Water-drive reservoir remaining oil distribution prediction method and device

By dividing seepage areas and initializing cellular automata models, using deduction rules to predict residual oil distribution of water-flooded reservoirs, the problem of high computing resources and time consumption in the existing technology is solved, and the rapid and accurate prediction of residual oil distribution and Demonstration of fine water injection strategies.

CN118940491APending Publication Date: 2024-11-12CHINA UNIV OF PETROLEUM (EAST CHINA)
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410980788.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

When using reservoir engineering methods to predict residual oil distribution in reservoir development, it is necessary to solve large-scale nonlinear equation systems, resulting in huge calculation resources and time consumption, making it difficult to obtain accurate predictions in a short time.

Method used

The effective seepage area and the invalid seepage area are divided by KH splitting method, and the data of production wells and injection wells are split, and the cellular automata model is initialized, and the geological attribute distribution and deduction rules are deduced to achieve the residual water-flooding reservoir. Oil distribution prediction avoids large-scale system of equations solving.

Benefits of technology

The residual oil distribution prediction in a short time saves calculation resources and time, can intuitively display the changes in residual oil distribution under different water injection strategies, help the reservoir to implement a fine water injection strategy, and quickly determine the fitting direction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940491A_ABST
    Figure CN118940491A_ABST
Patent Text Reader

Abstract

The invention discloses a water-drive reservoir remaining oil distribution prediction method and device. The method comprises the steps that S1, an effective seepage area and an invalid seepage area are divided; s2, splitting the production data of the production well and the water injection data of the water injection well; s3, initializing a cellular automaton model according to the geological attribute distribution of the effective seepage region and the split production data and water injection data; and S4, performing deduction according to a deduction rule by initializing the cellular automaton model to realize the distribution prediction of the remaining oil of the water-drive reservoir. By adopting the technical scheme of the invention, the residual oil distribution of the water-drive oil reservoir can be predicted in a short time by using an oil reservoir engineering mode.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of residual oil distribution prediction, and in particular relates to a method and a device for predicting residual oil distribution in a water drive oil reservoir. Background Art

[0002] Residual oil refers to the unextracted oil in the oil layer, reservoir or oil field that has been put into development. Water flooding is widely used in reservoir development, but it also leads to complex reservoir fluid distribution, which increases the difficulty of residual oil extraction. Rapid and accurate prediction of the distribution of residual oil is crucial for oil fields to formulate injection and production strategies, tap potential layers, improve oil displacement efficiency and even improve recovery. Common macroscopic residual oil research methods include development geological methods and reservoir engineering methods. Development geological methods analyze the residual oil enrichment area based on reservoir heterogeneity, geological characteristics and structural morphology. It has a very detailed description of reservoir geology, but the analysis of saturation can often only reflect the distribution of residual oil around the wellbore. Reservoir engineering methods use engineering theories such as seepage mechanics and rock and soil mechanics to describe the residual oil based on the dynamic development data of the oil field after highly matching geological modeling of the reservoir. Reservoir numerical simulation in reservoir engineering methods is the most widely used quantitative research method for residual oil in the current field of reservoir development. Currently, most reservoir numerical simulations are based on the finite difference method. In the process of predicting the remaining oil distribution, it is necessary to solve a large-scale nonlinear equation set, which consumes a lot of computing resources and time. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method and device for predicting the distribution of remaining oil in a water-driven oil reservoir, which can use reservoir engineering methods to predict the distribution of remaining oil in a water-driven oil reservoir in a relatively short time.

[0004] To achieve the above object, the present invention adopts the following technical solution:

[0005] A method for predicting the distribution of remaining oil in a water-driven oil reservoir, comprising:

[0006] Step S1, dividing the effective seepage area and the ineffective seepage area;

[0007] Step S2, splitting the production data of the production wells and the water injection data of the water injection wells;

[0008] Step S3, initializing the cellular automation model according to the geological attribute distribution of the effective seepage area and the split production data and water injection data;

[0009] Step S4, initialize the cellular automation model to perform deduction according to the deduction rules to achieve the prediction of the remaining oil distribution in the water drive reservoir.

[0010] Preferably, step S1 comprises:

[0011] Derive sand layer thickness from sand layer thickness contour map;

[0012] Use Kriging interpolation method to obtain the complete sand layer thickness distribution and set the sand layer thickness threshold;

[0013] The area where the sand layer thickness is greater than the sand layer thickness threshold is set as the effective seepage area, and the area where the sand layer thickness is not greater than the sand layer thickness threshold is set as the invalid seepage area.

[0014] Preferably, in step S2, the production data of production wells and the injection data of water injection wells are split using the KH splitting method to obtain the production and injection volume of the target reservoir at each deduction time point, and map them to the corresponding grids, and set the grid type to oil well node and water well node.

[0015] Preferably, the geological attribute distribution includes: sand layer thickness field, permeability field, porosity field, depth field, pressure field, and oil saturation field.

[0016] Preferably, the cellular automaton model consists of cellular nodes, cellular neighborhoods, neighbor nodes, cellular spaces and deduction rules; wherein the reservoir is divided into high-density grids, each grid is used as a cellular node, and the node state is characterized by oil saturation; a Moore-type neighborhood space is set, that is, eight nodes adjacent to the current node in physical position are used as neighbor nodes; the state of the cellular node at the next moment is determined by the current state of the node and the state of the neighbor nodes in the current neighborhood using reservoir engineering methods; all cellular nodes and their neighborhood spaces constitute the cellular space.

[0017] The present invention also provides a device for predicting the distribution of remaining oil in a water-driven oil reservoir, comprising:

[0018] A division module is used to divide effective seepage areas and ineffective seepage areas;

[0019] A splitting module is used to split the production data of production wells and the water injection data of water injection wells using the KH splitting method;

[0020] An initialization module is used to initialize the cellular automation model according to the geological attribute distribution of the effective seepage area and the split production data and water injection data;

[0021] The prediction module is used to initialize the cellular automation model to perform deduction according to the deduction rules to realize the prediction of the remaining oil distribution in the water-flooded reservoir.

[0022] Preferably, the geological attribute distribution includes: sand layer thickness field, permeability field, porosity field, depth field, pressure field, and oil saturation field.

[0023] Preferably, the cellular automaton model consists of cellular nodes, cellular neighborhoods, neighbor nodes, cellular spaces and deduction rules; wherein the reservoir is divided into high-density grids, each grid is used as a cellular node, and the node state is characterized by oil saturation; a Moore-type neighborhood space is set, that is, eight nodes adjacent to the current node in physical position are used as neighbor nodes; the state of the cellular node at the next moment is determined by the current state of the node and the state of the neighbor nodes in the current neighborhood using reservoir engineering methods; all cellular nodes and their neighborhood spaces constitute the cellular space.

[0024] The present invention uses lightweight deduction rules constructed by reservoir engineering methods, does not solve large-scale equations, and predicts the remaining oil distribution in a relatively short time, saving a lot of computing resources and computing time, which is conducive to intuitively displaying the changes in the remaining oil distribution under different water injection strategies in a short time, and helping the reservoir to implement a fine water injection strategy. At the same time, the remaining oil deduction with the current initial parameters can intuitively obtain the deduction of the remaining oil distribution in the current state, and compare it with the current actual situation, which is conducive to helping the reservoir numerical simulation to quickly determine the fitting direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0026] Figure 1 A flow chart of the method for predicting the distribution of remaining oil in a water-driven oil reservoir according to the present invention;

[0027] Figure 2 Schematic diagram for determining effective seepage area;

[0028] Figure 3 It is a schematic diagram of the cell space of the cell neighborhood;

[0029] Figure 4 A flow chart for predicting remaining oil distribution based on deduction rules. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] Embodiment 1:

[0033] like Figure 1 As shown, an embodiment of the present invention provides a method for predicting the distribution of remaining oil in a water-driven oil reservoir, comprising:

[0034] Step S1, dividing the effective seepage area and the ineffective seepage area;

[0035] Step S2, splitting the production data of the production wells and the water injection data of the water injection wells;

[0036] Step S3, initializing the cellular automation model according to the geological attribute distribution of the effective seepage area and the split production data and water injection data;

[0037] Step S4, initialize the cellular automation model to perform deduction according to the deduction rules to achieve the prediction of the remaining oil distribution in the water drive reservoir.

[0038] As an implementation method of the embodiment of the present invention, Figure 2 As shown, step S1 includes:

[0039] Derive sand layer thickness from sand layer thickness contour map;

[0040] Use Kriging interpolation method to obtain the complete sand layer thickness distribution and set the sand layer thickness threshold;

[0041] The area where the sand layer thickness is greater than the sand layer thickness threshold is set as the effective seepage area, and the area where the sand layer thickness is not greater than the sand layer thickness threshold is set as the invalid seepage area; wherein, the nodes in the effective seepage area undergo fluid seepage with their neighboring nodes, and the nodes in the invalid seepage area are set to have no fluid flow.

[0042] As an implementation method of an embodiment of the present invention, in step S2, the production data of the production well and the injection data of the water injection well are split using the KH splitting method to obtain the production and injection volume of the target reservoir at each deduction time point, and map them to the corresponding grids, and set the grid type to oil well node and water well node.

[0043] As an implementation method of an embodiment of the present invention, in step S3, the geological attribute distribution includes: sand layer thickness field, permeability field, porosity field, depth field, pressure field, and oil saturation field.

[0044] Furthermore, point data is derived from the contour map of permeability, depth and porosity of the target reservoir, mapped to the corresponding grid nodes, and the value of the node in the invalid seepage area is set to the current data point mean or set value to stabilize the interpolation effect, and then the permeability field, porosity field, depth field and pressure field are obtained based on the Kriging interpolation method; among which, the pressure field is converted from the depth field. The oil saturation value of the production well is obtained from the well logging data. Under the premise of ignoring the data interpretation date, the invalid seepage area is set to a fixed value and the oil saturation field is obtained using the Kriging interpolation method.

[0045] As an implementation of the embodiment of the present invention, in step S3, Figure 3 As shown in the figure, the cellular automaton model consists of cellular nodes, cellular neighborhoods, neighbor nodes, cellular spaces and deduction rules; the reservoir is divided into high-density grids, each grid is used as a cellular node, and the node state is characterized by oil saturation; a Moore-type neighborhood space is set, that is, the eight nodes adjacent to the current node in physical position are used as neighbor nodes; the state of the cellular node at the next moment is determined by the current state of the node and the state of the neighbor nodes in the current neighborhood using the reservoir engineering method; all cellular nodes and their neighborhood spaces constitute the cellular space.

[0046] As an implementation method of the embodiment of the present invention, Figure 4 As shown, step S4 is specifically as follows:

[0047] (1) Obtain the pressure of neighboring nodes and calculate the pressure difference:

[0048] Δp i =pp i

[0049] Among them, p is the current cell node pressure, p i is the pressure of the i-th neighbor node of the current cell node, Δp i is the pressure difference between the current cell node and the i-th neighbor node.

[0050] (2) Obtain the oil and water viscosity of the current reservoir and use Darcy's law formula to calculate the oil and water inflow and outflow volumes of the current node and neighboring nodes:

[0051]

[0052] Among them, m and n are the number of nodes in the cell neighborhood that are higher and lower than the current cell node, respectively, and k o and k w are the relative permeabilities of oil and water at the current cell node, and are the relative permeabilities of oil and water of the ith neighbor node of the current cell node, μ o and μ wThey are the viscosity of oil and water respectively.

[0053] (3) Based on the change in crude oil volume and the current node porosity, update the state of the cellular node at the next moment:

[0054]

[0055] in, and are the oil saturation at the next moment and the current moment, respectively. is the porosity of the node itself, V rock is the rock volume, that is, the cell node volume.

[0056] (4) Use the split flow equation and water content to obtain oil saturation from production data as a benchmark value for gradient descent adjustment of model parameters:

[0057]

[0058] Among them, f w is the moisture content, μ o and μ w are the viscosities of oil and water respectively, a and b are the fitting coefficients, S w is the water saturation.

[0059] The loss function is constructed with the deduced value and the benchmark value. That is, the loss function is set to the square of the difference between the deduced value and the benchmark value. The model parameters are optimized by gradient descent after a certain number of deduction steps.

[0060]

[0061] Among them, θ j+1 and θ j are the parameter sets of the cellular automaton model at the j+1th and jth steps respectively, α is the learning rate, and f(θ) is the loss function that measures the difference between the deduced value and the benchmark value.

[0062] Combining the oil-water relative permeability curve with the production data can avoid the need to calculate the fitting coefficient and derive the relationship between water cut and water saturation. Then, assuming that the sum of the oil saturation and water saturation in the water-flooded reservoir is 1, the oil saturation is obtained.

[0063] (5) Obtain physical parameters such as crude oil volume coefficient and crude oil compressibility coefficient to establish material balance equations, and update the pressure after fluid flow according to the cumulative change in fluid volume in the current deduction step:

[0064] N p B o +W p B w -W inj B w=NB oi C eff ΔP

[0065] Among them, N p and W p are the cumulative production of oil and water at the current node, W inj is the cumulative water injection volume of the current node.

[0066] If there is a production well in the current node, N p and W p Obtained from the split production data, if there is a water injection well in the current node, W inj Obtained from water injection data after splitting.

[0067] For cell nodes without production wells or injection wells, N p Represents the relative outflow of oil from the current cell node, W p -W inj represents the relative outflow of water at the current node, then the material balance equation is:

[0068]

[0069] in, and are the volume changes of oil and water at the current cell node at the i-th deduction moment, N is the crude oil reserves at the current node, and B o is the crude oil volume coefficient, B w is the formation water volume coefficient, B oi Current crude oil volume coefficient, ΔP The pressure drop at the current deduction time is the difference between the initial pressure and the current pressure, C eff is the rock compressibility coefficient.

[0070] Further, in step S4, first determine the nodes in the neighborhood of the current cell that are higher or lower than the pressure of the current cell node, and use Darcy's law to calculate the volume of oil and water flowing from the current cell node to the neighboring nodes with lower pressure than it. Similarly, calculate the volume of oil and water flowing into the current node from the neighboring nodes with higher pressure than the current node. Comprehensively calculate the inflow and outflow, and calculate the state of the cell node at the next moment according to the porosity of the current node. Calculate the oil saturation from the production data as the reference value based on the water content and the flow fraction equation, construct the loss function based on the difference between the deduced value and the reference value, reasonably set the learning rate, and optimize the model parameter set every specific deduction step interval. Determine the type of the current node, whether there are water wells and oil well nodes, determine the applicable equation form according to the node type of the current cell node, accumulate the volume changes of oil and water at the current node, and update the pressure after the fluid flow according to the material balance equation.

[0071] Embodiment 2:

[0072] The embodiment of the present invention further provides a device for predicting the distribution of remaining oil in a water-driven oil reservoir, comprising:

[0073] A division module is used to divide effective seepage areas and ineffective seepage areas;

[0074] A splitting module is used to split the production data of production wells and the water injection data of water injection wells using the KH splitting method;

[0075] An initialization module is used to initialize the cellular automation model according to the geological attribute distribution of the effective seepage area and the split production data and water injection data;

[0076] The prediction module is used to initialize the cellular automation model to perform deduction according to the deduction rules to realize the prediction of the remaining oil distribution in the water-flooded reservoir.

[0077] As an implementation method of an embodiment of the present invention, the geological attribute distribution includes: sand layer thickness field, permeability field, porosity field, depth field, pressure field, and oil saturation field.

[0078] As an implementation method of an embodiment of the present invention, a cellular automaton model is composed of cellular nodes, cellular neighborhoods, neighbor nodes, cellular spaces and deduction rules; wherein the reservoir is divided into high-density grids, each grid is used as a cellular node, and the node state is characterized by oil saturation; a Moore-type neighborhood space is set, that is, eight nodes adjacent to the current node in physical position are used as neighbor nodes; the state of the cellular node at the next moment is determined by the current state of the node and the state of the neighbor nodes in the current neighborhood using reservoir engineering methods; all cellular nodes and their neighborhood spaces constitute the cellular space.

[0079] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for predicting the distribution of remaining oil in a water-flooded reservoir, characterized in that: include: Step S1, dividing the effective seepage area and the ineffective seepage area; Step S2, splitting the production data of the production wells and the water injection data of the water injection wells; Step S3, initializing the cellular automation model according to the geological attribute distribution of the effective seepage area and the split production data and water injection data; Step S4, initialize the cellular automation model to perform deduction according to the deduction rules to achieve the prediction of the remaining oil distribution in the water drive reservoir.

2. The method for predicting the distribution of remaining oil in a water-flooding reservoir according to claim 1, characterized in that: Step S1 includes: Derive sand layer thickness from sand layer thickness contour map; Use Kriging interpolation method to obtain the complete sand layer thickness distribution and set the sand layer thickness threshold; The area where the sand layer thickness is greater than the sand layer thickness threshold is set as the effective seepage area, and the area where the sand layer thickness is not greater than the sand layer thickness threshold is set as the invalid seepage area.

3. The method for predicting the distribution of remaining oil in a water-flooding reservoir as claimed in claim 2, characterized in that: In step S2, the KH splitting method is used to split the production data of production wells and the injection data of water injection wells, and the production and injection volume of the target reservoir at each deduction time point are obtained and mapped to the corresponding grids, and the grid type is set to oil well node and water well node.

4. The method for predicting the distribution of remaining oil in a water-flooding reservoir as claimed in claim 3, characterized in that: The geological attribute distribution includes: sand layer thickness field, permeability field, porosity field, depth field, pressure field, and oil saturation field.

5. The method for predicting the distribution of remaining oil in a water-flooding reservoir according to claim 4, characterized in that: The cellular automaton model consists of cellular nodes, cellular neighborhoods, neighbor nodes, cellular spaces and deduction rules. The reservoir is divided into high-density grids, each grid is used as a cellular node, and the node state is characterized by oil saturation. A Moore-type neighborhood space is set, that is, the eight nodes physically adjacent to the current node are used as neighbor nodes. The state of the cellular node at the next moment is determined by the current state of the node and the state of the neighbor nodes in the current neighborhood using reservoir engineering methods. All cellular nodes and their neighborhood spaces constitute the cellular space.

6. A device for predicting the distribution of remaining oil in a water-driven oil reservoir, characterized in that: include: A division module is used to divide effective seepage areas and ineffective seepage areas; A splitting module is used to split the production data of production wells and the water injection data of water injection wells using the KH splitting method; An initialization module is used to initialize the cellular automation model according to the geological attribute distribution of the effective seepage area and the split production data and water injection data; The prediction module is used to initialize the cellular automation model to perform deduction according to the deduction rules to realize the prediction of the remaining oil distribution in the water-flooded reservoir.

7. The device for predicting the distribution of remaining oil in a water-flooding reservoir according to claim 6, characterized in that: The geological attribute distribution includes: sand layer thickness field, permeability field, porosity field, depth field, pressure field, and oil saturation field.

8. The device for predicting the distribution of remaining oil in a water-flooding reservoir according to claim 7, characterized in that: The cellular automaton model consists of cellular nodes, cellular neighborhoods, neighbor nodes, cellular spaces and deduction rules. The reservoir is divided into high-density grids, each grid is used as a cellular node, and the node state is characterized by oil saturation. A Moore-type neighborhood space is set, that is, the eight nodes physically adjacent to the current node are used as neighbor nodes. The state of the cellular node at the next moment is determined by the current state of the node and the state of the neighbor nodes in the current neighborhood using reservoir engineering methods. All cellular nodes and their neighborhood spaces constitute the cellular space.