CN108573320B - Method and system for calculating final recoverable reserves of shale gas reservoir - Google Patents
Method and system for calculating final recoverable reserves of shale gas reservoir Download PDFInfo
- Publication number
- CN108573320B CN108573320B CN201810191485.5A CN201810191485A CN108573320B CN 108573320 B CN108573320 B CN 108573320B CN 201810191485 A CN201810191485 A CN 201810191485A CN 108573320 B CN108573320 B CN 108573320B
- Authority
- CN
- China
- Prior art keywords
- shale gas
- gas reservoir
- data
- training
- recoverable reserves
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000010801 machine learning Methods 0.000 claims abstract description 41
- 238000005553 drilling Methods 0.000 claims abstract description 25
- 230000035699 permeability Effects 0.000 claims abstract description 14
- 238000012549 training Methods 0.000 claims description 58
- 238000003062 neural network model Methods 0.000 claims description 37
- 230000006870 function Effects 0.000 claims description 23
- 238000012706 support-vector machine Methods 0.000 claims description 20
- 230000009467 reduction Effects 0.000 claims description 16
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 210000002569 neuron Anatomy 0.000 claims description 12
- 239000011148 porous material Substances 0.000 claims description 11
- 238000010586 diagram Methods 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 8
- 238000012546 transfer Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000011156 evaluation Methods 0.000 abstract description 6
- 239000007789 gas Substances 0.000 description 109
- 239000011435 rock Substances 0.000 description 9
- 238000004364 calculation method Methods 0.000 description 6
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 6
- 238000004519 manufacturing process Methods 0.000 description 5
- 230000015572 biosynthetic process Effects 0.000 description 4
- 239000003345 natural gas Substances 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000003247 decreasing effect Effects 0.000 description 3
- 239000012530 fluid Substances 0.000 description 3
- 238000001179 sorption measurement Methods 0.000 description 3
- 229910052799 carbon Inorganic materials 0.000 description 2
- 239000003245 coal Substances 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- FAPWRFPIFSIZLT-UHFFFAOYSA-M Sodium chloride Chemical compound [Na+].[Cl-] FAPWRFPIFSIZLT-UHFFFAOYSA-M 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013398 bayesian method Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000003795 desorption Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 210000000653 nervous system Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000011780 sodium chloride Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Operations Research (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Biomedical Technology (AREA)
- Entrepreneurship & Innovation (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and a system for calculating final recoverable reserves of a shale gas reservoir, wherein the method comprises the following steps: s1, acquiring characteristic data of the shale gas reservoir, wherein the characteristic data comprises the porosity, the permeability, the well completion method, the drilling parameters, the fracture length and the flow conductivity of the shale gas reservoir; and S2, inputting the characteristic data into the trained machine learning model, and obtaining the predicted value of the final recoverable reserves of the shale gas reservoir. The method fully utilizes known historical data, establishes a machine learning model on the premise of not making any hypothesis, inputs characteristic data closely related to the final recoverable reserves of the shale gas reservoir into a trained support machine learning model, obtains the predicted value of the final recoverable reserves of the shale gas reservoir, and ensures that the obtained predicted value accords with the actual rule of the shale gas reservoir, so that the final recoverable reserves evaluation result is real and reliable and has high accuracy.
Description
Technical Field
The invention relates to the technical field of shale gas exploration, in particular to a method and a system for calculating final recoverable reserves of a shale gas reservoir.
Background
At present, the unconventional natural gas in the global range comprises shale gas, coal bed gas and dense gas, wherein the shale gas resource reserves are equivalent to the sum of the coal bed gas and the dense gas resources, and the reserves account for about half of the total amount of the three unconventional natural gases. Shale gas is an unconventional natural gas stored in dark or high carbon shale with typical in situ-reservoir-forming pattern characteristics. The shale gas reservoir is used for storing dark shale or high-carbon shale rich in organic matters, has the characteristics of low porosity and ultra-low permeability, and is mainly formed by gathering natural gas in an adsorption state or a free state. In contrast to conventional gas reservoirs, shale gas reservoirs are characterized as a "self-generating, self-storing" system, which is either a source rock, a reservoir or a cap rock, and migration of gases after production also occurs within the shale. The shale gas reserves calculation result provides important basis for formulating a shale gas development scheme, determining the investment scale of the shale gas industry and evaluating shale gas resources, and the reserves calculation process of the shale gas reservoir is different from that of a conventional gas reservoir due to the unique adsorption and desorption mechanism of the shale gas reservoir.
At present, two methods for calculating the reserves of shale gas reservoirs are mainly used: yield decline curves and methods of modeling flow. The yield decreasing curve method is only suitable for wells with the yield beginning to decrease and is not suitable for calculating the recoverable reserves of fixed-production wells which reach the quasi-stable flow but have no decreasing yield; for wells with decreasing production, the curve fitting method is carried out on the basis of a plurality of assumed conditions, and due to the idealization of the conditions, the result obtained by the method cannot be well applied to the actual production of the oil field, and has no guiding significance on the actual production. For the method for establishing the flow model, because the shale gas belongs to an unconventional gas reservoir, has the characteristics of low porosity and ultra-low permeability, the situation of pores is very complex, the slip effect and the adsorption effect of the gas exist, the flow in the pores can be still regarded as Darcy flow when the flow model is established at present, and the conclusion obtained by the method naturally has deviation according to the fact that the flow model does not accord with the actual law of the shale gas reservoir fundamentally.
In summary, when shale gas reservoir reserves are evaluated through the prior art, the reuse of historical data is not considered, so that a large amount of data are idle, the proposed method is mostly carried out under the assumed ideal condition and does not accord with the actual rule of the shale gas reservoir, and the final evaluation of the recoverable reserves has deviation and is inaccurate.
Disclosure of Invention
The present invention provides a method and system for calculating the final recoverable capacity of a shale gas reservoir that overcomes, or at least partially solves, the above problems.
According to one aspect of the invention, a method for calculating final recoverable reserves of a shale gas reservoir is provided, and comprises the following steps:
s1, acquiring characteristic data of the shale gas reservoir, wherein the characteristic data comprises the porosity, the permeability, the well completion method, the drilling parameters, the fracture length and the flow conductivity of the shale gas reservoir;
s2, inputting the characteristic data into a trained machine learning model, and obtaining a predicted value of the final recoverable reserves of the shale gas reservoir;
wherein the drilling parameters include weight-on-bit, rate-of-penetration, drilling cycle, rotational speed, pore pressure, burst pressure, and collapse pressure.
Preferably, before the step S1, the method further includes selecting the feature data:
and acquiring attribute data related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir as feature data according to an attribute reduction algorithm.
Preferably, the trained machine learning model in step S2 is built by:
acquiring characteristic data of a plurality of groups of shale gas reservoirs and corresponding recoverable reserves as sample data in a training sample set;
randomly selecting a plurality of groups of sample data with specific numbers from the training sample set, wherein the sample data with each specific group number forms a training subsample set;
and if the group number of the sample data in the training sample set is not more than the preset group number, inputting the sample data in each training sub-sample set into a support vector machine model for training, mapping the sample data to a higher-dimensional space through a Gaussian kernel function, and obtaining a plurality of trained support vector machine models.
Preferably, the method further comprises the following steps:
and if the group number of the sample data in the training sample set is greater than the preset group number, inputting the sample data in each training sub sample set into a neural network model for training, adjusting the connection weight of neurons in the neural network model according to the training result of each group of sample data in the training sub sample set until the error of the training result is less than a preset threshold value, and obtaining a plurality of trained neural network models.
Preferably, the neural network model is a BP neural network model:
the BP neural network model comprises 9 hidden layers;
neurons in the hidden layer employ sigmoid transfer functions.
Preferably, step S2 specifically includes:
inputting the characteristic data into a plurality of trained machine learning models, outputting a predicted value of the final recoverable reserves of a group of shale gas reservoirs by each machine learning model, and acquiring a plurality of groups of predicted values;
obtaining a normal distribution diagram of the predicted values according to a plurality of groups of the predicted values, and obtaining three predicted values of P10, P50 and P90 from the normal distribution diagram;
where P10 indicates that the predicted value has a 10% greater likelihood than the actual final recoverable reserve; p50 indicates that the predicted value has a 50% greater likelihood than the actual final recoverable reserve; p90 indicates that the predicted value has a 90% greater likelihood than the actual final recoverable amount.
According to another aspect of the present invention, there is also provided a system for calculating final recoverable reserves of a shale gas reservoir, comprising:
the system comprises a data acquisition module, a data acquisition module and a data processing module, wherein the data acquisition module is used for acquiring characteristic data of the shale gas reservoir, and the characteristic data comprises the porosity, the permeability, a well completion method, drilling parameters, fracture length and flow conductivity of the shale gas reservoir;
the predicted value obtaining module is used for inputting the characteristic data into a trained machine learning model and obtaining a predicted value of the final recoverable reserves of the shale gas reservoir;
wherein the drilling parameters include weight-on-bit, rate-of-penetration, drilling cycle, rotational speed, pore pressure, burst pressure, and collapse pressure.
Preferably, the system for calculating the final recoverable capacity of the shale gas reservoir further comprises:
and the characteristic selection module is used for acquiring attribute data related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir as characteristic data according to an attribute reduction algorithm.
According to the method and the system for calculating the final recoverable reserves of the shale gas reservoir, historical data of the shale gas reservoir are analyzed and processed, the characteristic data closely related to the final recoverable reserves of the shale gas reservoir are screened out from all the attribute data of the shale gas reservoir by using an attribute reduction algorithm of a rough set, and the characteristic data are input into a trained support vector machine model or a neural network model to obtain the predicted value of the final recoverable reserves of the shale gas reservoir. The method combines the calculation of the final recoverable reserves of the shale gas reservoir with the machine learning, fully utilizes the known historical data, establishes the machine learning model on the premise of not making any hypothesis, and ensures that the estimation result of the final recoverable reserves is real and reliable and has high accuracy because the predicted value obtained by the machine learning model conforms to the actual rule of the shale gas reservoir.
Drawings
FIG. 1 is a flow chart of a method for calculating a final recoverable reserve for a shale gas reservoir in accordance with an embodiment of the present invention;
fig. 2 is a functional block diagram of a system for calculating the final recoverable capacity of a shale gas reservoir in accordance with an embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
Referring to fig. 1, a flowchart of a method for calculating a final recoverable reserve of a shale gas reservoir according to an embodiment of the present invention includes: s1, acquiring characteristic data of the shale gas reservoir, wherein the characteristic data comprises the porosity, the permeability, the well completion method, the drilling parameters, the fracture length and the flow conductivity of the shale gas reservoir; s2, if the group number of the characteristic data is not larger than the preset group number, inputting the characteristic data into a trained machine learning model to obtain a predicted value of the final recoverable reserves of the shale gas reservoir; wherein the drilling parameters include weight-on-bit, rate-of-penetration, drilling cycle, rotational speed, pore pressure, burst pressure, and collapse pressure.
Specifically, one or more sets of characteristic data are collected that are closely related to the final recoverable reserves of the shale gas reservoir, each set of characteristic data including porosity, permeability, completion methodology, drilling parameters, fracture length, and conductivity of the shale gas reservoir.
The porosity is the ratio of the sum of all pore space volumes in the rock sample to the volume of the rock sample, and interconnected pores can not only store oil and gas, but also allow oil and gas to percolate therein, and is one of important parameters for reservoir evaluation. The permeability refers to the ability of rock to allow fluid to pass through under a certain pressure difference, and belongs to the physical foundation of oil reservoir rock or gas reservoir rock. The well completion method refers to a communication mode of a well bore of an oil and gas well and an oil and gas layer, and comprises perforation well completion, open hole well completion, liner well completion, gravel pack well completion and the like. The drilling parameters include weight-on-bit, rate-of-penetration, drilling cycle, rotational speed, pore pressure, burst pressure, and collapse pressure, among other parameters relevant to drilling a formation into a borehole from the surface. The fractured reservoir is a reservoir with fractures as main reservoir spaces and seepage channels, can increase effective porosity, generally has high permeability, and is beneficial to calculation of the recoverable reserves of shale gas reservoirs by collecting the lengths of the fractures. Furthermore, conductivity generally refers to fracture conductivity, i.e. the product of fracture closure width and fracture permeability under closure pressure, and the magnitude of the product is mainly related to fracture closure pressure, physical properties of proppant and the laying concentration of proppant in the fracture, and is also related to factors such as reservoir rock hardness, temperature, fluid properties, saline water environment, non-darcy flow conditions, bearing time, and damage of fracturing fluid to proppant layers.
Machine learning is used for studying how a computer simulates or realizes human learning behaviors to acquire new knowledge or skills, reorganizes an existing knowledge structure to continuously improve the performance of the computer, and comprises a support vector machine model, a neural network model, a Bayesian method model, a logistic regression model and the like. And inputting the characteristic data into the trained machine learning model, and outputting a corresponding predicted value of the final recoverable reserves of the shale gas reservoir by each trained machine learning model for each group of characteristic data.
The method fully utilizes known historical data, establishes a machine learning model on the premise of not making any hypothesis, inputs characteristic data closely related to the final recoverable reserves of the shale gas reservoir into the trained machine learning model, obtains the predicted value of the final recoverable reserves of the shale gas reservoir, and ensures that the obtained predicted value accords with the actual rule of the shale gas reservoir, so that the final recoverable reserves evaluation result is real and reliable, and has higher accuracy.
Based on the foregoing embodiment, as an optional embodiment, before the step S1, the method further includes selecting the feature data: and acquiring attribute data related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir as feature data according to an attribute reduction algorithm.
Specifically, all attribute data of the shale gas reservoir are preprocessed through an attribute reduction algorithm of the fuzzy set. The basic idea of the so-called fuzzy set attribute reduction algorithm is to find important attributes from known attribute data to form the best reduced attribute combination.
For a group of shale gas reservoirs, all attribute data include formation porosity, permeability, whether a special formation structure exists, rock properties, whether a special formation is drilled during drilling, a well completion method, the quality of well completion, drilling parameters, fracture lengths of natural fractures and artificial fractures, conductivity and the like. And for all attribute data of each group of shale gas reservoir, corresponding to a decision result, dividing all known attribute data into a plurality of attribute combinations, verifying whether each attribute combination is equivalent to the decision result corresponding to the corresponding attribute data, selecting one attribute combination retaining all attribute information as much as possible, removing all attribute data with uncertain factors, influencing machine learning model establishment and hindering regression prediction, obtaining effective reduction attributes, namely attribute data closely related to estimation of the shale gas reservoir collectable amount, and using the effective reduction attributes as feature data to enhance reliability and timeliness of the prediction result.
Based on the above embodiment, as an alternative embodiment, the trained machine learning model in step S2 is built by the following steps: acquiring characteristic data of a plurality of groups of shale gas reservoirs and corresponding final recoverable reserves as sample data in a training sample set; randomly selecting a plurality of groups of sample data with specific numbers from the training sample set, wherein the sample data with each specific group number forms a training subsample set; and if the group number of the sample data in the training sample set is not more than the preset group number, inputting the sample data in each training sub-sample set into a support vector machine model for training, mapping the sample data to a higher-dimensional space through a Gaussian kernel function, and obtaining a plurality of trained support vector machine models.
Specifically, multiple groups of characteristic data of the shale gas reservoir are obtained, each group of characteristic data has corresponding recoverable reserves, the characteristic data are used as sample data in a training sample set, multiple groups of sample data with specific groups are randomly selected from the training sample set, each group of sample data with specific groups forms a training sub-sample set, and each training sub-sample set can train a machine learning model.
And when the group number of all sample data in the training sample set is less than or equal to the preset group number, the group number of the sample data at the moment is known to be less, and the sample data is input into the trained support vector machine model. It should be noted that the support vector machine model is a machine learning model that is superior in solving the problems of small samples, nonlinearity and high-dimensional pattern recognition, and an optimal compromise can be found between the complexity and learning capability of the model according to limited sample information to obtain the best popularization capability. And after a certain amount of sample information is used for training the support vector machine model, for each group of characteristic data, the support vector machine model outputs a corresponding predicted value of the shale gas reservoir recoverable reserve.
After the sample data is input into the support vector machine model, the kernel function is used for mapping the sample data to a feature space with higher dimension or infinite dimension by adopting nonlinear mapping, and then corresponding linear operation is carried out on the feature space, so that the problem of inseparability of linearity is solved, and the sample is classified. In the feature space, when linearly separable samples are classified, the inner product of the samples needs to be calculated, but when the dimension of the samples is higher, dimension disaster is easily caused, so that the inner product of high-dimensional vectors can be converted into the inner product problem of solving low-dimensional vectors by introducing a kernel function. The kernel functions comprise linear kernel functions, p-order polynomial kernel functions, Gaussian kernel functions, multilayer perceptron kernel functions and the like.
Based on the above embodiment, as an optional embodiment, the trained neural network model is built by the following steps: and if the group number of the sample data in the training sample set is greater than the preset group number, inputting the sample data in each training sub sample set into a neural network model for training, adjusting the connection weight of neurons in the neural network model according to the training result of each group of sample data in the training sub sample set until the error of the training result is less than a preset threshold value, and obtaining a plurality of trained neural network models.
Specifically, when the number of groups of the feature data is greater than the preset number of groups, it is known that the number of groups of the feature data is large at this time, and the feature data is input into the trained neural network model. The neural network model simulates a human nervous system, and trains the relationship between characteristic data and the final recoverable reserves of the shale gas reservoir through transmission among an input layer, an intermediate transmission layer and an output layer.
Firstly, random values in the (-1,1) interval are given to the connection weight of the neurons, after each group of data is input into a neural network model, input signals act on output nodes through a hidden layer, output signals are generated through nonlinear transformation, and if actual output does not accord with expected output, the process of back propagation of errors is carried out. The error back transmission is to back transmit the output error to the input layer by layer through the hidden layer, and distribute the error to all units of each layer, and the error signal obtained from each layer is used as the basis for adjusting each connection weight. And (3) reducing the error along the gradient direction by adjusting the connection strength between the input node and the hidden layer node, the connection strength between the hidden layer node and the output node and the threshold value, repeatedly learning and training until the error of the training result, namely the output error is smaller than the preset threshold value, determining the connection weight corresponding to the output error, stopping training, and obtaining the trained neural network model at the moment.
And after a certain amount of sample information is used for training the neural network model, for each group of characteristic data, the neural network model outputs a corresponding predicted value of the final recoverable reserves of the shale gas reservoir.
Based on the above embodiment, as an optional embodiment, the neural network model is a BP neural network model: the BP neural network model comprises 9 hidden layers; neurons in the hidden layer employ sigmoid transfer functions.
Specifically, the neural network model adopted by the invention is a BP neural network model, the BP neural network model is a multilayer feedforward neural network trained according to an error back propagation algorithm, the two processes of signal forward propagation and error back propagation are included, namely, the error output is calculated according to the direction from input to output, and the weight and the threshold are adjusted according to the direction from output to input. In the forward propagation process, an input mode is processed layer by layer from an input layer through a hidden layer and is transferred to an output layer, and the state of each layer of neurons only affects the state of the next layer of neurons. If the expected output can not be obtained at the output layer, the reverse propagation is carried out, the error signal is returned along the original connecting path, and the error signal can be minimized by modifying the weight of each neuron.
The BP neural network model comprises 9 hidden layers, and neurons in the hidden layers adopt sigmoid transfer functions. The function of the transfer function is to control the activation of input to output, transform the possible infinite domain to the given output, and perform function conversion on the input and output to simulate the linear or nonlinear transfer characteristic of biological nerve. The types of the current common transfer functions comprise threshold logic, linear transfer functions, linear threshold functions, sigmoid functions, hyperbolic-tangent functions and the like.
Based on the above embodiment, as an optional embodiment, the method further includes: inputting the characteristic data into a plurality of trained machine learning models, outputting a predicted value of the final recoverable reserves of a group of shale gas reservoirs by each machine learning model, and acquiring a plurality of groups of predicted values; obtaining a normal distribution diagram of the predicted values according to a plurality of groups of the predicted values, and obtaining three predicted values P10, P50 and P90 from the normal distribution diagram; where P10 indicates that the predicted value has a 10% greater likelihood than the actual final recoverable reserve; p50 indicates that the predicted value has a 50% greater likelihood than the actual final recoverable reserve; p90 indicates that the predicted value has a 90% greater likelihood than the actual final recoverable amount.
Specifically, because a plurality of machine learning models are established, and each machine learning model outputs a predicted value, for the feature data of each group of shale gas reservoir, a plurality of groups of predicted values of the final recoverable reserves of the shale gas reservoir are output. At this time, according to the obtained multiple groups of predicted values, a corresponding normal distribution graph is made, and three symbolic predicted values of P10, P50 and P90 are obtained from the normal distribution graph, wherein P10 indicates that the predicted values have the probability of 10% being larger than the actual final recoverable reserves; p50 indicates that the predicted value has a 50% greater likelihood than the actual final recoverable reserve; p90 indicates that the predicted value has 90% probability of being larger than the actual final recoverable capacity, and the three marked predicted values of P10, P50 and P90 provide an evaluation criterion for the estimation of the final recoverable capacity of the shale gas reservoir.
The method comprises the steps of analyzing and processing historical data of the shale gas reservoir, screening out characteristic data closely related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir by utilizing an attribute reduction algorithm of a rough set, inputting the characteristic data into a trained support vector machine model or a neural network model, and obtaining the predicted value of the final recoverable reserves of the shale gas reservoir. The method combines the calculation of the final recoverable reserves of the shale gas reservoir with the machine learning, fully utilizes the known historical data, establishes the machine learning model on the premise of not making any hypothesis, and ensures that the estimation result of the final recoverable reserves is real and reliable and has high accuracy because the predicted value obtained by the machine learning model conforms to the actual rule of the shale gas reservoir.
Referring to fig. 2, a schematic structural diagram of a system for calculating final recoverable reserves of a shale gas reservoir according to an embodiment of the present invention includes: the data acquisition module 201 is configured to acquire characteristic data of a shale gas reservoir, where the characteristic data includes porosity, permeability, completion method, drilling parameters, fracture length, and conductivity of the shale gas reservoir; the predicted value obtaining module 202 is configured to input the feature data into a trained machine learning model, and obtain a predicted value of a final recoverable reserves of the shale gas reservoir; wherein the drilling parameters include weight-on-bit, rate-of-penetration, drilling cycle, rotational speed, pore pressure, burst pressure, and collapse pressure.
Specifically, the data acquisition module is used for acquiring feature data of machine learning models such as a support vector machine model or a neural network model which need to be input; the predicted value obtaining module is used for inputting the characteristic data into the support vector machine model so as to obtain the predicted value of the final recoverable reserves of the shale gas reservoir corresponding to the characteristic data. It should be noted that, the specific method steps for evaluating the final recoverable reserves of the shale gas reservoir by using the support vector machine model have been described in detail in the above method embodiments, and are not described herein again.
The method fully utilizes known historical data, establishes a machine learning model on the premise of not making any hypothesis, inputs characteristic data closely related to the final recoverable reserves of the shale gas reservoir into the trained machine learning model, obtains the predicted value of the final recoverable reserves of the shale gas reservoir, and ensures that the obtained predicted value accords with the actual rule of the shale gas reservoir, so that the final recoverable reserves evaluation result is real and reliable, and has higher accuracy.
Based on the above embodiment, as an optional embodiment, the method further includes: and the characteristic selection module is used for acquiring attribute data related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir as characteristic data according to an attribute reduction algorithm.
Specifically, the feature selection module is used for preprocessing all attribute data of the shale gas reservoir according to an attribute reduction algorithm, and eliminating all attribute data which hinder regression prediction, so that effective reduction attributes, namely attribute data which are closely related to estimation of the final recoverable reserve of the shale gas reservoir, are obtained and serve as feature data, and reliability and timeliness of prediction results are enhanced. It should be noted that, the specific method steps for processing attribute data into feature data by using the attribute reduction algorithm have been described in detail in the foregoing method embodiments, and are not described herein again.
The method comprises the steps of analyzing and processing historical data of the shale gas reservoir, screening out characteristic data closely related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir by utilizing an attribute reduction algorithm of a rough set, inputting the characteristic data into a trained support vector machine model or a neural network model, and obtaining the predicted value of the final recoverable reserves of the shale gas reservoir. The method combines the calculation of the final recoverable reserves of the shale gas reservoir with the machine learning, fully utilizes the known historical data, establishes the machine learning model on the premise of not making any hypothesis, and ensures that the estimation result of the final recoverable reserves is real and reliable and has high accuracy because the predicted value obtained by the machine learning model conforms to the actual rule of the shale gas reservoir.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (4)
1. A method for calculating final recoverable reserves of a shale gas reservoir is characterized by comprising the following steps:
s1, acquiring characteristic data of the shale gas reservoir, wherein the characteristic data comprises the porosity, the permeability, the well completion method, the drilling parameters, the fracture length and the flow conductivity of the shale gas reservoir;
s2, inputting the characteristic data into a trained machine learning model, and obtaining a predicted value of the final recoverable reserves of the shale gas reservoir;
wherein the drilling parameters include weight-on-bit, rate-of-penetration, drilling cycle, rotational speed, pore pressure, burst pressure, and collapse pressure;
before step S1, the method further includes selecting the feature data:
acquiring attribute data related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir as feature data according to an attribute reduction algorithm;
the trained machine learning model in step S2 is built by:
acquiring characteristic data of a plurality of groups of shale gas reservoirs and corresponding recoverable reserves as sample data in a training sample set;
randomly selecting a plurality of groups of sample data with specific numbers from the training sample set, wherein the sample data with each specific group number forms a training subsample set;
if the group number of the sample data in the training sample set is not larger than the preset group number, inputting the sample data in each training sub-sample set into a support vector machine model for training, mapping the sample data to a higher-dimensional space through a Gaussian kernel function, and obtaining a plurality of trained support vector machine models;
and if the group number of the sample data in the training sample set is greater than the preset group number, inputting the sample data in each training sub sample set into a neural network model for training, adjusting the connection weight of neurons in the neural network model according to the training result of each group of sample data in the training sub sample set until the error of the training result is less than a preset threshold value, and obtaining a plurality of trained neural network models.
2. The method for calculating final recoverable reserves of a shale gas reservoir as claimed in claim 1, wherein the neural network model is a BP neural network model:
the BP neural network model comprises 9 hidden layers;
neurons in the hidden layer employ sigmoid transfer functions.
3. The method for calculating the final recoverable capacity of the shale gas reservoir according to any one of claims 1 to 2, wherein the step S2 specifically comprises:
inputting the characteristic data into a plurality of trained machine learning models, outputting a predicted value of the final recoverable reserves of a group of shale gas reservoirs by each machine learning model, and acquiring a plurality of groups of predicted values;
obtaining a normal distribution diagram of the predicted values according to a plurality of groups of the predicted values, and obtaining three predicted values of P10, P50 and P90 from the normal distribution diagram;
where P10 indicates that the predicted value has a 10% greater likelihood than the actual final recoverable reserve; p50 indicates that the predicted value has a 50% greater likelihood than the actual final recoverable reserve; p90 indicates that the predicted value has a 90% greater likelihood than the actual final recoverable amount.
4. A system for calculating final recoverable reserves of a shale gas reservoir, comprising:
the system comprises a data acquisition module, a data acquisition module and a data processing module, wherein the data acquisition module is used for acquiring characteristic data of the shale gas reservoir, and the characteristic data comprises the porosity, the permeability, a well completion method, drilling parameters, fracture length and flow conductivity of the shale gas reservoir;
the predicted value obtaining module is used for inputting the characteristic data into a trained machine learning model and obtaining a predicted value of the final recoverable reserves of the shale gas reservoir;
wherein the drilling parameters include weight-on-bit, rate-of-penetration, drilling cycle, rotational speed, pore pressure, burst pressure, and collapse pressure;
the characteristic selection module is used for acquiring attribute data related to the final recoverable reserves of the shale gas reservoir from all attribute data of the shale gas reservoir as characteristic data according to an attribute reduction algorithm;
the trained machine learning model is built through the following steps:
acquiring characteristic data of a plurality of groups of shale gas reservoirs and corresponding recoverable reserves as sample data in a training sample set;
randomly selecting a plurality of groups of sample data with specific numbers from the training sample set, wherein the sample data with each specific group number forms a training subsample set;
if the group number of the sample data in the training sample set is not larger than the preset group number, inputting the sample data in each training sub-sample set into a support vector machine model for training, mapping the sample data to a higher-dimensional space through a Gaussian kernel function, and obtaining a plurality of trained support vector machine models;
and if the group number of the sample data in the training sample set is greater than the preset group number, inputting the sample data in each training sub sample set into a neural network model for training, adjusting the connection weight of neurons in the neural network model according to the training result of each group of sample data in the training sub sample set until the error of the training result is less than a preset threshold value, and obtaining a plurality of trained neural network models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810191485.5A CN108573320B (en) | 2018-03-08 | 2018-03-08 | Method and system for calculating final recoverable reserves of shale gas reservoir |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810191485.5A CN108573320B (en) | 2018-03-08 | 2018-03-08 | Method and system for calculating final recoverable reserves of shale gas reservoir |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108573320A CN108573320A (en) | 2018-09-25 |
| CN108573320B true CN108573320B (en) | 2021-06-29 |
Family
ID=63576083
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810191485.5A Active CN108573320B (en) | 2018-03-08 | 2018-03-08 | Method and system for calculating final recoverable reserves of shale gas reservoir |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108573320B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109740210B (en) * | 2018-12-21 | 2020-10-13 | 中国石油大学(北京) | Shale gas fracturing underground sand blocking accident real-time risk assessment method and device |
| CN110056348B (en) * | 2019-04-25 | 2021-05-11 | 中国海洋石油集团有限公司 | Method and system for determining formation fluid composition and properties |
| CN110837913B (en) * | 2019-09-24 | 2020-10-23 | 广州地理研究所 | Method and equipment for predicting land water reserves based on extreme gradient algorithm |
| CN110952970A (en) * | 2019-12-03 | 2020-04-03 | 西南石油大学 | Combined proppant flow conductivity prediction method |
| CN113011626B (en) * | 2019-12-19 | 2024-08-02 | 北京国双科技有限公司 | Formation pressure prediction model construction method, formation pressure prediction method and device |
| CN110952978B (en) * | 2019-12-20 | 2020-11-06 | 西南石油大学 | Drilling leakage fracture width prediction method based on neural network data mining |
| CN111520123B (en) * | 2020-03-27 | 2021-03-02 | 中国石油化工股份有限公司石油工程技术研究院 | Mechanical drilling speed prediction method, device and equipment |
| CN112819035B (en) * | 2021-01-12 | 2021-11-09 | 北京科技大学 | Method and device for judging gas channeling by utilizing PVT (physical vapor transport) experiment and machine learning |
| CN113297803B (en) * | 2021-06-17 | 2022-12-06 | 东北石油大学 | Intelligent simulation and estimation method and system for oil and gas yield |
| CN113488118A (en) * | 2021-08-12 | 2021-10-08 | 氢储(上海)能源科技有限公司 | Method for predicting performance of magnesium-based solid hydrogen storage material by using BP neural network model |
| CN113758851B (en) * | 2021-09-07 | 2022-07-15 | 中国石油大学(北京) | Compact core EUR determination experimental method |
| CN113738353B (en) * | 2021-09-28 | 2023-05-26 | 中国石油大学(北京) | Method, system, equipment and storage medium for predicting movable oil quantity of oil-containing shale |
| CN114358427B (en) * | 2022-01-07 | 2024-05-07 | 西南石油大学 | Method for predicting final recoverable reserves of shale gas well |
| CN115062029B (en) * | 2022-08-12 | 2022-11-04 | 北京安捷瑞软件有限公司 | Oil and gas reservoir exploration and development characteristic data result management system based on big data |
| CN115660221B (en) * | 2022-12-07 | 2023-03-21 | 西南石油大学 | Oil and gas reservoir economic recoverable reserve assessment method and system based on hybrid neural network |
| CN116911640B (en) * | 2023-09-11 | 2023-12-26 | 中国地质大学（北京） | Shale reservoir gas content prediction method based on machine learning |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104134101A (en) * | 2014-07-23 | 2014-11-05 | 中国石油集团川庆钻探工程有限公司 | Low-permeability reservoir natural gas productivity prediction method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8510242B2 (en) * | 2007-08-31 | 2013-08-13 | Saudi Arabian Oil Company | Artificial neural network models for determining relative permeability of hydrocarbon reservoirs |
-
2018
- 2018-03-08 CN CN201810191485.5A patent/CN108573320B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104134101A (en) * | 2014-07-23 | 2014-11-05 | 中国石油集团川庆钻探工程有限公司 | Low-permeability reservoir natural gas productivity prediction method |
Non-Patent Citations (4)
| Title |
|---|
| "利用SVM 方法计算中国南方X 盆地页岩气含气量";郭怀志 等;《世界地质》;20150930;第34卷(第3期);第757-763页 * |
| "基于ＲBF 神经网络的页岩气丰度预测";杨宇宁 等;《西安石油大学学报》;20151130;第30卷(第6期);第786-790页 * |
| "基于遗传算法改进ＢＰ神经网络的";田亚鹏 等;《中国科技论文》;20160831;第11卷(第15期);第1710-1715页 * |
| "页岩油气不确定性产量预测方法及应用";徐兵祥等;《兰州大学学报》;20171231;第45-49页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108573320A (en) | 2018-09-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108573320B (en) | Method and system for calculating final recoverable reserves of shale gas reservoir | |
| CN112901137B (en) | Deep well drilling mechanical drilling speed prediction method based on deep neural network Sequential model | |
| Xue et al. | A new fracture prediction method by combining genetic algorithm with neural network in low-permeability reservoirs | |
| CN110619353A (en) | Multi-scale logging curve automatic identification method based on deep learning | |
| CN113537585B (en) | Oil field production increasing measure recommendation method based on random forest and gradient lifting decision tree | |
| CN113792479B (en) | Coal bed methane reservoir fracturing effect evaluation method based on physical constraint | |
| CN112539054B (en) | Production optimization method for complex system of ground pipe network and underground oil reservoir | |
| Khoukhi | Hybrid soft computing systems for reservoir PVT properties prediction | |
| CN106326624A (en) | Method for predicating stratum fracture pressure | |
| CN117219180B (en) | Method and system for monitoring, evaluating and dynamically regulating and enhancing mineralization effect of carbon dioxide | |
| CN116384554A (en) | Method and device for predicting mechanical drilling speed, electronic equipment and computer storage medium | |
| CN110717270B (en) | Oil and gas reservoir simulation method based on data | |
| CN116910638A (en) | Dominant seepage channel intelligent identification method based on big data | |
| Wang et al. | An interpretable interflow simulated graph neural network for reservoir connectivity analysis | |
| CN115099406A (en) | Stratum pressure inversion method and device based on multivariate time sequence | |
| CN114065909A (en) | Well logging curve completion method based on CNN _ AB _ Bi-LSTM | |
| CN110671092A (en) | Oil gas productivity detection method and system | |
| CN117574755A (en) | Hierarchical multistage optimization method for fracturing parameters of horizontal well of shale reservoir well factory | |
| CN110118994B (en) | Continental facies hydrocarbon source rock quantitative prediction method based on seismic inversion and machine learning | |
| CN117172113A (en) | Method, system, equipment and medium for predicting rotary steerable drilling well track | |
| Anifowose et al. | A functional networks-type-2 fuzzy logic hybrid model for the prediction of porosity and permeability of oil and gas reservoirs | |
| CN116433059A (en) | Intelligent evaluation method and device for shale oil dessert | |
| Schiozer et al. | Integration of history matching and uncertainty analysis | |
| CN111706323A (en) | Water flooded layer fine interpretation and evaluation method based on GWO-LSSVM algorithm | |
| CN117633658B (en) | Rock reservoir lithology identification method and system |
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 |