CN111598444A - Well logging lithology identification method and system based on convolutional neural network - Google Patents
Well logging lithology identification method and system based on convolutional neural network Download PDFInfo
- Publication number
- CN111598444A CN111598444A CN202010410798.2A CN202010410798A CN111598444A CN 111598444 A CN111598444 A CN 111598444A CN 202010410798 A CN202010410798 A CN 202010410798A CN 111598444 A CN111598444 A CN 111598444A
- Authority
- CN
- China
- Prior art keywords
- lithology
- logging
- neural network
- convolutional neural
- data
- 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
Links
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 119
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000012549 training Methods 0.000 claims description 114
- 238000012360 testing method Methods 0.000 claims description 46
- 230000006870 function Effects 0.000 claims description 41
- 238000004590 computer program Methods 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 8
- 238000012804 iterative process Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 description 13
- 238000011176 pooling Methods 0.000 description 10
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000010606 normalization Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000004140 cleaning Methods 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 230000008034 disappearance Effects 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000011435 rock Substances 0.000 description 2
- 244000208734 Pisonia aculeata Species 0.000 description 1
- 230000003213 activating effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000009849 deactivation Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002779 inactivation Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000013507 mapping 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
- 230000004044 response Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 238000003325 tomography Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21B—EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
- E21B49/00—Testing the nature of borehole walls; Formation testing; Methods or apparatus for obtaining samples of soil or well fluids, specially adapted to earth drilling or wells
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Mining & Mineral Resources (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Quality & Reliability (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Geology (AREA)
- Molecular Biology (AREA)
- Entrepreneurship & Innovation (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Marketing (AREA)
- Mathematical Physics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Fluid Mechanics (AREA)
- Environmental & Geological Engineering (AREA)
- Game Theory and Decision Science (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geochemistry & Mineralogy (AREA)
- Agronomy & Crop Science (AREA)
- Animal Husbandry (AREA)
Abstract
The invention provides a well logging lithology identification method and system based on a convolutional neural network. The well logging lithology identification method based on the convolutional neural network comprises the following steps: acquiring current logging data; inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data; and determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data. The method can rapidly divide the lithology and improve the efficiency and the precision of lithology identification.
Description
Technical Field
The invention relates to the technical field of geophysical exploration of petroleum, in particular to a well logging lithology identification method and system based on a convolutional neural network.
Background
The geophysical well logging lithology identification is an important content of research on oil-gas-containing property evaluation, reservoir description and the like, and is a basis for solving various parameters of an oil-gas reservoir. Compared with other lithology identification methods (such as coring), the method for identifying the lithology of the stratum by using the logging information has the characteristics of high speed and low cost, and is widely adopted.
The conventional method for identifying the lithology of the stratum by using logging information mainly comprises a cross-plot method, a statistical method and imaging logging, but the traditional identification method is low in precision, low in efficiency and large in human factor influence, and the tomography logging is expensive and not beneficial to wide practical application, so that the research of the method for automatically identifying the lithology with high precision has important significance for the interpretation of logging data. Lithology can be automatically identified by establishing a logging interpretation model of rock type-logging parameters, but the method has low identification precision for well logging curves with unobvious characteristics.
With the development of well logging technology, well logging methods are increasing, accuracy is increasing, and data volume is increasing. Each logging curve is a special response to formation lithology information, if a plurality of logging curves are simultaneously synthesized for interpretation work, the problem of complex multidimensional nonlinearity is solved, and the problem of big data formed by the synthesis of a plurality of logging data is also solved, which cannot be solved by the traditional methods such as intersection mapping method and machine learning. Deep learning automatically extracts features from mass data and solves the problem of complex classification or prediction through feature change layer by layer.
Disclosure of Invention
The embodiment of the invention mainly aims to provide a well logging lithology identification method and system based on a convolutional neural network, so that lithology can be divided quickly, and the efficiency and the accuracy of lithology identification can be improved.
In order to achieve the above object, an embodiment of the present invention provides a well logging lithology identification method based on a convolutional neural network, including:
acquiring current logging data;
inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data;
and determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
The embodiment of the invention also provides a well logging lithology identification system based on the convolutional neural network, which comprises the following steps:
the first acquisition unit is used for acquiring current logging data;
the lithology probability unit is used for inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data;
and the lithology identification unit is used for determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
The embodiment of the invention also provides computer equipment which comprises a memory, a processor and a computer program stored on the memory and operated on the processor, wherein the processor realizes the steps of the well logging lithology identification method based on the convolutional neural network when executing the computer program.
The embodiment of the invention also provides a computer readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the steps of the method for identifying the lithology of the well logging based on the convolutional neural network.
According to the method and the system for identifying the lithology of the well logging based on the convolutional neural network, the current well logging data are input into the convolutional neural network framework optimal model, all lithology probabilities of the current well logging data are obtained, the lithology corresponding to the maximum value of the lithology probabilities is determined to be used as the lithology of the current well logging data, the lithology can be divided quickly, and the efficiency and the accuracy of lithology identification are 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 will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a flow chart of a well logging lithology identification method based on a convolutional neural network in an embodiment of the invention;
FIG. 2 is a flow chart of creating a convolutional neural network framework optimal model in an embodiment of the present invention;
fig. 3 is a flowchart of S206 in the embodiment of the present invention;
FIG. 4 is a schematic diagram of a convolutional neural network framework model in an embodiment of the present invention;
FIG. 5 is a schematic illustration of a well logging training curve and lithology in an embodiment of the present invention;
FIG. 6 is a block diagram of a well logging lithology recognition system based on a convolutional neural network in an embodiment of the present invention;
fig. 7 is a block diagram of a computer device in the embodiment of the present 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.
As will be appreciated by one skilled in the art, embodiments of the present invention may be embodied as a system, apparatus, device, method, or computer program product. Accordingly, the present disclosure may be embodied in the form of: entirely hardware, entirely software (including firmware, resident software, micro-code, etc.), or a combination of hardware and software.
In view of the fact that the lithology in the prior art is low in precision, low in efficiency and greatly influenced by human factors, the embodiment of the invention provides the well logging lithology identification method based on the convolutional neural network, which can be used for rapidly dividing the lithology and improving the efficiency and the precision of lithology identification. The present invention will be described in detail below with reference to the accompanying drawings.
FIG. 1 is a flow chart of a well logging lithology identification method based on a convolutional neural network in the embodiment of the invention. As shown in fig. 1, the well logging lithology identification method based on the convolutional neural network includes:
s101: and acquiring current logging data.
And the current logging data is a current logging curve subjected to data cleaning. The data cleansing includes data conversion that can convert the current log into tfrecrds data recognizable by tenserflow.
S102: and inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data.
S103: and determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
The execution subject of the well logging lithology identification method based on the convolutional neural network shown in FIG. 1 is a computer. As can be seen from the process shown in fig. 1, according to the well logging lithology identification method based on the convolutional neural network, the current well logging data is input into the convolutional neural network framework optimal model, so that each lithology probability of the current well logging data is obtained, and the lithology corresponding to the maximum value of the lithology probability is determined as the lithology of the current well logging data, so that the lithology can be divided quickly, and the efficiency and the accuracy of lithology identification can be improved.
FIG. 2 is a flow chart of creating an optimal model of a convolutional neural network framework in an embodiment of the present invention. As shown in fig. 2, the step of creating the optimal model of the convolutional neural network framework in advance includes:
the following iterative process is performed:
s201: and inputting the logging training data into the convolutional neural network framework model to obtain the lithology probabilities of the logging training data.
Before executing S201, the method further includes:
1. acquiring a logging training curve and a logging training depth range of a training well;
wherein the well logging training curve comprises one or more of a natural potential curve, a natural gamma curve, a sonic time difference curve, a porosity curve, a density curve, a bottom gradient resistivity curve, a micro-gradient resistivity curve, and a micro-potential resistivity curve. The logging training depth range comprises one depth point in the logging training curve and two depth line segments which are adjacent to the depth point up and down and have the same length.
2. And taking a logging training curve corresponding to the logging training depth range as logging training data, and taking the actual lithology of a depth point positioned in the center of the logging training depth range as the actual lithology of the logging training data.
And the logging training data is a logging training curve subjected to data cleaning.
For example, if the logging depth range is 97m-103m, the depth point located at the center of the logging training depth range is 100m, and two depth line segments with the same length, which are adjacent to the depth point, are 97m-99m and 101m-103m, respectively. The logging training data is a logging training curve which is located between 97m and 103m and is subjected to data cleaning, and the actual lithology corresponding to the depth of 100m is the actual lithology of the logging training data.
S202: and determining a loss function according to the actual lithology of the logging training data and the lithology probabilities of the logging training data.
S203: and judging whether the current iteration times reach the preset iteration times or not.
S204: and when the current iteration times reach the preset iteration times, taking the convolutional neural network framework model in the current iteration as the convolutional neural network framework optimal training model.
For example, the preset number of iterations is 1000. The invention also outputs the convolutional neural network framework model and the loss function every 100 times for the staff to view.
In specific implementation, the convolutional neural network framework model in the 1000 th iteration is used as the convolutional neural network framework optimal training model, then the iteration times are reset to zero, and the convolutional neural network framework optimal training model is output again when the iteration times reach 1000 again so as to obtain a plurality of convolutional neural network framework optimal training models.
S205: and when the current iteration times do not reach the preset iteration times, updating the convolutional neural network framework model according to the loss function, and continuously executing the iteration processing.
S206: and determining the convolutional neural network framework optimal model from the convolutional neural network framework optimal training models.
Fig. 3 is a flowchart of S206 in the embodiment of the present invention. As shown in fig. 3, S206 includes:
s301: and inputting the logging test data into the optimal training model of the convolutional neural network framework to obtain each lithology probability of the logging test data.
Before executing S301, the method further includes:
1. acquiring a logging test curve and a logging test depth range of a test well;
wherein the ratio of the training well to the testing well is 0.8: 0.2. For example, if there are 169 wells in a certain area, the number of training wells is 135, and the number of testing wells is 34; the number of well logging training data is 287515 and the number of well logging test data is 73075.
The well logging test curves include one or more of a natural potential curve, a natural gamma curve, a sonic time difference curve, a porosity curve, a density curve, a bottom gradient resistivity curve, a micro-gradient resistivity curve, and a micro-potential resistivity curve. The logging test depth range comprises one depth point in the logging test curve and two depth line segments which are adjacent to the depth point up and down and have the same length.
2. And taking a logging test curve corresponding to the logging test depth range as logging test data, and taking the actual lithology of a depth point positioned in the center of the logging test depth range as the actual lithology of the logging test data.
And the logging test data is a logging test curve subjected to data cleaning.
S302: and determining the accuracy of the convolutional neural network framework optimal training model according to the actual lithology of the logging test data and the lithology probabilities of the logging test data.
S303: and taking the convolutional neural network framework optimal training model corresponding to the maximum value in the accuracy as the convolutional neural network framework optimal model.
FIG. 4 is a diagram of a convolutional neural network framework model in an embodiment of the present invention. As shown in fig. 4, N is the category of the well logging training curve, the output data size is lithNum, and the number of lithology categories is lithNum. Convolution (Conv), ReLu activation functions, Batch Normalization (BN), Max pooling (Max Pool), Full Connectivity (FC), dropout (random inactivation regularization, not shown), and SoftMax activation functions were combined to define a convolutional neural network framework model. Random deactivation regularization is a regularization method that includes: each node of the full-connection layer is set with a probability to be eliminated, and part of nodes are eliminated randomly according to the probability in the subsequent training, so that the purposes of regularization and variance reduction are achieved.
As shown in fig. 4, the first step: the logging training data input to the first convolutional layer has a size of 7 × 1 × N, the convolution kernel used by the first convolutional layer has a size of 3 × 1, the step size is 1, 32 convolution kernels are used for convolution, and the output data of the first convolutional layer has a size of 7 × 1 × 32. 7 of the dimensions of the well logging training data is the well logging training depth range.
The convolution can extract more local features, and the invention introduces a plurality of convolution layers and can extract more local features of the logging training data.
FIG. 5 is a schematic representation of a well logging training curve and lithology in an embodiment of the present invention. As shown in fig. 5, the vertical axis in fig. 5 is depth in meters. The logging training curve comprises 7 curves including a natural gamma curve, a natural potential curve, a sound wave time difference curve, a 2.5 m bottom gradient resistivity curve, a micro-potential resistivity curve and a density curve, and N is 7 at the moment.
And a second convolution layer input data size is 7 × 1 × 32, a second convolution layer uses convolution kernel size of 3 × 1 and step size is 1, a total of 64 convolution kernels are convoluted, Batch Normalization (BN) and a ReLU activation function are used for carrying out nonlinear processing on convolution results, and a second convolution layer output data size is 7 × 1 × 64.
In batch normalization, through a certain normalization means, the distribution of the input values of any neuron of each layer of neural network is forcibly pulled back to the standard normal distribution with the mean value of 0 and the variance of 1, which is equivalent to the forced pull-back of more and more biased distribution to the standard distribution, so that the input values of the activation functions fall in the region where the nonlinear functions are sensitive to input, and at the moment, small changes of the input values of the activation functions can cause large changes of the loss functions, thereby avoiding the problem of gradient disappearance. And the gradient is increased, which means that the learning convergence speed is high, and the training speed can be greatly increased. The method solves the problem of gradient disappearance by introducing batch normalization, accelerates the convergence speed and accuracy of training, has the regularization capability similar to dropout, prevents overfitting to a certain extent, and relaxes certain parameter adjustment requirements.
In the neural network, the function of the activation function is to add some non-linear factors to the neural network, so that the neural network can better solve more complex problems. By activating the function, the data will be compressed to a certain range interval, and the size of the obtained data will determine whether the neuron is in an active state, i.e. activated. Commonly used activation functions include a ReLU function, a Sigmoid function, a tanh function, a SoftMax function, and the like. The ReLU activation function is linear, and therefore, is much faster than the Sigmod function and the tanh function regardless of forward propagation and backward propagation, and further, the ReLU activation function has no problem of gradient disappearance when input data is positive. The invention combines convolution, batch normalization and ReLU activation functions to better extract effective logging curve characteristics.
And thirdly, the input data size of the first maximum pooling layer is 7 multiplied by 1 multiplied by 64, the input data is subjected to dimensionality reduction by using the first maximum pooling layer, the pooling kernel size is 2 multiplied by 1, and the output data size of the first maximum pooling layer is 4 multiplied by 1 multiplied by 64.
The maximum pooling layer can reduce the dimensionality of a middle hidden layer of the neural network and reduce the computation of the next layers.
The fourth step is that the input data size of the third convolutional layer is 4 × 1 × 64, the convolution kernel size used by the third convolutional layer is 3 × 1, the step size is 1, 128 convolution kernels are convoluted, and the convolution result is subjected to nonlinear processing using Batch Normalization (BN) and ReLU activation functions, and the output data size of the third convolutional layer is 4 × 1 × 128.
Fifthly, the input data size of the fourth convolutional layer is 4 × 1 × 128, the convolution kernel size used by the fourth convolutional layer is 3 × 1, the step size is 1, 256 convolution kernels are convoluted, Batch Normalization (BN) and ReLU activation functions are used for performing nonlinear processing on convolution results, and the output data size of the fourth convolutional layer is 4 × 1 × 256.
Sixthly, the input data size of the second maximum pooling layer is 4 multiplied by 1 multiplied by 256, the dimension of the input data is reduced by using the second maximum pooling layer, the pooling kernel size is 4 multiplied by 1, the output data size of the second maximum pooling layer is 1 multiplied by 256, and the output data is transformed into a one-dimensional vector with the length of 256.
And seventhly, the input data size of the first full connection layer is 256, the local features of the logging training data are mapped to the known lithology class space by using full connection and dropout, and the output data size of the first full connection layer is 64. The fully-connected layer may map the learned "distributed feature representation" to a sample label space. The distributed feature expression in the invention is a local feature of logging training data, and the sample labeling space in the invention is a known lithology category space.
In the machine learning model, if the parameters of the model are too many and the training samples are too few, the trained model is easy to generate an overfitting phenomenon. The overfitting problem is often encountered when training the neural network, and the overfitting is specifically shown in the following steps: the loss function of the model on the training data is small, and the accuracy is high; but the loss function on the test data is larger and the accuracy is lower. Dropout may discard the neural network element from the network with a certain probability during training of the deep learning network, thereby preventing overfitting. According to the method, dropout is introduced into a full connection layer to solve the problems that parameters of a convolutional neural network framework model are too many, well logging training data are too little, and the trained model generates overfitting.
And fifthly, the input data size of the second fully-connected layer is 64, local features of the logging training data are mapped to a known lithologic class space by using fully-connected, dropout and softmax functions, and the output data size of the second fully-connected layer is lithNum. As shown in fig. 5, when both lithology of sandstone and mudstone is shared, lithNum is 2.
The SoftMax activation function is suitable for processing multi-classification problems, maps the neuron output of the last layer of the neural network into a (0, 1) interval, and maps the original severely changed numerical value into a certain interval to form a probability output form. According to the invention, a SoftMax function is introduced at the end of the network, local features of the logging training data are mapped to lithology classification space, and relative probability of each lithology classification can be obtained to determine lithology.
Logging test data is input into the convolutional neural network framework optimal model in the embodiment of the invention, the accuracy of the output result is 93.76%, and logging training data is input into the convolutional neural network framework optimal model in the embodiment of the invention, the accuracy of the output result is 93.45%. As shown in FIG. 5, the lithology of the current log data obtained by the present invention matches the actual lithology obtained by rock coring.
The specific process of the embodiment of the invention is as follows:
1. and acquiring a logging training curve and a logging training depth range of the training well, taking the logging training curve corresponding to the logging training depth range as logging training data, and taking the actual lithology of a depth point positioned in the center of the logging training depth range as the actual lithology of the logging training data.
2. And inputting the logging training data into the convolutional neural network framework model to obtain the lithology probabilities of the logging training data.
3. And determining a loss function according to the actual lithology of the logging training data and the lithology probabilities of the logging training data.
4. And judging whether the current iteration times reach the preset iteration times, when the current iteration times reach the preset iteration times, taking the convolutional neural network framework model in the current iteration as the optimal training model of the convolutional neural network framework, otherwise, updating the convolutional neural network framework model according to the loss function, and returning to the step 2.
5. And acquiring a logging test curve and a logging test depth range of the test well, taking the logging test curve corresponding to the logging test depth range as logging test data, and taking the actual lithology of a depth point positioned in the center of the logging test depth range as the actual lithology of the logging test data.
6. And inputting the logging test data into the optimal training model of the convolutional neural network framework to obtain each lithology probability of the logging test data.
7. And determining the accuracy of the convolutional neural network framework optimal training model according to the actual lithology of the logging test data and the lithology probabilities of the logging test data.
8. And taking the convolutional neural network framework optimal training model corresponding to the maximum value in the accuracy as the convolutional neural network framework optimal model.
9. And acquiring current logging data, and inputting the current logging data into the optimal model of the convolutional neural network framework to obtain each lithology probability of the current logging data.
10. And taking the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
In summary, according to the well logging lithology recognition method based on the convolutional neural network, the current well logging data are input into the convolutional neural network framework optimal model, all lithology probabilities of the current well logging data are obtained, the lithology corresponding to the maximum value of the lithology probabilities is determined to be used as the lithology of the current well logging data, the lithology can be divided rapidly, and the efficiency and the precision of lithology recognition are improved.
Based on the same inventive concept, the embodiment of the invention also provides a well logging lithology recognition system based on the convolutional neural network, and as the problem solving principle of the system is similar to that of the well logging lithology recognition method based on the convolutional neural network, the implementation of the system can refer to the implementation of the method, and repeated parts are not repeated.
FIG. 6 is a block diagram of a well logging lithology identification system based on a convolutional neural network in the embodiment of the present invention. As shown in fig. 6, the well-logging lithology recognition system based on the convolutional neural network comprises:
the first acquisition unit is used for acquiring current logging data;
the lithology probability unit is used for inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data;
and the lithology identification unit is used for determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
In one embodiment, the method further comprises the following steps:
an iteration unit for performing an iterative process of:
inputting the well logging training data into a convolutional neural network framework model to obtain each lithology probability of the well logging training data;
determining a loss function according to the actual lithology of the logging training data and the lithology probabilities of the logging training data;
judging whether the current iteration times reach preset iteration times or not;
when the current iteration times reach the preset iteration times, taking the convolutional neural network framework model in the current iteration as an optimal training model of the convolutional neural network framework, otherwise, updating the convolutional neural network framework model according to a loss function, and continuously executing iteration processing;
and the optimal model determining unit is used for determining the optimal model of the convolutional neural network framework from the optimal training models of the convolutional neural network frameworks.
In one embodiment, the optimal model determining unit is specifically configured to:
inputting the well logging test data into an optimal training model of a convolutional neural network framework to obtain each lithology probability of the well logging test data;
determining the accuracy of the convolutional neural network framework optimal training model according to the actual lithology of the logging test data and the lithology probabilities of the logging test data;
and taking the convolutional neural network framework optimal training model corresponding to the maximum value in the accuracy as the convolutional neural network framework optimal model.
In one embodiment, the method further comprises the following steps:
the second acquisition unit is used for acquiring a logging training curve and a logging depth range of a training well;
and the well logging training data determining unit is used for taking the well logging training curve corresponding to the well logging depth range as well logging training data and taking the actual lithology of the depth point positioned in the center of the well logging depth range as the actual lithology of the well logging training data.
In summary, the well logging lithology recognition system based on the convolutional neural network of the embodiment of the invention inputs the current well logging data into the convolutional neural network framework optimal model to obtain each lithology probability of the current well logging data, and determines the lithology corresponding to the maximum value of the lithology probability as the lithology of the current well logging data, so that the lithology can be divided quickly, and the efficiency and the precision of lithology recognition can be improved.
The embodiment of the invention also provides a specific implementation mode of computer equipment capable of realizing all the steps in the well logging lithology identification method based on the convolutional neural network in the embodiment. Fig. 7 is a block diagram of a computer device in an embodiment of the present invention, and referring to fig. 7, the computer device specifically includes the following:
a processor (processor)701 and a memory (memory) 702.
The processor 701 is configured to call a computer program in the memory 702, and the processor implements all the steps of the convolutional neural network based well-logging lithology recognition method in the above embodiments when executing the computer program, for example, the processor implements the following steps when executing the computer program:
acquiring current logging data;
inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data;
and determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
In summary, the computer device of the embodiment of the invention inputs the current logging data into the convolutional neural network framework optimal model to obtain each lithology probability of the current logging data, and determines the lithology corresponding to the maximum value of the lithology probability as the lithology of the current logging data, so that the lithology can be divided quickly, and the efficiency and the precision of lithology identification can be improved.
An embodiment of the present invention further provides a computer-readable storage medium capable of implementing all the steps in the method for identifying lithology of well logging based on convolutional neural network in the above embodiment, where the computer-readable storage medium stores a computer program, and the computer program, when executed by a processor, implements all the steps of the method for identifying lithology of well logging based on convolutional neural network in the above embodiment, for example, when the processor executes the computer program, implements the following steps:
acquiring current logging data;
inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data;
and determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
In summary, the computer-readable storage medium according to the embodiment of the present invention inputs the current logging data into the convolutional neural network framework optimal model to obtain each lithology probability of the current logging data, and determines the lithology corresponding to the maximum value of the lithology probability as the lithology of the current logging data, so that the lithology can be divided quickly, and the efficiency and accuracy of lithology identification can be improved.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Those of skill in the art will further appreciate that the various illustrative logical blocks, units, and steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate the interchangeability of hardware and software, various illustrative components, elements, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design requirements of the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present embodiments.
The various illustrative logical blocks, or elements, or devices described in connection with the embodiments disclosed herein may be implemented or performed with a general purpose processor, a digital signal processor, an Application Specific Integrated Circuit (ASIC), a field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, e.g., a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a digital signal processor core, or any other similar configuration.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may be stored in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. For example, a storage medium may be coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an ASIC, which may be located in a user terminal. In the alternative, the processor and the storage medium may reside in different components in a user terminal.
In one or more exemplary designs, the functions described above in connection with the embodiments of the invention may be implemented in hardware, software, firmware, or any combination of the three. If implemented in software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media that facilitate transfer of a computer program from one place to another. Storage media may be any available media that can be accessed by a general purpose or special purpose computer. For example, such computer-readable media can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to carry or store program code in the form of instructions or data structures and which can be read by a general-purpose or special-purpose computer, or a general-purpose or special-purpose processor. Additionally, any connection is properly termed a computer-readable medium, and, thus, is included if the software is transmitted from a website, server, or other remote source via a coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL), or wirelessly, e.g., infrared, radio, and microwave. Such discs (disk) and disks (disc) include compact disks, laser disks, optical disks, DVDs, floppy disks and blu-ray disks where disks usually reproduce data magnetically, while disks usually reproduce data optically with lasers. Combinations of the above may also be included in the computer-readable medium.
Claims (10)
1. A well logging lithology identification method based on a convolutional neural network is characterized by comprising the following steps:
acquiring current logging data;
inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data;
and determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
2. The convolutional neural network-based lithology log identification method of claim 1, wherein the step of pre-creating the convolutional neural network framework optimal model comprises:
the following iterative process is performed:
inputting well logging training data into a convolutional neural network framework model to obtain lithology probabilities of the well logging training data;
determining a loss function according to the actual lithology of the logging training data and the lithology probabilities of the logging training data;
judging whether the current iteration times reach preset iteration times or not;
when the current iteration times reach preset iteration times, taking the convolutional neural network framework model in the current iteration as an optimal training model of the convolutional neural network framework, otherwise, updating the convolutional neural network framework model according to the loss function, and continuously executing the iteration processing;
and determining the convolutional neural network framework optimal model from the convolutional neural network framework optimal training models.
3. The convolutional neural network-based lithology log identification method of claim 2, wherein determining a convolutional neural network framework optimal model from the respective convolutional neural network framework optimal training models comprises:
inputting well logging test data into the convolutional neural network framework optimal training model to obtain each lithology probability of the well logging test data;
determining the accuracy of the convolutional neural network framework optimal training model according to the actual lithology of the logging test data and the lithology probability of each logging test data;
and taking the convolutional neural network framework optimal training model corresponding to the maximum value in the accuracy as the convolutional neural network framework optimal model.
4. The convolutional neural network-based lithology log identification method of claim 2, further comprising:
acquiring a logging training curve and a logging depth range of a training well;
and taking the logging training curve corresponding to the logging depth range as logging training data, and taking the actual lithology of the depth point positioned in the center of the logging depth range as the actual lithology of the logging training data.
5. A well logging lithology recognition system based on a convolutional neural network, comprising:
the first acquisition unit is used for acquiring current logging data;
the lithology probability unit is used for inputting the current logging data into a pre-established convolutional neural network framework optimal model to obtain each lithology probability of the current logging data;
and the lithology identification unit is used for determining the lithology corresponding to the maximum lithology probability as the lithology of the current logging data.
6. The convolutional neural network-based well lithology identification system of claim 5, further comprising:
an iteration unit for performing an iterative process of:
inputting well logging training data into a convolutional neural network framework model to obtain lithology probabilities of the well logging training data;
determining a loss function according to the actual lithology of the logging training data and the lithology probabilities of the logging training data;
judging whether the current iteration times reach preset iteration times or not;
when the current iteration times reach preset iteration times, taking the convolutional neural network framework model in the current iteration as an optimal training model of the convolutional neural network framework, otherwise, updating the convolutional neural network framework model according to the loss function, and continuously executing the iteration processing;
and the optimal model determining unit is used for determining the optimal model of the convolutional neural network framework from the optimal training models of the convolutional neural network frameworks.
7. The convolutional neural network-based well logging lithology identification system of claim 6, wherein the optimal model determination unit is specifically configured to:
inputting well logging test data into the convolutional neural network framework optimal training model to obtain each lithology probability of the well logging test data;
determining the accuracy of the convolutional neural network framework optimal training model according to the actual lithology of the logging test data and the lithology probability of each logging test data;
and taking the convolutional neural network framework optimal training model corresponding to the maximum value in the accuracy as the convolutional neural network framework optimal model.
8. The convolutional neural network-based well lithology identification system of claim 6, further comprising:
the second acquisition unit is used for acquiring a logging training curve and a logging depth range of a training well;
and the well logging training data determining unit is used for taking the well logging training curve corresponding to the well logging depth range as well logging training data and taking the actual lithology of the depth point positioned in the center of the well logging depth range as the actual lithology of the well logging training data.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executed on the processor, wherein the processor when executing the computer program implements the steps of the convolutional neural network based well lithology identification method of any of claims 1 to 4.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the convolutional neural network-based well lithology recognition method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010410798.2A CN111598444A (en) | 2020-05-15 | 2020-05-15 | Well logging lithology identification method and system based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010410798.2A CN111598444A (en) | 2020-05-15 | 2020-05-15 | Well logging lithology identification method and system based on convolutional neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111598444A true CN111598444A (en) | 2020-08-28 |
Family
ID=72185648
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010410798.2A Pending CN111598444A (en) | 2020-05-15 | 2020-05-15 | Well logging lithology identification method and system based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111598444A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112651171A (en) * | 2020-12-15 | 2021-04-13 | 清华大学 | Data-driven rock physical modeling method and system |
| CN112990320A (en) * | 2021-03-19 | 2021-06-18 | 中国矿业大学(北京) | Lithology classification method and device, electronic equipment and storage medium |
| CN113033676A (en) * | 2021-03-30 | 2021-06-25 | 中铁第四勘察设计院集团有限公司 | Stratum dividing method, device, equipment and storage medium for combined error loss of data in hole |
| CN114528746A (en) * | 2020-11-04 | 2022-05-24 | 中国石油化工股份有限公司 | Complex lithology identification method, identification system, electronic device and storage medium |
| WO2024077538A1 (en) * | 2022-10-13 | 2024-04-18 | Saudi Arabian Oil Company | Methods and systems for predicting lithology and formation boundary ahead of the bit |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109614883A (en) * | 2018-11-21 | 2019-04-12 | 瑾逸科技发展扬州有限公司 | A kind of tight sand crack intelligent identification Method based on convolutional neural networks |
| WO2020082824A1 (en) * | 2018-10-25 | 2020-04-30 | 北京国双科技有限公司 | Model generation method and device and lithologic identification method and device |
-
2020
- 2020-05-15 CN CN202010410798.2A patent/CN111598444A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020082824A1 (en) * | 2018-10-25 | 2020-04-30 | 北京国双科技有限公司 | Model generation method and device and lithologic identification method and device |
| CN109614883A (en) * | 2018-11-21 | 2019-04-12 | 瑾逸科技发展扬州有限公司 | A kind of tight sand crack intelligent identification Method based on convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| 陈玉林等: "深度学习算法在测井曲线识别地层岩性中的应用实践", 《第六届数字油田国际学术会议论文集》 * |
| 陈钢花等: "卷积神经网络在岩性识别中的应用", 《测井技术》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114528746A (en) * | 2020-11-04 | 2022-05-24 | 中国石油化工股份有限公司 | Complex lithology identification method, identification system, electronic device and storage medium |
| CN112651171A (en) * | 2020-12-15 | 2021-04-13 | 清华大学 | Data-driven rock physical modeling method and system |
| CN112651171B (en) * | 2020-12-15 | 2024-05-14 | 清华大学 | Data-driven rock physical modeling method and system |
| CN112990320A (en) * | 2021-03-19 | 2021-06-18 | 中国矿业大学(北京) | Lithology classification method and device, electronic equipment and storage medium |
| CN113033676A (en) * | 2021-03-30 | 2021-06-25 | 中铁第四勘察设计院集团有限公司 | Stratum dividing method, device, equipment and storage medium for combined error loss of data in hole |
| WO2024077538A1 (en) * | 2022-10-13 | 2024-04-18 | Saudi Arabian Oil Company | Methods and systems for predicting lithology and formation boundary ahead of the bit |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111598444A (en) | Well logging lithology identification method and system based on convolutional neural network | |
| Zhu et al. | Intelligent logging lithological interpretation with convolution neural networks | |
| CN110346831B (en) | Intelligent seismic fluid identification method based on random forest algorithm | |
| CN107688201B (en) | RBM-based seismic prestack signal clustering method | |
| David et al. | Hierarchical data organization, clustering and denoising via localized diffusion folders | |
| CN104866578A (en) | Hybrid filling method for incomplete data | |
| CN108959322B (en) | Information processing method and device for generating image based on text | |
| CN112712025A (en) | Complex lithology identification method based on long-term and short-term memory neural network | |
| CN110909125B (en) | Detection method of media rumor of news-level society | |
| CN112861522B (en) | Aspect-level emotion analysis method, system and model based on dual-attention mechanism | |
| Zhu et al. | An automatic identification method of imbalanced lithology based on Deep Forest and K-means SMOTE | |
| WO2021127186A1 (en) | Methods and systems for subsurface modeling employing ensemble machine learning prediction trained with data derived from at least one external model | |
| CN113792594B (en) | Method and device for locating language fragments in video based on contrast learning | |
| CN113761805A (en) | Controllable source electromagnetic data denoising method, system, terminal and readable storage medium based on time domain convolution network | |
| Balamurugan et al. | Performance analysis of cart and C5. 0 using sampling techniques | |
| US20230176242A1 (en) | Framework for integration of geo-information extraction, geo-reasoning and geologist-responsive inquiries | |
| Li et al. | Digital construction of geophysical well logging curves using the LSTM deep-learning network | |
| CN111882042B (en) | Neural network architecture automatic search method, system and medium for liquid state machine | |
| CN115994608A (en) | Fracturing well yield prediction method and device based on bidirectional gating circulation unit | |
| Wang et al. | Landslide susceptibility analysis based on a PSO-DBN prediction model in an earthquake-stricken area | |
| CN112489689B (en) | Cross-database voice emotion recognition method and device based on multi-scale difference countermeasure | |
| US20240112113A1 (en) | Apparatus and method for fracking optimization | |
| CN116522070A (en) | Non-supervision intelligent fault diagnosis method and system for mechanical parts | |
| CN112559680B (en) | Text classification method based on multi-level feature extraction, model and storage medium | |
| CN112464483A (en) | Logging curve reconstruction method based on genetic neural network algorithm |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200828 |