CN104090307A - Method and system for processing nuclear magnetic resonance logging-while-drilling data - Google Patents

Abstract

The invention discloses a method and system for processing nuclear magnetic resonance logging-while-drilling data. The method comprises the steps that a sending end device compresses an obtained echo train and sends the compressed data to a receiving end device, the receiving end device decompresses the received compressed data to obtain a decompressed echo train, and then inversion and interpretation are conducted on the obtained decompressed echo train to obtain stratum geological information. Compared with an existing quick inversion algorithm, the method adopting the compressing and decompressing algorithms has the advantage that due to the fact that there is no nonnegativity restriction on the compressed data, calculation complexity and obtaining cost are reduced remarkably, and then market requirements are well met.

Description

A kind of realization with the method and system that bores nuclear magnetic resonance log data processing

Technical field

The present invention relates to boring NMR Logging Technology espespecially a kind of realization with the method and system that bores nuclear magnetic resonance log data processing.

Background technology

Along with the fast development of oil and gas prospect and production, the research of the correlation technique of geophysical survey, drilling well and well logging has been become to the study hotspot of this area.In order accurately, dynamically to instruct drilling well, need to improve the precision of the stratum geology relevant information of obtaining in well logging, in logging operation, extensively adopt at present NMR (Nuclear Magnetic Resonance) logging instrument to carry out well logging while drilling.Conventionally, the raw log data that NMR (Nuclear Magnetic Resonance) logging instrument once gathers when underground work (herein, abbreviation raw data) data volume is several thousand words (words of mentioning herein, length is 16 bits), and be 20 bits per second for the bandwidth of the mud channel of original data transmissions, therefore the data volume that in the unit interval, the data volume of mud transmission is transmitted much smaller than needs, mud channel cannot real-time Transmission raw data, thereby cause abovegroundly cannot obtaining in time raw data, and raw data is resolved and obtained for instructing the stratum geology relevant information of drilling well, this has hindered with the application of boring NMR Logging Technology.

Existingly with boring NMR Logging Technology, the echo string obtaining is carried out to fast inversion, obtain the data after fast inversion; Send and receive the data after fast inversion; Data after the fast inversion receiving are made an explanation, obtain stratum geology relevant information.Wherein, echo string is that the raw data of NMR (Nuclear Magnetic Resonance) logging instrument collection is carried out to the data that pre-service obtains, and is the vector that the range value by each echo forms.Because the data volume of the data after fast inversion is generally tens words, so the data of mud channel after can real-time Transmission fast inversion, guaranteed that real-time interpretation goes out a layer geology relevant information.

In sum, fast inversion algorithm is the existing key algorithm with boring in NMR Logging Technology, to simplify and obtain on the basis of normal inversion algorithm (being called inversion algorithm herein), therefore, the computational solution precision of inversion algorithm is higher than the computational solution precision of fast inversion algorithm.With the fast inversion class of algorithms seemingly, inversion algorithm be input as to NMR (Nuclear Magnetic Resonance) logging instrument adopt collection raw data carry out the echo string that pre-service obtains, the pre-service that raw data is carried out comprises phase alternation pulse train (PAPS, Phase Alternating Pulse Sequence) stack, progressive mean and rotation are processed, the specific implementation of these processing is those skilled in the art's conventional techniques means, repeats no more herein.

Echo is hydrogen nuclei transverse-relaxation signals.Specifically, NMR (Nuclear Magnetic Resonance) logging instrument is by being arranged on the permanent magnet in probe, in the stratum of wellbore, construct the uniform magnetic field region that a magnetic induction density is B0, according to the alternating electromagnetism ripple of CPMG (Carr-Purcell-Meiboom-Gill) pulse train transmitting Larmor frequency ω=γ * B0, in uniform magnetic field region, forming magnetic induction density is the B1 direction alternating magnetic field vertical with B0, wherein, CPMG is that the first letter abbreviations of four scientist's names forms.Hydrogen nuclei transverse-relaxation signals in can measuring fluid between 180 degree pulses in CPMG pulse train.Wherein, the magnetogyric ratio that γ is hydrogen nuclei.

In echo string, the expression of the range value M (ti) of echo is as shown in formula (1),

$M\left(t_i\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}P\left(T_{2j}\right)\exp (-t_i/T_{2j})---\left(1\right)$

In formula (1), M (ti) is the range value of i the echo that constantly ti gathers, the span of i be 1,2 ..., n}, n is the length of echo string, i.e. the quantity of the range value of echo in echo string.By ti and the interval between the moment (ti+1) are called sampling interval constantly.T2 is T2 constant, in theory T2 be one 0 to ∞ continually varying value.In reality, according to petrophysical correlation model, the value of T2 is between 0.3ms to 10s conventionally.Because the number of the range value of the echo comprising in an echo string is generally hundreds of to thousands of, therefore according to formula (1), ask for the system of equations of a Very Ill-conditioned of process need solution of P (T2j).Solving complexity for formula of reduction (1), conventionally by suitable layouting, namely set the span 0.3ms to 10s of T2, in General Inversion algorithm, within the scope of this, by logarithmic mean, getting 32,64 or 128 layouts, then, by inversion algorithm solution formula (1), obtain P (T2j).Wherein, P (T2j) is called as to T2 spectrum with the transformation curve of T2j, T2j is j T2 constant of layouting, the span of j be 1,2 ..., m}, m is that cloth is counted.

Further formula (1) is write as to the form of matrix, as shown in formula (2),

M＝A×P (2)

In formula (2), echo string M=[M (t ₁), M (t ₂) ..., M (t _n)] ^t, comprise the range value of n echo, respectively by moment t ₁the range value M (t of echo ₁), moment t ₂the range value M (t of echo ₂) ..., t constantly _nthe range value M (t of echo _n) represent, wherein n is the length of echo string.The calculating of A as shown in formula (3),

$A=\left[\begin{array}{cccc}{e}^{\frac{-t_1}{T_{21}}}& e^{\frac{-t_1}{T_{22}}}& ...& e^{\frac{-t_1}{T_{2m}}}\\ e^{\frac{-t_1}{T_{22}}}& e^{\frac{-t_2}{T_{22}}}& ...& e^{\frac{-t_2}{T_{2m}}}\\ ...& ...& ...& ...\\ e^{\frac{-t_n}{T_{21}}}& e^{\frac{-t_n}{T_{22}}}& ...& e^{\frac{-t_n}{T_{2m}}}\end{array}\right]---\left(3\right)$

In formula (3), m is that cloth is counted.Matrix A has represented a linear space, and the linear space that claims matrix A to represent is T2 spectral space, and matrix P can be understood as the coordinate of M under T2 spectral space.Correspondingly, the algorithm by echo string M compute matrix P is called inversion algorithm.

Although fast inversion algorithm has carried out some with respect to inversion algorithm, simplify, yet owing to matrix being had to the non-negative constrained of P, make to be difficult to effectively to simplify fast inversion algorithm complex.Like this, realize fast inversion algorithm still have very high complicacy in down-hole, simultaneously required to realize cost also higher, is difficult to meet the need of market.

Summary of the invention

In order to solve the problems of the technologies described above, the invention provides a kind of realization with the method and system that bores nuclear magnetic resonance log data processing, can effectively reduce the complexity of required algorithm, thereby meet well market demand.

In order to reach the object of the invention, the invention discloses a kind of realization with the method for boring nuclear magnetic resonance log data processing, comprising:

Transmitting terminal device compresses the echo string obtaining, and the data after compression are sent to receiving end device;

The data of receiving end device after to the compression receiving decompress, the echo string after being decompressed; And the echo string after the decompression obtaining is carried out to inverting and explanation, obtain stratum geological information.

Described compression comprises:

According to representing the matrix of T2 spectral space and the length of the data after compression, obtain condensation matrix;

The condensation matrix that utilization obtains compresses the echo string obtaining.

The matrix of described expression T2 spectral space is matrix A, and the linear space that matrix A represents is T2 spectral space; The described condensation matrix that obtains comprises:

According to matrix A and transposition compute matrix B thereof, wherein, the computing formula of matrix B is B=A * A ^t;

The eigenvalue matrix Σ of compute matrix B and eigenvectors matrix V; The pass of matrix B and its eigenvalue matrix Σ and its eigenvectors matrix V is B=V * Σ * V ^-1, wherein, each eigenvalue λ that matrix Σ is matrix B ₁, λ ₂..., λ _n, press the matrix that order from big to small forms, it can be expressed as $\mathrm{\Σ}=\left[\begin{array}{cccc}{\mathrm{\λ}}_1& & & \\ & {\mathrm{\λ}}_2& & \\ & & ...& \\ & & & {\mathrm{\λ}}_n\end{array}\right],$ Wherein, λ ₁>=λ ₂>=...>=λ _n, matrix V is each eigenvalue λ with matrix B ₁, λ ₂..., λ _nn * n orthogonal matrix that characteristic of correspondence vector forms for column vector;

Get matrix V ^-1in the capable formation of front k matrix V ^-1(k), obtain condensation matrix COMP=V ^-1(k).

Described compression comprises:

Calculate the product of condensation matrix COMP and echo string M, the data Dcomp after being compressed, wherein,

Data Dcomp=[D1 after compression, D2 ..., Dk] ^t, comprise k numerical value, respectively by D1, D2 ..., Dk represents, wherein, the length of the data of k after for compression;

Echo string M=[M (t1), M (t2), M (tn)] T, the range value that comprises n echo, respectively by the range value M (t1) of the echo of moment t1, constantly the range value M (t2) of the echo of t2 ..., constantly the range value M (tn) of the echo of tn represents, wherein n is the length of echo string.

Described decompression comprises: the product of the data Dcomp after calculating decompression matrix D COMP and compressing, the echo string M ' after being decompressed; Wherein,

Decompression matrix D COMP=V (k), wherein, the matrix of the capable formation of front k that V (k) is V;

Echo string M ' after decompression=[M ' (t ₁), M ' (t ₂) ..., M ' (t _n)] ^t, comprise the range value of n the echo after decompressing, respectively by moment t ₁decompression after range value the M ' (t of echo ₁), moment t ₂decompression after range value the M ' (t of echo ₂) ..., t constantly _ndecompression after range value the M ' (t of echo _n) represent, wherein n is the length of echo string.

Set in advance the corresponding relation of the length of the data after echo signal to noise ratio (S/N ratio) and described compression; Described obtain condensation matrix before, the method also comprises:

Calculate echo signal to noise ratio (S/N ratio); Wherein, echo signal to noise ratio (S/N ratio) is moment t ₃the range value M (t of echo ₃) with the ratio of the standard deviation of noise signal;

According to echo signal to noise ratio (S/N ratio) and described corresponding relation, obtain the length of the data after described compression.

The inventive method also discloses a kind of realization with the system of boring nuclear magnetic resonance log data processing, comprises transmitting terminal device and receiving end device, wherein, transmitting terminal transposition comprises compression unit, receiving end device comprises decompression unit and inverting and Interpretation unit, wherein

Compression unit, for the echo string obtaining is compressed, and sends to receiving end device by the data after compression;

Decompression unit, decompresses for the data after the compression to from transmitting terminal device, the echo string after being decompressed;

Inverting and Interpretation unit, carry out inverting and explanation for the echo string after the decompression to from decompression unit, obtains stratum geological information.

Described compression unit comprises condensation matrix acquisition module and compression module, wherein,

Condensation matrix acquisition module, for obtaining condensation matrix according to the length of the matrix of expression T2 spectral space and the data after compression;

Compression module, for utilizing the condensation matrix from condensation matrix acquisition module to compress the echo string obtaining.

The matrix of described expression T2 spectral space is matrix A, and the linear space that matrix A represents is T2 spectral space; Described condensation matrix acquisition module comprises can sparse matrix precomputation submodule, can sparse matrix calculating sub module and condensation matrix calculating sub module, wherein,

Can sparse matrix precomputation submodule, for according to matrix A and transposition compute matrix B thereof, wherein, the computing formula of matrix B is B=A * A ^t;

Can sparse matrix calculating sub module, for eigenvalue matrix Σ and the eigenvectors matrix V of compute matrix B; The pass of matrix B and its eigenvalue matrix Σ and its eigenvectors matrix V is B=V * Σ * V ^-1, wherein, each eigenvalue λ that matrix Σ is matrix B ₁, λ ₂..., λ _n, press the matrix that order from big to small forms, it can be expressed as $\mathrm{\Σ}=\left[\begin{array}{cccc}{\mathrm{\λ}}_1& & & \\ & {\mathrm{\λ}}_2& & \\ & & ...& \\ & & & {\mathrm{\λ}}_n\end{array}\right],$ Wherein, λ ₁>=λ ₂>=...>=λ _n, matrix V is each eigenvalue λ with matrix B ₁, λ ₂..., λ _nn * n orthogonal matrix that characteristic of correspondence vector forms for column vector;

Condensation matrix calculating sub module, for getting matrix V ^-1in the capable formation of front k matrix V ^-1(k), obtain condensation matrix COMP=V ^-1(k).

Described compression module specifically for: calculate the product of condensation matrix COMP and echo string M, the data Dcomp after being compressed, wherein,

Data Dcomp=[D1 after compression, D2 ..., Dk] ^t, comprise k numerical value, respectively by D1, D2 ..., Dk represents, wherein, the length of the data of k after for compression;

Echo string M=[M (t1), M (t2) ..., M (tn)] ^t, comprise the range value of n echo, respectively by the range value M (t1) of the echo of moment t1, constantly the range value M (t2) of the echo of t2 ..., constantly the range value M (tn) of the echo of tn represents, wherein n is the length of echo string.

Described decompression unit specifically for: calculate the product of the data Dcomp after decompression matrix D COMP and compression, the echo string M ' after being decompressed; Wherein,

Decompression matrix D COMP=V (k), wherein, the matrix of the capable formation of front k that V (k) is V;

Echo string M ' after decompression=[M ' (t ₁), M ' (t ₂) ..., M ' (t _n)] ^t, comprise the range value of n the echo after decompressing, respectively by moment t ₁decompression after range value the M ' (t of echo ₁), moment t ₂decompression after range value the M ' (t of echo ₂) ..., t constantly _ndecompression after range value the M ' (t of echo _n) represent, wherein n is the length of echo string.

Set in advance the corresponding relation of the length of the data after echo signal to noise ratio (S/N ratio) and described compression; Described condensation matrix acquisition module also for:

Calculate echo signal to noise ratio (S/N ratio); Wherein, echo signal to noise ratio (S/N ratio) is moment t ₃the range value M (t of echo ₃) with the ratio of the standard deviation of noise signal;

According to echo signal to noise ratio (S/N ratio) and described corresponding relation, obtain the length of the data after described compression.

Compared with prior art, technical scheme of the present invention comprises: transmitting terminal device compresses the echo string obtaining, and the data after compression are sent to receiving end device; The data of receiving end device after to the compression receiving decompress, the echo string after being decompressed; And the echo string after the decompression obtaining is carried out to inverting and explanation, obtain stratum geological information.By technical scheme provided by the invention, in down-hole by become hundred or the echo string of thousands of words compress, obtain the data after the compression of tens words, make the data after compression on the mud channel of little transmission bandwidth, to carry out real-time Transmission, and than fast inversion algorithm in existing method, compression and decompression algorithm in the inventive method, than fast inversion algorithm, because the data to after compression do not have the restriction of nonnegativity restrictions, make obviously to have reduced its computation complexity and realize cost, having met well the needs in market.

Other features and advantages of the present invention will be set forth in the following description, and, partly from instructions, become apparent, or understand by implementing the present invention.Object of the present invention and other advantages can be realized and be obtained by specifically noted structure in instructions, claims and accompanying drawing.

Accompanying drawing explanation

Accompanying drawing is used to provide the further understanding to technical solution of the present invention, and forms a part for instructions, is used from explanation technical scheme of the present invention with the application's embodiment mono-, does not form the restriction to technical solution of the present invention.

Fig. 1 is that the present invention realizes with the process flow diagram that bores the method for nuclear magnetic resonance log data processing;

Fig. 2 is the process flow diagram of the step compressed in the inventive method;

Fig. 3 is echo string after decompressing in one embodiment of the invention and the schematic diagram of the contrast of corresponding original echo string;

Fig. 4 is the contrast schematic diagram to the inversion result of the echo string of the echo string after decompressing and correspondence thereof in one embodiment of the invention;

Fig. 5 is that the present invention realizes with the composition structural representation that bores the system of nuclear magnetic resonance log data processing;

Fig. 6 is the composition structural representation of compression unit in system of the present invention.

Embodiment

For making the object, technical solutions and advantages of the present invention clearer, hereinafter in connection with accompanying drawing, embodiments of the invention are elaborated.It should be noted that, in the situation that not conflicting, the embodiment in the application and the feature in embodiment be combination in any mutually.

In the step shown in the process flow diagram of accompanying drawing, can in the computer system such as one group of computer executable instructions, carry out.And, although there is shown logical order in flow process, in some cases, can carry out shown or described step with the order being different from herein.

Fig. 1 is that the present invention realizes with the process flow diagram that bores the method for nuclear magnetic resonance log data processing, as shown in Figure 1, comprises the steps:

Step 110: transmitting terminal device compresses the echo string obtaining, and the data after compression are sent to receiving end device.

The echo string obtaining is compressed, be equivalent to calculate the echo string that the obtains coordinate on compression stroke, on mathematics, can be understood as, echo string is mapped on compression stroke, obtain the coordinate that represents that echo string each numerical value on compression stroke forms.Wherein,

Compression stroke can be understood as echo string can evacuated space subspace.Mode by mathematics is mapped to this by echo string can evacuated space, obtains the coordinate of echo string on this can evacuated space (hereinafter referred to as can evacuated space coordinate).Due to echo string on this can evacuated space, have can be sparse character, that is to say, comprise in can evacuated space coordinate in a large number close to 0 decimal or close to 0 minimum number.If suitably dwindled, can obtain compression stroke by evacuated space, make that the coordinate (hereinafter referred to as compression stroke coordinate) of echo string on compression stroke is non-in can evacuated space coordinate to be formed close to 0 decimal or non-minimum number close to 0, and compression stroke coordinate penetrates through reflection the data that obtain can be equivalent to that can evacuated space coordinate through reflection, to penetrate the data that obtain be echo string in the physical sense, can obtain so the effectively compression stroke of compression and back wave train.

Fig. 2 is the specific implementation process flow diagram of step 110, as shown in Figure 2, comprises the steps:

Step 210: obtain condensation matrix according to the length of the matrix of expression T2 spectral space and the data after compression.Wherein, represent that the matrix of T2 spectral space is matrix A, that is to say, the linear space that matrix A represents is T2 spectral space.This step specifically comprises:

Step 211: according to matrix A and transposition compute matrix B thereof, as shown in formula (4),

B＝A×A ^T (4)

Wherein, matrix B is n * n square formation, and n is the length of echo string.

Step 212: the eigenvalue matrix Σ of compute matrix B and eigenvectors matrix V, the relation of matrix B and its eigenvalue matrix Σ and its eigenvectors matrix V, as shown in formula (5),

B＝V×Σ×V ^-1 (5)

In formula (5), the eigenvalue matrix Σ of matrix B, is each eigenvalue λ of matrix B ₁, λ ₂..., λ _n, press the matrix that order from big to small forms, it can be expressed as $\mathrm{\Σ}=\left[\begin{array}{cccc}{\mathrm{\λ}}_1& & & \\ & {\mathrm{\λ}}_2& & \\ & & ...& \\ & & & {\mathrm{\λ}}_n\end{array}\right].$

The eigenvectors matrix V of matrix B is each eigenvalue λ with matrix B ₁, λ ₂..., λ _nn * n orthogonal matrix that characteristic of correspondence vector forms for column vector, V ^-1inverse matrix for V.

Known matrix B, according to the method for the eigenvectors matrix V of formula (5) compute matrix B, for those skilled in the art's the conventional number means that learn a skill, does not belong to protection scope of the present invention, repeats no more herein.

Step 213: get matrix V ^-1in the capable formation of front k matrix V ^-1(k), obtain condensation matrix COMP=V ^-1(k).Correspondingly, the space that condensation matrix COMP is corresponding is compression stroke.

Can be by V ^-1(k) reason as condensation matrix COMP is,

If by formula D=V ^-1* M, calculates matrix D, wherein, and echo string M=[M (t ₁), M (t ₂) ..., M (t _n)] ^t, comprise the range value of n echo, respectively by moment t ₁the range value M (t of echo ₁), moment t ₂the range value M (t of echo ₂) ..., t constantly _nthe range value M (t of echo _n) represent, wherein n is the length of echo string, D=[D ₁, D ₂..., D _n] ^t, comprise n numerical value, respectively by D ₁, D ₂..., Dn represents.Because the variation of the eigenwert of matrix B has the characteristic of index decreased, and matrix V is n * n orthogonal matrix that column vector forms for take each eigenwert characteristic of correspondence vector of matrix B, and correspondingly, in matrix D, the absolute value of each numerical value also presents the characteristic of index decreased.Give an example,

When sampling interval is 0.6ms, echo string length is 700, and it is 128 o'clock that cloth is counted, by formula D=V ^-1* M calculates matrix D, has been shown in the following Table 1 front 16 eigenvalue λ of matrix B ₁～λ ₁6, be shown in the following Table 2 front 16 numerical value D of corresponding matrix D ₁～D ₁₆.As shown in table 1, front 16 eigenwerts of matrix B, the characteristic that presents index decreased, and be the minimum number close to 0 since the 10th eigenwert, and be the decimal close to 0 since the 6th eigenwert, correspondingly, as shown in table 2, the absolute value of front 16 numerical value of matrix D also has similar rule.As seen through the above analysis, echo string is in matrix V ^-1on have can sparse property, can be by matrix V ^-1what corresponding space was called echo string can evacuated space.Further analyze,

Suppose to get matrix V ^-1in front 10 row form matrix V ^-1(10), i.e. k=10, and by formula D '=V ^-1(10) * M, calculates matrix D ', wherein, D '=[D ' ₁, D ' ₂..., D ' ₁₀] ^t, comprise 10 numerical value, respectively by D ₁, D ₂..., D ₁₀represent.By analysis above, can find out, matrix D ' comprised in matrix D other numerical value except the minimum number close to 0, because each numerical value close to 0 minimum number in D is minimum to the contribution of D " authenticity ", namely from D, removing these each numerical value close to 0 minimum number is acceptable on the impact of D " authenticity ", so D ' can reasonably replace D.Similarly,

Suppose to get matrix V ^-1in front 6 row form matrix V ^-1(6), i.e. k=6, and by formula D "=V ^-1(6) * M, calculates matrix D ", due to matrix D " comprised in matrix D non-close to 0 decimal or non-each numerical value close to 0 minimum number, similarly, D " can reasonably replace D.The rest may be inferred,

When k is any number in 6 to 16, by formula V ^-1(k) * M, the matrix calculating also can reasonably replace D.

In sum, by matrix V ^-1in the matrix V of the capable formation of front k ^-1(k) as condensation matrix COMP, be that reasonably wherein, the span of k can be 6 to 16.

λ 1	λ 2	λ3	λ4
				24166.9552	1039.329	146.9422	29.5004
λ5	λ6	λ7	λ8
				6.8157	1.6611	0.4108	0.1012
λ9	λ 10	λ 11	λ 12
				0.0246	0.0059	0.0014	0.0003
λ 13	λ 14	λ 15	λ 16
				7.1469e-05	1.5842e-05	3.451e-06	7.393e-07

Table 1

D 1	D 2	D3	D4
				-476.1993	-295.3678	-13.6003	-2.7955
D5	D6	D7	D8
				9.1832	-1.4301	0.7139	-0.2181
D9	D 10	D 11	D 12
				-0.0784	0.0215	-0.0021	-0.0024
D 13	D 14	D 15	D 16
				0.0001	-2.8892e-05	-1.2370e-05	5.2076e-06

Table 2

Step 220: utilize the condensation matrix obtaining to compress the echo string obtaining.Specifically,

Calculate the product of condensation matrix COMP and echo string M, the data Dcomp after being compressed.The calculating of data Dcomp after compression as shown in formula (6),

Dcomp＝COMP×M (6)

Wherein, the data Dcomp=[D after compression ₁, D ₂..., D _k] ^t, comprise k numerical value, respectively by D ₁, D ₂..., D _krepresent, wherein, k for compression after data length wherein, echo string M=[M (t ₁), M (t ₂) ..., M (t _n)] ^t, comprise the range value of n echo, respectively by moment t ₁the range value M (t of echo ₁), moment t ₂the range value M (t of echo ₂) ..., t constantly _nthe range value M (t of echo _n) represent, wherein n is the length of echo string.

Data Dcomp after being compressed, sends to receiving end device by the data after compression.Specifically, after data after being compressed, data after down-hole is to compression are carried out transmission processing, and on mud channel, transmit the data after compression, correspondingly, aboveground, from mud channel, the data after compression are carried out to reception & disposal, and the data after the decompression receiving are sent to step 120.

Wherein, the conventional techniques means that the specific implementation of transmission processing and reception & disposal is those skilled in the art, are not intended to limit the scope of the invention, and repeat no more herein.

Further,

The corresponding relation of the length k of the data after setting in advance echo signal to noise ratio snr and compressing, this corresponding relation can be the corresponding relation shown in following face table 3.

SNR scope	The span of k
		SNR≥100	8～14
100>SNR≥60	8～12
		60>SNR≥40	8～10
40>SNR≥30	8～9
		30>SNR≥20	7～8
SNR<20	6～7

Table 3

Before step 210, the inventive method also comprises: the length of determining the data after compression according to the corresponding relation of SNR and k.Specifically comprise:

Calculate echo signal to noise ratio (S/N ratio); Wherein, echo signal to noise ratio (S/N ratio) is moment t ₃the range value M (t of echo ₃) with the ratio of the standard deviation of noise signal; The length of the data after being compressed according to the corresponding relation of SNR and k.

Step 120: the data of receiving end device after to the compression receiving decompress, the echo string after being decompressed; And the echo string after the decompression obtaining is carried out to inverting and explanation, obtain stratum geological information.Wherein,

Data after received compression are decompressed, on mathematics, can be understood as, compression stroke coordinate reflection is mapped to echo string institute spatially, obtain can being equivalent in the physical sense the echo string after the decompression of echo string.

Specifically, calculate the product of the data Dcomp after decompression matrix D COMP and compression, the echo string M ' after being decompressed, it calculates as according to formula (7),

M’＝DCOMP×Dcomp (7)

In formula (7), the echo string M ' after decompression=[M ' (t ₁), M ' (t ₂) ..., M ' (t _n)] ^t, comprise the range value of n the echo after decompressing, respectively by moment t ₁decompression after range value the M ' (t of echo ₁), moment t ₂decompression after range value the M ' (t of echo ₂) ..., t constantly _ndecompression after range value the M ' (t of echo _n) represent.

The transposition that decompression matrix D COMP is COMP, COMP is for getting matrix V ^-1the matrix of the capable formation of front k, because matrix V has orthogonality, so the inverse matrix V of V ^-1=V ^t, therefore, DCOMP=V _(k).

In the compression and decompression algorithm for echo string M, the result of compression be there is no to the restriction of nonnegativity restrictions, therefore, to compare with fast inversion algorithm, the complexity of compression and decompression algorithm obviously reduces.

Next, the echo string after the decompression obtaining is carried out to inverting and explanation, obtain stratum geological information.In this step, the processing of inverting and explanation is identical with the processing of original echo string being carried out to inverting and explanation, and the conventional techniques means that its specific implementation is those skilled in the art, are not intended to limit the scope of the invention, and repeat no more herein.

Fig. 3 is echo string after decompressing in one embodiment of the invention and the schematic diagram of the contrast of corresponding original echo string, wherein, raw data is that echo string is the data in actual job, sampling interval is 0.6ms, echo string length is 700, signal to noise ratio snr is 34.52, and the length of the data after compression is 8.Fig. 3 horizontal ordinate is time value, and ordinate is range value.In Fig. 3+pattern represents that raw data is echo string, in Fig. 3-pattern represents the echo string after data after compressed and decompressed decompress.As shown in Figure 3, the echo string after decompression, is equivalent to echo string data to carry out " curve " in effect.

On the basis of Fig. 3, respectively the echo string after decompressing and echo to be ganged up and blocked SVD inversion algorithm and carry out inverting, the result obtaining is as shown in Figure 4.

Fig. 4 is the contrast schematic diagram to the inversion result of the echo string of the echo string after decompressing and correspondence thereof in one embodiment of the invention, and wherein, the span of T2 is 0.3ms to 10s, and it is 128 that cloth is counted.Fig. 4 horizontal ordinate is the numerical value of T2 constant T2, and Fig. 4 ordinate is range value.In Fig. 4, be that solid pattern represents that raw data is echo string, the T2 that obtains through inverting spectrum, and Fig. 4 hollow core delta pattern represents the echo string after data after compressed and decompressed decompress, the T2 obtaining through inverting spectrum.As shown in Figure 4, the T2 spectrum that clawback string inverting after decompressing is obtained with directly echo string is carried out to T2 that inverting obtains and composes and compare, except the size of crest place numerical value has fine distinction, T2 spectral line in two kinds of situations is substantially identical, has illustrated that compression and decompression algorithm of the present invention is reliably rationally.

Fig. 5 is that the present invention realizes with the composition structural representation that bores the system of nuclear magnetic resonance log data processing, as shown in Figure 5, comprises transmitting terminal device and receiving end device, wherein, transmitting terminal transposition comprises compression unit, and receiving end device comprises decompression unit and inverting and Interpretation unit, wherein

Compression unit, for the echo string obtaining is compressed, and sends to receiving end device by the data after compression.

Decompression unit, decompresses for the data after the compression to from transmitting terminal device, the echo string after being decompressed.

Inverting and Interpretation unit, carry out inverting and explanation for the echo string after the decompression to from decompression unit, obtains stratum geological information.

Fig. 6 is the composition structural representation of compression unit in system of the present invention, and as shown in Figure 6, compression unit comprises condensation matrix acquisition module and compression module, wherein,

Condensation matrix acquisition module, for obtaining condensation matrix according to the length of the matrix of expression T2 spectral space and the data after compression;

Compression module, for utilizing the condensation matrix from condensation matrix acquisition module to compress the echo string obtaining.

Further,

The matrix that represents T2 spectral space is matrix A, that is to say, the linear space that matrix A represents is T2 spectral space;

Condensation matrix acquisition module comprises can sparse matrix precomputation submodule, can sparse matrix calculating sub module and condensation matrix calculating sub module, wherein,

Can sparse matrix precomputation submodule, for according to matrix A and transposition compute matrix B thereof, wherein, the computing formula of matrix B is B=A * A ^t;

Can sparse matrix calculating sub module, for eigenvalue matrix Σ and the eigenvectors matrix V of compute matrix B; The pass of matrix B and its eigenvalue matrix Σ and its eigenvectors matrix V is B=V * Σ * V ^-1, wherein, each eigenvalue λ that matrix Σ is matrix B ₁, λ ₂..., λ _n, press the matrix that order from big to small forms, it can be expressed as $\mathrm{\Σ}=\left[\begin{array}{cccc}{\mathrm{\λ}}_1& & & \\ & {\mathrm{\λ}}_2& & \\ & & ...& \\ & & & {\mathrm{\λ}}_n\end{array}\right],$ Wherein, λ ₁>=λ ₂>=...>=λ _n, matrix V is each eigenvalue λ with matrix B ₁, λ ₂..., λ _nn * n orthogonal matrix that characteristic of correspondence vector forms for column vector;

Condensation matrix calculating sub module, for getting matrix V ^-1in the capable formation of front k matrix V ^-1(k), obtain condensation matrix COMP=V ^-1(k).

The corresponding relation of the length of the data after setting in advance echo signal to noise ratio (S/N ratio) and compressing; Condensation matrix acquisition module also for:

Calculate echo signal to noise ratio (S/N ratio); Wherein, echo signal to noise ratio (S/N ratio) is moment t ₃the range value M (t of echo ₃) with the ratio of the standard deviation of noise signal;

The length of the data after being compressed according to echo signal to noise ratio (S/N ratio) and corresponding relation.

Compression module specifically for:

Calculate the product of condensation matrix COMP and echo string M, the data Dcomp after being compressed, wherein, and the data Dcomp=[D1 after compression, D2 ..., Dk] ^t, comprise k numerical value, respectively by D1, D2 ..., Dk represents, wherein, k is the length of the data after compressing, echo string M=[M (t1), M (t2), M (tn)] T, comprise the range value of n echo, respectively by the range value M (t1) of the echo of moment t1, the range value M (t2) of the echo of moment t2 ..., constantly the range value M (tn) of the echo of tn represents, wherein n is the length of echo string.

Decompression unit specifically for: calculate the product of the data Dcomp after decompression matrix D COMP and compression, the echo string M ' after being decompressed; Wherein,

Decompression matrix D COMP=V (k), wherein, the matrix of the capable formation of front k that V (k) is V,

Echo string M ' after decompression=[M ' (t ₁), M ' (t ₂) ..., M ' (t _n)] ^t, comprise the range value of n the echo after decompressing, respectively by moment t ₁decompression after range value the M ' (t of echo ₁), moment t ₂decompression after range value the M ' (t of echo ₂) ..., t constantly _ndecompression after range value the M ' (t of echo _n) represent, wherein n is the length of echo string.

Although the disclosed embodiment of the present invention as above, the embodiment that described content only adopts for ease of understanding the present invention, not in order to limit the present invention.Those of skill in the art under any the present invention; do not departing under the prerequisite of the disclosed spirit and scope of the present invention; can in the form of implementing and details, carry out any modification and variation; but scope of patent protection of the present invention, still must be as the criterion with the scope that appending claims was defined.

Claims (12)

1. realization, with a method of boring nuclear magnetic resonance log data processing, is characterized in that, comprising:

Transmitting terminal device compresses the echo string obtaining, and the data after compression are sent to receiving end device;

The data of receiving end device after to the compression receiving decompress, the echo string after being decompressed; And the echo string after the decompression obtaining is carried out to inverting and explanation, obtain stratum geological information.

2. method according to claim 1, is characterized in that, described in compress and comprise:

According to representing the matrix of T2 spectral space and the length of the data after compression, obtain condensation matrix;

The condensation matrix that utilization obtains compresses the echo string obtaining.

3. method according to claim 2, is characterized in that, the matrix of described expression T2 spectral space is matrix A, and the linear space that matrix A represents is T2 spectral space; The described condensation matrix that obtains comprises:

According to matrix A and transposition compute matrix B thereof, wherein, the computing formula of matrix B is B=A * A ^t;

The eigenvalue matrix Σ of compute matrix B and eigenvectors matrix V; The pass of matrix B and its eigenvalue matrix Σ and its eigenvectors matrix V is B=V * Σ * V ^-1, wherein, each eigenvalue λ that matrix Σ is matrix B ₁, λ ₂..., λ _n, press the matrix that order from big to small forms, it can be expressed as $\mathrm{\Σ}=\left[\begin{array}{cccc}{\mathrm{\λ}}_1& & & \\ & {\mathrm{\λ}}_2& & \\ & & ...& \\ & & & {\mathrm{\λ}}_n\end{array}\right],$ Wherein, λ ₁>=λ ₂>=...>=λ _n, matrix V is each eigenvalue λ with matrix B ₁, λ ₂..., λ _nn * n orthogonal matrix that characteristic of correspondence vector forms for column vector;

Get matrix V ^-1in the capable formation of front k matrix V ^-1(k), obtain condensation matrix COMP=V ^-1(k).

4. according to the method in claim 2 or 3, it is characterized in that, described in compress and comprise:

Calculate the product of condensation matrix COMP and echo string M, the data Dcomp after being compressed, wherein,

Data Dcomp=[D1 after compression, D2 ..., Dk] ^t, comprise k numerical value, respectively by D1, D2 ..., Dk represents, wherein, the length of the data of k after for compression;

Echo string M=[M (t1), M (t2), M (tn)] T, the range value that comprises n echo, respectively by the range value M (t1) of the echo of moment t1, constantly the range value M (t2) of the echo of t2 ..., constantly the range value M (tn) of the echo of tn represents, wherein n is the length of echo string.

5. method according to claim 4, is characterized in that, described in decompress and comprise: calculate the product of the data Dcomp after decompression matrix D COMP and compression, the echo string M ' after being decompressed; Wherein,

Decompression matrix D COMP=V (k), wherein, the matrix of the capable formation of front k that V (k) is V;

Echo string M ' after decompression=[M ' (t ₁), M ' (t ₂) ..., M ' (t _n)] ^t, comprise the range value of n the echo after decompressing, respectively by moment t ₁decompression after range value the M ' (t of echo ₁), moment t ₂decompression after range value the M ' (t of echo ₂) ..., t constantly _ndecompression after range value the M ' (t of echo _n) represent, wherein n is the length of echo string.

6. method according to claim 2, is characterized in that, sets in advance the corresponding relation of the length of the data after echo signal to noise ratio (S/N ratio) and described compression; Described obtain condensation matrix before, the method also comprises:

Calculate echo signal to noise ratio (S/N ratio); Wherein, echo signal to noise ratio (S/N ratio) is moment t ₃the range value M (t of echo ₃) with the ratio of the standard deviation of noise signal;

According to echo signal to noise ratio (S/N ratio) and described corresponding relation, obtain the length of the data after described compression.

7. realization, with a system of boring nuclear magnetic resonance log data processing, is characterized in that, comprise transmitting terminal device and receiving end device, wherein, transmitting terminal transposition comprises compression unit, and receiving end device comprises decompression unit and inverting and Interpretation unit, wherein,

Compression unit, for the echo string obtaining is compressed, and sends to receiving end device by the data after compression;

Decompression unit, decompresses for the data after the compression to from transmitting terminal device, the echo string after being decompressed;

Inverting and Interpretation unit, carry out inverting and explanation for the echo string after the decompression to from decompression unit, obtains stratum geological information.

8. system according to claim 7, is characterized in that, described compression unit comprises condensation matrix acquisition module and compression module, wherein,

Condensation matrix acquisition module, for obtaining condensation matrix according to the length of the matrix of expression T2 spectral space and the data after compression;

Compression module, for utilizing the condensation matrix from condensation matrix acquisition module to compress the echo string obtaining.

9. system according to claim 8, is characterized in that, the matrix of described expression T2 spectral space is matrix A, and the linear space that matrix A represents is T2 spectral space; Described condensation matrix acquisition module comprises can sparse matrix precomputation submodule, can sparse matrix calculating sub module and condensation matrix calculating sub module, wherein,

Can sparse matrix precomputation submodule, for according to matrix A and transposition compute matrix B thereof, wherein, the computing formula of matrix B is B=A * A ^t;

Can sparse matrix calculating sub module, for eigenvalue matrix Σ and the eigenvectors matrix V of compute matrix B; The pass of matrix B and its eigenvalue matrix Σ and its eigenvectors matrix V is B=V * Σ * V ^-1, wherein, each eigenvalue λ that matrix Σ is matrix B ₁, λ ₂..., λ _n, press the matrix that order from big to small forms, it can be expressed as $\mathrm{\Σ}=\left[\begin{array}{cccc}{\mathrm{\λ}}_1& & & \\ & {\mathrm{\λ}}_2& & \\ & & ...& \\ & & & {\mathrm{\λ}}_n\end{array}\right],$ Wherein, λ ₁>=λ ₂>=...>=λ _n, matrix V is each eigenvalue λ with matrix B ₁, λ ₂..., λ _nn * n orthogonal matrix that characteristic of correspondence vector forms for column vector;

Condensation matrix calculating sub module, for getting matrix V ^-1in the capable formation of front k matrix V ^-1(k), obtain condensation matrix COMP=V ^-1(k).

10. system according to claim 8 or claim 9, is characterized in that, described compression module specifically for: calculate the product of condensation matrix COMP and echo string M, the data Dcomp after being compressed, wherein,

Data Dcomp=[D1 after compression, D2 ..., Dk] ^t, comprise k numerical value, respectively by D1, D2 ..., Dk represents, wherein, the length of the data of k after for compression;

Echo string M=[M (t1), M (t2) ..., M (tn)] ^t, comprise the range value of n echo, respectively by the range value M (t1) of the echo of moment t1, constantly the range value M (t2) of the echo of t2 ..., constantly the range value M (tn) of the echo of tn represents, wherein n is the length of echo string.

11. systems according to claim 10, is characterized in that, described decompression unit specifically for: calculate the product of the data Dcomp after decompression matrix D COMP and compression, the echo string M ' after being decompressed; Wherein,

Decompression matrix D COMP=V (k), wherein, the matrix of the capable formation of front k that V (k) is V;

Echo string M ' after decompression=[M ' (t ₁), M ' (t ₂) ..., M ' (t _n)] ^t, comprise the range value of n the echo after decompressing, respectively by moment t ₁decompression after range value the M ' (t of echo ₁), moment t ₂decompression after range value the M ' (t of echo ₂) ..., t constantly _ndecompression after range value the M ' (t of echo _n) represent, wherein n is the length of echo string.

12. systems according to claim 8, is characterized in that, set in advance the corresponding relation of the length of the data after echo signal to noise ratio (S/N ratio) and described compression; Described condensation matrix acquisition module also for:

Calculate echo signal to noise ratio (S/N ratio); Wherein, echo signal to noise ratio (S/N ratio) is moment t ₃the range value M (t of echo ₃) with the ratio of the standard deviation of noise signal;

According to echo signal to noise ratio (S/N ratio) and described corresponding relation, obtain the length of the data after described compression.