CN110554427A - Lithology combination prediction method based on forward modeling of seismic waveform - Google Patents
Lithology combination prediction method based on forward modeling of seismic waveform Download PDFInfo
- Publication number
- CN110554427A CN110554427A CN201910664786.XA CN201910664786A CN110554427A CN 110554427 A CN110554427 A CN 110554427A CN 201910664786 A CN201910664786 A CN 201910664786A CN 110554427 A CN110554427 A CN 110554427A
- Authority
- CN
- China
- Prior art keywords
- seismic
- data
- lithology
- speed
- combination
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000012549 training Methods 0.000 claims abstract description 39
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 20
- 238000004458 analytical method Methods 0.000 claims abstract description 12
- 238000013135 deep learning Methods 0.000 claims abstract description 10
- 238000010183 spectrum analysis Methods 0.000 claims abstract description 7
- 238000011160 research Methods 0.000 claims abstract description 5
- 238000005070 sampling Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 238000011176 pooling Methods 0.000 claims description 5
- 238000002790 cross-validation Methods 0.000 claims description 4
- 238000004088 simulation Methods 0.000 claims description 2
- 238000001228 spectrum Methods 0.000 claims description 2
- 238000009826 distribution Methods 0.000 abstract description 5
- 239000004576 sand Substances 0.000 description 7
- 238000005553 drilling Methods 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 208000035126 Facies Diseases 0.000 description 1
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
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
- G01V1/282—Application of seismic models, synthetic seismograms
-
- 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
- G01V1/30—Analysis
- G01V1/306—Analysis for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles
Abstract
The invention discloses a lithology combination prediction method based on seismic waveform forward modeling, which is characterized by carrying out classification statistics on types of lithology combinations of a research area by utilizing lithology combination relations and characteristics obtained on the well, and carrying out statistics on the ranges of speed, density and thickness of corresponding lithology combinations according to different lithology combinations; carrying out spectrum analysis on the seismic data to determine the effective frequency range of the seismic data, and randomly selecting seismic wavelets with fixed frequency in the frequency range; performing convolution operation on the lithologic combined speed and density data and the seismic wavelet data to obtain seismic waveform data, and converting the seismic waveform data into a two-dimensional picture through time-frequency analysis; forming a deep learning training sample by the two-dimensional picture and the corresponding lithological combination type, and training a model by adopting a convolutional neural network; the lithology combination type of the actual seismic waveform can be predicted by using the trained model. The method can quickly determine the distribution of lithological combinations, has high prediction accuracy and is convenient to popularize and use.
Description
Technical Field
The invention relates to the field of geophysical exploration seismic interpretation and comprehensive research, in particular to a lithology combination prediction method based on seismic waveform forward modeling.
Background
The lithological property combination refers to the combination arrangement relation of the lithological properties in the transverse direction or the longitudinal direction, and is one of the important marks reflecting the rock generation environment. In the field of petroleum exploration, the lithological combination obtained by drilling or logging mainly refers to the vertical combination relationship, and is an important basic data for carrying out geological feature and sedimentary facies analysis. However, lithology combination data obtained by drilling or logging can be obtained only by the position of a lower well, and the lithology combination data is a common hole. In conducting regional geological studies, it is more desirable to be able to obtain spatially lithological composition data.
the three-dimensional seismic data is an effective way for acquiring information such as spatial geological structure, lithology and the like, but the seismic waveform reflects the stacking characteristic of lithology combination of a large sleeve on the stratum under the influence of the resolution ratio, and has larger uncertainty. The traditional lithological combination classification method is mainly characterized in that seismic waveforms are divided into several types by a seismic waveform clustering method, and then well data are utilized for calibration to judge what lithological combination the seismic waveforms represent. The method directly utilizes seismic waveform data to perform clustering, characteristics of lithological combination data on the well cannot be added into a clustering process, meanwhile, clustering results are not controllable, only several types can be selected, the classification is too fine and cannot be clustered, and the classification is too coarse, so that different lithological combinations exist in the same type.
the artificial intelligence technology can mine potential correlation relations from a large amount of sample data, and some relation also exists between seismic waveform data and lithologic combination, but because actual drilling data are limited, the sample data which can be marked by the well data often cannot meet the application requirements of the artificial intelligence method. Therefore, the method combining the forward modeling of the seismic waveform and the artificial intelligence method can solve the problems of training samples and prediction accuracy.
Disclosure of Invention
the invention aims to solve the defects in the prior art and provides a lithology combination prediction method based on forward modeling of seismic waveforms.
In order to achieve the purpose, the invention is implemented according to the following technical scheme:
A lithology combination prediction method based on forward modeling of seismic waveforms comprises the following steps:
Step 1: firstly, carrying out classification statistics on the types of lithological combinations of a research area, and carrying out statistics on the ranges of speed, density and thickness of the corresponding lithological combinations according to different lithological combinations;
Step 2: according to the speed, density and thickness ranges of the lithological combination counted in the step 1, a Monte Carlo random simulation method is adopted to randomly select a speed value, a density value and a thickness value of the lithological combination, the speed value and the thickness value are converted into two-way travel time, and the two-way travel time value is sampled according to a set time interval so as to obtain speed and density data corresponding to the lithological combination;
and step 3: carrying out spectrum analysis on the seismic data to determine the effective frequency range of the seismic data, and randomly selecting seismic wavelets with fixed frequency in the frequency range;
And 4, step 4: performing convolution operation on the lithologic combination speed and density data obtained in the step (2) and the seismic wavelet data obtained in the step (3) to obtain seismic waveform data, and converting the seismic waveform data into a two-dimensional picture through time-frequency analysis;
And 5: forming a deep learning training sample by the two-dimensional picture obtained in the step (4) and the corresponding lithologic combination type, and training a model by adopting a convolutional neural network to obtain a trained convolutional neural network model;
Step 6: and converting the actual seismic data into a two-dimensional picture by the same time-frequency analysis method according to the length of the training sample, loading the two-dimensional picture into a trained convolutional neural network model for judging the lithology type, and taking the lithology combination type corresponding to the maximum probability value as a prediction result.
further, the calculation formula for converting the velocity value and the thickness value into the two-way travel time in the step 2 is as follows: 2 thickness/speed.
Further, in the step 2, when speed and density data conversion of lithology combination is performed, an initial mudstone is added as background speed and density, and the thickness of the mudstone is 20 m.
Further, in the step 3, a fourier transform is used to perform spectrum analysis to obtain a spectrum of the seismic data, and the effective frequency range is the minimum and maximum values of 2 frequencies corresponding to the half-position where the frequency energy is reduced to the maximum value, which correspond to the effective frequency range respectively.
Further, the step 4 specifically includes:
1) and (3) obtaining the reflection coefficient by the wave impedance calculation method according to the speed and density data of the lithological combination obtained in the step (2):
HNIs a certain depth or time value, pNρN-1Density values of upper and lower two points, VNVN-1The speed values of an upper point and a lower point are obtained;
2) performing convolution on the seismic wavelets and the reflection coefficients to obtain seismic waveform data;
3) The seismic waveform data comprise background data obtained by background speed and density, and the background data are deleted;
4) the time frequency analysis adopts a continuous wavelet transform method CWT, frequency division is carried out by setting different wavelet scales, the sampling interval is consistent with that of the synthesized seismic waveform, and two-dimensional data obtained by the time frequency analysis is a two-dimensional picture.
further, in step 4, the longitudinal coordinate of the two-dimensional picture is a time axis of the seismic waveform, and the transverse coordinate is a frequency axis of the seismic waveform.
Further, the convolutional neural network in the step 5 comprises an input layer, an output layer, 3 convolutional layers, 1 pooling layer and 1 full-link layer, and the total training parameters are more than 200 ten thousand.
Further, 70% of deep learning training samples are used for training, 20% of deep learning training samples are used for cross validation, 10% of deep learning training samples are used for testing, the convolutional neural network model is trained and optimized in a mode that small batches of samples are randomly selected for multiple iterations, and the prediction accuracy after training is over 90% and serves as a final convolutional neural network model.
compared with the prior art, the method has the advantages that the lithological combination relation and the characteristics thereof obtained on the well are utilized, and the complex lithological combination is simplified into a plurality of lithological combinations capable of reflecting the stratum sedimentary characteristics; randomly generating a large amount of reliable speed and density data according to the speed, density and thickness intervals of different lithologies obtained through statistics; converting the data into seismic waveform data according to a seismic convolution model, and converting the seismic waveform data into time-frequency data by a continuous wavelet transform method; the obtained time-frequency data and the corresponding lithologic combination category are utilized to form sample data for learning, and the sample data can generate tens of thousands of samples through random extraction; the samples can be learned and trained through a designed convolutional neural network to establish a training model, and the training model can meet the prediction requirement after multiple iterations; and finally, inputting the actual seismic channels into the model according to the length of the training data for prediction, and obtaining a prediction graph of lithologic combination type distribution of the whole region. The method can quickly determine the distribution of the lithological combination, has high prediction accuracy and is convenient to popularize and use.
Drawings
FIG. 1 is a sand shale velocity statistic of an embodiment of the present invention.
fig. 2 is a sand shale density statistic of an embodiment of the present invention.
FIG. 3 is a seismic data spectral analysis of an embodiment of the invention.
FIG. 4 is a lithology combination convolution model calculation according to an embodiment of the present invention.
FIG. 5 is a partial lithology combined waveform of an embodiment of the present invention.
FIG. 6 is a diagram of a seismic waveform conversion to a time-frequency image, in accordance with an embodiment of the present invention.
FIG. 7 is a graph illustrating model training errors and accuracy in accordance with an embodiment of the present invention.
FIG. 8 is a profile of lithology portfolio type distributions along a layer in accordance with an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. The specific embodiments described herein are merely illustrative of the invention and do not limit the invention.
the lithology combination prediction method based on forward modeling of seismic waveforms comprises the following steps:
Step 1: lithology combination statistics of drilling data
and carrying out classification statistics on the lithological combination types of the research area according to well data, and mainly classifying through the arrangement combination relation of lithological thickness change and lithological property. The lithological combinations of the region are divided into 4 types, namely thick sand and thin mud combinations: sandstone thickness range [30, 35%]m, thickness range of mudstone [3,10]m, thick mud and thick sand combination: sandstone thickness range [20, 30%]m, mudstone thickness range [30, 35%]m, thick mud and thin sand combination: sandstone thickness range 10,20]m mudstone thickness range [40, 50%]m and mud + sandy mud combination: sandstone thickness range [5,10]m, mudstone thickness range [40, 60%]And m, counting the ranges of the speed, the density and the thickness of the corresponding lithology according to different lithology combinations. FIG. 1 and FIG. 2 are statistical interaction graphs of velocity and density of sand shale, sandstone velocity range [3400,4500]m/s, mudstone velocity range [2500,3500]m/s, sand density range [2, 2.6%]g/m3Mudstone density range [1.8,2.8 ]]g/m3。
step 2: and (2) according to the different lithological combination types obtained in the step (1), respectively counting the speed, density and thickness ranges of the type combinations, randomly selecting the speed, density and thickness values of each lithological combination in the ranges by adopting a Monte Carlo method, obtaining a two-way travel time value of the lithological combination through 2 x thickness/speed, and sampling the time value according to 0.002s to obtain corresponding speed and density data. The specific implementation steps are as follows:
a, setting forward sampling points and sampling intervals of seismic waveforms, wherein if one sampling point is 50, the sampling interval is 2ms and corresponds to a seismic wave length of 100 ms;
B, selecting a lithology combination type and a corresponding speed, density and thickness range thereof, setting an initial mudstone as a background speed and density, wherein the length is 20;
C, obtaining sampling points of the sandstone according to the thickness and the speed of the sandstone which are randomly extracted and a calculation formula (2 x thickness/speed), and obtaining sampling points of the mudstone according to the thickness and the speed of the mudstone which are randomly extracted and the calculation formula (2 x thickness/speed);
D, repeating the step C until the number of sampling points reaches 50;
And E, actually outputting the time domain speed and density data with the length of 70, wherein the former 20 are background data.
and step 3: and (3) carrying out spectrum analysis on the seismic data to determine the effective frequency range of the seismic data, and randomly selecting seismic wavelets with fixed frequency in the frequency range. The frequency range corresponding to half the energy of the seismic spectrogram obtained from FIG. 3 was set to [7,42] Hz. The seismic wavelets are Rake wavelets, with a length of 50 and a frequency randomly chosen among [7,42] Hz.
And 4, step 4: carrying out convolution operation on the velocity and density data generation reflection coefficient obtained in the step (2) and the wavelet data obtained in the step (3) to obtain seismic waveform data, and converting the seismic waveform data into a two-dimensional picture through time-frequency analysis, wherein the method comprises the following specific steps:
And A, calculating the speed and density data obtained in the step 2 by the following formula to form a reflection coefficient curve:
HNIs a certain depth or time value, pNρN-1density values of upper and lower two points, VNVN-1The speed values of an upper point and a lower point are obtained;
B: after the obtained reflection coefficient curve and the seismic wavelets with randomly selected frequencies are subjected to convolution operation, a group of seismic waveform data can be generated, as shown in FIG. 4;
generating a large amount of synthetic seismic waveform data by randomly extracting different speeds, densities, thicknesses and wavelet frequencies, wherein the seismic waveform length is 50 sampling points and is 100ms, and the seismic waveform data is shown in figure 5;
d: the obtained seismic waveform data is subjected to time-frequency decomposition by a continuous wavelet transform method, so that a corresponding time-frequency image can be obtained, wherein the frequency decomposition number is 60, and the corresponding frequency range is 5-65Hz, as shown in FIG. 6. Each time-frequency image is 50 x 60 in size, and the corresponding lithology combination type can be obtained through tracing, so that a large amount of labeled sample data is constructed.
And 5: a large amount of randomly generated two-dimensional data and corresponding lithology combination types can be obtained through the step 4, so that a sample library which can meet deep learning is formed, a convolutional neural network method is adopted for model training, and finally a better prediction model is obtained, wherein the specific implementation steps are as follows:
A: training sample data is randomly extracted and disordered, and 70% of samples are used for training, 20% of samples are used for cross validation, and 10% of samples are used for testing. The total number of 30000 samples was divided as follows, 21000 samples for training, 6000 samples for cross validation, 3000 samples for testing;
B: 21000 training samples, wherein 128 (minbatch) samples are randomly selected each time to serve as a batch of one training, and one epochs is used for iterating 21000 times for training;
and C, adopting a 7-layer convolutional neural network model as a training model, wherein the convolutional neural network model comprises an input layer, an output layer, 3 convolutional layers, 1 pooling layer and 1 full-connection layer, and the total training parameters are generally more than 200 ten thousand. The input sample data size is 50 × 60 — 3000, 50 represents the length of the seismic waveform, and 60 is the number of frequency decompositions. The output layer is a length-4 vector representing the probability of each lithology combination. The convolution kernel of 3 x 3 of convolution layer sampling is hidden in the middle, and the pooling layer is maximum pooling of 2 x 2;
D: and evaluating the training effect by comparing errors and precision on the training set and the test set after training, and considering that the convolutional neural network model can achieve the expected effect when the prediction precision on the test set is more than 85 percent, wherein the convolutional neural network model is used as a trained convolutional neural network model. FIG. 7 is a graph of model training error versus accuracy.
Step 6: and (3) converting the actual seismic data into a two-dimensional picture with the size of 50 x 60 according to 50 sampling points of the length of the training sample by the same time-frequency analysis parameters and methods. And loading the lithology type into a trained model for judgment, and taking the lithology type corresponding to the maximum probability value as a prediction result. FIG. 8 is a diagram of seismic waveform data taken along a layer of length 50, the predicted results of each seismic trace forming a lithology portfolio type distribution map on a surface.
The technical solution of the present invention is not limited to the limitations of the above specific embodiments, and all technical modifications made according to the technical solution of the present invention fall within the protection scope of the present invention.
Claims (8)
1. A lithology combination prediction method based on forward modeling of seismic waveforms is characterized by comprising the following steps:
step 1: firstly, carrying out classification statistics on the types of lithological combinations of a research area, and carrying out statistics on the ranges of speed, density and thickness of the corresponding lithological combinations according to different lithological combinations;
Step 2: according to the speed, density and thickness ranges of the lithological combination counted in the step 1, a Monte Carlo random simulation method is adopted to randomly select a speed value, a density value and a thickness value of the lithological combination, the speed value and the thickness value are converted into two-way travel time, and the two-way travel time value is sampled according to a set time interval so as to obtain speed and density data corresponding to the lithological combination;
And step 3: carrying out spectrum analysis on the seismic data to determine the effective frequency range of the seismic data, and randomly selecting seismic wavelets with fixed frequency in the frequency range;
And 4, step 4: performing convolution operation on the lithologic combination speed and density data obtained in the step (2) and the seismic wavelet data obtained in the step (3) to obtain seismic waveform data, and converting the seismic waveform data into a two-dimensional picture through time-frequency analysis;
and 5: forming a deep learning training sample by the two-dimensional picture obtained in the step (4) and the corresponding lithologic combination type, and training a model by adopting a convolutional neural network to obtain a trained convolutional neural network model;
Step 6: and converting the actual seismic data into a two-dimensional picture by the same time-frequency analysis method according to the length of the training sample, loading the two-dimensional picture into a trained convolutional neural network model for judging the lithology type, and taking the lithology combination type corresponding to the maximum probability value as a prediction result.
2. the lithology combination prediction method based on seismic waveform forward modeling according to claim 1, wherein the calculation formula of converting the velocity value and the thickness value into the two-way travel time in step 2 is as follows: 2 thickness/speed.
3. the lithology combination prediction method based on seismic waveform forward modeling according to claim 2, characterized in that: in the step 2, when the speed and density data of the lithology combination are converted, an initial mudstone is added as the background speed and density, and the thickness of the mudstone is 20 m.
4. the lithology combination prediction method based on seismic waveform forward modeling according to claim 1, characterized in that: in the step 3, Fourier transform is adopted to perform spectrum analysis to obtain the spectrum of the seismic data, and the effective frequency range is the minimum value and the maximum value of the effective frequency range corresponding to 2 frequencies respectively at the position where the frequency energy is reduced to the half of the maximum value.
5. The lithology combination prediction method based on seismic waveform forward modeling according to claim 3, wherein the step 4 specifically includes:
1) and (3) obtaining the reflection coefficient by the wave impedance calculation method according to the speed and density data of the lithological combination obtained in the step (2):
HNIs a certain depth or time value, pNρN-1density values of upper and lower two points, VNVN-1The speed values of an upper point and a lower point are obtained;
2) Performing convolution on the seismic wavelets and the reflection coefficients to obtain seismic waveform data;
3) The seismic waveform data comprise background data obtained by background speed and density, and the background data are deleted;
4) The time frequency analysis adopts a continuous wavelet transform method CWT, frequency division is carried out by setting different wavelet scales, the sampling interval is consistent with that of the synthesized seismic waveform, and two-dimensional data obtained by the time frequency analysis is a two-dimensional picture.
6. the lithology combination prediction method based on seismic waveform forward modeling as claimed in claim 5, wherein: in the step 4, the longitudinal coordinate of the two-dimensional picture is a time axis of the seismic waveform, and the transverse coordinate of the two-dimensional picture is a frequency axis of the seismic waveform.
7. The lithology combination prediction method based on seismic waveform forward modeling according to claim 1, characterized in that: the convolutional neural network in the step 5 comprises an input layer, an output layer, a convolutional layer with 3 layers, a pooling layer with 1 layer and a full-connection layer with 1 layer, and the total training parameters are more than 200 ten thousand.
8. the lithology combination prediction method based on seismic waveform forward modeling according to claim 7, characterized in that: 70% of deep learning training samples are used for training, 20% of deep learning training samples are used for cross validation, 10% of deep learning training samples are used for testing, a convolutional neural network model is trained and optimized in a mode that small batches of samples are randomly selected for multiple iterations, and the prediction accuracy after training is over 90% and serves as a final convolutional neural network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910664786.XA CN110554427B (en) | 2019-07-23 | Lithology combination prediction method based on forward modeling of seismic waveforms |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910664786.XA CN110554427B (en) | 2019-07-23 | Lithology combination prediction method based on forward modeling of seismic waveforms |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110554427A true CN110554427A (en) | 2019-12-10 |
| CN110554427B CN110554427B (en) | 2024-04-26 |
Family
ID=
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112434878A (en) * | 2020-12-09 | 2021-03-02 | 同济大学 | Cascade sample equalization-based seismic fluid prediction method |
| CN113159136A (en) * | 2021-03-30 | 2021-07-23 | 中铁第四勘察设计院集团有限公司 | In-hole data fusion stratum dividing method, device, equipment and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016041189A1 (en) * | 2014-09-19 | 2016-03-24 | 杨顺伟 | Method for evaluating shale gas reservoir and seeking desert area |
| CN106707340A (en) * | 2016-12-13 | 2017-05-24 | 中国石油天然气股份有限公司大港油田分公司 | Method for predicting volcanic rock facies |
| CN106842307A (en) * | 2015-12-04 | 2017-06-13 | 中国石油化工股份有限公司 | It is a kind of based on just drilling the subtle hydrocarbon reservoir method that the lower waveform separation of constraint is retrieved again |
| KR20180120413A (en) * | 2017-04-27 | 2018-11-06 | (주)지오룩스 | System for processing and analyzing magnetotellurics monitoring data using deep learning algorithm |
| CN108802812A (en) * | 2017-04-28 | 2018-11-13 | 中国石油天然气股份有限公司 | A kind of formation lithology inversion method of well shake fusion |
| CN109271898A (en) * | 2018-08-31 | 2019-01-25 | 电子科技大学 | Solution cavity body recognizer based on optimization convolutional neural networks |
| EP3441795A1 (en) * | 2017-08-09 | 2019-02-13 | Sick AG | Sensor for detecting an object and method for adjusting a switching point |
| US20190064389A1 (en) * | 2017-08-25 | 2019-02-28 | Huseyin Denli | Geophysical Inversion with Convolutional Neural Networks |
| CN109828304A (en) * | 2019-03-08 | 2019-05-31 | 中国海洋石油集团有限公司 | A method of lithological sequence model is predicted using seismic data based on deep learning |
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016041189A1 (en) * | 2014-09-19 | 2016-03-24 | 杨顺伟 | Method for evaluating shale gas reservoir and seeking desert area |
| CN106842307A (en) * | 2015-12-04 | 2017-06-13 | 中国石油化工股份有限公司 | It is a kind of based on just drilling the subtle hydrocarbon reservoir method that the lower waveform separation of constraint is retrieved again |
| CN106707340A (en) * | 2016-12-13 | 2017-05-24 | 中国石油天然气股份有限公司大港油田分公司 | Method for predicting volcanic rock facies |
| KR20180120413A (en) * | 2017-04-27 | 2018-11-06 | (주)지오룩스 | System for processing and analyzing magnetotellurics monitoring data using deep learning algorithm |
| CN108802812A (en) * | 2017-04-28 | 2018-11-13 | 中国石油天然气股份有限公司 | A kind of formation lithology inversion method of well shake fusion |
| EP3441795A1 (en) * | 2017-08-09 | 2019-02-13 | Sick AG | Sensor for detecting an object and method for adjusting a switching point |
| US20190064389A1 (en) * | 2017-08-25 | 2019-02-28 | Huseyin Denli | Geophysical Inversion with Convolutional Neural Networks |
| CN109271898A (en) * | 2018-08-31 | 2019-01-25 | 电子科技大学 | Solution cavity body recognizer based on optimization convolutional neural networks |
| CN109828304A (en) * | 2019-03-08 | 2019-05-31 | 中国海洋石油集团有限公司 | A method of lithological sequence model is predicted using seismic data based on deep learning |
Non-Patent Citations (4)
| Title |
|---|
| 吕金龙等: "岩性波阻抗正演模拟差异方法在储层预测中的应用", 《大庆石油地质与开发》 * |
| 吕金龙等: "岩性波阻抗正演模拟差异方法在储层预测中的应用", 《大庆石油地质与开发》, vol. 35, no. 04, 31 August 2016 (2016-08-31), pages 135 - 139 * |
| 陈志刚等: "低频地震资料分频反演在滨里海盆地中区块的应用", 《石油地球物理勘探》 * |
| 陈志刚等: "低频地震资料分频反演在滨里海盆地中区块的应用", 《石油地球物理勘探》, vol. 53, 31 December 2018 (2018-12-31), pages 261 - 265 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112434878A (en) * | 2020-12-09 | 2021-03-02 | 同济大学 | Cascade sample equalization-based seismic fluid prediction method |
| CN112434878B (en) * | 2020-12-09 | 2022-09-20 | 同济大学 | Cascade sample equalization-based seismic fluid prediction method |
| CN113159136A (en) * | 2021-03-30 | 2021-07-23 | 中铁第四勘察设计院集团有限公司 | In-hole data fusion stratum dividing method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mousavi et al. | Deep-learning seismology | |
| CN113759425B (en) | Method and system for evaluating filling characteristics of deep paleo-karst reservoir stratum by well-seismic combination | |
| CN111239802B (en) | Deep learning speed modeling method based on seismic reflection waveform and velocity spectrum | |
| EP3254142B1 (en) | A method for determining sedimentary facies using 3d seismic data | |
| CN108931814B (en) | Multi-attribute fusion based bedrock fracture prediction method | |
| CN109061765B (en) | Trap evaluation method for heterogeneous thin sandstone interbed reservoir | |
| US11802985B2 (en) | Method and system for analyzing filling for karst reservoir based on spectrum decomposition and machine learning | |
| CN104122581B (en) | A kind of poststack sound impedance inversion method | |
| Wang et al. | Quantitative thickness prediction of tectonically deformed coal using Extreme Learning Machine and Principal Component Analysis: a case study | |
| Eidsvik et al. | Estimation of geological attributes from a well log: an application of hidden Markov chains | |
| CN110609320A (en) | Pre-stack seismic reflection pattern recognition method based on multi-scale feature fusion | |
| Zhang et al. | Seismic lithology/fluid prediction via a hybrid ISD-CNN | |
| CN108226997A (en) | A kind of seismic facies analysis method based on earthquake data before superposition | |
| CN110118994B (en) | Continental facies hydrocarbon source rock quantitative prediction method based on seismic inversion and machine learning | |
| CN109425889B (en) | Method for depicting ancient karst underground river | |
| Zeng et al. | Impact of sedimentary facies on machine learning of acoustic impedance from seismic data: Lessons from a geologically realistic 3D model | |
| CN110554427B (en) | Lithology combination prediction method based on forward modeling of seismic waveforms | |
| CN110554427A (en) | Lithology combination prediction method based on forward modeling of seismic waveform | |
| CN113419278B (en) | Well-seismic joint multi-target simultaneous inversion method based on state space model and support vector regression | |
| CN106291675A (en) | A kind of geological data reconstructing method based on base tracer technique | |
| Dai et al. | Stratigraphic automatic correlation using SegNet semantic segmentation model | |
| Li* et al. | VMD based sedimentary cycle division for unconventional facies analysis | |
| CN114462703A (en) | Acoustic parameter curve prediction method, logging curve prediction method and electronic equipment | |
| CN113514884A (en) | Compact sandstone reservoir prediction method | |
| Yazid et al. | Machine learning-based techniques for reservoir characterization using rock-typing from well logs |
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 |