Visual flow field representation method based on data mining

Abstract

The present invention belongs to the technical field of oilfield water drive development, and specifically relates to a method for visualizing flow field characterization based on data mining. It includes the following steps: Step 1, establish a typical block geological model, connect with the streamline numerical simulator through Matlab, extract streamline feature data, and obtain the streamline field; Step 2, perform attribute dimension reduction on the extracted streamline data, and reconstruct the flow field; Step 3, perform streamline clustering analysis to realize the visualization of the flow field; Step 4, comprehensively consider the influencing factors of the reservoir flow field, establish flow field intensity and streamline density characterization parameters to characterize the influence of fluid cumulative effect and instantaneous effect on the flow field; Step 5, quantitative characterization of flow field intensity; Step 6, quantitative characterization of streamline density. This method solves the problem that the existing technical methods cannot accurately reconstruct and visualize the reservoir flow field.

Description

Technical Field

[0001] The invention belongs to the technical field of oilfield water flooding development, and in particular relates to a visualized flow field characterization method based on data mining. Background Art

[0002] At present, domestic old oil fields have generally entered the development stage of medium-high water content and ultra-high water content, which has led to rising oil field development costs and serious decline in benefits. At present, flow field adjustment is one of the important means to stabilize production under low oil prices. In the process of water drive development, due to the influence of geological characteristics, fluids, well networks and other conditions, especially the horizontal and vertical heterogeneity of the reservoir, the displacement of injected water in the underground flow field is uneven, the distribution of residual oil is becoming more and more complex and scattered, the flow field of the reservoir is becoming more and more complex, and the flow field characteristics are unclear. For water drive development reservoirs, the development of the flow field is closely related to the effect of water drive. Therefore, visual representation of the reservoir flow field can effectively guide the adjustment of the flow field of water drive reservoirs, the exploration of residual oil potential, and the formulation of water control and oil stabilization measures.

[0003] In order to fully tap underground resources and improve crude oil recovery, it is necessary to improve the understanding of underground flow fields, clarify the distribution of reservoir flow fields, and formulate targeted injection and production adjustment measures and directions for tapping remaining oil potential. However, the existing technical methods mostly use traditional reservoir numerical simulation technology to simulate flow field distribution through numerical tracer methods, and the simulation accuracy needs to be further improved. In addition, the existing methods cannot visualize the reservoir flow field. Summary of the invention

[0004] The purpose of the present invention is to provide a visual flow field characterization method based on data mining, which solves the problem that the existing technical methods cannot accurately reconstruct and visualize the reservoir flow field.

[0005] To solve the above problem, the present invention adopts the following technical solutions:

[0006] The present invention provides a visual flow field characterization method based on data mining, which comprises the following steps:

[0007] Step 1: Establish a typical block geological model, connect it with the streamline numerical simulator through Matlab, extract streamline characteristic data, and obtain the streamline field;

[0008] Step 2, perform attribute dimension reduction on the extracted streamline data and reconstruct the flow field;

[0009] Step 3, perform streamline clustering analysis to visualize the flow field;

[0010] Step 4, comprehensively considering the factors affecting the reservoir flow field, establishing flow field intensity and streamline density characterization parameters to characterize the impact of fluid accumulation effect and transient effect on the flow field;

[0011] Step 5, quantitative characterization of flow field intensity;

[0012] Step 6: Quantitative characterization of streamline density.

[0013] In step 1, the streamline reservoir numerical simulation method is used to establish a streamline model and extract relevant data for quantitatively characterizing the flow field. Compared with traditional reservoir numerical simulation technology, it has the computational advantage of processing data of larger orders of magnitude, while maintaining the advantages of an obvious displacement front and reducing the influence of grid direction during the displacement process. It can transform abstract streamline feature data into intuitive underground fluid migration laws and residual oil distribution characteristics, and more accurately characterize the flow field.

[0014] Preferably, in step 2, principal component analysis is used to perform attribute dimension reduction; N P ×N A attributes are linearly combined to form N P ×N A principal components, and select the first N principal components with cumulative contributions exceeding 90% as new attributes to describe a streamline; select appropriate principal components to characterize all attributes of the original flow field, and use the selected principal components to reconstruct the flow field; preferably, the accuracy of reconstruction is used as the criterion for selecting the principal components.

[0015] Further preferably, the more principal components are selected, the greater the principal component contributions are, and the finer the reconstructed flow field is.

[0016] Preferably, the number of principal components that enables the flow field reconstruction accuracy to reach 90% is the optimal number of principal components.

[0017] Preferably, in step 3, the flow field is divided and clustered according to the streamline features extracted by the principal component analysis method, and streamlines with similar features are clustered into one category. According to the clustering criterion, each streamline is represented by a point on a two-dimensional plane and then restored to three-dimensional space to achieve clustering and visualization of the streamlines.

[0018] Further preferably, the K_Means clustering method is selected to perform cluster analysis on the streamline features, and the silhouette coefficient method is applied to evaluate the flow field clustering effect. The closer the silhouette coefficient is to 1, the better the clustering effect.

[0019] In step 4, appropriate parameter indicators are selected by comprehensively considering the influencing factors of the reservoir flow field, and a flow field intensity parameter is established to characterize the influence of the fluid cumulative effect on the flow field, reflecting the influence of the historical flow field on the reservoir production conditions; a streamline density parameter is established to characterize the influence of the fluid instantaneous effect on the flow field, intuitively reflecting the flow field's influence on the fluid displacement capacity and degree of spread.

[0020] Preferably, in step 5, the quantitative characterization method of the flow field intensity includes:

[0021] Selecting a characterization parameter that characterizes the flow field intensity;

[0022] Establishment of flow field intensity characterization formula: Apply the hierarchical analysis method to establish a hierarchical model, divide the preferred flow field intensity characterization parameters, dynamic flow field intensity, and comprehensive flow field intensity into factor layer, transition layer, and target layer according to their mutual relationship, and establish a hierarchical model; establish a pairwise comparison discriminant matrix, obtain the maximum eigenvalue corresponding to the matrix, use the eigenvector corresponding to the maximum eigenvalue as the weight vector, perform consistency test, and output the parameter weight coefficient; obtain the membership matrix according to the membership function, multiply it by the weight coefficient, and then normalize it to obtain the flow field intensity.

[0023] Preferably, the water flow multiple and the dynamic index of water content are selected from the flow field intensity characterization parameters to comprehensively characterize the flow field intensity.

[0024] Further preferably, the flow field intensity calculation formula is:

[0025]

[0026] E j =(b1, b2…b n ) T / max(b1,b2…b n )

[0027]

[0028]

[0029] Among them, E d -Flow field intensity; R d -Membership of each parameter; ω d -Weight coefficient of each parameter; E j -normalized flow field intensity; r 1i - water flow multiple membership function; r i2 -Moisture content membership function.

[0030] Preferably, the quantitative characterization method of the streamline density in step 6 is: the streamline density at any point can be expressed as D n =N, where N represents the number of points at the nth point. The K_means clustering algorithm is used to reduce the number of points, improve the calculation speed, and perform normalization. The streamline density at the nth point can be expressed as:

[0031] d n =D n / max(D n ).

[0032] Beneficial effects achieved by the present invention:

[0033] Compared with traditional reservoir numerical simulation technology, the present invention applies fine streamline numerical simulation technology to analyze flow field characteristics, and adopts data mining technologies such as principal component analysis and clustering methods to establish flow field intensity and streamline density parameters to visualize and quantitatively characterize the flow field. This method has the computational advantage of processing data of a larger order of magnitude, while maintaining the advantages of a clear displacement front and reducing the influence of grid direction during the displacement process. It can transform abstract streamline feature data into intuitive underground fluid migration laws and residual oil distribution characteristics, and more accurately characterize the flow field. This method can provide directional guidance for flow field optimization and adjustment and residual oil potential mining.

[0034] This method was applied in the test of Chengbei 11 well area in the main Guantao Formation of Chengdao Oilfield in Shengli Offshore, and the flow field optimization adjustment of the well area was successfully guided according to this method. After the application of this method on-site implementation, the water drive reserves were increased by 1.86 million tons, the daily oil increase was 44 tons, the cumulative oil increase was 5,467 tons, and the water cut increase rate was controlled within 2%, achieving good results, and achieving good economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0036] Figure 1 It is a flowchart of a specific embodiment of the visual flow field characterization method based on data mining of the present invention;

[0037] Figure 2 It is a flowchart of principal component analysis attribute dimension reduction in a specific embodiment of the present invention;

[0038] Figure 3 Selecting the best principal component flow field reconstruction map in a specific embodiment of the present invention;

[0039] Figure 4 A flow chart of a K-means clustering algorithm in a specific embodiment of the present invention;

[0040] Figure 5 It is a schematic diagram of the transformation of a two-dimensional plane into a three-dimensional space by streamline clustering representation in a specific embodiment of the present invention;

[0041] Figure 6 A schematic diagram of selecting the optimal number of clusters using the silhouette coefficient method in a specific embodiment of the present invention;

[0042] Figure 7 It is a streamline point isometric processing diagram in a specific embodiment of the present invention;

[0043] Figure 8It is a schematic diagram of streamline intensity + streamline density flow field characterization in a specific embodiment of the present invention;

[0044] Fig. 9 It is a schematic diagram of visualizing flow field representation in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0045] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0046] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations and / or combinations thereof.

[0047] In order to enable those skilled in the art to more clearly understand the technical solution of the present invention, the technical solution of the present invention will be described in detail below in conjunction with specific embodiments.

[0048] like Figure 1 As shown, Figure 1 The flowchart of the visual flow field characterization method based on data mining is shown in Figure 1. Taking Chengbei 11 well area of ​​the main Guantao formation of Chengdao oilfield in Shengli offshore as an example, the method includes the following steps:

[0049] Step 1, establish a typical block geological model, connect it with the streamline numerical simulator through Matlab, extract streamline feature data, and obtain the streamline field; the streamline feature extraction data diagram is described in Table 1 below.

[0050] Table 1 Extracted streamline feature data

[0051] Extract keywords Streamline features GEOMETRY Streamline location SWAT Saturation TIME OF FIGHT Streamline flight time FLOOIL Oil phase velocity FLOWAT Water phase velocity PFRACOIL Oil production per flow line PFRACWAT Water production per flow line

[0052] Step 2: Apply principal component analysis to reduce the attribute dimension of the extracted streamline data; select appropriate principal components to characterize all the attributes of the original flow field, and use the selected principal components to reconstruct the flow field; convert N P ×N A attributes are linearly combined to form N P ×N A The first N principal components with cumulative contributions exceeding 90% are selected as new "attributes" to describe a streamline. The principal component analysis attribute dimension reduction flow chart is as follows: Figure 2As shown. Select appropriate principal components to characterize all the properties of the original flow field, and use the selected principal components to reconstruct the flow field. In this embodiment, the optimal number of principal components is 4, and the accuracy of flow field reconstruction reaches 90%; the principal component flow field reconstruction diagram is shown in Figure 3 shown.

[0053] Step 3: Based on the extracted streamline features, the K_Means clustering method suitable for flow field characterization is selected to perform cluster analysis on the streamline features. The K_means clustering algorithm flow chart is shown in Figure 4 As shown, the streamline clustering represents the schematic diagram of the transformation of two-dimensional plane into three-dimensional space as shown in Figure 5 As shown; and the silhouette coefficient method is used to evaluate the clustering effect of the flow field. The closer the silhouette coefficient is to 1, the better the clustering effect. The optimal clustering number of the flow field in this embodiment is 6, which can ensure the effective clustering of the flow field and realize the visualization of the flow field, as shown in FIG. Figure 6 As shown;

[0054] Step 4, comprehensively considering the factors affecting the reservoir flow field, establishing flow field intensity and streamline density characterization parameters to characterize the impact of fluid accumulation effect and transient effect on the flow field;

[0055] Step 5, quantitative characterization of flow field intensity, select the characterization parameters of water flow rate and water content from dynamic and static indicators, apply hierarchical analysis method, establish a hierarchical model, divide the water flow rate and water content, dynamic flow field intensity, and comprehensive flow field intensity into factor layer, transition layer and target layer according to their mutual relationship, and establish a hierarchical model; establish a paired comparison discriminant matrix, calculate the maximum eigenvalue corresponding to the matrix, use the eigenvector corresponding to the maximum eigenvalue as the weight vector, perform consistency test, and output the parameter weight coefficient; obtain the membership matrix according to the membership function, the water flow rate and water content are the membership degrees of the first column and the second column of the matrix respectively, and after multiplying by the weight coefficient, normalize and obtain the flow field intensity.

[0056] Calculation of flow field intensity:

[0057] E d =R d ·ω d

[0058] Among them, E d -Flow field intensity;

[0059] R d -Membership of each parameter;

[0060] ω d -Weight coefficients of each parameter;

[0061]

[0062]

[0063] Among them, r 1i - Water flow multiple membership function;

[0064] r i2 -Moisture content membership function;

[0065]

[0066] E j =(b1, b2…b n ) T / max(b1,b2…b n )

[0067] Among them, E j -Flow field intensity (normalized);

[0068] Step 6, streamline density is the number of streamlines per unit area. The streamline density calculation method improves the traditional calculation method of statistical target objects and clustering algorithms. Since the points on the original streamline are not equidistant, the point density cannot fully represent the line density. Therefore, the points on a streamline are equidistant, and then the point density on each streamline is counted to represent it, such as Figure 7 shown.

[0069] Calculation of streamline density: The streamline density at any point can be expressed as D n =N, where N represents the number of points at the nth point. The K_means clustering algorithm is used to reduce the number of points, improve the calculation speed, and perform normalization. The streamline density at the nth point can be expressed as:

[0070] d n =D n / max(D n )

[0071] Through the above steps, the visualization field diagram of flow field intensity + streamline density between well groups in the target block is obtained, such as Figure 8 As shown, further, the flow field intensity + streamline density clustering representation diagram of the reservoir flow field can be obtained by applying the visualization flow field characterization method of the present invention, such as Fig. 9 As shown in the figure, the thickness of the mainstream line is used to represent the density of the streamlines between wells, and the streamline color is used as a measure of the flow field strength. It can intuitively reflect the flow and distribution state of the fluid between wells, quantitatively characterize the difference in the degree of fluid displacement between wells, and provide a basis for the optimization and adjustment of the flow field.

[0072] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A visual flow field characterization method based on data mining, characterized in that: It includes the following steps: Step 1: Establish a typical block geological model, connect it with the streamline numerical simulator through Matlab, extract streamline characteristic data, and obtain the flow field; Step 2, performing attribute dimension reduction on the extracted streamline feature data and reconstructing the flow field; Step 3, perform streamline clustering analysis to visualize the flow field; Step 4, comprehensively considering the factors affecting the reservoir flow field, establishing flow field intensity and streamline density characterization parameters to characterize the impact of fluid accumulation effect and transient effect on the flow field; Step 5, quantitative characterization of flow field intensity; Step 6, quantitative characterization of streamline density; In step 3, based on the streamline features extracted by the principal component analysis method, the flow field is divided and clustered, streamlines with similar features are clustered into one category, and each streamline is represented by a point on a two-dimensional plane according to the clustering criterion, and then restored to three-dimensional space to achieve clustering and visualization of streamlines; The K_Means clustering method was selected to perform cluster analysis on streamline features, and the silhouette coefficient method was used to evaluate the clustering effect of the flow field. The closer the silhouette coefficient is to 1, the better the clustering effect is. In step 5, the quantitative characterization method of the flow field intensity includes: Selecting a characterization parameter that characterizes the flow field intensity; Establishment of flow field intensity representation: Applying the analytic hierarchy process, the flow field intensity representation parameters, dynamic flow field intensity, and comprehensive flow field intensity are divided into factor layer, transition layer, and target layer according to their mutual relationship, and a hierarchical structure model is established; a pairwise comparison discriminant matrix is ​​established, the maximum eigenvalue corresponding to the matrix is ​​obtained, and the eigenvector corresponding to the maximum eigenvalue is used as the weight vector for consistency test, and the parameter weight coefficient is output; the membership matrix is ​​obtained according to the membership function, and after multiplication by the weight coefficient, it is normalized to obtain the flow field intensity; Select water flow multiple and water content dynamic index from the flow field intensity characterization parameters to comprehensively characterize the flow field intensity; The calculation formula of flow field intensity is: E j =(b1,b2…b n ) T / max(b1,b2…b n ) Among them, E d —Flow field intensity; R d —Membership degree of each parameter; ω d —weight coefficient of each parameter; E j —normalized flow field intensity; r i1 — water flow multiple membership function; r i2 —Moisture content membership function; Quantitative characterization method of streamline density: The points on a streamline are equally spaced, and then the density of points on each streamline is counted to represent the points. Calculation of streamline density: The streamline density at any point is expressed as D n =N, where N represents the number of points at the nth point. The K_means clustering algorithm is used to reduce the number of points, improve the calculation speed, and perform normalization. The streamline density at the nth point is expressed as: d n =D n / max(D n )。 2. The method according to claim 1, characterized in that: In step 2, principal component analysis is used to reduce the attribute dimension; N P ×N A attributes are linearly combined to form N P ×N A The first N principal components with cumulative contributions exceeding 90% are selected as new attributes to describe a streamline; Appropriate principal components are selected to characterize all properties of the original flow field, and the selected principal components are used to reconstruct the flow field; the accuracy of the reconstruction is used as the criterion for selecting the principal components.

3. The method according to claim 2, characterized in that: The more principal components are selected, the greater the principal component contribution is, and the finer the reconstructed flow field is.

4. The method according to claim 2, characterized in that: The number of principal components that makes the flow field reconstruction accuracy reach 90% is the optimal number of principal components.

Images