CN105893674A - Method for performing geological attribute prediction with global covariance - Google Patents

Abstract

The invention provides a method for performing geological attribute prediction with global covariance. The method includes the following steps that a three-dimensional structural model is established according to collected geological data, wherein a logging curve of an oil reservoir to be predicted serves as input data of the three-dimensional structural model, a geological attribute prediction result serves as output data of the three-dimensional structural model, and the three-dimensional structural model comprises multiple grids; the input logging curve is coarsened into the grids to serve as known data points of a global covariance method, wherein the geological attribute values of the known data points on the logging curve are known; the geological attribute value of the oil reservoir is predicted with the global covariance method according to the known data points. By means of the method, the speed of solving a geological attribute prediction equation can be greatly increased compared with that in an original Kriging solving method and is increased by 3 times or above.

Description

Method for predicting geological property by using global covariance

Technical Field

The invention relates to the technical field of petroleum exploration and development, in particular to a method for predicting geological properties by using global covariance.

Background

Geological prediction is the early work in the petroleum exploration process, and the smooth petroleum exploration and development can be realized only on the basis of obtaining a relatively accurate geological prediction result.

In oil exploration and development, a kriging method is generally adopted in the existing geological property prediction method. Of these, the kriging method is a geostatistical gridding method that is useful in many fields. Kriging attempts to express trends implicit in the data, for example, high points would be connected along a ridge, rather than isolated by bullseye contours. The kriging process involves several factors: change map model, drift type and blockiness. The central idea of the kriging method is to describe the continuity and the anisotropy of data by a variogram, and to calculate the values of weights and covariances by the variogram, which measure the spatial correlation of an attribute at different positions. Is generally applied to any phenomenon that requires estimation of its spatial distribution with point data.

However, when the kriging method is adopted for geological prediction, the problem of poor continuity of a geological prediction result often exists, so that the geological prediction result is not in accordance with geological significance, and the subsequent petroleum exploration and development process is influenced.

Patent CN 104252549a discloses an analysis well placement method based on kriging interpolation, which adopts the kriging method to perform structural analysis on a space field, proposes a variation function model, and finally performs kriging calculation according to the variation graph function model to realize analysis on a well placement area. However, the crikin method is still used for analysis, and the continuity ratio of the analysis result is poor, so that the accuracy of the analysis result is not high.

Disclosure of Invention

The object of the present invention is to solve at least one of the technical drawbacks mentioned.

Therefore, the invention aims to provide a method for predicting the geological property by using the global covariance, which can greatly improve the speed of solving the geological property prediction equation compared with the original kriging solution method and improve the speed of solving the geological property prediction equation by more than 3 times.

In order to achieve the above object, an embodiment of the present invention provides a method for performing geologic attribute prediction by using global covariance, including the following steps:

step S1, establishing a three-dimensional structure model according to the collected geological data, wherein the input data of the three-dimensional structure model is a logging curve of an oil deposit to be detected, the output data is a geological attribute prediction result, and the three-dimensional structure model comprises a plurality of grids;

step S2, coarsening the input logging curve into the grid as the known data point of the global covariance method, wherein the geological attribute value of the known data point on the logging curve is known;

and step S3, according to the known data points in the step S2, predicting the geological attribute value of the oil deposit by using a global covariance method.

Further, in step S1, the building a three-dimensional structure model according to the collected geological data includes the following steps:

interpreting and analyzing the collected geological data to obtain corresponding interpreted geological data, wherein the interpreted geological data comprises: fault data, horizon data and layering data of the detected oil reservoir;

and establishing the three-dimensional construction model according to the interpreted geological data.

Further, the geological property prediction result comprises: and the porosity, saturation and permeability corresponding to the detected oil reservoir.

Further, in the step S3,

firstly, let the geological property prediction equation be as follows:

$\underset{j=1}{\overset{M}{\Σ}}{w}_j\·C(u_i,u_j)=C(u,u_i),i=1,\mathrm{...},M,$

wherein u is_iAnd u_jThe geologic attribute values, C (u), for the ith and jth known data points, respectively_i,u_j) Is u_iAnd u_jCovariance of (d), w_jFor weight value, M is the number of the known data points;

then, solving the geological property prediction equation by adopting a global covariance method, and rewriting the geological property prediction equation into the following form:

LUw＝r，

wherein, (LU)_ij＝C(u_i,u_j)，r_i＝C(u,u_i) L and U are a lower triangular matrix and an upper triangular matrix respectively;

then, the LU matrix is solved, and after the L matrix and the U matrix are obtained, the weight w is solved_j，

Finally, knowing the weight w_jAnd predicting the geological property value of the unknown data point in the three-dimensional structure model, and recording as a geological property prediction result.

Further, the L and U matrices are:

wherein,

$U_{rj}=a_{rj}-\frac{{\Σ}_{k=1}^{r-1}{l}_{ik}{U}_{kj}}{l_{rr}},(r=0,1,2,\mathrm{....}n-1;j=r,\mathrm{....}n-1),$

$L_{ir}=a_{ir}-\frac{{\mathrm{\Σ}}_{k=1}^{r-1}{l}_{ik}{u}_{kr}}{u_{rr}},(r=0,1,\mathrm{2...},n-1;i=r+1,\mathrm{...},n-1).$

according to the method for predicting the geological property by using the global covariance, the geological property equation is solved by using the global covariance method, when the number of geological property grids is large, the speed of solving the geological property prediction equation is greatly improved compared with the speed of solving the original Krigin solution method, and the speed of solving the geological property prediction equation is improved by more than 3 times. The global covariance advancing method aims at the field of geological attribute prediction, the prediction effect is more geological significance than the original local Krigin, and the method can provide technical support for reservoir exploration and development research.

Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.

Drawings

The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:

FIG. 1 is a flow chart of a method for geologic property prediction using global covariance, according to an embodiment of the invention;

FIG. 2 is a schematic diagram of a three-dimensional construction model according to an embodiment of the invention;

FIG. 3 is a schematic diagram illustrating the effect of a log upscaling to a three-dimensional formation model according to an embodiment of the present invention;

FIG. 4 is a schematic illustration of a geological property prediction according to an embodiment of the present invention;

FIGS. 5(a) and (b) are schematic diagrams of geological property predictions for the local kriging and global covariance methods, respectively;

fig. 6(a) and (b) are respectively an effect diagram of predicting geological properties by the local kriging method and the global covariance method when the variance is large.

Detailed Description

Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.

The invention provides a method for predicting geological properties by using global covariance. The global covariance method can be understood as a solution method of a geological property prediction equation optimized by combining the geological property prediction field on the basis of a kriging algorithm.

As shown in fig. 1, the method for predicting geologic properties by using global covariance according to the embodiment of the present invention includes the following steps:

and step S1, establishing a three-dimensional structure model according to the collected geological data.

Specifically, the collected geological data is first explained and analyzed to obtain corresponding explained geological data. Wherein the interpreted geological data comprises: fault data, horizon data and stratification data of the measured reservoir. Then, a three-dimensional formation model is built based on the interpreted geological data.

As shown in FIG. 2, the three-dimensional model may represent the geometric features of a subsurface reservoir, including multiple networks. For example, three-dimensional construction models are typically composed of millions or tens of millions of meshes. Specifically, input data of the three-dimensional structure model is a logging curve of the oil reservoir to be tested, and output data is a geological property prediction result.

And step S2, coarsening the input logging curve into the grid of the three-dimensional structure model, wherein the geological property value of the known data point on the logging curve is known as the known data point of the global covariance method.

As shown in fig. 3, the characteristic points on each log are known data points, and the geologic property values of the known data points are known. In practice, often thousands of known data are available in reservoir models, and the number of grids can be as many as the order of millions.

And step S3, according to the known data points in the step S2, predicting the geological attribute value of the oil deposit by using a global covariance method.

Firstly, it is to be noted that the global covariance method optimizes the kriging algorithm as a solution method of the geological property prediction equation based on the kriging algorithm and in combination with the geological property prediction field, so that the algorithm efficiency is improved, and the effect is more significant than that of the original local kriging method in terms of the geological property prediction field. And predicting the geological attribute value of the unknown data point according to the known data point by using a global covariance method.

Step S31, the geological property prediction equation based on the kriging method is to adopt values of linear prediction properties at positions to be estimated of a series of properties, assuming that expected values of the properties at all positions are constants, and taking the variance of the minimized error as a target, so as to solve the weight of data, and through a series of deductions, the geological property prediction equation is set as follows:

${\mathrm{\Σ}}_{j=1}^M{w}_j\·C(u_i,u_j)=C(u,u_i),i=1,\mathrm{...},M,---\left(1\right)$

wherein u is_iAnd u_jThe geologic attribute values, C (u), for the ith and jth known data points, respectively_i,u_j) Is u_iAnd u_jCovariance of (d), w_jFor weight values, M is the number of known data points. To obtain M weights wj, an M-dimensional linear system of equations (1) needs to be solved.

The system of equations (1) is rewritten as the following matrix expression:

Aw＝r, (2)

wherein Aij ═ C (u)_i,u_j)，r_i＝C(u,u_i)。

For the solution of the geological property prediction equation (1) or (2), in order to better accord with the geological property rule, the global covariance method is adopted for the solution in the step. Since the global covariance method uses all known data to predict the data of the point to be estimated, i.e. when the known point is determined, the matrix a is invariant.

Step S32, solving the geological attribute prediction equation by adopting a global covariance method, and rewriting the geological attribute prediction equation into the following form:

LUw＝r， (3)

wherein, (LU)_ij＝C(u_i,u_j)，r_i＝C(u,u_i) And L and U are a lower triangular matrix and an upper triangular matrix respectively.

Step S33, solving the LU matrix, and after obtaining the L and U matrices, solving the weight w_j。

Specifically, the L and U matrices are:

wherein,

$U_{rj}=a_{rj}-\frac{{\mathrm{\Σ}}_{k=1}^{r-1}{l}_{ik}{U}_{kj}}{l_{rr}},(r=0,1,2,\mathrm{....}n-1;j=r,\mathrm{....}n-1),$

$L_{ir}=a_{ir}-\frac{{\mathrm{\Σ}}_{k=1}^{r-1}{l}_{ik}{u}_{kr}}{u_{rr}},(r=0,1,\mathrm{2...},n-1;i=r+1,\mathrm{...},n-1).$

it should be noted that, the calculation of the L and U matrices is a cyclic iterative process, the U matrix is iterated from row 1 to row n, the L matrix is iterated from column 1 to column n, and the U matrix is calculated before the L matrix in each iteration.

After calculating the L and U matrices according to the above process, the weight w is solved by equation (3)_j

In step S34, the weight w is learned_jAnd predicting the geological property value of the unknown data point in the three-dimensional structure model, and recording as a geological property prediction result.

In one embodiment of the invention, the geological property prediction may include: and the porosity, saturation and permeability corresponding to the detected oil reservoir. For example, FIG. 4 shows a predictive value diagram of a reservoir porosity attribute.

In equation (3), compared to O (M) required for conventional solution Aw ═ r³) The operand, global covariance method solution can be in O (M)²) Is completed within the sub-operation. O (M) decomposing LU³) Taking time consumption into consideration, in a grid containing N unknown data points, the operation complexity of the geological attribute prediction equation is calculated by using a global covariance method, wherein the operation complexity is O (N.M)³) Reduced to O (M)³+N·M²)。

The following describes a comparison between the conventional local kriging method and the global covariance method of the present invention with reference to fig. 5(a) and 5(b), and fig. 6(a) and 6 (b).

Fig. 5(a) and (b) are schematic diagrams of geological property predictions for the local kriging and global covariance methods, respectively. Where fig. 5(a) is generated from local kriging using near 200 known data. As can be seen from fig. 5(a), there are many discontinuities (circles) in the three-dimensional structure model. The three-dimensional structure model shown in fig. 5(b) is obtained by the global covariance method, and the continuity is better than that of the local kriging method.

Fig. 6(a) and (b) are respectively an effect diagram of predicting geological properties by the local kriging method and the global covariance method when the variance is large. As can be seen by comparing the two graphs, the global covariance method adopted in FIG. 6(b) has stronger continuity of geological properties and better conforms to geological rules.

The following describes the comparison of the efficiency of the existing global and local kriging method and the global covariance method of the present invention with reference to tables 1 and 2. Wherein table 1 gives the complexity of the three methods. As can be seen from table 1, the global covariance approach of the present invention is the least complex.

Global covariance method	Local kriging algorithm	Global kriging algorithm
			O(NM)	O(NM log M+NK3)	O(NM3)

TABLE 1

Table 2 gives the algorithm running times of the global and local kriging methods and the global covariance method on three-dimensional grids of different scales.

Known point number	Size of the grid	Global covariance method	Local kriging algorithm	Global kriging algorithm
					281	40 × 64 × 8 ═ 2 ten thousand	<1sec	3sec	143sec
584	80 × 128 × 16 ═ 16 ten thousand	4sec	59sec	2.6hr
					1166	160 × 256 × 32 ═ 130 ten thousand	64sec	1076sec	142hr
2315	320 × 512 × 64 ═ 1000 ten thousand	944sec	5.0hr	/

TABLE 2

As can be seen from table 2, all the methods run with a linear increase in the number of grids, given that the number of data points remains the same. But under the same conditions, the global covariance method of the invention takes the shortest time. When the number of the three-dimensional grids is more than ten million, the kriging method takes more than 5 hours, the global covariance method only takes 944 seconds, and the algorithm efficiency is greatly improved. Therefore, the geological attribute prediction equation solved by using the global covariance method not only meets the requirement of geological attribute prediction practical problems, but also can greatly improve the solving speed.

According to the method for predicting the geological property by using the global covariance, the geological property equation is solved by using the global covariance method, when the number of geological property grids is large, the speed of solving the geological property prediction equation is greatly improved compared with the speed of solving the original Krigin solution method, and the speed of solving the geological property prediction equation is improved by more than 3 times. The global covariance method provided by the invention aims at the field of geological attribute prediction, the prediction effect is more geological significance than the original local kriging, and a technical support can be provided for reservoir exploration and development research.

In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made in the above embodiments by those of ordinary skill in the art without departing from the principle and spirit of the present invention. The scope of the invention is defined by the appended claims and their full range of equivalents.

Claims (5)

1. A method for predicting geological properties by using global covariance is characterized by comprising the following steps:

step S1, establishing a three-dimensional structure model according to the collected geological data, wherein the input data of the three-dimensional structure model is a logging curve of an oil deposit to be detected, the output data is a geological attribute prediction result, and the three-dimensional structure model comprises a plurality of grids;

step S2, coarsening the input logging curve into the grid as the known data point of the global covariance method, wherein the geological attribute value of the known data point on the logging curve is known;

and step S3, according to the known data points in the step S2, predicting the geological attribute value of the oil deposit by using a global covariance method.

2. The method for geologic property prediction using global covariance of claim 1 wherein, in step S1, the step of building a three-dimensional structure model from the collected geologic data comprises the steps of:

interpreting and analyzing the collected geological data to obtain corresponding interpreted geological data, wherein the interpreted geological data comprises: fault data, horizon data and layering data of the detected oil reservoir;

and establishing the three-dimensional construction model according to the interpreted geological data.

3. The method of claim 1, wherein the geologic property prediction result comprises: and the porosity, saturation and permeability corresponding to the detected oil reservoir.

4. The method for geologic property prediction employing global covariance of claim 1 wherein, in step S3,

firstly, let the geological property prediction equation be as follows:

$\underset{j=1}{\overset{M}{\Σ}}{w}_j\·C(u_i,u_j)=C(u,u_i),i=1,...,M,$

wherein u is_iAnd u_jThe geologic attribute values, C (u), for the ith and jth known data points, respectively_i,u_j) Is u_iAnd u_jCovariance of (d), w_jFor weight value, M is the number of the known data points;

then, solving the geological property prediction equation by adopting a global covariance method, and rewriting the geological property prediction equation into the following form:

LUw＝r，

wherein, (LU)_ij＝C(u_i,u_j)，r_i＝C(u,u_i) L and U are a lower triangular matrix and an upper triangular matrix respectively;

then, the LU matrix is solved, and after the L matrix and the U matrix are obtained, the weight w is solved_j，

Finally, knowing the weight w_jAnd predicting the geological property value of the unknown data point in the three-dimensional structure model, and recording as a geological property prediction result.

5. The method for geologic property prediction employing global covariance of claim 4 wherein the L and U matrices are:

wherein, U_0j＝a_0j,(j＝0,1,2,….n-1)，

$U_{rj}=a_{rj}-\frac{{\mathrm{\Σ}}_{k=1}^{r-1}{l}_{ik}{U}_{kj}}{l_{rr}},(r=0,1,2,\mathrm{....}n-1;j=r,\mathrm{....}n-1),$

$L_{ir}=a_{ir}-\frac{{\mathrm{\Σ}}_{k=1}^{r-1}{l}_{ik}{u}_{kr}}{u_{rr}},(r=0,1,2...,n-1;i=r+1,...,n-1).$