CN111580181B - Water guide collapse column identification method based on multi-field multi-feature information fusion - Google Patents
Water guide collapse column identification method based on multi-field multi-feature information fusion Download PDFInfo
- Publication number
- CN111580181B CN111580181B CN202010322895.6A CN202010322895A CN111580181B CN 111580181 B CN111580181 B CN 111580181B CN 202010322895 A CN202010322895 A CN 202010322895A CN 111580181 B CN111580181 B CN 111580181B
- Authority
- CN
- China
- Prior art keywords
- seismic
- attribute
- neural network
- pso
- water guide
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V11/00—Prospecting or detecting by methods combining techniques covered by two or more of main groups G01V1/00 - G01V9/00
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geophysics (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses a water guide collapse column identification method based on multi-field multi-feature information fusion, which comprises the following steps of: s1, extracting seismic attributes of coal mine seismic exploration and resistivity attributes of electromagnetic exploration; s2, preprocessing the seismic attribute features, and further optimizing the features to obtain a set of optimized seismic attribute features; s3, training and testing the PSO-BP neural network; s4, carrying out classification and identification on the seismic attribute and the resistivity attribute on the two-dimensional section to be tested by adopting the trained PSO-BP neural network; and S5, reconstructing the three-dimensional surface of the water guide trapping column according to the recognition result of the two-dimensional section, and recognizing the form of the trapping column and the water distribution state in the three-dimensional space. The method selects various seismic attributes and resistivity attributes as characteristics, and uses the PSO-BP neural network to quickly identify the water guide collapse column, so that the identification accuracy of the coal mine water guide collapse column can be greatly improved.
Description
Technical Field
The invention relates to the technical field of seismic data interpretation, in particular to a water guide collapse column identification method based on multi-field multi-feature information fusion.
Background
Coal seams in China are complex in occurrence conditions, deep in burial, complex in geological structure, high in gas, high in ground stress, high in ground temperature and other unsafe factors, and are prone to cause gas explosion, water penetration, roof fall and other geological disasters, wherein water inrush from mines is one of the main disasters threatening safety production of coal mines, and in major accidents of coal mines in China, casualties and economic losses caused by water inrush accidents are great, and damage is quite serious. Therefore, how to improve the accuracy of water channel detection is a hot spot of the current coal mine safety mining research.
The coal mine water guide channel mainly comprises a water guide trapping column, a fault, a mining empty water accumulation area and the like, and the water guide trapping column is the main water guide channel. The technical methods commonly used for detecting the water guide channel at present are as follows: the seismic prospecting method comprises the steps of seismic prospecting, a transient electromagnetic method, a geoelectromagnetic method, a geological radar, nuclear magnetic resonance and the like, wherein the seismic prospecting and the electromagnetic prospecting (the transient electromagnetic method and the geoelectromagnetic method) have good effects. The methods have respective characteristics and advantages according to the field detection conditions and the characteristics of the target body, but certain limitations still exist in the aspects of methods and technologies, and the water guide collapse columns are difficult to identify on a two-dimensional section of a single exploration method.
The electromagnetic exploration has relatively remarkable effect on identifying and finding out the water-rich property of the water guide channel, and has the advantages of strong penetration high-resistance capacity, sensitive response to a low-resistance water-rich area and the like, but due to the volume effect, a low-resistance abnormal body on a resistivity section diagram is often presented as a low-resistance area, and the specific boundary and space occurrence scale of the water guide channel is difficult to define, namely, the electromagnetic exploration can determine the water distribution state of the water guide channel, but cannot accurately determine the position and the form of the water guide channel.
The seismic exploration method has high-resolution characteristics and can detect the position and the shape of the water guide channel. In recent years, with the common application of intelligent technology and visualization technology, the seismic attribute analysis method is developed rapidly and widely applied to structural interpretation, lithology interpretation, reservoir evaluation and oil reservoir description, so that coherent body technology and multi-attribute analysis are also developed greatly. When earthquake interpretation is carried out on water guide channels such as collapse columns, faults and goafs, in order to effectively identify the positions and the forms of the water guide channels and improve the quantitative analysis precision, the fine interpretation is carried out by using a multi-attribute comprehensive analysis technology, which is the most appropriate means. However, the seismic exploration method cannot identify the water distribution state of the water guide channel.
Therefore, how to provide a water diversion collapse column identification method based on multi-field multi-feature information fusion, which can reduce errors caused by a single exploration method, is a problem to be solved by those skilled in the art.
Disclosure of Invention
In view of the above, the invention provides a water guide trapping column identification method based on multi-field multi-feature information fusion, which can accurately identify and position the water guide trapping column by combining a seismic wave field of seismic exploration and electromagnetic field features of electromagnetic exploration, and solves the problems that the position, the form and the water distribution state of the water guide trapping column cannot be comprehensively identified and have errors by using one detection method alone.
In order to achieve the purpose, the invention adopts the following technical scheme:
a water guide collapse column identification method based on multi-field multi-feature information fusion comprises the following steps:
s1, extracting seismic attribute characteristics and electromagnetic resistivity attribute characteristics respectively aiming at two kinds of physical field data of a seismic wave field of coal mine seismic exploration and an electromagnetic field of electromagnetic method exploration;
s2, preprocessing all the acquired seismic attribute features and the acquired electromagnetic resistivity attribute features, and further optimizing the preprocessed seismic attribute features to obtain an optimized seismic attribute feature set;
s3, establishing a PSO-BP neural network; taking the optimal seismic attribute feature and the electromagnetic resistivity attribute feature as input data of the PSO-BP neural network for training and testing to obtain a trained PSO-BP neural network;
s4, carrying out classification and identification on the seismic attribute and the resistivity attribute on the two-dimensional section to be tested by adopting the trained PSO-BP neural network;
and S5, reconstructing the three-dimensional surface of the water guide trapping column according to the recognition result of the two-dimensional section, and recognizing the form of the water guide trapping column and the water distribution state in the three-dimensional space.
Preferably, S1 specifically includes the following:
processing and interpreting coal mine seismic wave field data to obtain a depth domain stacking seismic data volume, and extracting the seismic attribute characteristics; and performing time-depth conversion on the electromagnetic field data of the coal mine, and calculating the resistivity attribute characteristics.
Preferably, S2 specifically includes the following:
s21, normalizing all extracted feature data;
and S22, performing correlation analysis on the normalized seismic attribute feature data, calculating a correlation coefficient, performing cluster analysis according to the correlation coefficient, and performing optimization on the seismic attribute features by combining the geological significance and the correlation coefficient of each attribute to obtain a set of optimized seismic attribute features.
Preferably, the establishing of the PSO-BP neural network in S3 specifically includes the following contents:
the PSO-BP neural network comprises a PSO algorithm and a BP neural network, the PSO algorithm optimizes an initial connection weight and a threshold of the BP neural network, and the BP neural network comprises an input layer, a hidden layer and an output layer;
determining the number of neurons of the hidden layer of the BP neural network, and determining an activation function, a training algorithm, a weight learning function, a performance function, a training target minimum error and a learning rate; determining PSO algorithm parameters.
Preferably, the training and testing of the PSO-BP neural network in S3 specifically includes the following contents:
(4) acquiring preferred seismic attribute characteristic data and resistivity attribute characteristic data of the collapse column as a sample data set;
(5) randomly selecting partial samples from the sample data set as a training set, and training the PSO-BP neural network; the residual sample data is used as a test set, and the trained network is input to calculate the accuracy;
(6) and if the accuracy does not meet the requirement, further adjusting the neural network parameters until the requirement is met.
Preferably, S4 specifically includes the following:
and acquiring the optimal seismic attribute feature and the resistivity attribute feature data aiming at different two-dimensional sections to be tested, and respectively inputting the optimal seismic attribute feature and the resistivity attribute feature data into the trained PSO-BP neural network for classification and identification, thereby realizing the identification of the position, the shape and the water distribution condition of the water guide collapse column on each two-dimensional section.
Preferably, S5 specifically includes the following:
and according to the water guide trapping columns recognized on different two-dimensional sections, the three-dimensional surface morphology of the water guide trapping columns is reconstructed through a Poisson surface reconstruction algorithm, and the three-dimensional morphology and the water distribution scale of the water guide trapping columns are recognized.
According to the technical scheme, compared with the prior art, the water guide collapse column identification method based on multi-field multi-feature information fusion is provided, the method combines two kinds of physical field data of a seismic wave field and an electromagnetic field, selects multiple seismic attributes and resistivity attributes as features, and uses a PSO-BP neural network to quickly identify the water guide collapse column, so that accurate identification of the position, the three-dimensional form and the water distribution scale of the water guide collapse column in a coal mine can be realized, the identification error of the water guide collapse column caused by the defect of a single geophysical method is effectively avoided, and the identification accuracy of the water guide collapse column is greatly improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a schematic flow chart of a water guide trapping column recognition method based on multi-field multi-feature information fusion, provided by the invention;
FIG. 2 is a schematic diagram of a three-layer BP neural network water guide channel prediction structure model in a water guide trapping column recognition method based on multi-field multi-feature information fusion, provided by the invention;
fig. 3 is a flow chart of a PSO-BP neural network algorithm in the water diversion trapping column identification method based on multi-field multi-feature information fusion provided by the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention discloses a water guide collapse column identification method based on multi-field multi-feature information fusion, which comprises the following steps of:
s1, extracting seismic attribute characteristics and resistivity attribute characteristics respectively aiming at two kinds of physical field data of a seismic wave field of coal mine seismic exploration and an electromagnetic field of electromagnetic method exploration;
s2, preprocessing all acquired seismic attribute features and resistivity attribute features, and further optimizing the preprocessed seismic attribute features to obtain a set of optimized seismic attribute features;
s3, establishing a PSO-BP neural network; training and testing the optimized seismic attribute characteristics and resistivity attribute characteristics as input data of the PSO-BP neural network to obtain a trained PSO-BP neural network;
s4, carrying out classification and identification on the seismic attribute and the resistivity attribute on the two-dimensional section to be tested by adopting the trained PSO-BP neural network;
and S5, reconstructing the three-dimensional surface of the water guide trapping column according to the recognition result of the two-dimensional section, and recognizing the form of the water guide trapping column and the water distribution state in the three-dimensional space.
In order to further implement the above technical solution, S1 specifically includes the following contents:
processing and interpreting coal mine seismic wave field data, acquiring a depth domain stacking seismic data volume, and extracting seismic attribute characteristics; the seismic attribute characteristics comprise 26 seismic attributes including absorption attenuation, envelope, instantaneous bandwidth, reflection intensity, relative acoustic impedance, root-mean-square amplitude, chaotic body, phase cosine, dip deviation, main frequency, instantaneous gradient magnitude, instantaneous frequency, curvature consistency evaluation, variance, instantaneous phase, "sweet spot", wave impedance, three-dimensional edge reinforcement, amplitude ratio, three-dimensional curvature, dip consistency evaluation, dip illumination, pattern equalization, median filtering, deviation removal and seismic trace gradient;
performing time-depth conversion on the electromagnetic field data of the coal mine, and calculating the resistivity attribute characteristics; wherein the resistivity attribute characteristics include apparent resistivity values and inverted resistivity values.
In order to further implement the above technical solution, S2 specifically includes the following contents:
s21, extracting 26 kinds of seismic attribute feature data of water-conducting collapse columns at coal mine drilling positions and roadway positions and 2 kinds of attribute feature data of apparent resistivity values and inversion resistivity values, normalizing the 28 kinds of feature data to be in a (0, 1) range, and then calculating correlation coefficients for the 26 kinds of seismic attributes; the calculation formula is as follows:
Rijin the range of (-1,1), the correlation coefficient of any two attributes is closer to 1 or-1, the higher the correlation is, the closer to 0, the lower the correlation is; m is the number of attributes; n is the total number of each attribute; x is the number ofki、xkjThe ith and jth attribute values of the kth number respectively;
are respectively of the ith kind anda j-th attribute mean;
and S22, performing correlation analysis on the normalized seismic attribute feature data, calculating a correlation coefficient, performing cluster analysis according to the correlation coefficient, and optimizing the seismic attribute features by combining the geological significance and the correlation coefficient of each attribute to obtain an optimal seismic attribute feature set.
Specifically, during cluster analysis, the lower the correlation coefficient between attributes is, the larger the value after clustering is; the higher the correlation coefficient between attributes, the smaller the value after clustering.
According to the clustering result, the 26 seismic attribute features can be divided into the following different attribute clusters according to the difference of the correlation coefficients:
the four attributes of graph balance, median filtering, deviation removal and seismic channel gradient have high correlation; the envelope, the reflection intensity, the root mean square amplitude, the instantaneous gradient magnitude, the three-dimensional edge enhancement and the 'sweet spot' attribute correlation are high; the variance, the amplitude ratio, the dip angle consistency failure meter and the dip angle illumination attribute correlation are high; the evaluation attribute correlation of the main frequency, the instantaneous frequency and the curvature consistency is high; the correlation among the attributes of absorption attenuation, chaotic bodies, phase cosines, dip angle deviation, wave impedance, instantaneous phase, three-dimensional curvature, instantaneous bandwidth and relative wave impedance is low.
Through comparison, in combination with the geological significance of each attribute, one attribute is selected from the attribute clusters with high correlation and high correlation as an optimal seismic attribute feature, a proper amount of attributes are selected from the attributes with low correlation as the optimal seismic attribute feature, and the following 10 optimal seismic attribute features can be selected: absorption attenuation, root mean square amplitude, chaos, phase cosine, inclination deviation, instantaneous frequency, variance, instantaneous phase, wave impedance and median filtering.
In order to further implement the above technical solution, a PSO-BP neural network is established in S3, where the PSO-BP neural network includes a PSO algorithm and a BP neural network, the PSO algorithm optimizes an initial connection weight and a threshold of the BP neural network, and the BP neural network includes an input layer, a hidden layer, and an output layer; the method specifically comprises the following steps:
a neural network prediction model of the method adopts a three-layer BP network structure, namely an input layer, a hidden layer and an output layer. The three-layer BP neural network water guide channel prediction structure model is shown in figure 2.
Determining the number of hidden layer neurons of the BP neural network, and selecting the appropriate number of hidden neurons by adopting the following expression:
in the formula: m represents the number of cryptic neurons; n represents the number of neurons in the input layer; l represents the number of neurons in the output layer; a [1,10] is set.
According to the above formula, the input features of the water trapping column prediction are 12 (10 seismic attribute features and 2 resistivity attribute features), that is, n is 12, and the output is 1, that is, l is 1. The range of the number m of the hidden layer neurons is 4-14. And (3) carrying out tests on m under the condition that parameters such as a sample set, iteration times, target precision and an activation function are the same, and preferably determining that the number of the hidden layer neurons is 13 through calculation.
3. The selection of the activation function is to combine three commonly used activation functions tanh, sigmoid and purelin (linearity) to form 9 groups of combinations of the input layer activation function and the output layer activation function, and on the basis of the selection of the default function, the accuracy is compared through multiple tests to find that preferably, the hidden layer activation function is a sigmoid function, and the output layer activation function is a tanh function.
4. And (3) selecting a training algorithm, and calculating the accuracy of the test set for 11 matlab common training functions under the condition of a certain neural network structure. Preferably, the Levenberg-Marquardt algorithm with higher accuracy is selected by the training algorithm. The 11 common training functions include: the method comprises the following steps of a steepest descent BP algorithm, a momentum BP algorithm, an elastic BP algorithm, a Fletcher-Reeves correction algorithm, a Polak _ Ribier correction algorithm, a Powell-Belle reset algorithm, a BFGS quasi-Newton algorithm, an OSS algorithm, a Levenberg-Marquardt algorithm, a normalized conjugate gradient method and a Bayesian rule method.
Selecting other parameters of the BP neural network, preferably selecting a learnd function by a weight learning function, selecting a mean square error (mse) as a performance function of the network, setting a minimum error of a training target to be 0.001 and setting a learning rate to be 0.01.
And 6, determining parameters of a PSO algorithm, and optimizing the initial connection weight and the threshold of the BP neural network by the PSO algorithm. FIG. 3 is a flow chart of PSO-BP neural network algorithm, and the fitness curve of the particle swarm algorithm is decreased gradually when the evolution algebra is 20. Preferably, the evolution algebra M is 20, the population size N is 20, and the inertial weight ω ismax＝0.9、ωmin0.4, acceleration constant c1 c2 1.49445, speed margin [ -1,1]Particle boundary [ -5,5 [)]。
In order to further implement the above technical solution, the training and testing of the PSO-BP neural network in S3 specifically includes the following contents:
(1) acquiring preferable seismic attribute characteristic data and resistivity attribute characteristic data of the collapse columns at the drilling position and the roadway as sample data sets, and marking the sample data into 3 types according to actual conditions, namely no collapse column, no collapse column containing water and collapse column containing water; constructing a PSO-BP neural network by utilizing matlab, wherein the input is the optimized 12 characteristics, and the number of input layer neurons of BP is 12; the number of hidden layer neurons is 13; the number of neurons in an output layer is 1, the output data comprises 3 types, the non-trapping column is 1, the trapping column contains no water and is 2, and the trapping column contains water and is 3.
(2) Randomly selecting 80% of samples from the sample data set as a training set, and training the PSO-BP neural network; the residual 20% of sample data is used as a test set, the trained network is input, and the accuracy between the output value and the true value is calculated;
(3) and if the accuracy does not meet the requirement, further adjusting the neural network parameters until the requirement is met.
In order to further implement the above technical solution, S4 specifically includes the following contents:
the trained PSO-BP neural network carries out classification and identification on the two-dimensional section, and on the two-dimensional section, the 12 characteristic parameters in the seismic wave field and the electromagnetic field data are extracted and input into the trained PSO-BP neural network to calculate an output value. And mapping the output to obtain the position, the form and the water distribution condition recognition result of the trapping column on the two-dimensional section.
In order to further implement the above technical solution, S5 specifically includes the following contents:
according to the water guide trapping column data on the identified two-dimensional sections, firstly, converting the two-dimensional data into three-dimensional data, then, establishing an octree space by using a Poisson surface reconstruction algorithm on a three-dimensional point set, calculating a vector field, solving an indication function by using a Poisson equation, and further extracting a proper isosurface to obtain the three-dimensional surface form and the water distribution scale of the water guide trapping column.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (6)
1. A water guide collapse column identification method based on multi-field multi-feature information fusion is characterized by comprising the following steps:
s1, extracting seismic attribute characteristics and resistivity attribute characteristics respectively aiming at two kinds of physical field data of a seismic wave field of coal mine seismic exploration and an electromagnetic field of electromagnetic method exploration;
s2, preprocessing all the acquired seismic attribute features and the acquired resistivity attribute features, and further optimizing the preprocessed seismic attribute features to obtain an optimized seismic attribute feature set;
s21, extracting 26 kinds of seismic attribute feature data of water-conducting collapse columns at coal mine drilling positions and roadway positions and 2 kinds of attribute feature data of apparent resistivity values and inversion resistivity values, normalizing the 28 kinds of feature data to be in a (0, 1) range, and then calculating correlation coefficients for the 26 kinds of seismic attributes; the calculation formula is as follows:
Rijin the range of (-1,1), the correlation coefficient of any two attributes is closer to 1 or-1, the higher the correlation is, the closer to 0, the lower the correlation is; m is the number of attributes; n is the total number of each attribute; x is the number ofki、xkjThe ith and jth attribute values of the kth number respectively;
the average values of the ith and jth attributes are respectively;
s22, performing correlation analysis on the normalized seismic attribute feature data, calculating a correlation coefficient, performing cluster analysis according to the correlation coefficient, and optimizing the seismic attribute features by combining the geological significance and the correlation coefficient of each attribute to obtain an optimized seismic attribute feature set;
s3, establishing a PSO-BP neural network; taking the preferred seismic attribute feature and the resistivity attribute feature as input data of the PSO-BP neural network for training and testing to obtain a trained PSO-BP neural network;
s4, carrying out classification and identification on the seismic attribute and the resistivity attribute on the two-dimensional section to be tested by adopting the trained PSO-BP neural network;
and S5, reconstructing the three-dimensional surface of the water guide trapping column according to the recognition result of the two-dimensional section, and recognizing the form of the trapping column and the water distribution state in the three-dimensional space.
2. The method for identifying the water guide trapping column based on the fusion of the multi-field and multi-feature information as claimed in claim 1, wherein the S1 specifically comprises the following contents:
processing and interpreting coal mine seismic wave field data to obtain a depth domain stacking seismic data volume, and extracting the seismic attribute characteristics; and performing time-depth conversion on the electromagnetic field data of the coal mine, and calculating the resistivity attribute characteristics.
3. The method for identifying the water diversion trapping column based on the multi-field multi-feature information fusion as claimed in claim 1, wherein the establishing of the PSO-BP neural network in S3 specifically includes the following steps:
the PSO-BP neural network comprises a PSO algorithm and a BP neural network, the PSO algorithm optimizes an initial connection weight and a threshold of the BP neural network, and the BP neural network comprises an input layer, a hidden layer and an output layer;
determining the number of neurons of the hidden layer of the BP neural network, and determining an activation function, a training algorithm, a weight learning function, a performance function, a training target minimum error and a learning rate; determining PSO algorithm parameters.
4. The method for identifying the water diversion trapping column based on the fusion of the multi-field and multi-feature information as claimed in claim 1, wherein the training and testing of the PSO-BP neural network in S3 specifically comprises the following steps:
(1) acquiring preferred seismic attribute characteristic data and resistivity attribute characteristic data of the collapse column as a sample data set;
(2) randomly selecting partial samples from the sample data set as a training set, and training the PSO-BP neural network; the residual sample data is used as a test set, and the trained network is input to calculate the accuracy;
(3) and if the accuracy does not meet the requirement, further adjusting the neural network parameters until the requirement is met.
5. The method for identifying the water guide trapping column based on the fusion of the multi-field and multi-feature information as claimed in claim 1, wherein the S4 specifically comprises the following contents:
and acquiring the optimal seismic attribute feature and the resistivity attribute feature data aiming at different two-dimensional sections to be tested, and respectively inputting the optimal seismic attribute feature and the resistivity attribute feature data into the trained PSO-BP neural network for classification and identification, thereby realizing the identification of the position, the shape and the water distribution condition of the water guide collapse column on each two-dimensional section.
6. The method for identifying the water guide trapping column based on the fusion of the multi-field and multi-feature information as claimed in claim 1, wherein the S5 specifically comprises the following contents:
and according to the water guide trapping columns recognized on different two-dimensional sections, the three-dimensional surface morphology of the water guide trapping columns is reconstructed through a Poisson surface reconstruction algorithm, and the three-dimensional morphology and the water distribution scale of the water guide trapping columns are recognized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010322895.6A CN111580181B (en) | 2020-04-22 | 2020-04-22 | Water guide collapse column identification method based on multi-field multi-feature information fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010322895.6A CN111580181B (en) | 2020-04-22 | 2020-04-22 | Water guide collapse column identification method based on multi-field multi-feature information fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111580181A CN111580181A (en) | 2020-08-25 |
| CN111580181B true CN111580181B (en) | 2021-07-20 |
Family
ID=72124522
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010322895.6A Active CN111580181B (en) | 2020-04-22 | 2020-04-22 | Water guide collapse column identification method based on multi-field multi-feature information fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111580181B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111880241B (en) * | 2020-09-07 | 2022-12-27 | 中国煤炭地质总局地球物理勘探研究院 | Seismoelectric data multi-attribute fusion technology |
| CN112305614B (en) * | 2020-10-20 | 2024-03-29 | 中海石油(中国)有限公司 | Method and system for describing space spreading range of gas cloud area |
| CN113640876B (en) * | 2021-07-09 | 2023-05-30 | 中国煤炭地质总局地球物理勘探研究院 | Method for carrying out fine recognition on collapse column by utilizing chaos attribute |
| CN113742995B (en) * | 2021-07-28 | 2024-07-09 | 淄博矿业集团有限责任公司 | Mine water inflow prediction method and system based on coal mine big data |
| CN113805250B (en) * | 2021-08-23 | 2024-01-12 | 鞍钢集团矿业有限公司 | High-density earthquake and electric method combined detection and identification method for goaf of strip mine |
| CN114000827B (en) * | 2021-11-05 | 2023-07-14 | 中国矿业大学 | Collapse column exploration method based on pressure seepage principle and electromagnetic response characteristics |
| CN115577616A (en) * | 2022-09-15 | 2023-01-06 | 福瑞升(成都)科技有限公司 | Carbonatite fracture-cave earthquake depicting method and device based on deep learning |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3975674A (en) * | 1972-09-29 | 1976-08-17 | Mceuen Robert B | Geothermal exploration method utilizing electrical resistivity and seismic velocity |
| US7023213B2 (en) * | 2002-12-10 | 2006-04-04 | Schlumberger Technology Corporation | Subsurface conductivity imaging systems and methods |
| EP2204673A2 (en) * | 2009-01-05 | 2010-07-07 | PGS Geophysical AS | Combined electromagnetic and seismic acquisition system and method |
| CN105223612A (en) * | 2015-06-10 | 2016-01-06 | 中国矿业大学 | A kind of coal mine flood prediction and evaluation method based on earthquake information |
| CN107102379A (en) * | 2016-02-19 | 2017-08-29 | 师素珍 | A kind of method that seat earth watery prediction is carried out based on many attribution inversions |
| EP2024891B1 (en) * | 2006-04-28 | 2019-01-09 | KJT Enterprises, Inc. | Integrated earth formation evaluation method using controlled source electromagnetic survey data and seismic data |
| CN110968826A (en) * | 2019-10-11 | 2020-04-07 | 重庆大学 | Magnetotelluric deep neural network inversion method based on spatial mapping technology |
| CN111042866A (en) * | 2019-12-30 | 2020-04-21 | 安徽惠洲地质安全研究院股份有限公司 | Multi-physical-field cooperative water inrush monitoring method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9519072B2 (en) * | 2006-05-11 | 2016-12-13 | Schlumberger Technology Corporation | Method and apparatus for locating gas hydrate |
| US8787111B2 (en) * | 2011-02-02 | 2014-07-22 | Westerngeco L.L.C. | Devices and methods for positioning TOWs in marine seismic systems |
| US9110195B2 (en) * | 2011-04-14 | 2015-08-18 | Wen J. Whan | Electromagnetic and its combined surveying apparatus and method |
| CN103833275A (en) * | 2014-01-07 | 2014-06-04 | 山东大学 | Similar material for combined detection of physical model test and preparation method thereof |
| US9939548B2 (en) * | 2014-02-24 | 2018-04-10 | Saudi Arabian Oil Company | Systems, methods, and computer medium to produce efficient, consistent, and high-confidence image-based electrofacies analysis in stratigraphic interpretations across multiple wells |
| CN104237970B (en) * | 2014-09-23 | 2017-07-07 | 中国石油天然气集团公司 | Electromagnetism of Earthquake joint exploration system and its data acquisition device and collecting method |
| CN105607147B (en) * | 2015-12-21 | 2018-07-13 | 中南大学 | A kind of method and system of inverting shale gas reservoir resistivity |
-
2020
- 2020-04-22 CN CN202010322895.6A patent/CN111580181B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3975674A (en) * | 1972-09-29 | 1976-08-17 | Mceuen Robert B | Geothermal exploration method utilizing electrical resistivity and seismic velocity |
| US7023213B2 (en) * | 2002-12-10 | 2006-04-04 | Schlumberger Technology Corporation | Subsurface conductivity imaging systems and methods |
| EP2024891B1 (en) * | 2006-04-28 | 2019-01-09 | KJT Enterprises, Inc. | Integrated earth formation evaluation method using controlled source electromagnetic survey data and seismic data |
| EP2204673A2 (en) * | 2009-01-05 | 2010-07-07 | PGS Geophysical AS | Combined electromagnetic and seismic acquisition system and method |
| CN105223612A (en) * | 2015-06-10 | 2016-01-06 | 中国矿业大学 | A kind of coal mine flood prediction and evaluation method based on earthquake information |
| CN107102379A (en) * | 2016-02-19 | 2017-08-29 | 师素珍 | A kind of method that seat earth watery prediction is carried out based on many attribution inversions |
| CN110968826A (en) * | 2019-10-11 | 2020-04-07 | 重庆大学 | Magnetotelluric deep neural network inversion method based on spatial mapping technology |
| CN111042866A (en) * | 2019-12-30 | 2020-04-21 | 安徽惠洲地质安全研究院股份有限公司 | Multi-physical-field cooperative water inrush monitoring method |
Non-Patent Citations (2)
| Title |
|---|
| 基于粒子群神经网络的底板突水预测研究;薛茹 等;《能源技术与管理》;20101231;第6-8页 * |
| 综合物探技术在工作面导水构造探测中的应用;李江华 等;《煤矿安全》;20180331;第49卷(第3期);第129-132页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111580181A (en) | 2020-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111580181B (en) | Water guide collapse column identification method based on multi-field multi-feature information fusion | |
| CN113759425B (en) | Method and system for evaluating filling characteristics of deep paleo-karst reservoir stratum by well-seismic combination | |
| CN109709603B (en) | Seismic horizon identification and tracking method and system | |
| CN111239802B (en) | Deep learning speed modeling method based on seismic reflection waveform and velocity spectrum | |
| CN105259572B (en) | The seismic facies computational methods classified automatically based on seismic multi-attribute parametrical nonlinearity | |
| CN104502997B (en) | A kind of method of utilization fracture spacing curve prediction fracture spacing body | |
| CN110333551A (en) | Dolostone reservoirs prediction technique, system and the storage medium combined based on well shake | |
| CN106707340A (en) | Method for predicting volcanic rock facies | |
| CN111596978A (en) | Web page display method, module and system for lithofacies classification by artificial intelligence | |
| CN112883564B (en) | Water body temperature prediction method and prediction system based on random forest | |
| CN103293551A (en) | Model constraint based impedance inversion method and model constraint based impedance inversion system | |
| CN102478668A (en) | Method for applying seismic multiattribute parameters to predicting coal seam thickness | |
| CN109085663A (en) | A kind of tight sandstone reservoir stratification seam recognition methods | |
| CN103487832A (en) | Method for classifying supervised waveforms in three-dimensional seismic signal | |
| CN109763814B (en) | Stratum matching visual analysis method based on multi-dimensional logging data | |
| CN107436452A (en) | Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm | |
| CN115291281A (en) | Real-time micro-earthquake magnitude calculation method and device based on deep learning | |
| CN108226997A (en) | Seismic facies division method based on pre-stack seismic data | |
| CN106033127B (en) | Crustal stress azimuthal seismic Forecasting Methodology based on shear wave velocity rate of change | |
| CN107218033A (en) | A kind of inversion method for the method and formation parameter for recognizing attitude of stratum | |
| CN110609327B (en) | Carbonate reservoir facies prediction method and device based on pre-stack seismic attributes | |
| CN116047583A (en) | Adaptive wave impedance inversion method and system based on depth convolution neural network | |
| CN108680950B (en) | A kind of desert seismic signal method for detecting position based on Self-adaptive Block Matching | |
| CN105676288B (en) | A kind of sand body stacked relation method of discrimination based on reflectance factor analysis | |
| CN113219531A (en) | Method and device for identifying gas-water distribution of tight sandstone |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |