CN108416816B

CN108416816B - Compression processing method and device for multi-dimensional nuclear magnetic resonance logging data

Info

Publication number: CN108416816B
Application number: CN201810200332.2A
Authority: CN
Inventors: 谢然红; 谷明宣; 郭江峰; 金国文; 高伦
Original assignee: China University of Petroleum Beijing
Current assignee: China University of Petroleum Beijing
Priority date: 2018-03-12
Filing date: 2018-03-12
Publication date: 2021-01-29
Anticipated expiration: 2038-03-12
Also published as: CN108416816A

Abstract

The invention provides a compression processing method and a compression processing device for multi-dimensional nuclear magnetic resonance logging data, wherein the method comprises the following steps: acquiring multi-dimensional nuclear magnetic resonance logging echo string data and a multi-dimensional nuclear magnetic resonance logging nuclear matrix; carrying out preliminary compression processing on echo string data by adopting a window averaging method, and carrying out preliminary compression processing on a kernel matrix by adopting the window averaging method; performing one-dimensional discrete cosine transform on the preliminarily compressed echo string data to obtain a first discrete cosine transform coefficient, and acquiring first low-frequency data; performing two-dimensional discrete cosine transform on the preliminarily compressed kernel matrix to obtain a second discrete cosine transform coefficient and obtain second low-frequency data; and performing one-dimensional inverse discrete cosine transform on the first low-frequency data, and performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain compressed multi-dimensional nuclear magnetic resonance logging data. The compression method is high in compression ratio, the calculated amount of the inversion process of the nuclear magnetic resonance logging data is reduced, and the inversion speed is improved.

Description

Compression processing method and device for multi-dimensional nuclear magnetic resonance logging data

Technical Field

The invention relates to the technical field of oil and gas exploration, in particular to a compression processing method and device for multidimensional nuclear magnetic resonance logging data.

Background

In oil and gas exploration, a nuclear magnetic resonance logging instrument is required to measure an oil well or a stratum so as to obtain reservoir pore fluid information. The multi-dimensional nuclear magnetic resonance logging, such as two-dimensional nuclear magnetic resonance logging or three-dimensional nuclear magnetic resonance logging, can simultaneously measure the transverse relaxation time T2, the longitudinal relaxation time T1 and the fluid diffusion coefficient D of reservoir pore fluid, is obviously superior to the one-dimensional nuclear magnetic resonance logging in the aspect of reservoir evaluation, but the multi-dimensional nuclear magnetic resonance logging data volume is huge, the inversion calculation volume of the multi-dimensional nuclear magnetic resonance logging data is very large, and the inversion speed is further influenced. Therefore, in order to realize real-time processing of the multidimensional nuclear magnetic resonance logging data, the multidimensional nuclear magnetic resonance logging data needs to be compressed before inversion.

In the prior art, a window averaging method may be adopted to compress multidimensional nuclear magnetic resonance logging data, specifically, multiple sets of echo string data of the multidimensional nuclear magnetic resonance logging are separately compressed, that is, a single set of echo string data is divided into several windows, then the echo data in each window are summed, and further the multidimensional nuclear magnetic resonance logging echo string data is compressed, and finally a multidimensional nuclear magnetic resonance logging matrix is compressed.

However, in the prior art, the window averaging method has a low compressibility for the multi-dimensional nuclear magnetic resonance logging data, and in order to ensure that useful information of the formation is not lost in the compressed data, a large compression value needs to be set, so that the inversion speed of the compressed multi-dimensional nuclear magnetic resonance logging data is low.

Disclosure of Invention

The invention provides a compression processing method and a compression processing device for multidimensional nuclear magnetic resonance logging data, which are used for solving the problems that the compression rate of the compressed multidimensional nuclear magnetic resonance logging data is low and the inversion speed of the compressed multidimensional nuclear magnetic resonance logging data is low.

In one aspect, the present invention provides a method for compressing multidimensional nuclear magnetic resonance logging data, including:

acquiring multi-dimensional nuclear magnetic resonance logging echo string data, and constructing a multi-dimensional nuclear magnetic resonance logging nuclear matrix according to the multi-dimensional nuclear magnetic resonance logging echo string data;

performing preliminary compression processing on the multidimensional nuclear magnetic resonance logging echo string data by adopting a window averaging method to obtain preliminarily compressed multidimensional nuclear magnetic resonance logging echo string data, and performing preliminary compression processing on the multidimensional nuclear magnetic resonance logging matrix by adopting the window averaging method to obtain the preliminarily compressed multidimensional nuclear magnetic resonance logging matrix;

performing one-dimensional discrete cosine transform on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data to obtain a first discrete cosine transform coefficient, and acquiring low-frequency data in the first discrete cosine transform coefficient to obtain first low-frequency data;

performing two-dimensional discrete cosine transform on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix to obtain a second discrete cosine transform coefficient, and acquiring low-frequency data in the second discrete cosine transform coefficient to obtain second low-frequency data;

and performing one-dimensional inverse discrete cosine transformation on the first low-frequency data, and performing two-dimensional inverse discrete cosine transformation on the second low-frequency data to obtain compressed multi-dimensional nuclear magnetic resonance logging data, wherein the compressed multi-dimensional nuclear magnetic resonance logging data comprises data obtained by performing one-dimensional inverse discrete cosine transformation on the first low-frequency data and data obtained by performing two-dimensional inverse discrete cosine transformation on the second low-frequency data.

In another aspect, the present invention provides a device for compressing multidimensional nmr logging data, comprising:

the acquisition module is used for acquiring multi-dimensional nuclear magnetic resonance logging echo string data and constructing a multi-dimensional nuclear magnetic resonance logging nuclear matrix according to the multi-dimensional nuclear magnetic resonance logging echo string data;

the first compression module is used for performing preliminary compression processing on the multi-dimensional nuclear magnetic resonance logging echo string data by adopting a window averaging method to obtain preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data, and performing preliminary compression processing on the multi-dimensional nuclear magnetic resonance logging nuclear matrix by adopting the window averaging method to obtain the preliminarily compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix;

the first transformation module is used for performing one-dimensional discrete cosine transformation on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data to obtain a first discrete cosine transformation coefficient, and acquiring low-frequency data in the first discrete cosine transformation coefficient to obtain first low-frequency data;

the second transformation module is used for performing two-dimensional discrete cosine transformation on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix to obtain a second discrete cosine transformation coefficient, and acquiring low-frequency data in the second discrete cosine transformation coefficient to obtain second low-frequency data;

and the second compression module is used for performing one-dimensional inverse discrete cosine transform on the first low-frequency data and performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain compressed multi-dimensional nuclear magnetic resonance logging data, wherein the compressed multi-dimensional nuclear magnetic resonance logging data comprise data obtained by performing one-dimensional inverse discrete cosine transform on the first low-frequency data and data obtained by performing two-dimensional inverse discrete cosine transform on the second low-frequency data.

The invention provides a compression processing method and a device of multidimensional nuclear magnetic resonance logging data, which utilize a window averaging method to carry out primary compression on original multidimensional nuclear magnetic resonance logging echo string data and a multidimensional nuclear magnetic resonance logging matrix, then transform the data after the primary compression to a frequency domain space to obtain a discrete cosine transform coefficient, intercept the low-frequency part of the discrete cosine transform coefficient to reserve the main characteristics of the original multidimensional nuclear magnetic resonance logging data, then carry out inverse discrete cosine transform on the low-frequency part to finally obtain the compressed multidimensional nuclear magnetic resonance logging data, and realize secondary compression on the multidimensional nuclear magnetic resonance logging data; thereby completing the compression process of the multi-dimensional nuclear magnetic resonance logging data. According to the scheme provided by the embodiment, the multidimensional nuclear magnetic resonance logging data can be well compressed, and the compression rate is high, so that when the compressed multidimensional nuclear magnetic resonance logging data is inverted, the calculated amount in the inversion process is reduced, and the inversion speed is increased; in addition, compared with the existing singular value truncation method, the scheme provided by the embodiment has the advantages of small calculation amount and fast compression process.

Drawings

Fig. 1 is a schematic flowchart of a method for compressing multidimensional nmr logging data according to an embodiment of the present invention;

FIG. 2 is a schematic flow chart of another method for compressing multi-dimensional NMR logging data according to an embodiment of the present invention;

FIG. 3 is a first diagram illustrating a simulated T1-T2-D distribution in another method for compressing data from a multi-dimensional NMR well log according to an embodiment of the invention;

FIG. 4 is a second diagram illustrating simulated T1-T2-D distribution in another method for compressing data from a multi-dimensional NMR well log according to an embodiment of the invention;

FIG. 5 is a third diagram illustrating a simulated T1-T2-D distribution in another method for compressing data from a multi-dimensional NMR well log according to an embodiment of the invention;

fig. 6 is a schematic diagram of a plurality of echo trains of three-dimensional nmr logging data in another method for compressing the nmr logging data according to the embodiment of the present invention;

fig. 7 is a schematic diagram of three-dimensional nmr logging echo train data after preliminary compression of three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to an embodiment of the present invention;

fig. 8 is a schematic diagram of first low-frequency data of three-dimensional nmr logging data in another method for compressing the multidimensional nmr logging data according to the embodiment of the invention;

fig. 9 is a schematic diagram of three-dimensional nmr logging echo train data after final compression of three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to an embodiment of the present invention;

fig. 10 is a first diagram illustrating inversion results of original three-dimensional nmr logging data in another method for compressing nmr logging data according to an embodiment of the present invention;

fig. 11 is a second graph of an inversion result of original three-dimensional nmr logging data in another method for compressing nmr logging data according to the embodiment of the present invention;

fig. 12 is a third diagram illustrating an inversion result of original three-dimensional nmr logging data in another method for compressing nmr logging data according to an embodiment of the present invention;

fig. 13 is a first graph of inversion results of compressed three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to the embodiment of the present invention;

fig. 14 is a second graph of an inversion result of compressed three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to the embodiment of the present invention;

fig. 15 is a third diagram illustrating an inversion result of compressed three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to the embodiment of the present invention;

FIG. 16 is a schematic diagram of a simulated T2-D distribution in another method for compressing multi-dimensional NMR logging data according to an embodiment of the invention;

fig. 17 is a schematic diagram of a plurality of echo trains of two-dimensional nmr logging data in another method for compressing the multidimensional nmr logging data according to the embodiment of the present invention;

fig. 18 is a schematic diagram of two-dimensional nmr logging echo train data after preliminary compression of two-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to an embodiment of the present invention;

fig. 19 is a schematic diagram of first low-frequency data of two-dimensional nmr log data in another method for compressing md log data according to an embodiment of the present invention;

fig. 20 is a schematic diagram of two-dimensional nmr logging echo train data after final compression of two-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to an embodiment of the present invention;

fig. 21 is a diagram illustrating an inversion result of original two-dimensional nmr logging data in another method for compressing nmr logging data according to an embodiment of the present invention;

fig. 22 is a diagram of an inversion result of compressed two-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to the embodiment of the present invention;

fig. 23 is a schematic structural diagram of a multi-dimensional nmr logging data compression apparatus according to an embodiment of the present invention;

fig. 24 is a schematic structural diagram of another apparatus for compressing multidimensional nmr log data according to an embodiment of the present invention.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

The specific application scenario of the present invention is as follows. The invention relates to the technical field of well logging data processing in oil and gas exploration. In oil and gas exploration, a nuclear magnetic resonance device is needed to measure an oil well or a stratum, namely, a nuclear magnetic resonance logging process is carried out to obtain multidimensional nuclear magnetic resonance logging data, wherein the multidimensional nuclear magnetic resonance logging data is two-dimensional nuclear magnetic resonance logging data or three-dimensional nuclear magnetic resonance logging data, and then the multidimensional nuclear magnetic resonance logging data is subjected to inversion processing to analyze information of the stratum. However, hundreds of echo data are acquired in the nuclear magnetic resonance logging process, and particularly, the data volume of multi-dimensional nuclear magnetic resonance logging data is huge; when the acquired multi-dimensional nuclear magnetic resonance logging data are directly inverted, the calculated amount is very large, the inversion speed is further influenced, a very large computer memory is needed, and the calculation speed is very low; moreover, the acquired multi-dimensional nuclear magnetic resonance logging data contains a lot of redundant information, so that the data volume of the multi-dimensional nuclear magnetic resonance logging data is further increased. Therefore, the multidimensional nuclear magnetic resonance logging data needs to be compressed, and then the compressed multidimensional nuclear magnetic resonance logging data needs to be inverted. If data compression is not carried out, a large amount of multidimensional nuclear magnetic resonance logging data are directly inverted, so that a very large computer memory is needed, and the calculation speed is seriously influenced. Data compression is necessary before inverting the acquired multi-dimensional nmr logging data. At present, the compression method of multidimensional nuclear magnetic resonance logging data mainly comprises the following steps: window averaging, singular value truncation. The window averaging method is to divide the echo string data of the multi-dimensional nuclear magnetic resonance logging into a plurality of windows, then sum the echo string data in each window respectively, and further compress the echo string data. Specifically, echo strings in multidimensional nuclear magnetic resonance logging data are in an exponential decay rule, and the echo strings are usually divided into a plurality of windows at logarithmic equal intervals in a time domain; the first 3 echoes are not compressed normally when the window is segmented because the first echoes play a very important role in short relaxation and porosity evaluation; when a plurality of echo strings are compressed, each echo string is compressed by a window averaging method, and then all the compressed echoes are stacked together. The window averaging method has lower compressibility for multidimensional nuclear magnetic resonance logging data. The singular value truncation method is to perform singular value truncation on a nuclear matrix formed by multi-dimensional nuclear magnetic resonance logging echo string data so as to realize data compression. Specifically, most of singular values of a kernel matrix of the multi-dimensional nuclear magnetic resonance logging echo string data tend to be zero, so that only the first few large singular values need to be reserved, and then compression processing is completed. However, when the singular value truncation method is adopted to compress the multidimensional nuclear magnetic resonance logging data, the singular value decomposition of the nuclear matrix is required, so that the calculated amount is large, and the compression processing process is slow.

The invention provides a compression processing method and a compression processing device for multi-dimensional nuclear magnetic resonance logging data, and aims to solve the technical problems in the prior art.

The following describes the technical solution of the present invention and how to solve the above technical problems with specific examples. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present invention will be described below with reference to the accompanying drawings.

Fig. 1 is a schematic flow chart of a method for compressing multidimensional nuclear magnetic resonance logging data according to an embodiment of the present invention. As shown in fig. 1, the method includes:

step 101, obtaining multi-dimensional nuclear magnetic resonance logging echo string data, and constructing a multi-dimensional nuclear magnetic resonance logging nuclear matrix according to the multi-dimensional nuclear magnetic resonance logging echo string data.

In this embodiment, the execution subject of this embodiment may be a compression processing device of multidimensional nuclear magnetic resonance logging data, a computer, or any other device that can execute the method of this embodiment.

Firstly, acquiring multidimensional nuclear magnetic resonance logging data of a stratum by using a nuclear magnetic resonance logging instrument, wherein the multidimensional nuclear magnetic resonance logging data are two-dimensional nuclear magnetic resonance logging data, or the multidimensional nuclear magnetic resonance logging data are three-dimensional nuclear magnetic resonance logging data. The multidimensional nuclear magnetic resonance logging data are multidimensional nuclear magnetic resonance logging echo string data, and a multidimensional nuclear magnetic resonance logging nuclear matrix can be constructed according to the multidimensional nuclear magnetic resonance logging echo string data; the multi-dimensional nuclear magnetic resonance logging nuclear matrix is A, the multi-dimensional nuclear magnetic resonance logging nuclear matrix is a matrix with m rows and n columns, and n is the number of multi-dimensional nuclear magnetic resonance logging echoes; the multi-dimensional nuclear magnetic resonance logging echo string data is b, and the echo string data b comprises m rows of data; m is a positive integer and n is a positive integer. For example, two-dimensional nuclear magnetic resonance logging echo string data is acquired, and a two-dimensional nuclear magnetic resonance logging nuclear matrix is constructed according to the two-dimensional nuclear magnetic resonance logging echo string data; and acquiring three-dimensional nuclear magnetic resonance logging echo string data, and constructing a three-dimensional nuclear magnetic resonance logging nuclear matrix according to the three-dimensional nuclear magnetic resonance logging echo string data.

And 102, performing primary compression processing on the multi-dimensional nuclear magnetic resonance logging echo string data by adopting a window averaging method to obtain the multi-dimensional nuclear magnetic resonance logging echo string data after the primary compression, and performing primary compression processing on the multi-dimensional nuclear magnetic resonance logging nuclear matrix by adopting the window averaging method to obtain the multi-dimensional nuclear magnetic resonance logging nuclear matrix after the primary compression.

In an alternative embodiment, the first dimension of the preliminarily compressed multi-dimensional NMR logging echo train data bc1i elements of

The element in the ith row and the jth column in the preliminarily compressed multidimensional nuclear magnetic resonance logging nuclear matrix Ac1 is

Wherein N is_iDividing the multi-dimensional nuclear magnetic resonance logging echo string data into s windows and then counting the number of echoes in the ith window, N_iIs a positive integer, s is a positive integer greater than or equal to 1, i belongs to [1, s ]]I is an integer; the total number of echoes in the s windows is m, and m is N₁+…+N_i-1+N_i+…+N_sM is a positive integer; when i is 0, r₁When 0 ≦ i ≦ s, r_i＝N₁+…+N_i-1；k∈[r_i+1,r_i+N_i]K is an integer; a. the_kjThe method is characterized in that elements in the kth row and the jth column in a multidimensional nuclear magnetic resonance logging core matrix A are included, the multidimensional nuclear magnetic resonance logging core matrix A is a matrix with m rows and n columns, n is the number of multidimensional nuclear magnetic resonance logging echoes, n is a positive integer, j belongs to [1, n ∈]J is an integer; b_kThe method comprises the steps that elements on the kth line in multi-dimensional nuclear magnetic resonance logging echo string data b are obtained, and the echo string data b comprises m-line data; epsilon_kThe preset noise data E includes m lines of data as elements on the k-th line in the preset noise data E. The preset noise data E is white gaussian noise having a noise level of 0.1 pu.

In this embodiment, specifically, first, the multi-dimensional nmr log echo train data has m echoes, the m echoes are divided into s windows, and the number of the echoes in the ith window is N_i，N_iIs a positive integer, i.e. N in the ith window_iAn echo, and m is equal to N₁+…+N_i-1+N_i+…+N_sS is a positive integer greater than or equal to 1, i belongs to [1, s ]]And i is an integer. And, preset noise data E including m line data is set in advance. The echoes in each window are then summed, followed by the preliminary compressed multi-dimensional nuclear magnetic resonance logging echo train data bc1, anAnd (4) carrying out multi-dimensional nuclear magnetic resonance logging on a nuclear matrix Ac1 after the initial compression.

In the formula

In (A)_kjFor the elements in the kth row and the jth column in the multi-dimensional nuclear magnetic resonance logging nuclear matrix A, j belongs to [1, n ]]And j is an integer. In the formula

In (b)_kFor the elements on the k-th line, epsilon, in the multi-dimensional NMR well-logging echo train data b_kThe elements on the k-th line in the noise data E are preset.

Before compressing the multidimensional nuclear magnetic resonance logging data by a window averaging method, the number of rows of a multidimensional nuclear magnetic resonance logging nuclear matrix A is m, and the number of rows in multidimensional nuclear magnetic resonance logging echo string data b is m; after compressing the multi-dimensional nuclear magnetic resonance logging data by a window average method, the number of rows of the preliminarily compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix Ac1 is s, and the number of rows of the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data bc1 is s; and compressing the number of rows of the multi-dimensional nuclear magnetic resonance logging nuclear matrix and the multi-dimensional nuclear magnetic resonance logging echo string data from m to s. When inverting compressed multi-dimensional NMR logging data, the inverse problem can be converted from A_m×nf_n×1＝b_m×1+ε_m×1Conversion to Ac1_s×nf_n×1＝bc1_s×1Carry out the solution A_m×nLogging nuclear matrices A, b for multi-dimensional NMR_m×1For multi-dimensional NMR well-logging echo-string data b, epsilon_m×1For presetting the noise data E, Ac1_s×nIs a multi-dimensional nuclear magnetic resonance logging nuclear matrix Ac1, bc1 after primary compression_s×1Is the primarily compressed multi-dimensional nuclear magnetic resonance logging echo string data bc1, f_n×1Is an inverse model.

103, performing one-dimensional discrete cosine transform on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data to obtain a first discrete cosine transform coefficient, and obtaining low-frequency data in the first discrete cosine transform coefficient to obtain first low-frequency data.

In this embodiment, specifically, one-dimensional discrete cosine transform is performed on the preliminarily compressed multidimensional nuclear magnetic resonance logging echo string data bc1, so that a transform coefficient of the multidimensional nuclear magnetic resonance logging echo string data compressed by a window averaging method can be obtained, that is, a first discrete cosine transform coefficient is obtained; and then intercepting a low-frequency part in a transformation coefficient of the multi-dimensional nuclear magnetic resonance logging echo string data compressed by the window averaging method to obtain first low-frequency data.

And 104, performing two-dimensional discrete cosine transform on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging core matrix to obtain a second discrete cosine transform coefficient, and acquiring low-frequency data in the second discrete cosine transform coefficient to obtain second low-frequency data.

In this embodiment, specifically, two-dimensional discrete cosine transform is performed on the preliminarily compressed multidimensional nuclear magnetic resonance logging matrix Ac1 to obtain a transform coefficient of the multidimensional nuclear magnetic resonance logging matrix compressed by a window averaging method, so as to obtain a second discrete cosine transform coefficient; and then intercepting a low-frequency part in the transformation coefficient of the multidimensional nuclear magnetic resonance logging nuclear matrix compressed by the window averaging method to obtain second low-frequency data. The execution order of step 103 and step 104 is not limited.

And 105, performing one-dimensional inverse discrete cosine transform on the first low-frequency data, and performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain compressed multi-dimensional nuclear magnetic resonance logging data, wherein the compressed multi-dimensional nuclear magnetic resonance logging data comprises data obtained by performing one-dimensional inverse discrete cosine transform on the first low-frequency data and data obtained by performing two-dimensional inverse discrete cosine transform on the second low-frequency data.

In this embodiment, specifically, a one-dimensional inverse discrete cosine transform is performed on the first low-frequency data to obtain finally compressed multi-dimensional nuclear magnetic resonance logging echo string data, that is, the one-dimensional inverse discrete cosine transform-processed first low-frequency data is obtained; meanwhile, performing two-dimensional inverse discrete cosine transformation on the second low-frequency data to obtain a finally compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix, and thus obtaining the second low-frequency data subjected to the two-dimensional inverse discrete cosine transformation; therefore, compressed multi-dimensional nuclear magnetic resonance logging data can be obtained, wherein the compressed multi-dimensional nuclear magnetic resonance logging data comprise the finally compressed multi-dimensional nuclear magnetic resonance logging echo string data and the finally compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix, namely the compressed multi-dimensional nuclear magnetic resonance logging data comprise data obtained by performing one-dimensional inverse discrete cosine transform on the first low-frequency data and data obtained by performing two-dimensional inverse discrete cosine transform on the second low-frequency data.

In the embodiment, the original multi-dimensional nuclear magnetic resonance logging echo string data and the multi-dimensional nuclear magnetic resonance logging nuclear matrix are preliminarily compressed by using a window averaging method, then the preliminarily compressed data are converted into a frequency domain space to obtain a discrete cosine transform coefficient, then the low frequency part of the discrete cosine transform coefficient is intercepted to retain the main characteristics of the original multi-dimensional nuclear magnetic resonance logging data, then the low frequency part is subjected to inverse discrete cosine transform, and finally the compressed multi-dimensional nuclear magnetic resonance logging data are obtained, so that the secondary compression of the multi-dimensional nuclear magnetic resonance logging data is realized; thereby completing the compression process of the multi-dimensional nuclear magnetic resonance logging data. According to the scheme provided by the embodiment, the multidimensional nuclear magnetic resonance logging data can be well compressed, and the compression rate is high, so that when the compressed multidimensional nuclear magnetic resonance logging data is inverted, the calculated amount in the inversion process is reduced, and the inversion speed is increased; in addition, compared with the existing singular value truncation method, the scheme provided by the embodiment has the advantages of small calculation amount and fast compression process.

Fig. 2 is a schematic flow chart of another method for compressing multidimensional nuclear magnetic resonance logging data according to an embodiment of the present invention. As shown in fig. 2, the method includes:

step 201, obtaining multi-dimensional nuclear magnetic resonance logging echo string data, and constructing a multi-dimensional nuclear magnetic resonance logging nuclear matrix according to the multi-dimensional nuclear magnetic resonance logging echo string data.

In this embodiment, specifically, this step may refer to step 101 in fig. 1, and is not described again.

Step 202, performing preliminary compression processing on the multidimensional nuclear magnetic resonance logging echo string data by using a window averaging method to obtain preliminarily compressed multidimensional nuclear magnetic resonance logging echo string data, and performing preliminary compression processing on a multidimensional nuclear magnetic resonance logging nuclear matrix by using the window averaging method to obtain the preliminarily compressed multidimensional nuclear magnetic resonance logging nuclear matrix.

In this embodiment, specifically, this step may refer to step 102 in fig. 1, and is not described again.

Step 203, performing one-dimensional discrete cosine transform on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data to obtain a first discrete cosine transform coefficient, and intercepting the first q data in the first discrete cosine transform coefficient to obtain first low-frequency data, wherein q is not more than s, and q is an integer.

Wherein, when u is 0, the first discrete cosine transform coefficient is bc1'(u) bc1' (0), wherein,

when u is 1,2, … s-1, the first discrete cosine transform coefficient is

Wherein bc1(x) is the x-th element in the multi-dimensional nuclear magnetic resonance logging echo string data bc1 after preliminary compression, and x belongs to [0, s-1]]X is an integer, pi is a circumference ratio, u belongs to [0, s-1]]And u is an integer.

In this embodiment, specifically, the one-dimensional discrete cosine transform is performed on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data bc1 to obtain a transform coefficient of the multi-dimensional nuclear magnetic resonance logging echo string data compressed by the window averaging method, and the transform coefficient of the multi-dimensional nuclear magnetic resonance logging echo string data compressed by the window averaging method may be referred to as a first discrete cosine transform coefficient bc1'_s×1. When u is 0, the first discrete cosine transform coefficient is

When u is 1,2, …, s-1, the first discrete cosine transform coefficient is

Wherein s is the stepThe number of windows in step 202, s is also the number of echoes after the multi-dimensional nuclear magnetic resonance logging echo string data is compressed by using a window averaging method in step 202; bc1(x) is the x-th element in the multi-dimensional nuclear magnetic resonance logging echo string data bc1 after preliminary compression, and x belongs to [0, s-1]]X is an integer, pi is a circumference ratio, u belongs to [0, s-1]]And u is an integer. Moreover, the primarily compressed multi-dimensional nuclear magnetic resonance logging echo string data bc1 can adopt bc1_s×1And (4) showing.

Then, the first q data in the first discrete cosine transform coefficient bc1' (u) are intercepted, and the first q data in the first discrete cosine transform coefficient bc1' (u) are low-frequency data, so that the first q low-frequency parts in the first discrete cosine transform coefficient bc1' (u) are intercepted, and the first low-frequency data are obtained, wherein q is less than or equal to s, and q is an integer. The first low frequency data may be bc2'_q×1This is expressed, or the first low frequency data is bc2'(u ″) (bc 1' (u ″), where u ″ -0, 1, …, q-1, q ≦ s, and q is an integer.

And 204, performing two-dimensional discrete cosine transform on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging kernel matrix to obtain a second discrete cosine transform coefficient, and intercepting the first q rows of data in the second discrete cosine transform coefficient to obtain second low-frequency data, wherein q is not more than s, and q is an integer.

Wherein the second discrete cosine transform coefficient is Ac1'_s×nAnd the second discrete cosine transform coefficient Ac1'_s×nThe element in the u-th row and the v-th column in (1) is

Pi is the circumference ratio, x belongs to [0, s-1]]X is an integer, u is an element of [0, s-1]]U is an integer, y is an element [0, n-1 ]]Y is an integer, v is an element [0, n-1 ]]V is an integer, and Ac1(x, y) is an element on the y column of the x row in the preliminarily compressed multidimensional NMR logging nuclear matrix Ac 1; when u is equal to 0, the reaction is carried out,

when u is 1,2, …, s-1, c (u) is 1; when v is equal to 0, the voltage is set to 0,

when v is 1,2, …, n-1, c (v) is 1.

In this embodiment, specifically, two-dimensional discrete cosine transform is performed on the preliminarily compressed multidimensional nmr logging core matrix Ac1 to obtain a transform coefficient of the multidimensional nmr logging core matrix compressed by the window averaging method, and the transform coefficient of the multidimensional nmr logging core matrix compressed by the window averaging method may be referred to as a second discrete cosine transform coefficient Ac1'_s×n。

Wherein the second discrete cosine transform coefficient Ac1'_s×nThe element in the u-th row and the v-th column in (1) is

Pi is the circumference ratio, x belongs to [0, s-1]]X is an integer, u is an element of [0, s-1]]U is an integer, y is an element [0, n-1 ]]Y is an integer, v is an element [0, n-1 ]]V is an integer, and Ac1(x, y) is the element on the y column at the x row in the preliminarily compressed multidimensional NMR log nuclear matrix Ac 1. It is known that x is 0,1, …, s-1, u is 0,1, …, s-1, y is 0,1, …, n-1, v is 0,1, …, n-1.

Wherein, when u is 0,

when v is 1,2, …, n-1, c (v) is 1.

Then, the second discrete cosine transform coefficient Ac1 'is truncated'_s×nThe front q line data of (1) is the second discrete cosine transform coefficient Ac1'_s×nThe front q line data in the sequence is low-frequency data, so that a second discrete cosine transform coefficient Ac1 'is intercepted'_s×nThe low frequency part of the front q lines in the sequence is obtained as second low frequency data Ac2'_q×n(ii) a It can be known that the truncated is the second discrete cosine transform coefficient Ac1'_s×nThe first q rows, all columns of data in (1). The second low-frequency data is Ac2'(u ", v) ═ Ac1' (u", v), where u ═ 0,1, …, q-1, q ≦ s, q is an integer, and v ═ 0,1, …, n-1.

And 205, performing one-dimensional inverse discrete cosine transform on the first low-frequency data, and performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain compressed multi-dimensional nuclear magnetic resonance logging data, wherein the compressed multi-dimensional nuclear magnetic resonance logging data comprises data obtained by performing one-dimensional inverse discrete cosine transform on the first low-frequency data and data obtained by performing two-dimensional inverse discrete cosine transform on the second low-frequency data.

Step 205 specifically includes: step 2051, performing one-dimensional inverse discrete cosine transform on the first low-frequency data to obtain finally compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2_q×1Wherein the final compressed multi-dimensional NMR logging echo string data bc2_q×1The data of the first low-frequency data after one-dimensional inverse discrete cosine transform and the finally compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2_q×1The x' +1 th element in (A) is

x′∈[0,q-1]X ' is an integer, bc2' (u ') is the u ' +1 th element in the first low frequency data, u ' is e [0, q-1]U' is an integer; when u' is equal to 0, the reaction is carried out,

when u 'is 1,2, …, q-1, C (u') is 1. Step 2052, performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain a final compressed multidimensional nuclear magnetic resonance logging nuclear matrix Ac2_q×nWherein the final compressed multi-dimensional NMR logging nuclear matrix Ac2_q×nThe data of the second low-frequency data after two-dimensional inverse discrete cosine transform and the finally compressed multidimensional nuclear magnetic resonance logging nuclear matrix Ac2_q×nThe element in the x '+ 1 th row and y' +1 th column in (1) is

x′∈[0,q-1]X 'is an integer, y' belongs to [0, n-1 ]]Y 'is an integer, u' belongs to [0, q-1 ]]U 'is an integer, v' is an element [0, n-1 ]]V 'is an integer, Ac2' (u ', v') is the secondElements on the u '+ 1 th row and v' +1 th column in the low frequency data; when u' is equal to 0, the reaction is carried out,

when u 'is 1,2, …, q-1, C (u') is 1; when v' is equal to 0, the reaction is carried out,

when v 'is 1,2, …, q-1, C (v') is 1.

In this embodiment, specifically, the first low frequency data bc2'_q×1Performing one-dimensional inverse discrete cosine transformation to obtain first low-frequency data subjected to one-dimensional inverse discrete cosine transformation, wherein the data subjected to one-dimensional inverse discrete cosine transformation of the first low-frequency data is finally compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2_q×1. Wherein the final compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2_q×1The x' +1 th element in (A) is

x′∈[0,q-1]X ' is an integer, bc2' (u ') is the u ' +1 th element in the first low frequency data, u ' is e [0, q-1]And u' is an integer. Wherein, when u' is 0,

when u 'is 1,2, …, q-1, C (u') is 1.

And, for the second low frequency data Ac2'_q×nPerforming two-dimensional inverse discrete cosine transformation to obtain second low-frequency data subjected to two-dimensional inverse discrete cosine transformation, wherein the data subjected to two-dimensional inverse discrete cosine transformation of the second low-frequency data is a finally compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix Ac2_q×n。

Wherein the finally compressed multi-dimensional NMR logging nuclear matrix Ac2_q×nThe element in the x '+ 1 th row and y' +1 th column in (1) is

Ac2' (u ', v ') is the thAnd the u '+ 1 th row and v' +1 th column in the two low-frequency data. Wherein, when u' is 0,

when v 'is 1,2, …, q-1, C (v') is 1.

It can be seen that after step 202, the number of rows of the preliminarily compressed multi-dimensional nmr logging kernel matrix Ac1 is s, and after step 205 is completed, the finally compressed multi-dimensional nmr logging echo string data bc2_q×1The number of lines of the multi-dimensional nuclear magnetic resonance logging echo string data is q, so that the number of lines of the multi-dimensional nuclear magnetic resonance logging echo string data is compressed from s to q; after the step 202, the number of rows of the preliminarily compressed multi-dimensional NMR logging nuclear matrix Ac1 is s, and after the step 205 is finished, the finally compressed multi-dimensional NMR logging nuclear matrix Ac2_q×nQ, such that the number of rows in the multi-dimensional nmr log kernel matrix is again compressed from s to q. It can be seen that after the above steps, the number of rows of the multidimensional nuclear magnetic resonance logging echo string data is compressed from m to q, and the number of rows of the multidimensional nuclear magnetic resonance logging kernel matrix is compressed from m to q.

For example, a model of a three-dimensional Nuclear Magnetic Resonance (NMR) formation with bound water, mobile water and light oil was modeled at T1-T2-D, where D is the Diffusion coefficient (Diffusion), T2 is the transverse relaxation time, and T1 is the longitudinal relaxation time. To facilitate visualization of T1-T2-D spectra, fig. 3 is a first simulated T1-T2-D distribution diagram in another compression processing method of multi-dimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, fig. 4 is a second simulated T1-T2-D distribution diagram in another compression processing method of multi-dimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, fig. 5 is a third simulated T1-T2-D distribution diagram in another compression processing method of multi-dimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, as shown in fig. 3-5, fig. 3 is a result of respectively projecting a T1-T2-D model onto two-dimensional T2-D, and the abscissa of fig. 3 is T2, which is expressed in milliseconds (ms) and the ordinate is a T2Is the diffusion coefficient in square centimeters per second (cm)²s^-1) (ii) a FIG. 4 is a result of projecting a T1-T2-D model onto a two-dimensional T1-D, respectively, FIG. 4 having the abscissa T1 in milliseconds (ms) and the ordinate diffusion coefficient in square centimeters per second (cm)²s^-1) (ii) a FIG. 5 is the result of projecting T1-T2-D models into two dimensions T2-T1, respectively; the abscissa of fig. 5 is T2 in milliseconds (ms), and the ordinate is T1 in milliseconds (ms).

Fig. 6 is a schematic diagram of multiple echo trains of three-dimensional nmr logging data in another method for compressing nmr logging data according to an embodiment of the present invention, as shown in fig. 6, fig. 6 includes raw echo data with noise and without noise, and fig. 6 shows time in units of milliseconds (ms) on the abscissa and Porosity (Porosity) in units of pu on the ordinate; for a single set of echo strings of three-dimensional nmr logging data, the relevant information of the three-dimensional nmr logging data is as follows. Number of groups of echo train: 14. the echo intervals TE are changed to 0.45, 0.45, 0.45, 0.45, 0.45, 0.45, 0.45, 0.45, 0.9, 1.8, 3.6, 7.2, 9.6, 12.5ms, respectively. The change waiting times WT are 12.0, 3.0, 1.0, 0.3, 0.1, 0.03, 0.01, 0.003, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0s, respectively. The magnetic field gradient G is 30 Gs/cm. T1 number of spots (T1_ num): 20, the number of the cells is 20; t2 number of spots (T2_ num): 20, the number of the cells is 20; d number of spots (D _ num): 20 pieces of the feed. T1, T2, and D point count product T1_ num × T2_ num × D _ num is 20 × 20 × 20. Thus, the constructed kernel matrix is a 18041 × 8000 matrix.

Fig. 7 is a schematic diagram of three-dimensional nuclear magnetic resonance logging echo string data after preliminary compression of three-dimensional nuclear magnetic resonance logging data in another method for compressing multidimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, and as shown in fig. 7, the result in fig. 7 is obtained after the echo string data in fig. 6 is subjected to preliminary compression processing in step 202.

Fig. 8 is a schematic diagram of first low-frequency data of three-dimensional nuclear magnetic resonance logging data in another compression processing method for multi-dimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, and as shown in fig. 8, after obtaining a discrete cosine transform coefficient and intercepting a low-frequency portion of the three-dimensional nuclear magnetic resonance logging echo string data after the preliminary compression shown in fig. 7 in step 203, a result shown in fig. 8 is obtained. The diagram a in fig. 8 shows all the first discrete cosine transform coefficients, i.e. all the transform coefficients; fig. 8 b is a diagram of the first q data of the first discrete cosine transform coefficient, i.e., the transform coefficient of the low frequency part; the abscissa of each of the graphs a and b in fig. 8 is frequency, and the ordinate is a transform coefficient.

Fig. 9 is a schematic diagram of three-dimensional nuclear magnetic resonance logging echo string data after final compression of three-dimensional nuclear magnetic resonance logging data in another multi-dimensional nuclear magnetic resonance logging data compression processing method according to an embodiment of the present invention, and as shown in fig. 9, step 205 is adopted to perform one-dimensional inverse discrete cosine transform on the first q data of the first discrete cosine transform coefficient shown in fig. 8, so as to obtain first low-frequency data after one-dimensional inverse discrete cosine transform, that is, the final compressed three-dimensional nuclear magnetic resonance logging echo string data is obtained. The abscissa in FIG. 9 represents the number of echoes, and the ordinate represents the porosity (unit: pu).

Fig. 10 is a first inversion result diagram of original three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to the embodiment of the present invention, fig. 11 is a second inversion result diagram of original three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to the embodiment of the present invention, fig. 12 is a third inversion result diagram of original three-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to the embodiment of the present invention, as shown in fig. 10 to 12, after inverting the original three-dimensional nmr logging data, the results of fig. 10 to 12 are obtained, where the abscissa of fig. 10 is T2 and the unit is millisecond (ms), the ordinate is a diffusion coefficient and the unit is square centimeter per second (cm/sec), and fig. 10 is a diagram of fig. 12²s^-1) (ii) a The abscissa of FIG. 11 is T1 in milliseconds (ms) and the ordinate is the diffusion coefficient in square centimeters per second (cm)²s^-1) (ii) a The abscissa of fig. 12 is T2 in milliseconds (ms), and the ordinate is T1 in milliseconds (ms).

FIG. 13 is another multi-dimensional kernel provided by embodiments of the present inventionFig. 14 is a graph of a result of inversion of compressed three-dimensional nuclear magnetic resonance logging data in another method for compressing multi-dimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, and fig. 15 is a graph of a result of inversion of compressed three-dimensional nuclear magnetic resonance logging data in another method for compressing multi-dimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, as shown in fig. 13 to 15, after the three-dimensional nuclear magnetic resonance logging data obtained by the compression method according to the present embodiment is inverted, the results of fig. 13 to 15 are obtained, where the abscissa of fig. 13 is T2, the unit is millisecond (ms), the ordinate is diffusion coefficient, and the unit is square centimeter/second (cm/sec), and fig. 13 is a graph of fig. 15²s^-1) (ii) a The abscissa of FIG. 14 is T1 in milliseconds (ms) and the ordinate is the diffusion coefficient in square centimeters per second (cm)²s^-1) (ii) a The abscissa of fig. 15 is T2 in milliseconds (ms), and the ordinate is T1 in milliseconds (ms). As can be seen from comparison between fig. 10-12 and fig. 13-15, the inversion result of the compressed three-dimensional nmr logging data obtained by the present technical solution is almost the same as the inversion result of the original data.

As another example, a T2-D formation model was simulated for two-dimensional nuclear magnetic resonance containing bound water, mobile water, and light oil, where D is the diffusion coefficient and T2 is the transverse relaxation time. FIG. 16 is a schematic diagram of a simulated T2-D distribution in another method for compressing data from a multi-dimensional NMR well log according to an embodiment of the invention, as shown in FIG. 16, with the abscissa being T2 in milliseconds (ms) and the ordinate being the diffusion coefficient in square centimeters per second (cm) in cm²s^-1)。

Fig. 17 is a schematic diagram of multiple echo trains of two-dimensional nmr logging data in another method for compressing nmr logging data according to an embodiment of the present invention, as shown in fig. 17, fig. 17 includes raw echo data with noise and without noise, and the abscissa in fig. 17 is time in milliseconds (ms) and the ordinate is porosity in pu; for a single set of echo trains of two-dimensional nuclear magnetic resonance logging data, the relevant information of the two-dimensional nuclear magnetic resonance logging data is as follows. The number of echo series is 7; the echo intervals are respectively: 0.45, 0.9, 1.8, 3.6, 7.2, 9.6 and 12.5 ms; the number of the echo waves is respectively as follows: 2000, 1000, 500, 250, 125, 94, 72; the magnetic field gradient G is 30 Gs/cm. T2 number of spots (T2_ num): 45, the number of the cells is 45; d number of spots (D _ num): 45 are provided. The product of T2 and the D distribution number is T2_ num × D _ num ═ 45 × 45. Thus, the constructed core matrix is a 4041 × 2025 matrix.

Fig. 18 is a schematic diagram of two-dimensional nuclear magnetic resonance logging echo train data after preliminary compression of two-dimensional nuclear magnetic resonance logging data in another method for compressing multidimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, and as shown in fig. 18, the result of fig. 18 is obtained after the echo train data in fig. 17 is subjected to preliminary compression processing in step 202.

Fig. 19 is a schematic diagram of first low-frequency data of two-dimensional nuclear magnetic resonance logging data in another compression processing method for multi-dimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, and as shown in fig. 19, after acquiring a discrete cosine transform coefficient and intercepting a low-frequency portion of the two-dimensional nuclear magnetic resonance logging echo string data after the preliminary compression shown in fig. 18 in step 203, a result shown in fig. 19 is obtained. The diagram a in fig. 19 shows all the first discrete cosine transform coefficients, i.e., all the transform coefficients; fig. 19 is a diagram b showing the first q data of the first discrete cosine transform coefficient, i.e., the transform coefficient of the low frequency part; in fig. 19, the abscissa of each of the graphs a and b is frequency, and the ordinate is a transform coefficient.

Fig. 20 is a schematic diagram of two-dimensional nuclear magnetic resonance logging echo string data after final compression of two-dimensional nuclear magnetic resonance logging data in another multi-dimensional nuclear magnetic resonance logging data compression processing method according to an embodiment of the present invention, and as shown in fig. 20, step 205 is adopted to perform one-dimensional inverse discrete cosine transform on the first q data of the first discrete cosine transform coefficient shown in fig. 19, so as to obtain first low-frequency data after one-dimensional inverse discrete cosine transform, that is, the final compressed two-dimensional nuclear magnetic resonance logging echo string data is obtained. The abscissa in FIG. 20 represents the number of echoes, and the ordinate represents the porosity (unit: pu).

Fig. 21 is a diagram of an inversion result of original two-dimensional nmr logging data in another method for compressing md nmr logging data according to an embodiment of the present invention, and as shown in fig. 21, after the original two-dimensional nmr logging data is inverted, the result of fig. 21 is obtained, where an abscissa is T2 and has a unit of millisecond (ms), and an ordinate is a diffusion coefficient and has a unit of square centimeter per second (cm/sec)²s^-1)。

Fig. 22 is a diagram of an inversion result of compressed two-dimensional nmr logging data in another method for compressing multidimensional nmr logging data according to an embodiment of the present invention, and as shown in fig. 22, after the compressed two-dimensional nmr logging data obtained by the compression method of this embodiment is inverted, a result of fig. 22 is obtained, where an abscissa of fig. 22 is T2 and a unit is milliseconds (ms), and an ordinate is a diffusion coefficient and a unit is square centimeters per second (cm/sec), and a horizontal coordinate of fig. 22 is T2 and a vertical coordinate of the same is a diffusion coefficient²s^-1). As can be seen from comparison between fig. 21 and 22, the inversion result of the compressed two-dimensional nmr logging data obtained by the present technical solution is almost the same as the inversion result of the original data.

According to the scheme provided by the embodiment, the multidimensional nuclear magnetic resonance logging data can be well compressed, and the compression rate is high, so that when the compressed multidimensional nuclear magnetic resonance logging data is inverted, the calculated amount in the inversion process is reduced, and the inversion speed is increased; in addition, the scheme provided by this embodiment intercepts the low-frequency portion of the discrete cosine transform coefficient to retain the main features of the original multidimensional nuclear magnetic resonance echo data, so that the main features of the original multidimensional nuclear magnetic resonance logging data can be still ensured under the condition of high compression ratio, the inversion speed is effectively increased and the calculation memory is reduced under the condition of no accuracy loss. In addition, compared with the existing singular value truncation method, the scheme provided by the embodiment has the advantages of small calculation amount and fast compression process.

Fig. 23 is a schematic structural diagram of a compressing device for multi-dimensional nmr logging data according to an embodiment of the present invention, and as shown in fig. 23, the compressing device of this embodiment may include:

the obtaining module 31 is configured to obtain multi-dimensional nuclear magnetic resonance logging echo string data, and construct a multi-dimensional nuclear magnetic resonance logging nuclear matrix according to the multi-dimensional nuclear magnetic resonance logging echo string data.

The first compression module 32 is configured to perform preliminary compression processing on the multidimensional nuclear magnetic resonance logging echo string data by using a window averaging method to obtain preliminarily compressed multidimensional nuclear magnetic resonance logging echo string data, and perform preliminary compression processing on the multidimensional nuclear magnetic resonance logging matrix by using the window averaging method to obtain the preliminarily compressed multidimensional nuclear magnetic resonance logging matrix.

The first transform module 33 is configured to perform one-dimensional discrete cosine transform on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data to obtain a first discrete cosine transform coefficient, and obtain low-frequency data in the first discrete cosine transform coefficient to obtain first low-frequency data.

The second transform module 34 is configured to perform two-dimensional discrete cosine transform on the preliminarily compressed multidimensional nuclear magnetic resonance logging core matrix to obtain a second discrete cosine transform coefficient, and obtain low-frequency data in the second discrete cosine transform coefficient to obtain second low-frequency data.

The second compression module 35 is configured to perform one-dimensional inverse discrete cosine transform on the first low-frequency data and perform two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain compressed multidimensional nuclear magnetic resonance logging data, where the compressed multidimensional nuclear magnetic resonance logging data includes data of the first low-frequency data after the one-dimensional inverse discrete cosine transform and data of the second low-frequency data after the two-dimensional inverse discrete cosine transform.

The device for compressing multidimensional nuclear magnetic resonance logging data provided in this embodiment may refer to the method for compressing multidimensional nuclear magnetic resonance logging data provided in fig. 1, and the principle and the technical effect are the same and are not described again.

Fig. 24 is a schematic structural diagram of another apparatus for compressing multidimensional nuclear magnetic resonance logging data according to an embodiment of the present invention, and based on the embodiment shown in fig. 23, as shown in fig. 24, in the apparatus according to the embodiment of the present invention, the ith element in the preliminarily compressed multidimensional nuclear magnetic resonance logging echo train data bc1 is

Wherein N is_iDividing the multi-dimensional nuclear magnetic resonance logging echo string data into s windows and then counting the number of echoes in the ith window, N_iIs a positive integer, s is a positive integer greater than or equal to 1, i belongs to [1, s ]]I is an integer; the total number of echoes in the s windows is m, and m is N₁+…+N_i-1+N_i+…+N_sM is a positive integer; when i is 0, r₁When 0 ≦ i ≦ s, r_i＝N₁+…+N_i-1；k∈[r_i+1,r_i+N_i]K is an integer; a. the_kjThe method is characterized in that elements in the kth row and the jth column in a multidimensional nuclear magnetic resonance logging core matrix A are included, the multidimensional nuclear magnetic resonance logging core matrix A is a matrix with m rows and n columns, n is the number of multidimensional nuclear magnetic resonance logging echoes, n is a positive integer, j belongs to [1, n ∈]J is an integer; b_kThe method comprises the steps that elements on the kth line in multi-dimensional nuclear magnetic resonance logging echo string data b are obtained, and the echo string data b comprises m-line data; epsilon_kThe preset noise data E includes m lines of data as elements on the k-th line in the preset noise data E.

A first transformation module 33 comprising:

the first transform submodule 331 is configured to perform one-dimensional discrete cosine transform on the preliminarily compressed dimensional nuclear magnetic resonance logging echo string data to obtain a first discrete cosine transform coefficient; when u is 0, the first discrete cosine transform coefficient is bc1'(u) bc1' (0), wherein,

when u is 1,2, …, s-1, the first discrete cosine transform coefficient is

bc1(x) is in the multi-dimensional nuclear magnetic resonance logging echo string data bc1 after primary compressionX is the [0, s-1] of]X is an integer, pi is a circumference ratio, u belongs to [0, s-1]]And u is an integer.

The first truncating submodule 332 is configured to truncate the first q data in the first discrete cosine transform coefficient to obtain first low-frequency data, where q is not greater than s, and q is an integer.

A second transformation module 34 comprising:

the second transform submodule 341 is configured to perform two-dimensional discrete cosine transform on the preliminarily compressed multidimensional nuclear magnetic resonance logging nuclear matrix to obtain a second discrete cosine transform coefficient; the second discrete cosine transform coefficient is Ac1'_s×nAnd the second discrete cosine transform coefficient Ac1'_s×nThe element in the u-th row and the v-th column in (1) is

when v is 1,2, …, n-1, c (v) is 1.

The first truncating submodule 342 is configured to truncate the first q rows of data in the second discrete cosine transform coefficient to obtain second low-frequency data, where q is equal to or less than s, and q is an integer.

A second compression module 35 comprising:

a first inverse transformation submodule 351, configured to perform one-dimensional inverse discrete cosine transformation on the first low-frequency data to obtain final compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2_q×1Wherein the final compressed multi-dimensional NMR logging echo string data bc2_q×1The first low-frequency data is subjected to one-dimensional inverse discrete cosine transform, and finally compressed multi-dimensional nuclear magnetismResonance logging echo train data bc2_q×1The x' +1 th element in (A) is

when u 'is 1,2, …, q-1, C (u') is 1.

A second inverse transformation submodule 352, configured to perform two-dimensional inverse discrete cosine transformation on the second low-frequency data to obtain a final compressed multidimensional nuclear magnetic resonance logging kernel matrix Ac2_q×nWherein the final compressed multi-dimensional NMR logging nuclear matrix Ac2_q×nThe data of the second low-frequency data after two-dimensional inverse discrete cosine transform and the finally compressed multidimensional nuclear magnetic resonance logging nuclear matrix Ac2_q×nThe element in the x '+ 1 th row and y' +1 th column in (1) is

x′∈[0,q-1]X 'is an integer, y' belongs to [0, n-1 ]]Y 'is an integer, u' belongs to [0, q-1 ]]U 'is an integer, v' is an element [0, n-1 ]]V 'is an integer, and Ac2' (u ', v') is an element on the u '+ 1 th row and v' +1 th column in the second low frequency data; when u' is equal to 0, the reaction is carried out,

when v 'is 1,2, …, q-1, C (v') is 1.

The device for compressing multidimensional nuclear magnetic resonance logging data provided in this embodiment can refer to the method for compressing multidimensional nuclear magnetic resonance logging data provided in fig. 2, and the principle and the technical effect are the same and are not described again.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A compression processing method of multi-dimensional nuclear magnetic resonance logging data is characterized by comprising the following steps:

performing one-dimensional inverse discrete cosine transform on the first low-frequency data, and performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain compressed multi-dimensional nuclear magnetic resonance logging data, wherein the compressed multi-dimensional nuclear magnetic resonance logging data comprises data obtained by performing one-dimensional inverse discrete cosine transform on the first low-frequency data and data obtained by performing two-dimensional inverse discrete cosine transform on the second low-frequency data;

the multi-dimensional nuclear magnetic resonance logging nuclear matrix is A, the multi-dimensional nuclear magnetic resonance logging nuclear matrix is a matrix with m rows and n columns, and n is the number of multi-dimensional nuclear magnetic resonance logging echoes; the multi-dimensional nuclear magnetic resonance logging echo string data is b, and the echo string data b comprises m rows of data; m is a positive integer and n is a positive integer.

2. The method according to claim 1, wherein the ith element in the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo train data bc1 is

Wherein N is_iDividing the multi-dimensional nuclear magnetic resonance logging echo string data into the number of echoes in the ith window after s windows, N_iIs a positive integer, s is a positive integer greater than or equal to 1, i belongs to [1, s ]]I is an integer; the total number of echoes in the s windows is m, and m is N₁+…+N_i-1+N_i+…+N_sM is a positive integer; when i is 0, r₁When 0 ≦ i ≦ s, r_i＝N₁+…+N_i-1；k∈[r_i+1,r_i+N_i]K is an integer; a. the_kjThe multidimensional nuclear magnetic resonance logging nuclear matrix A is a matrix with m rows and n columns as elements in the kth row and the jth column in the multidimensional nuclear magnetic resonance logging nuclear matrix AN is the number of the multi-dimensional nuclear magnetic resonance logging echoes, n is a positive integer, and j belongs to [1, n ]]J is an integer; b_kElements on the kth line in the multi-dimensional nuclear magnetic resonance logging echo string data b are acquired, wherein the echo string data b comprises m rows of data; epsilon_kThe preset noise data E includes m lines of data as elements on the k-th line in the preset noise data E.

3. The method of claim 2, wherein the first discrete cosine transform coefficient is bc1'(u) ═ bc1' (0) when u ═ 0, wherein,

when u is 1,2, …, s-1, the first discrete cosine transform coefficient is

Wherein bc1(x) is the x-th element in the preliminarily compressed multi-dimensional nuclear magnetic resonance logging echo string data bc1, x belongs to [0, s-1], x is an integer, pi is a circumference rate, u belongs to [0, s-1], and u is an integer;

correspondingly, obtaining low-frequency data in the first discrete cosine transform coefficient to obtain first low-frequency data, including:

and intercepting the first q data in the first discrete cosine transform coefficient to obtain the first low-frequency data, wherein q is not more than s, and q is an integer.

4. The method of claim 2, wherein the second discrete cosine transform coefficient is Ac1'_s×nAnd the second discrete cosine transform coefficient Ac1'_s×nThe element in the u-th row and the v-th column in (1) is

Pi is the circumference ratio, x belongs to [0, s-1]]X is an integer, u is an element of [0, s-1]]U is an integer, y is an element [0, n-1 ]]Y is an integer, v is an element [0, n-1 ]]V is an integer, and Ac1(x, y) is the preliminarily compressed multidimensional NMR log coreThe x-th row and y-th column elements in the matrix Ac 1; when u is equal to 0, the reaction is carried out,

when v is 1,2, …, n-1, c (v) is 1;

correspondingly, obtaining low-frequency data in the second discrete cosine transform coefficient to obtain second low-frequency data, including:

and intercepting the first q rows of data in the second discrete cosine transform coefficient to obtain the second low-frequency data, wherein q is not more than s, and q is an integer.

5. The method of claim 3 or 4, wherein performing a one-dimensional inverse discrete cosine transform on the first low frequency data and performing a two-dimensional inverse discrete cosine transform on the second low frequency data comprises:

performing one-dimensional inverse discrete cosine transformation on the first low-frequency data to obtain finally compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2_q×1Wherein the final compressed multi-dimensional NMR logging echo train data bc2_q×1The final compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2 is the data of the first low-frequency data after one-dimensional inverse discrete cosine transformation_q×1The x' +1 th element in (A) is

when u 'is 1,2, …, q-1, C (u') is 1;

performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain the final dataCompressed multidimensional NMR logging nuclear matrix Ac2_q×nWherein the final compressed multi-dimensional NMR logging nuclear matrix Ac2_q×nThe final compressed multidimensional nuclear magnetic resonance logging nuclear matrix Ac2 is the data of the second low-frequency data after two-dimensional inverse discrete cosine transform_q×nThe element in the x '+ 1 th row and y' +1 th column in (1) is

when v 'is 1,2, …, q-1, C (v') is 1.

6. A device for compressing multi-dimensional nuclear magnetic resonance logging data, comprising:

the second compression module is used for performing one-dimensional inverse discrete cosine transform on the first low-frequency data and performing two-dimensional inverse discrete cosine transform on the second low-frequency data to obtain compressed multi-dimensional nuclear magnetic resonance logging data, wherein the compressed multi-dimensional nuclear magnetic resonance logging data comprise data obtained by performing one-dimensional inverse discrete cosine transform on the first low-frequency data and data obtained by performing two-dimensional inverse discrete cosine transform on the second low-frequency data;

7. The apparatus of claim 6, wherein the ith element in the preliminarily compressed multi-dimensional NMR log echo train data bc1 is

Wherein N is_iDividing the multi-dimensional nuclear magnetic resonance logging echo string data into the number of echoes in the ith window after s windows, N_iIs a positive integer, s is a positive integer greater than or equal to 1, i belongs to [1, s ]]I is an integer; total number of echoes in s windowsThe number is m, and m is N₁+…+N_i-1+N_i+…+N_sM is a positive integer; when i is 0, r₁When 0 ≦ i ≦ s, r_i＝N₁+…+N_i-1；k∈[r_i+1,r_i+N_i]K is an integer; a. the_kjThe elements in the kth row and the jth column in the multidimensional nuclear magnetic resonance logging matrix A are included, the multidimensional nuclear magnetic resonance logging matrix A is a matrix with m rows and n columns, n is the number of the multidimensional nuclear magnetic resonance logging echoes, n is a positive integer, j belongs to [1, n ∈ ]]J is an integer; b_kElements on the kth line in the multi-dimensional nuclear magnetic resonance logging echo string data b are acquired, wherein the echo string data b comprises m rows of data; epsilon_kThe preset noise data E includes m lines of data as elements on the k-th line in the preset noise data E.

8. The apparatus of claim 7, wherein the first transformation module comprises:

the first transformation submodule is used for carrying out one-dimensional discrete cosine transformation on the preliminarily compressed dimensional nuclear magnetic resonance logging echo string data to obtain a first discrete cosine transformation coefficient; when u is 0, the first discrete cosine transform coefficient is bc1'(u) bc1' (0), wherein,

when u is 1,2, …, s-1, the first discrete cosine transform coefficient is

bc1(x) is the x-th element in the initially compressed multi-dimensional nuclear magnetic resonance logging echo string data bc1, and x belongs to [0, s-1]]X is an integer, pi is a circumference ratio, u belongs to [0, s-1]]U is an integer;

and the first truncation submodule is used for truncating the first q data in the first discrete cosine transform coefficient to obtain the first low-frequency data, wherein q is not more than s, and q is an integer.

9. The apparatus of claim 7, wherein the second transformation module comprises:

the second transformation submodule is used for carrying out two-dimensional discrete cosine transformation on the preliminarily compressed multi-dimensional nuclear magnetic resonance logging nuclear matrix to obtain a second discrete cosine transformation coefficient; the second discrete cosine transform coefficient is Ac1'_s×nAnd the second discrete cosine transform coefficient Ac1'_s×nThe element in the u-th row and the v-th column in (1) is

Pi is the circumference ratio, x belongs to [0, s-1]]X is an integer, u is an element of [0, s-1]]U is an integer, y is an element [0, n-1 ]]Y is an integer, v is an element [0, n-1 ]]V is an integer, and Ac1(x, y) is an element on the x-th row and the y-th column in the preliminarily compressed multi-dimensional NMR logging nuclear matrix Ac 1; when u is equal to 0, the reaction is carried out,

when v is 1,2, …, n-1, c (v) is 1;

and the first truncation submodule is used for truncating the first q rows of data in the second discrete cosine transform coefficient to obtain the second low-frequency data, wherein q is not more than s, and q is an integer.

10. The apparatus of claim 8 or 9, wherein the second compression module comprises:

a first inverse transformation submodule, configured to perform one-dimensional inverse discrete cosine transformation on the first low-frequency data to obtain finally compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2_q×1Wherein the final compressed multi-dimensional NMR logging echo train data bc2_q×1The final compressed multi-dimensional nuclear magnetic resonance logging echo string data bc2 is the data of the first low-frequency data after one-dimensional inverse discrete cosine transformation_q×1The x' +1 th element in (A) is

when u 'is 1,2, …, q-1, C (u') is 1;

a second inverse transformation submodule, configured to perform two-dimensional inverse discrete cosine transformation on the second low-frequency data to obtain a final compressed multidimensional nuclear magnetic resonance logging kernel matrix Ac2_q×nWherein the final compressed multi-dimensional NMR logging nuclear matrix Ac2_q×nThe final compressed multidimensional nuclear magnetic resonance logging nuclear matrix Ac2 is the data of the second low-frequency data after two-dimensional inverse discrete cosine transform_q×nThe element in the x '+ 1 th row and y' +1 th column in (1) is

when v 'is 1,2, …, q-1, C (v') is 1.