CN114428342A - Markenko Green function reconstruction method and system based on MPI - Google Patents

Markenko Green function reconstruction method and system based on MPI Download PDF

Info

Publication number
CN114428342A
CN114428342A CN202011076701.5A CN202011076701A CN114428342A CN 114428342 A CN114428342 A CN 114428342A CN 202011076701 A CN202011076701 A CN 202011076701A CN 114428342 A CN114428342 A CN 114428342A
Authority
CN
China
Prior art keywords
green function
mpi
focusing
function
green
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202011076701.5A
Other languages
Chinese (zh)
Inventor
韩冬
李博
白英哲
崔月
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Petroleum and Chemical Corp
Sinopec Geophysical Research Institute
Original Assignee
China Petroleum and Chemical Corp
Sinopec Geophysical Research Institute
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China Petroleum and Chemical Corp, Sinopec Geophysical Research Institute filed Critical China Petroleum and Chemical Corp
Priority to CN202011076701.5A priority Critical patent/CN114428342A/en
Publication of CN114428342A publication Critical patent/CN114428342A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/28—Processing seismic data, e.g. analysis, for interpretation, for correction
    • G01V1/36—Effecting static or dynamic corrections on records, e.g. correcting spread; Correlating seismic signals; Eliminating effects of unwanted energy
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/50—Corrections or adjustments related to wave propagation
    • G01V2210/51—Migration
    • G01V2210/512—Pre-stack

Abstract

The invention provides a Markenko Green function reconstruction method and system based on MPI, and belongs to the field of seismic data prestack depth migration imaging. The method is based on MPI, an initial Green function and corrected seismic data are used as input, parallel task division is carried out according to coordinates of underground focusing points, and an iterative inversion method is used for carrying out numerical solution on a Marchenko equation to obtain an uplink Green function and a downlink Green function. The method greatly improves the computational efficiency of the Markenko Green function reconstruction.

Description

Markenko Green function reconstruction method and system based on MPI
Technical Field
The invention belongs to the field of seismic data prestack depth migration imaging, and particularly relates to a Markhenko Green function reconstruction method and system based on MPI.
Background
The Green function is a response function of a single pulse point source and is important for seismic imaging and seismic data reconstruction. The conventional green function reconstruction method represented by a seismic interference method is a data-driven method, which can reconstruct a green function only through interaction between seismic records without medium information, but the method needs to set a detection point at a certain underground position and set a seismic source point around a medium.
The current exploration earthquake mainly takes unilateral illumination as the main part and does not accord with the basic assumption condition of the earthquake interference method. Compared with the conventional seismic interference method, the Marchenko Green function reconstruction method does not need to arrange a real wave detection point in a medium, and can carry out Green function reconstruction only by unilateral illumination. Meanwhile, uplink and downlink green functions obtained by the Marchenko green function reconstruction method not only contain primary reflection wave information, but also contain interlayer multiple wave information.
Chinese patent publication CN107958266A discloses a method for discretizing continuous attributes based on MPI parallel, which includes: firstly, reading data of an information system, horizontally dividing the information system into m sample data subsets, and distributing the sample data subsets to n nodes through communication; secondly, each node performs normalization processing on the data in parallel to obtain new data, then performs clustering on attributes in parallel, and merges clustering results through communication; finally, interval division is carried out according to the clustering result, and attribute coding is carried out on different intervals, so that a continuous attribute discretization result is obtained, and an information system after attribute discretization is constructed, so that follow-up work such as attribute reduction can be carried out by using rough set knowledge; chinese patent publication CN110927921A discloses a digital grating assisted auto-focusing method and apparatus, which generates a digital grating pattern by setting a digital micromirror device DMD module and irradiating it with an illumination light source, and projects the digital grating pattern onto a three-dimensional platform by setting a projection lens, and by setting an observation module and a control module, can judge whether the projection lens is out of focus, and then control the three-dimensional platform to complete coarse focusing, and by setting a sensor module, can feed back position information of the three-dimensional platform in real time, and adjust the three-dimensional platform by the control module to complete fine focusing; chinese patent publication CN109490425A discloses a passive material low-frequency reflection coefficient measuring method based on green function reconstruction technology, which includes: 1) obtaining a reflection time reversal term in the multi-dimensional Markov equation by using a time reversal technology; 2) reconstructing a scattering wave Green function from the hydrophone to the surface of the sample; 3) realizing scattered wave focusing in a numerical value mode; 4) and (4) calculating a reflection coefficient measurement value. According to the method, a time reversal technology is used for obtaining a reflection time reversal term in a multi-dimensional Markov equation, then the reflection time reversal term is substituted into the simplified multi-dimensional Markov equation to obtain a scattering wave Green function from a hydrophone to the surface of a sample, then the reconstructed scattering wave Green function and the reflection time reversal term corresponding to the position of the hydrophone are convolved, and all array element convolution terms are summed to obtain the focusing of the scattering wave on the surface of the sample.
With the aim of seismic migration imaging, a plurality of focus points are required to be arranged in a target area for Marchenkok Green function reconstruction, the process is large in calculation amount and high in memory consumption, and a single calculation node is difficult to meet the industrial production requirements. The Green function is of great importance to seismic imaging, the reconstruction precision of the Green function is directly related to the seismic imaging precision, and the method is oriented to the high-precision Green function reconstruction requirement and the problem of large multi-focus Marhenko Green function reconstruction calculation amount.
Disclosure of Invention
The invention aims to solve the problems in the prior art, and provides a Markenko Green function reconstruction method and a Markenko Green function reconstruction system based on MPI, so that the calculation efficiency is greatly improved.
The invention is realized by the following technical scheme:
the invention provides a Markhenko Green function reconstruction method based on MPI, which is based on MPI, takes an initial Green function and corrected seismic data as input, performs parallel task division according to coordinates of an underground focus point, and numerically solves a Marchenko equation by an iterative inversion method to obtain an uplink Green function and a downlink Green function.
A further development of the invention is that the method comprises:
the method comprises the following steps: determining the number of focusing points and the coordinates of each focusing point;
step two: starting an MPI parallel environment;
step three: setting the number of the computing nodes, numbering the computing nodes from 0, taking the computing node with the number 0 as a main node, and taking other computing nodes as slave nodes;
step four: the master node distributes the task to each slave node, and each slave node finishes the task to obtain a focusing wave field;
step five: each slave node calculates by utilizing a focusing wave field to obtain a Green function;
step six: each slave node outputs a Green function;
step seven: ending the MPI parallel environment.
A further refinement of the invention is that the operation of the master node in step four to assign a task to each slave node comprises:
the master node reads in the coordinates of each focus point in the focus target area and issues the coordinates to each slave node.
In a further development of the invention, each of the fourth steps of obtaining a focused wavefield from a node completion task comprises:
each slave node obtains a focused wavefield using the following formula:
Figure BDA0002717024170000031
Figure BDA0002717024170000032
Figure BDA0002717024170000033
Figure BDA0002717024170000034
wherein the content of the first and second substances,
Figure BDA0002717024170000035
and
Figure BDA0002717024170000036
are respectively undergroundxiA down-line focusing function and an up-line focusing function of the k iteration of focusing;
Figure BDA0002717024170000037
scattering tail waves obtained for the kth iteration;
R(x0,x'0and t) is in x'0Excitation x at a point0Receiving the obtained surface reflection response;
Figure BDA0002717024170000038
Figure BDA0002717024170000039
is the inverse of the transmission response from the surface excitation point to the subsurface focus point;
Gd(x′0，xi-t) is an initial green function Gd(x'0,xiT) the result of the inverse temporal transformation;
θ is a window function, as follows:
Figure BDA0002717024170000041
where t is time, tdThe travel time is directly reached from the focal point to the surface seismic source point.
The invention is further improved in that the operation of the step five comprises the following steps:
each slave node calculates an uplink Green function and a downlink Green function by using the following formula:
Figure BDA0002717024170000042
wherein G is-(xi,x0T) and G+(xi,x0T) are respectively the earth's surface x0Excited at point and in the subsurface xiReceiving the obtained uplink Green function and downlink Green function;
f1 -(x0,xit) and f1 +(x0,xiT) are each underground xiFocusing up-and down-focusing functions, the values of which are taken
Figure BDA0002717024170000043
And
Figure BDA0002717024170000044
the value of (c).
The invention is further improved in that the second step is realized by calling MPI _ Init;
in the sixth step, each slave node outputs the Green function by calling a file output function;
the seventh step is to finish the parallel environment by calling MPI _ Finailze.
In a further development of the invention, the operation of step one comprises:
dividing a focusing target area according to an imaging target and computing power;
determining discrete intervals of focus points according to geological structure characteristics, and setting the number of total focus points as n;
and gridding the focusing target area, and determining the x and y coordinates of each focusing point.
The invention further improves the method, and the method also comprises the following steps before the step one:
acquiring conventional seismic records;
and correcting the acquired seismic records to meet the Marchenko Green function reconstruction requirement.
In a second aspect of the present invention, there is provided an MPI-based Marchenko green function reconstruction system, including:
the data preparation unit is used for determining the number of the focus points and the coordinates of each focus point;
the MPI processing unit is connected with the data preparation unit and used for starting an MPI parallel environment, calculating to obtain a Green function and outputting the Green function;
the MPI processing unit performs the following operations:
starting an MPI parallel environment;
setting the number of the computing nodes, numbering the computing nodes from 0, taking the computing node with the number 0 as a main node, and taking other computing nodes as slave nodes;
the main node distributes the task to each slave node according to the coordinate of each focusing point of the focusing target area, and each slave node finishes the task to obtain a focusing wave field;
each slave node calculates by utilizing a focusing wave field to obtain a Green function;
each slave node outputs a Green function;
ending the MPI parallel environment.
A third aspect of the invention. There is provided a computer-readable storage medium storing at least one program executable by a computer, the at least one program, when executed by the computer, causing the computer to perform the steps in the MPI-based Marchenko green function reconstruction method.
Compared with the prior art, the invention has the beneficial effects that: the method greatly improves the computational efficiency of the Markenko Green function reconstruction.
Drawings
FIG. 1 is a block diagram of the steps of the method of the present invention;
FIG. 2 is a concave model;
FIG. 3 is a descending Green's function at a plurality of focal points of the concave model;
FIG. 4 shows the ascending Green's function at a plurality of focal points of the concave model;
FIG. 5 salt dome model;
FIG. 6 is a descending Green's function at a plurality of focal point locations of the salt dome model;
FIG. 7 is an ascending Green's function at a plurality of focal point locations of the salt dome model;
FIG. 8 is a schematic diagram of the system of the present invention.
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings:
MPI is a parallel programming tool widely applied at present, and has the characteristics of good portability, powerful functions, high use efficiency and the like. The invention discloses a Markenko Green function reconstruction method based on MPI (Multi-point information interface), which aims at solving the problems of high-precision Green function reconstruction demand and large amount of computation of multi-focus Markenko Green function reconstruction, establishes the relation between seismic records received from the earth surface and the Green function formed by the excitation of underground virtual seismic source points through the Markenko equation, takes the initial Green function and the corrected seismic records as input, and numerically solves the Markenko equation by an iterative inversion method to obtain the uplink and downlink Green functions; and meanwhile, parallel tasks are divided according to the focus point positions, so that frequent communication among nodes is avoided, and the calculation efficiency is greatly improved.
(1) Principle of method
In Marchenko reconstruction, two mutually independent wavefields can be expressed as a focusing function and a green function, respectively. The correlation between the focus function and the green function can be expressed as
Figure BDA0002717024170000061
In the above formula, G-(xi,x0T) and G+(xi,x0And t) respectively represent the earth's surface x0Excited at point and in the subsurface xiReceiving the obtained uplink Green function and downlink Green function, f1 -(x0,xiT) and f1 +(x0,xiAnd t) is expressed as x in the groundiUp-and down-going focusing functions of focus, R (x)0,x'0And t) is represented by x'0Excitation x at a point0The resulting surface reflection response is received. Denotes convolution symbols, t denotes time (-t is time reversal). In the time-space domain, the distribution ranges of the focusing function and the Green function are different, a window function theta symmetrical in the time direction can be introduced to strip the focusing function and the Green function, and the formula (1) is obtained according to the causal characteristics of signals) Is solved, and can be expressed as
Figure BDA0002717024170000062
Figure BDA0002717024170000071
Figure BDA0002717024170000072
Figure BDA0002717024170000073
D0Representing the surface and the integral is the surface integral. x is the number of0Is the spatial position of the shot point, x'0For the spatial position of the detection point, xiIs the focal point spatial location.
In equations (2) to (5), k is the number of iterations (k.gtoreq.1).
Figure BDA0002717024170000074
For the inverse of the transmission response from the surface excitation point to the subsurface focus point, where the interface is smoother or transmission loss is not a consideration, it can be expressed as
Figure BDA0002717024170000075
In the above formula, Gd(x′0，xi-t) is the initial tracking function Gd(x'0,xiT), the result of the inverse temporal transformation of t),
Figure BDA0002717024170000076
in order to scatter the wake wave,
Figure BDA0002717024170000077
and
Figure BDA0002717024170000078
x in the subsurface for the k-th iteration respectivelyiAn up-line focusing function and a down-line focusing function of the focus,
Figure BDA0002717024170000079
is the initial up-going focusing function. The window function theta is expressed in a specific form
Figure BDA00027170241700000710
In the above formula, tdThe travel time is directly reached from the focal point to the surface seismic source point. Window function theta, time t, spatial position x of underground focus pointiAnd the spatial position x of the earth's surface seismic source point0In correlation, the window function θ mainly ensures that only the wave field information related to the focusing function participates in the calculation process in the iterative calculation process. After multiple iterations, the convergence is gradually realized, and the expression of the focusing function on the scattering tail wave is more accurate. The final iteration result of the focusing function is substituted into the formula 1, and the uplink Green function G can be obtained respectively-(xi,x0T) and a downstream Green function G+(xi,x0T); the uplink Green function and the downlink Green function can accurately present the interlayer multiple information.
The conventional Green function reconstruction method represented by a seismic interference method is limited by the assumed conditions of underground detectors and medium peripheral illumination, is not suitable for the current seismic exploration, can reconstruct an uplink and downlink Green function which contains primary reflection waves and interlayer multiples by a Marchenko Green function reconstruction method only by unilateral illumination, and aims at solving the problems of high precision Green function reconstruction requirements and large calculation amount of multi-focus Marhenko Green function reconstruction.
(2) Technical idea and technical implementation
The method is based on MPI, an initial Green function and a corrected seismic record are used as input, parallel task division is carried out according to the coordinates of a focusing point, and a Marchenko equation is subjected to numerical solution to obtain an uplink Green function and a downlink Green function.
Specifically, a target area is defined facing the imaging requirement, discretization is carried out on the target area, and the total number of focus points is set to be n; solving an equation of a solved equation of an equation before reconstructing a Marchenko of a function of a Green (the equation of an equation of a equation of an equation of a direct wave of a equation of an equation of a solved equation of a equation of an equation of a equation of an equation of a equation of an equation of ad(x'0,xiT), the method for constructing the initial green function may be the existing method, and is not described herein again. ) And obtaining the target area by using the formula (6)
Figure BDA0002717024170000081
(for the invention, "input seismic data and Gd in Gd" in FIG. 1 are input data, which is a directly obtained physical quantity, meanwhile, the seismic data are processed to meet the reconstruction requirement of the Marchenko Green function, specifically, conventional surface seismic records are collected and corrected, the correction method can adopt various existing methods, for example, SRME and the like are used for suppressing surface multiple waves, the wavelet is compressed by a deconvolution method, if the seismic data lack the near-channel data, the reconstruction of the near-channel seismic data and the like are needed, and finally the reconstruction requirement of the Marchenko Green function is met.
Initial green's function G in target aread(x'0,xiT) (initial Green function G, Gd in "input seismic data and Gd" in FIG. 1d(x'0,xiT), which is input data for the present invention, is a physical quantity directly obtained. ) And corrected seismic data R (x)0,x'0And t) as input data reconstructed by a Marchenko Green function.
As shown in fig. 1, an embodiment of a Marchenko green function reconstruction method according to the present invention is as follows:
[ EXAMPLES one ]
The method comprises the following steps:
the method comprises the following steps: dividing a focusing target area according to an imaging target and computing power, determining discrete intervals of focusing points according to geological structure characteristics, and setting the number of total focusing points as n; and gridding the focusing target area, and determining the x and y coordinates of each focusing point. These are all realized by the existing method, and are not described herein again.
Step two: calling MPI _ Init and starting a parallel environment;
step three: setting the number of the computing nodes, numbering the computing nodes from 0, taking the computing node with the number 0 as a master node, and taking other computing nodes as slave nodes; the number of the computing nodes is set by a user according to the condition of the cluster nodes;
step four: the master node distributes tasks to each slave node according to the coordinates of each focusing point in the focusing target area, namely the master node reads in the coordinates of each focusing point in the focusing target area and issues coordinate information to each slave node, and each slave node obtains seismic records corresponding to a work area;
each slave node (a master node distributes tasks, and each slave node is responsible for completing the tasks) acquires task information, a subprogram is called, the subprogram solves a focusing function according to an iterative method shown in formulas (2) to (6), and construction of a focusing wave field is completed;
Figure BDA0002717024170000091
Figure BDA0002717024170000092
Figure BDA0002717024170000093
Figure BDA0002717024170000094
Figure BDA0002717024170000095
step five: solving the green function by each slave node, specifically, substituting the focusing wave field obtained by each slave node through multiple iterations into a formula (1) to obtain an uplink green function and a downlink green function:
Figure BDA0002717024170000096
step six: each slave node calls a file output function and outputs a calculation result of each slave node, namely a Marchenko reconstruction Green function of a plurality of focusing points is output;
step seven: calling MPI _ Finailize, and ending the parallel environment.
The invention also provides a Markenko Green function reconstruction system, and the embodiment of the system is as follows:
[ example two ]
As shown in fig. 8, the system includes:
a data preparation unit 10 for determining the number of focus points and the coordinates of each focus point;
the MPI processing unit 20 is connected with the data preparation unit 10 and used for starting an MPI parallel environment, calculating to obtain a Green function and outputting the Green function;
the MPI processing unit performs the following operations:
starting an MPI parallel environment;
setting the number of the computing nodes, numbering the computing nodes from 0, taking the computing node with the number 0 as a master node, and taking other computing nodes as slave nodes; the number of the computing nodes is set by a user according to the cluster node condition;
the main node distributes the task to each slave node according to the coordinate of each focusing point of the focusing target area, and each slave node finishes the task to obtain a focusing wave field;
each slave node calculates by utilizing a focusing wave field to obtain a Green function;
each slave node outputs a Green function;
ending the MPI parallel environment.
The invention also provides a computer-readable storage medium, which has the following embodiments:
[ EXAMPLE III ]
The computer-readable storage medium stores at least one program executable by a computer, the at least one program causing the computer to perform the steps in the Marchenko green function reconstruction method when executed by the computer.
[ EXAMPLE IV ]
FIG. 2 is a concave model, the size of the model grid is 801 x 401, the sampling interval in the horizontal direction and the vertical direction is 5m, the maximum distance in the horizontal direction is 4005m, the maximum depth in the vertical direction is 2005m, 400 shot points are arranged on the ground surface, a synthetic seismic record is obtained by using a finite difference wave field numerical simulation method, an initial Green function is obtained by solving an equation of equation, the initial Green function and the synthetic seismic record are used as input, the whole model is used as a target area to carry out Marchenko Green function reconstruction, parallel task division is carried out on the target area, the invention totally uses 30 nodes, the CPU model of each node is Intel (R) Xeon (R) E5-2670v2, task division is carried out according to the position of a focus point each time, the iteration times are 15 times when Marchenko Green function reconstruction is carried out, the focus point is selected as (1000 ) m, (1500,1000) m, the downlink green's function and the uplink green's function at (2000,1000) m and (2500,1000) m are shown, and the results are shown in fig. 3 and fig. 4, it can be seen that the uplink green's function and the downlink green's function both represent the primary reflected wave and the inter-layer multiple wave correctly, and the details in the uplink green's function can be found to be richer by comparing the uplink green's function with the downlink green's function.
[ example two ]
Fig. 5 is a salt dome model with a grid size of 120 x 251, the difficulties in green's function reconstruction of the model are salt dome at the middle and thin inter-layer at the top and bottom of the salt dome. The sampling interval of the model in the horizontal direction and the vertical direction is 5m, the maximum distance in the horizontal direction is 6000m, the maximum depth in the vertical direction is 1255m, 600 shot points are arranged on the earth surface in total, a finite difference wave field numerical simulation method is utilized to obtain a synthetic seismic record, 5m grids are adopted for the wave field numerical simulation in the horizontal direction and the vertical direction, the time sampling interval is 0.5ms, meanwhile, an initial green function is obtained by solving a equation of a function and the like, the initial green function and the synthetic seismic record are used as input, the whole model is used as a target area to carry out Marchenko green function reconstruction, parallel task division is carried out on the target area, 30 nodes are used in total, task division is carried out according to the position of a focus point, the iteration times are 15 times when each Marchenko green function reconstruction is carried out, the focus point is selected to be (1000,900) m, (2000,900) m, (3000,9000) m and (4000,900) m, and downlink green function and uplink green function are displayed (as shown in figures 6 and 7), the model test shows that the Marchenko Green function reconstruction method provided by the invention meets the requirements of complex model Green function reconstruction.
The model test can show that the Markhenko Green function reconstruction method based on MPI can meet the requirement of high-precision Green function reconstruction.
The invention belongs to the technical field of seismic data prestack depth migration imaging. Aiming at the problem that the high-precision Green function reconstruction requirement and the multi-focus Markenko Green function reconstruction in the industry have large calculation amount, the invention develops a Markenko Green function reconstruction method based on MPI, which takes an initial Green function and a corrected seismic record as input, numerically solves a Markenko equation by an iteration or direct inversion method to obtain an uplink green function and a downlink green function, and simultaneously performs parallel task division according to focus points, thereby avoiding frequent communication among nodes and greatly improving the calculation efficiency.
Finally, it should be noted that the above-mentioned technical solution is only one embodiment of the present invention, and it will be apparent to those skilled in the art that various modifications and variations can be easily made based on the application method and principle of the present invention disclosed, and the method is not limited to the above-mentioned specific embodiment of the present invention, so that the above-mentioned embodiment is only preferred, and not restrictive.

Claims (10)

1. A Markhenko Green function reconstruction method based on MPI is characterized in that: the method is based on MPI, an initial Green function and corrected seismic data are used as input, parallel task division is carried out according to coordinates of underground focusing points, and an iterative inversion method is used for carrying out numerical solution on a Marchenko equation to obtain an uplink Green function and a downlink Green function.
2. The MPI-based Marchenko green function reconstruction method of claim 1, wherein: the method comprises the following steps:
the method comprises the following steps: determining the number of focusing points and the coordinates of each focusing point;
step two: starting an MPI parallel environment;
step three: setting the number of the computing nodes, numbering the computing nodes from 0, taking the computing node with the number 0 as a main node, and taking other computing nodes as slave nodes;
step four: the master node distributes the task to each slave node, and each slave node finishes the task to obtain a focusing wave field;
step five: each slave node calculates by utilizing a focusing wave field to obtain a Green function;
step six: each slave node outputs a Green function;
step seven: ending the MPI parallel environment.
3. The MPI-based Marchenko green function reconstruction method of claim 2, wherein: the operation of the master node in step four assigning a task to each slave node comprises:
the master node reads in the coordinates of each focus point in the focus target area and issues the coordinates to each slave node.
4. The MPI-based Marchenko Green's function reconstruction method of claim 3, characterized in that: the operation of obtaining a focused wave field from the node completion task in each of the fourth steps comprises:
each slave node obtains a focused wavefield using the following formula:
Figure FDA0002717024160000021
Figure FDA0002717024160000022
Figure FDA0002717024160000023
Figure FDA0002717024160000024
wherein the content of the first and second substances,
Figure FDA0002717024160000025
and
Figure FDA0002717024160000026
are respectively underground xiA down-line focusing function and an up-line focusing function of the k iteration of focusing;
Figure FDA0002717024160000027
scattering tail waves obtained for the kth iteration;
R(x0,x'0and t) is in x'0Excitation x at a point0Receiving the obtained surface reflection response;
Figure FDA0002717024160000028
Figure FDA0002717024160000029
is the inverse of the transmission response from the surface excitation point to the subsurface focus point;
Gd(x′0，xi-t) is an initial green function Gd(x'0,xiT) the result of the inverse temporal transformation;
θ is a window function, as follows:
Figure FDA00027170241600000210
where t is time, tdThe travel time is directly reached from the focal point to the surface seismic source point.
5. The MPI-based Marchenko Green's function reconstruction method of claim 4, characterized in that: the operation of the step five comprises the following steps:
each slave node calculates an uplink Green function and a downlink Green function by using the following formula:
Figure FDA00027170241600000211
wherein G is-(xi,x0T) and G+(xi,x0T) are respectively the earth's surface x0Excited at point and in the subsurface xiReceiving the obtained uplink Green function and downlink Green function;
f1 -(x0,xit) and f1 +(x0,xiT) are each underground xiFocusing up-and down-focusing functions, the values of which are taken
Figure FDA0002717024160000031
And
Figure FDA0002717024160000032
the value of (c).
6. The MPI-based Marchenko Green's function reconstruction method of claim 5, characterized in that: the second step is realized by calling MPI _ Init;
in the sixth step, each slave node outputs the Green function by calling a file output function;
the seventh step is to finish the parallel environment by calling MPI _ Finailze.
7. The MPI-based Marchenko green function reconstruction method of claim 2, wherein: the operation of the first step comprises the following steps:
dividing a focusing target area according to an imaging target and computing power;
determining discrete intervals of focus points according to geological structure characteristics, and setting the number of total focus points as n;
and gridding the focusing target area, and determining the x and y coordinates of each focusing point.
8. The MPI-based Marchenko green function reconstruction method of claim 2, wherein: before the step one, the method further comprises the following steps:
acquiring conventional seismic records;
and correcting the acquired seismic records to meet the Marchenko Green function reconstruction requirement.
9. A Markhenko Green function reconstruction system based on MPI is characterized in that: the system comprises:
the data preparation unit is used for determining the number of the focus points and the coordinates of each focus point;
the MPI processing unit is connected with the data preparation unit and used for starting an MPI parallel environment, calculating to obtain a Green function and outputting the Green function;
the MPI processing unit performs the following operations:
starting an MPI parallel environment;
setting the number of the computing nodes, numbering the computing nodes from 0, taking the computing node with the number 0 as a main node, and taking other computing nodes as slave nodes;
the main node distributes the task to each slave node according to the coordinate of each focusing point of the focusing target area, and each slave node finishes the task to obtain a focusing wave field;
each slave node calculates by utilizing a focusing wave field to obtain a Green function;
each slave node outputs a Green function;
ending the MPI parallel environment.
10. A computer-readable storage medium characterized by: the computer-readable storage medium stores at least one program executable by a computer, the at least one program causing the computer to perform the steps in the MPI-based Marchenko green function reconstruction method according to any one of claims 1 to 8.
CN202011076701.5A 2020-10-10 2020-10-10 Markenko Green function reconstruction method and system based on MPI Pending CN114428342A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011076701.5A CN114428342A (en) 2020-10-10 2020-10-10 Markenko Green function reconstruction method and system based on MPI

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011076701.5A CN114428342A (en) 2020-10-10 2020-10-10 Markenko Green function reconstruction method and system based on MPI

Publications (1)

Publication Number Publication Date
CN114428342A true CN114428342A (en) 2022-05-03

Family

ID=81309162

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011076701.5A Pending CN114428342A (en) 2020-10-10 2020-10-10 Markenko Green function reconstruction method and system based on MPI

Country Status (1)

Country Link
CN (1) CN114428342A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115184999A (en) * 2022-07-06 2022-10-14 吉林大学 Marchenko imaging focusing function correction method based on deep learning

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115184999A (en) * 2022-07-06 2022-10-14 吉林大学 Marchenko imaging focusing function correction method based on deep learning
CN115184999B (en) * 2022-07-06 2024-04-19 吉林大学 Marchenko imaging focusing function correction method based on deep learning

Similar Documents

Publication Publication Date Title
US4964103A (en) Three dimensional before stack depth migration of two dimensional or three dimensional seismic data
WO2018129844A1 (en) Seismic diffracted wave separation method and device
CA2616379C (en) Method for tomographic inversion by matrix transformation
EP2715405A1 (en) Method of processing seismic data by providing surface offset common image gathers
EP3259619A1 (en) Method and system of processing seismic data by providing surface aperture common image gathers
CN102944896A (en) Model method static correction method for surface survey data
CN116774292B (en) Seismic wave travel time determining method, system, electronic equipment and storage medium
CN114839673B (en) Separation method, separation system and computer equipment for multi-seismic-source efficient acquisition wave field
Gjøystdal et al. Improved applicability of ray tracing in seismic acquisition, imaging, and interpretation
Ma et al. Topography-dependent eikonal traveltime tomography for upper crustal structure beneath an irregular surface
CN103119472A (en) Hybrid method for full waveform inversion using simultaneous and sequential source method
EP3612863B1 (en) Post-stack kirchhoff depth de-migration method for tilted transverse isotropic (tti) and heterogeneous media based on ray tracing on migrated data
CN114428342A (en) Markenko Green function reconstruction method and system based on MPI
US5587942A (en) 3D wave equation migration of a 2D grid of seismic data
US5764514A (en) Method for modelling kinematic seismic data processed with at least one motion operator
AU739128B2 (en) A method of seismic processing, and in particular a 3D seismic prospection method implementing seismic data migration
Hardy High resolution tomographic MVA with automation
Colavitti et al. A new approach to construct 3-D crustal shear-wave velocity models: method description and application to the Central Alps
Snopek Inversion of gravity data with application to density modeling of the Hellenic subduction zone
Verdel et al. Non-linear Gaussian beam tomography-a new method capable of determining highly complex velocity models
Decker et al. A variational approach for picking optimal surfaces from semblance-like panels
WO2024087126A1 (en) Method and system for determining migration data using cross-spread gathers and parallel processing
CN110954956B (en) Method for evaluating acquisition trace of observation system and computer-readable storage medium
US20160238727A1 (en) Removal of an Estimated Acquisition Effect from a Marine Survey Measurement
RU2794968C1 (en) Simulation of seismic velocities

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination