CN112906760A - Horizontal well fracturing segment segmentation method, system, equipment and storage medium - Google Patents
Horizontal well fracturing segment segmentation method, system, equipment and storage medium Download PDFInfo
- Publication number
- CN112906760A CN112906760A CN202110130584.4A CN202110130584A CN112906760A CN 112906760 A CN112906760 A CN 112906760A CN 202110130584 A CN202110130584 A CN 202110130584A CN 112906760 A CN112906760 A CN 112906760A
- Authority
- CN
- China
- Prior art keywords
- graph
- horizontal well
- matrix
- neural network
- segmentation
- 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
- 230000011218 segmentation Effects 0.000 title claims abstract description 98
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000003860 storage Methods 0.000 title claims abstract description 7
- 238000013528 artificial neural network Methods 0.000 claims abstract description 86
- 239000011159 matrix material Substances 0.000 claims abstract description 84
- 238000004364 calculation method Methods 0.000 claims abstract description 42
- 239000013598 vector Substances 0.000 claims abstract description 19
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 16
- 238000005457 optimization Methods 0.000 claims abstract description 13
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims abstract description 10
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 10
- 230000003595 spectral effect Effects 0.000 claims abstract description 7
- 230000001902 propagating effect Effects 0.000 claims abstract description 3
- 239000010410 layer Substances 0.000 claims description 43
- 238000012549 training Methods 0.000 claims description 28
- 230000006870 function Effects 0.000 claims description 22
- 230000008569 process Effects 0.000 claims description 20
- 238000010586 diagram Methods 0.000 claims description 12
- 238000010276 construction Methods 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 5
- 230000002238 attenuated effect Effects 0.000 claims description 3
- 239000002356 single layer Substances 0.000 claims description 3
- 238000001228 spectrum Methods 0.000 claims description 3
- 206010017076 Fracture Diseases 0.000 description 12
- 208000010392 Bone Fractures Diseases 0.000 description 10
- 239000002131 composite material Substances 0.000 description 9
- 238000011161 development Methods 0.000 description 9
- 230000018109 developmental process Effects 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 9
- 238000013507 mapping Methods 0.000 description 5
- 238000009826 distribution Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000005086 pumping Methods 0.000 description 4
- 239000011435 rock Substances 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000005553 drilling Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002347 injection Methods 0.000 description 2
- 239000007924 injection Substances 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- E21B43/00—Methods or apparatus for obtaining oil, gas, water, soluble or meltable materials or a slurry of minerals from wells
- E21B43/25—Methods for stimulating production
- E21B43/26—Methods for stimulating production by forming crevices or fractures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Mining & Mineral Resources (AREA)
- Geology (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Environmental & Geological Engineering (AREA)
- Fluid Mechanics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geochemistry & Mineralogy (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses a horizontal well fracturing segment segmentation method, a system, equipment and a storage medium, wherein the weight of a graph neural network and the weight of a full-connection layer are initialized; constructing an input graph according to the geological information of the horizontal well; creating a self-connection weighted adjacency matrix of an input graph, and performing Laplace spectral decomposition to obtain a Laplace matrix; solving a characteristic vector matrix of the Laplace matrix, inputting the characteristic vector matrix into a neural network of the graph to perform forward propagation calculation, and obtaining a segmentation scheme of network prediction; calculating a difference value between a network predicted segmentation scheme and an artificial segmentation scheme by using a cross entropy loss function, and then reversely propagating and updating the weight by using an Adam optimization algorithm; until all the horizontal well geological information is trained completely, obtaining a final trained graph neural network; and inputting the geological information of the horizontal well to be segmented into a neural network of the graph to obtain a horizontal well fracturing segment segmentation scheme. The segmentation efficiency is greatly improved, and the segmentation quality is kept stable.
Description
Technical Field
The invention belongs to the field of unconventional oil and gas exploitation, and relates to a horizontal well fracturing section segmenting method, a system, equipment and a storage medium.
Background
Along with the development of shale gas and compact oil gas in China, unconventional oil gas resources become a new development hot spot in recent years. At present, the horizontal well staged multi-cluster fracturing technology is one of core technologies of unconventional oil and gas resource development, and a staged fracturing development mode of clustered perforation-composite bridge plug combined operation is widely applied in horizontal well construction. Compared with other development modes, the method has the advantages of large-displacement injection, cluster perforation, sectional volume fracturing, high operation efficiency and the like. Through dividing the horizontal well into a plurality of fracturing sections, carry out many clusters perforation in the section, can form several hydraulic fractures simultaneously under single pump injection, effectively reduce construction cost.
The staged fracturing technology of the cluster perforation-composite bridge plug needs to divide a horizontal well section into a plurality of sections (the control distance of one section is 100-150 m), the first section adopts an oil pipe, a continuous oil pipe and a cable crawler to carry out fracturing after perforation, and other sections adopt a cluster perforation-composite bridge plug combined process technology for construction. The combined instrument string is lowered into the well by a cable and is pushed in a hydraulic pumping mode at a large-inclination horizontal well section, namely a hydraulic pumping process technology. The method comprises the steps of plugging a previous section by using a composite bridge plug, performing clustering perforation on the section to form a connecting instrument string, and performing volume fracturing construction on the section.
The first step of the staged fracturing technology of the clustered perforation-composite bridge plug is horizontal well staging, the geological exploration is firstly carried out on a logging well for horizontal well staging, geological information such as a gas measurement value, porosity, rock stratum distribution and a predicted fracture section of the horizontal well is obtained through measurement, and then horizontal well sections which are located in the same small layer, have natural fracture development sections, similar reservoir parameters, similar rock mechanics parameters and similar geological structures are divided into the same staged fracturing section according to the information, so that the horizontal well is divided into a plurality of fracturing sections. The grade of the subsection level directly determines the perforation quality, the effective degree of hydraulic fracture and the volume fracturing construction effect, and the quality directly influences the oil gas output efficiency and the output, so the horizontal well subsection is a key link of the clustering perforation-composite bridge plug subsection fracturing technology.
At present, the fracturing section of the horizontal well is segmented manually, the horizontal well is segmented manually according to geological information, the manual segmentation efficiency is low, the labor cost is high, the quality of the manual segmentation is unstable, some production logging data show that the segmentation of the fracturing section is not reasonable enough, effective hydraulic fractures cannot be formed in partial perforating clusters, effective fracturing cannot be formed in partial fracturing sections, the oil and gas production efficiency is reduced, and the waste of oil and gas resources is caused. Therefore, how to improve the segmentation efficiency, reduce the segmentation cost and improve the segmentation quality becomes a problem which needs to be solved urgently.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a horizontal well fracturing segment segmentation method, a horizontal well fracturing segment segmentation system, horizontal well fracturing segment segmentation equipment and a storage medium, which can automatically segment efficiently and quickly under the condition of keeping the segmentation quality similar to that of manual segmentation, thereby greatly improving the segmentation efficiency, reducing the labor cost and keeping the segmentation quality stable.
In order to achieve the purpose, the invention adopts the following technical scheme to realize the purpose:
a horizontal well fracturing segment segmenting method comprises the following steps;
step 5, acquiring a network-predicted manual segmentation scheme of the horizontal well fracturing section, calculating a difference value between the network-predicted segmentation scheme and the manual segmentation scheme by using a cross entropy loss function, and then reversely propagating and updating the weight by using an Adam optimization algorithm;
step 6, the steps 2 to 5 are circulated until all the horizontal well geological information is trained completely, and a finally trained graph neural network is obtained;
and 7, inputting geological information of the horizontal well to be segmented into a neural network of the graph to obtain a horizontal well fracturing segment segmentation scheme.
Preferably, in step 2, the specific process of constructing the input map is as follows: determining the minimum segment length of the horizontal well and the total length of the horizontal well, enabling the minimum segment length to be an initial value of graph nodes, enabling each segment to be an input graph node, and calculating the number of the graph nodes; and according to the horizontal well geological information of the segments represented by the graph nodes and the distance relation between each segment, creating weighted edges between the graph nodes to represent the weights of the two graph nodes and the edges between the graph nodes, wherein the set of all the graph nodes and all the weighted edges forms an input graph of the graph neural network.
Further, the calculation process of the number of the graph nodes is as follows:
wherein total _ length represents the total length of the horizontal well, min _ length represents the minimum segment length, and XnumRepresenting the number of graph nodes.
Preferably, in step 3, the elements in ith row and jth column of the matrix with the weighted adjacency matrix are connected
The values of (A) are:
wherein X (i, j) represents the weighted edge weight between the ith node and the jth node obtained in the step 3;
the laplace matrix is:
in the formula
The self-join weighted adjacency matrix for the graph obtained in step 4,
is composed of
Degree matrix of (L) is a pair
And carrying out Laplace spectrum decomposition and normalization to obtain a Laplace matrix.
Preferably, the specific process in step 4 is to calculate an eigenvector matrix of the laplacian matrix, and input the ith layer eigenvector matrix of the hidden state of the input graph and the normalized laplacian matrix into the ith layer of the graph neural network for calculation to obtain a calculation result; performing nonlinear activation processing on the calculation result by using a Relu activation function to complete the calculation of the single-layer graph neural network until the feature vector matrix of each layer of the graph neural network is updated; and inputting all updating results of the graph neural network into a full-connection layer for node classification, wherein the nodes classified into one class are one section, and thus a horizontal well segmentation scheme predicted by the graph neural network is obtained.
Preferably, in step 5, the cross entropy loss function is:
where Y represents the manual segmentation scheme,
segmentation scheme, Y, representing network predictioniAnd
respectively representing each item element in the two vectors;
the calculation process of the Adam optimization algorithm is as follows:
t←t+1
mt←β1·mt-1+(1-β1)·gt
wherein t represents a time step t; f (theta) represents an objective function to be optimized, namely a cross entropy loss function in the invention; theta denotes the network weight parameter to be updated, thetat-1Parameter, θ, representing the last time steptA parameter representing a current time step updated after an iteration; beta is a1And beta2The two important hyper-parameters of the algorithm generally take values of 0.9 and 0.999 respectively; alpha is the learning rate, generally the initial value is 0.01 or 0.001, and the learning rate is attenuated after 10-20 iterations; ε represents the optimization algorithm bias and is typically 10-8。
A horizontal well frac segment staging system comprising:
the weight initialization module is used for initializing the weight of the graph neural network and the weight of the full connection layer;
the input diagram constructing module is used for constructing an input diagram according to the geological information of the horizontal well;
the Laplace matrix calculation module is used for creating a self-connection weighted adjacency matrix of the input graph and carrying out Laplace spectral decomposition on the self-connection weighted adjacency matrix to obtain a Laplace matrix;
the network prediction segmentation scheme calculation module is used for solving a characteristic vector matrix of the Laplace matrix, inputting the characteristic vector matrix into a neural network of a graph for forward propagation calculation, and obtaining a network prediction segmentation scheme;
the back propagation updating weight module is used for obtaining a network predicted artificial segmentation scheme of the horizontal well fracturing section, calculating a difference value between the network predicted segmentation scheme and the artificial segmentation scheme by using a cross entropy loss function, and then performing back propagation updating weight by using an Adam optimization algorithm;
the graph neural network training completion acquisition module is used for repeatedly inputting the calculation processes of the graph construction module, the Laplace matrix calculation module, the network prediction segmentation scheme calculation module and the back propagation updating weight module until all the horizontal well geological information is trained completely, and obtaining a finally trained graph neural network;
and the horizontal well fracturing section segmenting module is used for inputting the geological information of the horizontal well to be segmented into the neural network of the graph to obtain a horizontal well fracturing section segmenting scheme.
A computer apparatus comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor when executing the computer program implements the steps of any of the horizontal well fracture staged methods described above.
A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of any of the horizontal well fracture zone segmentation methods described above.
Compared with the prior art, the invention has the following beneficial effects:
according to the non-Euclidean space characteristics of the logging geological information, the logging geological data is subjected to graphic data modeling and serves as a graph neural network training set, a graph neural network is constructed according to the powerful capability of the graph neural network in the aspect of processing the characteristics of the non-Euclidean space data, and the graph neural network is made to learn the relation between the geological data and the segmentation scheme by utilizing the conventional artificial segmentation scheme. After the training of the neural network of the graph is finished, the network is used for automatically segmenting the horizontal well which is not subjected to manual segmentation without manual intervention, so that the horizontal well can be automatically segmented efficiently and quickly under the condition that the segmentation quality is close to that of the manual segmentation, the segmentation efficiency is greatly improved, the labor cost is reduced, and the segmentation quality is kept stable.
Drawings
FIG. 1 is a schematic diagram of a neural network architecture used in the present invention;
FIG. 2 is a schematic illustration of a horizontal well block configuration to be performed by the present invention;
FIG. 3 is a schematic diagram of a neural network training process provided by the present invention;
FIG. 4 is a schematic diagram illustrating a process for automatically segmenting the trained neural network provided by the present invention;
fig. 5 is a flow chart of an implementation of the present invention.
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings:
FIG. 1 is a schematic diagram of a neural network structure for computing according to the present invention. The graph neural network is mainly divided into two parts, namely a hidden layer and an output layer, wherein the hidden layer is used for hiding the output hidden state H of the previous layerlFeature vector normalized Laplace matrix L of graph nodes and weight W of the layerlPerforming matrix multiplication operation, and then performing nonlinear activation processing on the result by using a relu function, thereby realizing the updating of a hidden state;
Hl+1＝σ(LHlWl)
the output layer is used for receiving the final calculation result output by the hidden layer, namely the final updated graph information, and then performing operation on the graph information to obtain the final calculation result. The form of the output layer and the meaning of the calculation result depend on specific downstream tasks, the output layer is a full-connection layer in the invention, the classification operation is carried out on the graph information, and the final output data represents a horizontal well segmentation scheme.
In the formula
For the final calculation of the neural network of the graph obtained in step 8, WfFor full-link weights, softmax denotes the full-link classification operation,
representing the segmentation result finally output by the network.
Fig. 2 is a schematic diagram of a horizontal well block configuration to be performed by the present invention. Along with the development of shale gas and compact oil gas in China, the staged fracturing development mode of the cluster perforation-composite bridge plug combined operation is widely applied in the horizontal well construction. The method comprises the steps of firstly drilling a vertical well on the ground surface, obtaining the approximate position of an oil-gas layer through well logging detection, then drilling a horizontal well, measuring geological information such as a gas logging value, porosity, rock stratum distribution and a predicted fracture section of the horizontal well, dividing horizontal well sections which are located in the same small layer, have natural fracture development sections, similar reservoir parameters, similar rock mechanical parameters and similar geological structures into the same perforation fracture section according to the information, and accordingly dividing the horizontal well into a plurality of fracture sections. After the segmentation is finished, the cluster perforation-composite bridge plug combined operation technology can be carried out for construction, the first section adopts an oil pipe, a continuous oil pipe and a cable crawler for perforation and then fracturing, the other sections use cables to enable the combined operation instrument to be strung into the well, and the combined operation instrument is propelled in a hydraulic pumping mode at a large-inclination horizontal well section, namely the hydraulic pumping technology. The method comprises the steps of plugging a previous section by using a composite bridge plug, performing clustering perforation on the section to form a connecting instrument string, and performing volume fracturing construction on the section.
FIG. 3 is a schematic diagram of a neural network training process provided by the present invention. Firstly, acquiring data of horizontal well data which is manually segmented before, using the data as a training set by a graph neural network, and carrying out graph modeling on geological data before training. And then creating weighted edges among the nodes of the graph according to geological information such as gas measurement values, porosity and the like and the distance relationship among the segments, so that the creation of the input graph is finished. And performing the operation on all horizontal well data in the training set to obtain an input graph training set. Before training, the graph neural network weights and the full-connection layer weights are initialized, and then training can be started. The training of a graph is mainly divided into 5 steps: performing Laplace spectral decomposition and normalization on the weighted adjacent matrix of the input image to obtain a normalized Laplace matrix; solving a characteristic vector matrix of the input graph nodes; performing matrix multiplication on an input graph feature vector matrix, a normalized Laplace matrix and a graph neural network weight matrix, and performing nonlinear activation processing by using a relu function to update a hidden state of a graph node; repeating the previous process until all hidden layers of the graph neural network are subjected to iterative computation; inputting the final calculation result of the hidden layer into a full-connection layer for node classification, wherein the nodes are divided into one section, and a horizontal well segmentation scheme predicted by a graph neural network is obtained; calculating a difference value between a network predicted segmentation scheme and an artificial segmentation scheme by using a cross entropy loss function, and performing back propagation by using an Adam optimization algorithm to update the network weight to a direction of reducing the loss function;
and after one training is finished, judging whether the training is finished or not, if not, inputting a next graph to repeat the operation for next training, and if all the graphs in the training set are trained, finishing the training of the graph neural network.
FIG. 4 is a schematic diagram of a process of automatically segmenting the trained neural network provided by the present invention. The method comprises the steps of firstly obtaining horizontal well logging geological data with segments, then carrying out graph modeling on the horizontal well according to the method for manufacturing the training set, then inputting the graph into a trained graph neural network for forward propagation calculation, and finally obtaining a segmentation scheme of network prediction, thereby realizing automatic prediction of the graph neural network.
The horizontal well fracturing segment segmentation task aims at performing perforation fracturing segment segmentation on a horizontal well according to logging geological information, and belongs to a typical task of manually extracting characteristics depending on manual experience and manpower, so that the mapping relation between logging data and a segmentation scheme can be learned through a neural network, and automatic segmentation of the horizontal well is realized. For data processed by a traditional neural network, the data must be Euclidean space data with a regular structure, data nodes are ordered nodes with constant sizes, and the number of adjacent nodes of any one data node is the same. However, the segmented logging data for the horizontal well segmentation does not have the characteristic, the data does not have spatial translation, different data nodes are not independent, and the logging data belongs to typical graph data of non-Euclidean space, so that the invention proposes that a graph neural network specially processing the characteristics of the non-Euclidean space data is used for automatically segmenting the horizontal well fracturing segment. To achieve this goal, the following 4 basic processes are required:
carrying out proper mathematical abstraction on the logging data, and transforming the logging data into a graph node matrix X;
processing the input data X by using a graph neural network, and learning a mapping relation between a segmentation scheme and the input data;
Comparing the network segmentation scheme with the manual segmentation scheme, continuously learning the network, and continuously updating the network weight to ensure that the quality of the network segmentation scheme is continuously close to that of the manual segmentation scheme;
y is a manual segmentation scheme that is,
is a segmentation scheme of the output of the network,
is the difference between the two schemes, the network f () is updated in the direction of the smallest possible difference.
And obtaining the trained graph neural network. Because the training target is to reduce the difference between the network output segmentation scheme and the artificial segmentation scheme, the finally obtained graph neural network can output the segmentation scheme with the quality similar to that of the artificial segmentation scheme
A horizontal well fracturing section segmenting method based on a graph neural network comprises two parts, wherein the first part is training of the graph neural network, and the second part is reasoning of the graph neural network, and the method is characterized in that:
and training the neural network of the graph is used for realizing the mapping from the geological information of the horizontal well to the segmentation scheme. Constructing the existing horizontal well which is segmented manually into graphs, taking the graphs as training data sets, training a graph neural network through forward propagation and backward propagation, and learning a mapping relation between horizontal well geological information and a segmentation scheme, so that the difference between the network segmentation scheme and the manual segmentation scheme is as small as possible, and finally obtaining the trained graph neural network.
And the inference of the graph neural network is used for automatically segmenting the horizontal well which is not segmented manually through the trained graph neural network without manual intervention. And constructing the horizontal well to be segmented into an input graph, inputting the graph into the trained graph neural network, and outputting the network to be the segmentation scheme.
As shown in fig. 5, the horizontal well fracture section segmenting method based on the neural network comprises the following work flows:
wherein total _ length represents the total length of the horizontal well, min _ length represents the minimum segment length, and XnumRepresenting the number of graph nodes, which is equal to the total horizontal well length divided by the minimum segment length rounded up.
And 4, creating a self-connection weighted adjacency matrix of the graph according to the input graph obtained in the step 3. The number of rows and columns of the matrix is the same and is the number of nodes in the figure; elements in the matrix at row i and column j
The value of (A) is shown by the following formula:
wherein X (i, j) represents the weighted edge weight between the ith node and the jth node obtained in step 3.
Step 5, performing Laplace spectral decomposition and normalization on the self-connection weighted adjacency matrix obtained in the step 4 by using the following formula to obtain a normalized Laplace matrix L;
in the formula
The self-join weighted adjacency matrix for the graph obtained in step 4,
is composed of
Degree matrix of (L) is a pair
And carrying out Laplace spectrum decomposition and normalization to obtain a Laplace matrix.
Step 6, solving the eigenvector matrix H of the Laplace matrix L obtained in the step 50The solving process is as follows:
|λI-L|＝0
(λI-L)α＝0
H0＝[α1，α2，...，αn]
in the formula, I is an identity matrix, λ is a matrix eigenvalue to be solved, L is an input graph laplacian matrix obtained in step 4, and α is an eigenvector corresponding to the eigenvalue λ. All possible values of alpha together form the eigenvector matrix H0. The matrix represents the initial hidden state of the node;
Zl＝LHlWl
in the formula WlRepresenting the l-th layer diagram neural network weight.
Step 8, carrying out nonlinear activation processing on the result of the step 7 by using a Relu activation function to complete the calculation of the single-layer graph neural network and realize the hidden stateState from HlTo Hl+1And updating by adding 1 to l;
Hl+1＝Relu(Zl)
l＝l+1
and 9, if l is less than the total layer number of the neural network of the graph after updating in the step 8, skipping to the step 7, executing the step 7 and the step 8 again, and otherwise skipping to the step 10.
And step 10, inputting the final result obtained by calculating the graph neural network into a full-connected layer for node classification, wherein the nodes classified into one class are a section, and thus the horizontal well segmentation scheme predicted by the graph neural network is obtained. The full connection layer calculation formula is as follows:
in the formula
For the final calculation of the neural network of the graph obtained in step 9, WfFor full-link weights, softmax denotes the full-link classification operation,
and the segmented result which represents the final network output is divided into nodes of the same class as one segment.
And 11, calculating the difference value between the network prediction segmentation scheme and the artificial segmentation scheme by using a cross entropy loss function. The cross entropy loss function is as follows:
where Y represents the manual segmentation scheme,
represents the network segmentation scheme, Y, obtained in step 10iAnd
representing each item element in the two vectors separately.
After the cross entropy is obtained, backward propagation is carried out by using an Adam optimization algorithm, so that the network weight is updated in the direction of reducing the loss function. The calculation process of the Adam optimization algorithm is as follows:
t←t+ 1
mt←β1·mt-1+(1-β1)·gt
wherein t represents a time step t; f (theta) represents an objective function to be optimized, namely a cross entropy loss function in the invention; theta denotes the network weight parameter to be updated, thetat-1Parameter, θ, representing the last time steptA parameter representing a current time step updated after an iteration; beta is a1And beta2The two important hyper-parameters of the algorithm generally take values of 0.9 and 0.999 respectively; alpha is the learning rate, generally the initial value is 0.01 or 0.001, and the learning rate is attenuated after 10-20 iterations; ε represents the optimization algorithm bias and is typically 10-8。
Step 12, if horizontal well data to be trained still exist, training is not finished, the step 2 is skipped, and the step 2 to the step 12 are repeated; otherwise, all training is finished, and the step 13 is skipped;
step 13, finishing the training to obtain a trained graph neural network;
step 14, measuring geological information of a horizontal well to be segmented;
step 15, repeating the steps 2 to 11;
step 16, outputting a segmentation scheme of the horizontal well to be segmented by the neural network;
the above-mentioned contents are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modification made on the basis of the technical idea of the present invention falls within the protection scope of the claims of the present invention.
Claims (9)
1. A horizontal well fracturing segment segmenting method is characterized by comprising the following steps;
step 1, initializing weights and full-connection layer weights of a graph neural network;
step 2, constructing an input graph according to the geological information of the horizontal well;
step 3, creating a self-connection weighted adjacency matrix of the input graph, and performing Laplace spectral decomposition on the self-connection weighted adjacency matrix to obtain a Laplace matrix;
step 4, solving a characteristic vector matrix of the Laplace matrix, inputting the characteristic vector matrix into a neural network of a graph to perform forward propagation calculation, and obtaining a segmentation scheme of network prediction;
step 5, acquiring a network-predicted manual segmentation scheme of the horizontal well fracturing section, calculating a difference value between the network-predicted segmentation scheme and the manual segmentation scheme by using a cross entropy loss function, and then reversely propagating and updating the weight by using an Adam optimization algorithm;
step 6, the steps 2 to 5 are circulated until all the horizontal well geological information is trained completely, and a finally trained graph neural network is obtained;
and 7, inputting geological information of the horizontal well to be segmented into a neural network of the graph to obtain a horizontal well fracturing segment segmentation scheme.
2. The horizontal well fracturing segment segmenting method according to claim 1, wherein in the step 2, the specific process of constructing the input graph is as follows: determining the minimum segment length of the horizontal well and the total length of the horizontal well, enabling the minimum segment length to be an initial value of graph nodes, enabling each segment to be an input graph node, and calculating the number of the graph nodes; and according to the horizontal well geological information of the segments represented by the graph nodes and the distance relation between each segment, creating weighted edges between the graph nodes to represent the weights of the two graph nodes and the edges between the graph nodes, wherein the set of all the graph nodes and all the weighted edges forms an input graph of the graph neural network.
3. The horizontal well fracture section segmentation method according to claim 2, wherein the calculation process of the number of graph nodes is as follows:
wherein total _ length represents the total length of the horizontal well, min _ length represents the minimum segment length, and XnumRepresenting the number of graph nodes.
4. The horizontal well fracture section segmentation method according to claim 1, wherein in step 3, the elements in the ith row and the jth column in the matrix of the weighted adjacency matrix are connected in a self-connection manner
The values of (A) are:
wherein X (i, j) represents the weighted edge weight between the ith node and the jth node obtained in the step 3;
the laplace matrix is:
in the formula
The self-join weighted adjacency matrix for the graph obtained in step 4,
is composed of
Degree matrix of (L) is a pair
And carrying out Laplace spectrum decomposition and normalization to obtain a Laplace matrix.
5. The horizontal well fracturing segment segmenting method according to claim 1, wherein the specific process of the step 4 is to calculate an eigenvector matrix of a Laplace matrix, and input the I layer eigenvector matrix of the hidden state of the input graph and the normalized Laplace matrix into the I layer of a graph neural network for calculation to obtain a calculation result; performing nonlinear activation processing on the calculation result by using a Relu activation function to complete the calculation of the single-layer graph neural network until the feature vector matrix of each layer of the graph neural network is updated; and inputting all updating results of the graph neural network into a full-connection layer for node classification, wherein the nodes classified into one class are one section, and thus a horizontal well segmentation scheme predicted by the graph neural network is obtained.
6. The horizontal well fracture section segmentation method according to claim 1, wherein in step 5, the cross entropy loss function is:
where Y represents the manual segmentation scheme,
segmentation scheme, Y, representing network predictioniAnd
respectively representing each item element in the two vectors;
the calculation process of the Adam optimization algorithm is as follows:
t←t+1
mt←β1·mt-1+(1-β1)·gt
wherein t represents a time step t; f (theta) represents an objective function to be optimized, namely a cross entropy loss function in the invention; theta denotes the network weight parameter to be updated, thetat-1Parameter, θ, representing the last time steptA parameter representing a current time step updated after an iteration; beta is a1And beta2The two important hyper-parameters of the algorithm generally take values of 0.9 and 0.999 respectively; alpha is the learning rate, generally the initial value is 0.01 or 0.001, and the learning rate is attenuated after 10-20 iterations; ε represents the optimization algorithm bias and is typically 10-8。
7. A horizontal well fracture section segmentation system, comprising:
the weight initialization module is used for initializing the weight of the graph neural network and the weight of the full connection layer;
the input diagram constructing module is used for constructing an input diagram according to the geological information of the horizontal well;
the Laplace matrix calculation module is used for creating a self-connection weighted adjacency matrix of the input graph and carrying out Laplace spectral decomposition on the self-connection weighted adjacency matrix to obtain a Laplace matrix;
the network prediction segmentation scheme calculation module is used for solving a characteristic vector matrix of the Laplace matrix, inputting the characteristic vector matrix into a neural network of a graph for forward propagation calculation, and obtaining a network prediction segmentation scheme;
the back propagation updating weight module is used for obtaining a network predicted artificial segmentation scheme of the horizontal well fracturing section, calculating a difference value between the network predicted segmentation scheme and the artificial segmentation scheme by using a cross entropy loss function, and then performing back propagation updating weight by using an Adam optimization algorithm;
the graph neural network training completion acquisition module is used for repeatedly inputting the calculation processes of the graph construction module, the Laplace matrix calculation module, the network prediction segmentation scheme calculation module and the back propagation updating weight module until all the horizontal well geological information is trained completely, and obtaining a finally trained graph neural network;
and the horizontal well fracturing section segmenting module is used for inputting the geological information of the horizontal well to be segmented into the neural network of the graph to obtain a horizontal well fracturing section segmenting scheme.
8. A computer apparatus comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor when executing the computer program performs the steps of the horizontal well fracture zone segmentation method of any one of claims 1 to 6.
9. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the horizontal well fracture staged method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110130584.4A CN112906760B (en) | 2021-01-29 | 2021-01-29 | Horizontal well fracturing segment segmentation method, system, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110130584.4A CN112906760B (en) | 2021-01-29 | 2021-01-29 | Horizontal well fracturing segment segmentation method, system, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112906760A true CN112906760A (en) | 2021-06-04 |
| CN112906760B CN112906760B (en) | 2024-05-03 |
Family
ID=76121685
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110130584.4A Active CN112906760B (en) | 2021-01-29 | 2021-01-29 | Horizontal well fracturing segment segmentation method, system, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112906760B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114218862A (en) * | 2021-12-15 | 2022-03-22 | 河北工业大学 | HHO-RCNN-based petroleum logging hydrocarbon reservoir identification method |
| CN116291336A (en) * | 2023-02-14 | 2023-06-23 | 电子科技大学 | Automatic segmentation clustering system based on deep self-attention neural network |
| CN118673368A (en) * | 2024-08-23 | 2024-09-20 | 西南石油大学 | Shale gas horizontal well segmentation clustering method based on graph convolution neural network |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130264056A1 (en) * | 2012-03-21 | 2013-10-10 | Oiltool Engineering Services, Inc. | Multizone Frac System |
| US20170284181A1 (en) * | 2014-08-26 | 2017-10-05 | Gas Technology Institute | Hydraulic fracturing system and method |
| CN109799533A (en) * | 2018-12-28 | 2019-05-24 | 中国石油化工股份有限公司 | A kind of method for predicting reservoir based on bidirectional circulating neural network |
| CN109812247A (en) * | 2019-02-21 | 2019-05-28 | 中国石油天然气股份有限公司 | Perforation and fracturing method for improving transformation degree of compact oil horizontal well |
| CN109993203A (en) * | 2019-02-23 | 2019-07-09 | 中国海洋石油集团有限公司 | A kind of well logging oil-gas-water layer intelligent interpretation evaluation method |
| CN110287516A (en) * | 2019-05-13 | 2019-09-27 | 西南石油大学 | A kind of shale gas reservoir pressure break horizontal well compressibility integrated evaluating method |
| CN110400006A (en) * | 2019-07-02 | 2019-11-01 | 中国石油化工股份有限公司 | Oil well output prediction technique based on deep learning algorithm |
| CN112085161A (en) * | 2020-08-20 | 2020-12-15 | 清华大学 | Graph neural network method based on random information transmission |
-
2021
- 2021-01-29 CN CN202110130584.4A patent/CN112906760B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130264056A1 (en) * | 2012-03-21 | 2013-10-10 | Oiltool Engineering Services, Inc. | Multizone Frac System |
| US20170284181A1 (en) * | 2014-08-26 | 2017-10-05 | Gas Technology Institute | Hydraulic fracturing system and method |
| CN109799533A (en) * | 2018-12-28 | 2019-05-24 | 中国石油化工股份有限公司 | A kind of method for predicting reservoir based on bidirectional circulating neural network |
| CN109812247A (en) * | 2019-02-21 | 2019-05-28 | 中国石油天然气股份有限公司 | Perforation and fracturing method for improving transformation degree of compact oil horizontal well |
| CN109993203A (en) * | 2019-02-23 | 2019-07-09 | 中国海洋石油集团有限公司 | A kind of well logging oil-gas-water layer intelligent interpretation evaluation method |
| CN110287516A (en) * | 2019-05-13 | 2019-09-27 | 西南石油大学 | A kind of shale gas reservoir pressure break horizontal well compressibility integrated evaluating method |
| CN110400006A (en) * | 2019-07-02 | 2019-11-01 | 中国石油化工股份有限公司 | Oil well output prediction technique based on deep learning algorithm |
| CN112085161A (en) * | 2020-08-20 | 2020-12-15 | 清华大学 | Graph neural network method based on random information transmission |
Non-Patent Citations (3)
| Title |
|---|
| 刘全稳: "测井神经网络技术综述", 石油地球物理勘探, vol. 31, no. 1, pages 64 - 70 * |
| 王忠东等: "页岩气水平井产量主控因素分析及产能预测", 测井技术, vol. 41, no. 5, pages 577 - 582 * |
| 高东伟;: "页岩气水平井压裂细分段密分簇优化设计及应用", 长江大学学报(自然科学版), vol. 16, no. 04, pages 40 - 44 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114218862A (en) * | 2021-12-15 | 2022-03-22 | 河北工业大学 | HHO-RCNN-based petroleum logging hydrocarbon reservoir identification method |
| CN114218862B (en) * | 2021-12-15 | 2024-10-18 | 河北工业大学 | Petroleum logging hydrocarbon reservoir identification method based on HHO-RCNN |
| CN116291336A (en) * | 2023-02-14 | 2023-06-23 | 电子科技大学 | Automatic segmentation clustering system based on deep self-attention neural network |
| CN116291336B (en) * | 2023-02-14 | 2024-05-24 | 电子科技大学 | Automatic segmentation clustering system based on deep self-attention neural network |
| CN118673368A (en) * | 2024-08-23 | 2024-09-20 | 西南石油大学 | Shale gas horizontal well segmentation clustering method based on graph convolution neural network |
| CN118673368B (en) * | 2024-08-23 | 2024-10-22 | 西南石油大学 | Shale gas horizontal well segmentation clustering method based on graph convolution neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112906760B (en) | 2024-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shi et al. | Development of subsurface geological cross-section from limited site-specific boreholes and prior geological knowledge using iterative convolution XGBoost | |
| CN108334949B (en) | Image classifier construction method based on optimized deep convolutional neural network structure fast evolution | |
| CN112906760B (en) | Horizontal well fracturing segment segmentation method, system, equipment and storage medium | |
| CN110807544B (en) | Oil field residual oil saturation distribution prediction method based on machine learning | |
| Karkevandi-Talkhooncheh et al. | Application of hybrid adaptive neuro-fuzzy inference system in well placement optimization | |
| CN110619353B (en) | Multi-scale logging curve automatic identification method based on deep learning | |
| Wang et al. | A critical review on intelligent optimization algorithms and surrogate models for conventional and unconventional reservoir production optimization | |
| CN113610945B (en) | Ground stress curve prediction method based on hybrid neural network | |
| CN112541572A (en) | Residual oil distribution prediction method based on convolutional encoder-decoder network | |
| Raji et al. | A new approach in well placement optimization using metaheuristic algorithms | |
| CN115204064B (en) | Gas injection huff-puff parameter optimization method and system in shale oil exploitation | |
| Kim et al. | Construction of prior models for ES-MDA by a deep neural network with a stacked autoencoder for predicting reservoir production | |
| Wang et al. | A novel surrogate-assisted multi-objective optimization method for well control parameters based on tri-training | |
| CN111058840A (en) | Organic carbon content (TOC) evaluation method based on high-order neural network | |
| Huang et al. | A deep-learning-based graph neural network-long-short-term memory model for reservoir simulation and optimization with varying well controls | |
| Deng et al. | A hybrid machine learning optimization algorithm for multivariable pore pressure prediction | |
| Chen et al. | Interpretable machine learning optimization (InterOpt) for operational parameters: A case study of highly-efficient shale gas development | |
| CN116612664A (en) | Ship traffic flow prediction method based on improved space-time diagram attention neural network | |
| Avalos et al. | Geological modeling using a recursive convolutional neural networks approach | |
| CN111461348A (en) | Deep network embedded learning method based on graph core | |
| CN115099406A (en) | Stratum pressure inversion method and device based on multivariate time sequence | |
| CN116911216B (en) | Reservoir oil well productivity factor assessment and prediction method | |
| Islam et al. | A surrogate assisted quantum-behaved algorithm for well placement optimization | |
| CN110619382A (en) | Convolution depth network construction method suitable for seismic exploration | |
| CN117574755A (en) | Hierarchical multistage optimization method for fracturing parameters of horizontal well of shale reservoir well factory |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |