considering the above two factors, the difference between the original oil saturation and current oil saturation divided by the original oil saturation is defined as the production index of the conglomerate reservoir. Fow=So−So′So14 Where Fow—production index (decimal); So—original oil saturation (decimal); So′—current oil saturation (decimal). The production index represents the dynamic flooding parameter of the oil and gas layer, and it can reflect the flooding level of the current oil layer. By using the relative value principle the production index can eliminate the conglomerate lithological influence, and avoid the weakness of using the current oil saturation as the single value to quantitatively evaluate the water flooding level. Alternatively, the application of the water production rate model needs to calculate the oil–water relative permeability ratio and oil–water viscosity, but these two types of data also are too limited in the actual oilfield interpretation, so the accuracy and efficiency of the application are not suitable for the evaluation requirement. The produced index can effectually avoid these problems in evaluating the flooding level and has a good application effect for the conglomerate reservoir. Based on the longitudinal invasion models of the water flooded reservoir, the theoretical relationship between the original reservoir resistivity and current reservoir resistivity have been established according to the petrophysical volume model. As the analysis on the influential factors of resistivity the computational models of original reservoir resistivity have also been respectively established for three kinds of conglomerate lithologies, finally realizing the accurate calculation and inversion of the original reservoir resistivity. Additionally, the injected water constantly washes the reservoir mud and leads to an increase of the effective porosity—the increased volume is equal to the washed mud volume. Therefore, the effective porosity of the original reservoir can be calculated by using the following method which is the difference between the effective porosity of the current reservoir and the amount of clay change (ϕ = ϕ′  -  ΔVsh). From the above analysis, both parameters of the original reservoir resistivity and original effective porosity can be calculated accurately by various data in the current reservoir, then the established multiple linear regression models are used to calculate the original reservoir oil saturation. Combined with the current oil saturation the production index can be determined to quantitatively evaluate the water flooded levels of a conglomerate reservoir. 4.2. The effects of application The above methods are used to evaluate the water flooded levels of a conglomerate reservoir and gain a good application effect in actual interpretation. Figure 11 shows the quantitative interpretation results of water flooded levels for oil well X which belongs to a conglomerate reservoir in the Karamay oilfield. From figure 11 we can see that the blue line represents the original oil saturation of the conglomerate reservoir and the red line represents the current oil saturation in the saturation evaluation channel, and the difference between the two saturations represents the displacement volume of crude oil in the pore space. The ratio between this difference and the original oil saturation is the production index which can represent the water flooded levels of the conglomerate reservoir from water injection development to the current reservoir state. The application of the production index in the conglomerate reservoir greatly improves the quantitative interpretation precision of the water-flooded layers. Based on the above methods the water flooded levels come from coring well 22 where layer 146 has been evaluated, and the evaluation results agree with the testing oil conclusions. From the comparative results, the production index can accurately interpret the water flooded levels of layer 121 and the precision can reach 82%, but the precision of oil saturation and the water production rate can only reach respectively 57% and 63%, which are lower than the production index in interpretation accuracy because of the serious heterogeneity impact of the conglomerate reservoir. Figure 11. Open in new tabDownload slide Graph of quantitative interpretation results of flooded levels. Figure 11. Open in new tabDownload slide Graph of quantitative interpretation results of flooded levels. 5. Conclusions The main factors which affect the resistivity change of the water flooded reservoir have two aspects. On the one hand, the injected water constantly washes the mud and leads to the decreasing of clay volume, which can influence reservoir resistivity change. On the other hand, the injected water displaces crude oil in the reservoir pore space and leads to an increase of reservoir water volume. In addition, the injected water and original formation water will exchange various ions because of the difference between two kinds of water and will finally reach a balance. These factors can also influence reservoir resistivity change. The variable quantity of clay volume is the relationship between the injected water volume and the pore structure. With the increase of the water injected volume, the variable quantity of reservoir clay which has different lithologies and different pore structures increases with a different slope. When the reservoir reaches a middle-high water flooded level, the effect of the injected water washing the reservoir mud becomes weaker because the main flow network in the rock is formed, and the curves of the clay change volume and appear to have a breakpoint value. The clay resistivity is the relationship between the CEC and the internal factors, which include the type of clay mineral and its content. Because the main water type is NaHCO3 in the research area and its pH value is about 7.8, the montmorillonite formed by terrigenous clastic translates into illite and an illite–montmorillonite mixture layer under an alkaline environment. In addition, the aluminum silicate minerals are weathered and leached by the acidic medium and finally form kaolinite. Therefore, the main types of clay mineral in the research area are illite and an illite–montmorillonite mixture layer together with a little kaolinite. The production index used by the principle of relative value eliminates the influence of the conglomerate lithologies and heterogeneity and evaluates the quantitatively water flooded layer, and greatly improves the interpretation accuracy of the water flooded levels. For the high waterflood oilfield, this parameter has achieved a good application effect in secondary adjustment and development, and has provided technical support on avoiding the perforation of a high water-flooded layer and enhancing oil recovery. Acknowledgments This research was supported by the Young Teacher's Research Starting Foundation of the University of Chinese Academy of Sciences and the Project of Karst Groundwater Resources Exploration and Assessment in Beijing (BJYRS-ZT-03). References Amidu S A , Dunbar J A . , 2008 An evaluation of the electrical-resistivity method for water-reservoir salinity studies , Geophysics , vol. 73 (pg. 39 - 49 ) 10.1190/1.2938994 Google Scholar Crossref Search ADS WorldCat Crossref Annunziata D C , Claudio M C , Liliana G , Antonio V . , 1999 Effect of Ph, exchange cations and hydrolytic species of Al and Fe on formation and properties of montmorillonite-protein complexes , Effect of Mineral-Organic-Microorganism Interactions on Soil and Freshwater Environments New York, NY Kluwer Academic/Plenum (pg. 159 - 166 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Archie G E . , 1942 The electrical resistivity log as an aid in determining some reservoir characteristics , Trans. AIME , vol. 146 (pg. 54 - 62 ) 10.2118/942054-G Google Scholar Crossref Search ADS WorldCat Crossref Argaud . , 1989 Formation factor and resistivity index of heterogeneous media Proc. of the 12th SPWLA Annual Int. Formation Evaluation Symp. (pg. 132 - 141 ) Argaud M , Aquitaine E , Giouse H . , 1989 Salinity and saturation effects on shaly sandstone conductivity , SPE , vol. 29 (pg. 49 - 60 ) OpenURL Placeholder Text WorldCat Asquith G B . , 1990 BASIC program for determining effective water saturation in shaly sands using percentage of clay-bound water-filled microporosity , Geobyte , vol. 5 (pg. 3 - 14 ) OpenURL Placeholder Text WorldCat Chanh C M , Pablo S , Ali R B , Leak W O , Mohammed A A . , 2011 Combining resistivity and capture sigma logs for formation evaluation in unknown water salinity-a case study in a mature carbonate field , SPE Reservoir Eval. Eng. , vol. 14 (pg. 446 - 456 ) 10.2118/135160-PA Google Scholar Crossref Search ADS WorldCat Crossref Clavier C , Coates G , Dumanoir J . , 1977 Theoretical and experimental basis for the dual-water model for interpretation of shaly sands , SPE , vol. 17 (pg. 59 - 68 ) OpenURL Placeholder Text WorldCat Clavier C , Coates G , Dumanoir J . , 1984 The theoretical and experimental bases for the dual-water model for interpretation of shaly sands , J. Soc. Pet. Eng. , vol. 24 (pg. 153 - 168 ) 10.2118/6859-PA Google Scholar Crossref Search ADS WorldCat Crossref Cosenza P , Pret D , Zamora M . , 2015 Effect of the local clay distribution in the effective electrical conductivity of clay rocks , J. Geophys. Res. , vol. 120 (pg. 145 - 168 ) 10.1002/2014JB011429 Google Scholar Crossref Search ADS WorldCat Crossref Crane S D . , 1990 Impacts of microporosity, rough pore surface and conductive minerals on saturation calculations form electric measurements: an extended Archie's Law , SPWLA , vol. 1 pg. 990 OpenURL Placeholder Text WorldCat Geng C X , Zhao C Y , Ma D H , Yang J B , Song X H . , 2005 Method for estimating original oil saturation of watered out layers , Daqing Pet. Geol. Dev. , vol. 24 (pg. 67 - 69 ) (in Chinese) OpenURL Placeholder Text WorldCat Givens W W . , 1986 Formation factor, resistivity index and related equations based upon a conductive rock matrix model (CRMM) 27th SPWLA (pg. 25 - 31 ) Hamdi N , Srasra E . , 2013 Hydraulic conductivity study of compacted clay soils used as landfill for an acidic waste , Waste Manage , vol. 33 (pg. 6 - 60 ) 10.1016/j.wasman.2012.08.012 Google Scholar Crossref Search ADS WorldCat Crossref Hamilton D S , Holtz M H , Ryles P , Lonergan T , Hillger M . , 1998 Approaches to identifying reservoir heterogeneity and reserve growth opportunities in a continental-scale bed-load fluvial system: Hutton sandstone, Jackson field, Australia , AAPG Bull , vol. 82 (pg. 2192 - 2219 ) OpenURL Placeholder Text WorldCat Han J , Kamber M . , 2006 , Data Mining: Concepts and Techniques (The Morgan Kaufmann Series in Data Management Systems) 2nd edn San Francisco, CA Morgan Kaufmann (pg. 137 - 152 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hill H J , Milburn J D . , 1956 Effect of clay and water salinity on electrochemical behavior of reservoir rocks , Trans. AMIE , vol. 207 (pg. 65 - 72 ) OpenURL Placeholder Text WorldCat Jakosky J J , Hopper R H . , 1937 The effect of moisture on the direct current resistivities of oil sands and rocks , Geophysics , vol. 2 (pg. 33 - 54 ) 10.1190/1.1438064 Google Scholar Crossref Search ADS WorldCat Crossref Mclachlan D S , Button M B , Adams S R , Gorringe V M , Kneen J D , Muoe J , Wedepohl E . , 1987 Formation resistivity factors for a compressible solid-brine mixture , Geophysics , vol. 52 (pg. 194 - 203 ) 10.1190/1.1442295 Google Scholar Crossref Search ADS WorldCat Crossref Naidja A , Violante A , Huang P M . , 1995 Adsorption of tyrosinase on montmorillonite as influenced by hydroxyaluminum coatings , Clays Clay Miner , vol. 43 (pg. 647 - 655 ) 10.1346/CCMN.1995.0430601 Google Scholar Crossref Search ADS WorldCat Crossref Okotie S , Ikporo B . , 2014 Adjustment to oil saturation estimate due to various reservoir drive mechanisms , Int. J. Adv. Pet. Eng. Technol. , vol. 1 (pg. 17 - 26 ) OpenURL Placeholder Text WorldCat Pan H P , Wang X , Fan Z J , Ma Y , Li M Q . , 2000 Computing method of reservoir originality oil saturation , Geoscience , vol. 14 (pg. 451 - 453 ) (in Chinese) OpenURL Placeholder Text WorldCat Pardo D , Torres-Verdin C . , 2014 Fast 1D inversion of logging-while-drilling resistivity measurements for improved estimation of formation resistivity in high-angle and horizontal wells , Geophysics , vol. 8 (pg. 111 - 124 ) 10.1190/geo2014-0211.1 OpenURL Placeholder Text WorldCat Crossref Scott B J , Dani O . , 2005 Thermal and geometrical effects on bulk permittivity of porous mixtures containing bound water , Electromagnetic Aquametry Berlin Springer (pg. 63 - 87 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Silva L P , Bassioni Z . , 1985 A shaly sand conductivity model based on variable equivalent counterion conductivity and dual water concepts Proc. of the 26th SPW LA Annual Logging Symp. (pg. 17 - 20 ) Tan F Q , Li H Q , Liu H T , Jiang F C , Yu H Y . , 2009 Micro-geological causes and macro-geological controlling factors of low-resistivity oil layers in the Puao Oilfield , Pet. Sci. , vol. 6 (pg. 246 - 253 ) 10.1007/s12182-009-0039-3 Google Scholar Crossref Search ADS WorldCat Crossref Tan F Q , Li H Q , Meng Z X , Guo H F , Li X Y . , 2010a Research on application of data mining method in petroleum exploration and development , Oil Geophys. Prospect , vol. 45 (pg. 91 - 97 ) (in Chinese) OpenURL Placeholder Text WorldCat Tan F Q , Li H Q , Xu C F , Li Q Y , Peng S C . , 2010b Quantitative evaluation methods for water-flooded layers of conglomerate reservoir based on well logging data , Pet. Sci. , vol. 7 (pg. 485 - 493 ) 10.1007/s12182-010-0092-y Google Scholar Crossref Search ADS WorldCat Crossref Usama F , Gregory P S , Padhraic S . , 1996 , Knowledge Discovery and Data Mining: Towards a Unifying Framework Portland, Oregon AAAI Press (pg. 82 - 88 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Vladimir S , Aleksandr M , Albert R , Omar D R . , 2007 Estimation of clay content in soil based on resistivity modeling and laboratory measurements , Geophys. Prospect , vol. 55 (pg. 265 - 275 ) 10.1111/j.1365-2478.2007.00599.x Google Scholar Crossref Search ADS WorldCat Crossref Walid M M , Khaled S S . , 2014 A numerical technique for an accurate determination of formation resistivity factors using FR-RO overlays method , Arab. J. Geosci , vol. 8 (pg. 1291 - 1297 ) OpenURL Placeholder Text WorldCat Waxman M , Smits L J M . , 1968 Electrical conductivities in oil bearing shaly sand , Soc. Pet. Eng. , vol. 42 (pg. 107 - 121 ) 10.2118/1863-A Google Scholar Crossref Search ADS WorldCat Crossref Winsaure W O , McCardell W M . , 1956 Ionic double –layer conductivity in reservoir rock , Trans. AIME , vol. 198 (pg. 129 - 134 ) OpenURL Placeholder Text WorldCat Witten I H , Frank E . , 2005 , Data Mining: Practical Machine Learning Tools and Techniques 2nd edn Boston, MA Morgan Kaufmann (pg. 17 - 28 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © 2015 Sinopec Geophysical Research Institute
TI - Evaluation on an original resistivity inversion method of water flooding a conglomerate reservoir based on petrophysical analysis
JF - Journal of Geophysics and Engineering
DO - 10.1088/1742-2132/12/5/780
DA - 2015-10-01
UR - https://www.deepdyve.com/lp/oxford-university-press/evaluation-on-an-original-resistivity-inversion-method-of-water-EXcTLP3Nwd
SP - 780
VL - 12
IS - 5
DP - DeepDyve
ER -