Coal-bed gas well productivity evaluation method based on artificial intelligence

Abstract

The invention discloses a coal-bed gas well productivity evaluation method based on artificial intelligence, and the method comprises the steps: collecting the production data of an existing coal-bed gas well, carrying out the preprocessing of the production data and basic data, inputting the basic data into a coal-bed gas well production condition prediction model, and obtaining the predicted production data, comparing the predicted production data with actual production data to obtain an actual-to-actual proportion, performing influence analysis on the basic data according to a change correlation degree to obtain a predicted change coefficient, and constructing a production capacity function according to the actual-to-actual proportion and the predicted change coefficient, and constructing a coal-bed gas well productivity prediction model according to the production capacity function, optimizing the coal-bed gas well productivity evaluation model, inputting to-be-evaluated data into the coal-bed gas well productivity evaluation model, and outputting an evaluation result. The method not only can improve the precision of coal-bed gas well productivity evaluation based on artificial intelligence, but also has good interpretability, and can be directly applied to coal-bed gas well productivity prediction of field production.

Description

Technical Field

[0001] The present invention relates to the field of evaluation, and in particular to a coalbed methane well productivity evaluation method based on artificial intelligence. Background Art

[0002] With the adjustment of global energy structure and the improvement of environmental protection awareness, the development and utilization of coalbed methane as a clean and efficient energy source has received more and more attention. However, the production capacity of coalbed methane wells is affected by factors such as complex and changeable geological conditions and different drilling conditions. Traditional evaluation methods often rely on existing data in the mining area and are not accurate enough in predicting unmined areas, which makes it difficult to meet the needs of modern coalbed methane exploration and development.

[0003] Through machine learning, deep learning and other technologies, it is possible to automatically process and analyze a large amount of geological data, and improve the accuracy and efficiency of the evaluation. At present, although some geological evaluation methods based on artificial intelligence have been proposed, their application in the evaluation of coalbed methane well productivity still has some limitations and uncertainties. Therefore, it is necessary to conduct in-depth research on the evaluation method of coalbed methane well productivity based on artificial intelligence and optimize the current evaluation method. Summary of the invention

[0004] The purpose of the present invention is to provide a coalbed methane well productivity evaluation method based on artificial intelligence.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] The present invention comprises the following steps:

[0007] Collect production data and basic data of existing coalbed methane wells, and pre-process the production data and basic data; the basic data include the average thickness, occurrence, burial depth, porosity, permeability, gas content, caprock thickness, lithology, development well type, fracturing conditions, etc. of coal seams; the production data include actual annual gas production, maximum daily gas production, 20-year gas production, etc.;

[0008] Inputting the production data into a coalbed methane production prediction model to obtain coalbed methane well predicted production data, and comparing the predicted production data with actual production data to obtain an actual-prediction ratio;

[0009] Performing an impact analysis on the basic data according to the change correlation to obtain a production data change coefficient, and constructing a production capacity function according to the actual-to-predict ratio and the production data change coefficient;

[0010] Constructing a coalbed methane well productivity evaluation model according to the production capacity function, and optimizing the coalbed methane well productivity evaluation model;

[0011] The data to be evaluated is input into the coalbed methane well productivity evaluation model, and the evaluation result is output.

[0012] Furthermore, the method of inputting the basic data and production data into the coalbed methane well productivity evaluation model to obtain predicted production data includes:

[0013] The coalbed methane well productivity prediction model is based on the prediction model of time series analysis algorithm and BP neural network algorithm;

[0014] The decision tree is used to extract the characteristics of production data, obtain the production capacity characteristics, and input the production capacity characteristics into the gas layer production capacity prediction model;

[0015] The time series analysis algorithm learns the time change law of production characteristics, simulates the time change law to predict production data and obtains the first prediction value; the BP neural network algorithm uses back propagation to supervise the production capacity characteristics and obtain the second prediction value; the Gaussian mixture is used to fuse the first prediction value and the second prediction value to obtain the predicted production data.

[0016] Furthermore, a method for performing an impact analysis on the impact data according to the change correlation to obtain a coalbed methane well production change coefficient includes:

[0017] Calculate the sensitivity of the changes affecting the data:

[0018]

[0019] The impact data The sensitivity of The fuzzy coefficient is The fuzzy neighborhood particle that affects the data attribute τ is Affects the data sensitive attribute set The fuzzy neighborhood particle is Impact data The adjustment constant is q;

[0020] Taking the influencing data with sensitivity greater than 0 as the key data, a fuzzy similarity relationship matrix is ​​established, and the expression is:

[0021]

[0022] The key data in row v and column j is r zj , the key data center of the p-th feature in the v-th row is The number of features is N, and the genetic factor in the zth row and jth column is The change sensitivity of the key data in row v and column j is

[0023] To solve the fuzzy equivalent matrix of the fuzzy similarity relationship matrix, the expression is:

[0024] W * ={h vj *}

[0025] The fuzzy equivalence matrix is ​​W * , the fuzzy equivalent element is h vj * , the fuzzy equivalent elements are ranked in descending order, and the key data whose elements in the matrix are 1 are taken as category 1;

[0026] Calculate the change correlation of key data in category 1:

[0027]

[0028] The relevant data are Related data The number of related data is n. The correlation degree with the production data change w is Related data The control factor is The correlation constant is L;

[0029] According to the change correlation, the relationship between the basic data and production data of coalbed methane wells is captured through machine learning algorithm to obtain the production data change coefficient.

[0030] Furthermore, the method of constructing a production capacity function according to the actual-to-predicted ratio and the production data variation coefficient includes:

[0031]

[0032] The basic parameters of the i-th coalbed methane well are: The actual prediction ratio of the i-th coalbed methane well is The coefficient of variation of production data of the i-th coalbed methane well is The i+1th The satisfactory deviation value is The δ-1th The satisfactory deviation value is No. The maximum satisfactory deviation is No. The minimum satisfactory deviation is The contribution weight of the basic parameters of the i-th coalbed methane well is θ i , the mass multiplier is The proportional control coefficient is The constant of change is

[0033] Furthermore, the method for constructing a coalbed methane well productivity evaluation model using the production capacity function includes:

[0034] The objective function is constructed based on the production capacity function, and the expression is:

[0035]

[0036] The predicted production capacity of the i-th coalbed methane well is The first linear weight is ζ1, the second linear weight is ζ2, and the i-th production capacity evaluation factor is g i , production capacity evaluation factor g i The maximum information coefficient is f i , the L2 norm function is ||·||2, production capacity and predicting production capacity The loss function is

[0037] The coalbed methane well productivity evaluation model includes random forest algorithm, fuzzy C-means clustering algorithm, sensitivity screening algorithm, neural network fusion algorithm, and recurrent neural network algorithm;

[0038] The random forest algorithm divides the data to be evaluated into a training set and a test set in a ratio of 4:1;

[0039] The fuzzy C-means clustering algorithm calculates the membership of the training set data, clusters the training set data according to the membership, and obtains the classified data;

[0040] The sensitivity screening algorithm calculates the sensitivity of the categorical data and outputs the categorical data with a sensitivity greater than 0.416 as important factors;

[0041] The neural network fusion algorithm fuses the important factors of the same cluster through the connection between neurons to obtain the cluster index;

[0042] The recurrent neural network algorithm compares the cluster index and the actual evaluation result, obtains the corresponding relationship between the cluster index and the actual evaluation result, and predicts the evaluation value according to the corresponding relationship to obtain the predicted evaluation value;

[0043] Calculate the mean square error between the predicted evaluation value and the actual evaluation value of the test set data, and adjust the hyperparameters of the recurrent neural network algorithm based on the mean square error.

[0044] Furthermore, the method for optimizing the coalbed methane well productivity evaluation model includes:

[0045] Create an experience pool, collect actions based on the policy network, and calculate the probability of the action occurring:

[0046]

[0047] The probability of action c occurring is The first network weight parameter is ψ 1 , the state is y, the action c is in state y and the first network weight parameter ψ 1 The policy network under 1 ), the logarithmic function with base 2 is log2(·), and the natural constant is e;

[0048] Execute the action, get the new state and immediate reward. If the new state exceeds the boundary, the task is terminated and the experience data is stored in the experience pool.

[0049] If the task termination is 1, all data is obtained from the experience pool, input into the evaluation network to obtain the state value function, and calculate the advantage function:

[0050]

[0051] The time step of the tth iteration is l t , the second network weight parameter of the t+1th iteration is The second network weight parameter for the tth iteration is The time step of the t+1th iteration is l t+1 , the state value of the second network weight parameter in the tth iteration state y is The state value of the second network weight parameter in the t+1th iteration state y is The decay rate is η, and the maximum number of iterations is t max , the reward function for the tth iteration is s t , t max The step size of the -1 iteration is No. max The decay rate of -t+1 iterations is

[0052] Calculate the value:

[0053]

[0054] The value of the tth iteration is The task progress value is The value of task completion is The task penalty value is The decay rate of the at-th evaluation is η a-t , the control factor of the ath evaluation is b a ;

[0055] Sample small batches of data, calculate the probability of the action occurring in the time step, and calculate the objective function value:

[0056]

[0057] The entropy function is B(·), the entropy weight coefficient is ρ, the clipping factor is x, and the evaluation error of the tth iteration is θ t , the number of evaluation indicators is u, and the first network weight parameter of the tth iteration is The round number function of the tth iteration is Z t [·];

[0058] Update the first network weight parameter, the expression is:

[0059]

[0060] The updated first network weight parameter is The first network weight parameter ψ 1 The differential of

[0061] Calculate the loss function:

[0062]

[0063] The loss function is θ t , the loss constant is υ;

[0064] Update the second network weight parameter, the expression is:

[0065]

[0066] The updated weight parameter of the second network is The second network weight parameter ψ 2 The differential of

[0067] Continue iterating until the maximum number of iterations is reached.

[0068] In a second aspect, an embodiment of the present application further provides an electronic device, including:

[0069] A processor; and a memory arranged to store computer executable instructions, which when executed cause the processor to perform the method steps described in the first aspect.

[0070] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0071] The beneficial effects of the present invention are:

[0072] The present invention is a coalbed methane well productivity evaluation method based on artificial intelligence. Compared with the prior art, the present invention has the following technical effects:

[0073] The present invention can improve the accuracy of coalbed methane well productivity evaluation based on artificial intelligence through preprocessing, predicting production data, obtaining the actual-prediction ratio, obtaining the production capacity variation coefficient, constructing the production capacity function, building the model and the model optimization steps, thereby improving the accuracy of coalbed methane well productivity evaluation based on artificial intelligence. The optimization of coalbed methane well productivity evaluation based on artificial intelligence can greatly save resources and improve work efficiency. It can realize the intelligent evaluation of coalbed methane well productivity and automatically correct the coalbed methane well productivity evaluation in real time, which is of great significance to coalbed methane development and can adapt to the coalbed methane well productivity evaluation of various well types and different development blocks, and has a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 This is a flow chart of the steps of a coalbed methane well productivity evaluation method based on artificial intelligence of the present invention;

[0075] Figure 2 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0076] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0077] The present invention provides a coalbed methane well productivity evaluation method based on artificial intelligence, comprising the following steps:

[0078] like Figure 1 As shown, in this embodiment, the following steps are included:

[0079] Collect basic data and production data of coalbed methane wells, and pre-process the basic data and production data; the basic data include average thickness, occurrence, burial depth, porosity, permeability, gas content, caprock thickness, lithology, development well type, fracturing conditions, etc. of coal seams; the production data include actual annual gas production, maximum daily gas production, 20-year EUR, etc.;

[0080] In the actual evaluation, the basic parameters of gas production of XX coalbed methane well are: coal seam thickness 10m, nearly horizontal occurrence, burial depth 500m, porosity 10.2%, permeability 0.5 millidarcy, gas content 15 cubic meters / ton, cap rock thickness 30m, cap rock lithology mudstone, good sealing, mining method L-type staged fracturing horizontal well, hydraulic fracturing; coal metamorphism degree: anthracite, vitrinite content 75.1%, ash content 15%, cleat fracture development density medium, pressure 20MPa, coal seam floor lithology sandstone, faults are not developed;

[0081] Production data: average daily gas production 6,000 cubic meters, EUR 22 million cubic meters in 20 years;

[0082] Inputting the basic parameters into a gas layer well productivity prediction model to obtain predicted production data, and comparing the predicted production data with actual production data to obtain an actual-prediction ratio;

[0083] In the actual assessment, the average daily gas production was 6055 cubic meters, the actual EUR for 20 years was 22.2 million cubic meters, and the actual-to-prediction ratio was 0.991;

[0084] Performing an impact analysis on the impact data according to the change correlation to obtain a production data change coefficient, and constructing a production capacity function according to the actual-prediction ratio and the production data change coefficient;

[0085] In the actual evaluation, the coefficient of variation of production data was 0.013;

[0086] Constructing a coalbed methane well productivity evaluation model according to the production capacity function, and optimizing the coalbed methane well productivity evaluation model;

[0087] Inputting the data to be evaluated into the coalbed methane well productivity evaluation model and outputting the evaluation results;

[0088] In the actual evaluation, the basic parameters of the data to be evaluated are coal seam thickness of 8m, burial depth of 800m, porosity of 8.02%, permeability of 0.3 mD, gas content of 12 cubic meters / ton, shale lithology of coal seam roof, mudstone lithology of floor, coal metamorphism degree of bituminous coal, vitrinite content of 59.91%, ash content of 20%, low density of cleat fracture development, temperature of 30℃, pressure of 25MPa, effective burial depth of coal seam of 781m, effective thickness of 6m, and directional well development method; the predicted production results are: average daily gas production of 1,856 cubic meters and EUR4.85 million in 20 years.

[0089] In this embodiment, the method of inputting the basic data and production data of the coalbed methane well into the coalbed methane well productivity prediction model to obtain the predicted coalbed methane well productivity includes:

[0090] The gas layer well productivity prediction model is based on the prediction model of time series analysis algorithm and BP neural network algorithm;

[0091] A decision tree is used to extract the characteristics of basic parameters and production data of coalbed methane wells, obtain production capacity characteristics, and input the production capacity characteristics into the coalbed methane well capacity prediction model;

[0092] The time series analysis algorithm learns the time change law of production data characteristics, simulates the time change law to predict the production data and obtains the first prediction value; the BP neural network algorithm uses back propagation to supervise the production capacity characteristics and obtain the second prediction value; the Gaussian mixture is used to fuse the first prediction value and the second prediction value to obtain the predicted production data.

[0093] In this embodiment, the method of performing impact analysis on the impact data according to the change correlation to obtain the production change coefficient includes:

[0094] Calculate the sensitivity of changes affecting the data:

[0095]

[0096] The basic data The sensitivity of the change is The fuzzy coefficient is The fuzzy neighborhood particle of the basic data attribute τ is Basic data sensitive attribute set The fuzzy neighborhood particle is Basic data The adjustment constant is q;

[0097] The basic data with change sensitivity greater than 0 is used as the key data to establish the fuzzy similarity relationship matrix, which is expressed as follows:

[0098]

[0099] The key data in row v and column j is r zj , the key data center of the p-th feature in the v-th row is The number of features is N, and the genetic factor in the zth row and jth column is The change sensitivity of the key data in row v and column j is

[0100] To solve the fuzzy equivalent matrix of the fuzzy similarity relationship matrix, the expression is:

[0101] W * ={h vj *}

[0102] The fuzzy equivalence matrix is ​​W * , the fuzzy equivalent element is h vj * , the fuzzy equivalent elements are ranked in descending order, and the key data whose elements in the matrix are 1 are taken as category 1;

[0103] Calculate the change correlation of key data in category 1:

[0104]

[0105] The relevant data are Related data The number of related data is n. The correlation degree with the production data change w is Related data The control factor is The correlation constant is L;

[0106] According to the change correlation, the relationship between the basic data and production data of coalbed methane wells is captured through machine learning algorithm to obtain the production data change coefficient.

[0107] In this embodiment, the method for constructing a production capacity function according to the actual-to-predict ratio and the production data variation coefficient includes:

[0108]

[0109] The basic parameters of the i-th coalbed methane well are: The actual prediction ratio of the i-th coalbed methane well is The coefficient of variation of production data of the i-th coalbed methane well is The i+1th The satisfactory deviation value is The δ-1th The satisfactory deviation value is No. The maximum satisfactory deviation is No. The minimum satisfactory deviation is The contribution weight of the basic parameters of the i-th coalbed methane well is θ i , the mass multiplier is The proportional control coefficient is The constant of change is

[0110] In this embodiment, the method for constructing a coalbed methane well productivity evaluation model according to the production capacity function includes:

[0111] The objective function is constructed based on the production capacity function, and the expression is:

[0112]

[0113] The predicted production capacity of the i-th coalbed methane well is The first linear weight is ζ1, the second linear weight is ζ2, and the i-th production capacity evaluation factor is g i , production capacity evaluation factor g i The maximum information coefficient is f i , the L2 norm function is ||·||2, production capacity and predicting production capacity The loss function is

[0114] The coalbed methane well productivity evaluation model includes random forest algorithm, fuzzy C-means clustering algorithm, sensitivity screening algorithm, neural network fusion algorithm, and recurrent neural network algorithm;

[0115] The random forest algorithm divides the data to be evaluated into a training set and a test set in a ratio of 4:1;

[0116] The fuzzy C-means clustering algorithm calculates the membership of the training set data, clusters the training set data according to the membership, and obtains the classified data;

[0117] The sensitivity screening algorithm calculates the sensitivity of the categorical data and outputs the categorical data with a sensitivity greater than 0.416 as important factors;

[0118] The neural network fusion algorithm fuses the important factors of the same cluster through the connection between neurons to obtain the cluster index;

[0119] The recurrent neural network algorithm compares the cluster index and the actual evaluation result, obtains the corresponding relationship between the cluster index and the actual evaluation result, and predicts the evaluation value according to the corresponding relationship to obtain the predicted evaluation value;

[0120] Calculate the mean square error between the predicted evaluation value and the actual evaluation value of the test set data, and adjust the hyperparameters of the recurrent neural network algorithm based on the mean square error.

[0121] In this embodiment, the method for optimizing the coalbed methane well productivity evaluation model includes:

[0122] Create an experience pool, collect actions based on the policy network, and calculate the probability of the action occurring:

[0123]

[0124] The probability of action c occurring is The first network weight parameter is ψ 1 , the state is y, the action c is in state y and the first network weight parameter ψ 1 The policy network under 1 ), the logarithmic function with base 2 is log2(·), and the natural constant is e;

[0125] Execute the action, get the new state and immediate reward. If the new state exceeds the boundary, the task is terminated and the experience data is stored in the experience pool.

[0126] If the task termination is 1, all data is obtained from the experience pool, input into the evaluation network to obtain the state value function, and calculate the advantage function:

[0127]

[0128] The time step of the tth iteration is l t , the second network weight parameter of the t+1th iteration is The second network weight parameter for the tth iteration is The time step of the t+1th iteration is l t+1 , the state value of the second network weight parameter in the tth iteration state y is The state value of the second network weight parameter in the t+1th iteration state y is The decay rate is η, and the maximum number of iterations is t max , the reward function for the tth iteration is s t , t max The step size of the -1 iteration is No. max The decay rate of -t+1 iterations is

[0129] Calculate the value:

[0130]

[0131] The value of the tth iteration is The task progress value is The value of task completion is The task penalty value is The decay rate of the at-th evaluation is η a-t , the control factor of the ath evaluation is b a ;

[0132] Sample small batches of data, calculate the probability of the action occurring in the time step, and calculate the objective function value:

[0133]

[0134] The entropy function is B(·), the entropy weight coefficient is ρ, the clipping factor is x, and the evaluation error of the tth iteration is θ t , the number of evaluation indicators is u, and the first network weight parameter of the tth iteration is The round number function of the tth iteration is Z t [·];

[0135] Update the first network weight parameter, the expression is:

[0136]

[0137] The updated first network weight parameter is The first network weight parameter ψ1 The differential of

[0138] Calculate the loss function:

[0139]

[0140] The loss function is θ t , the loss constant is υ;

[0141] Update the second network weight parameter, the expression is:

[0142]

[0143] The updated weight parameter of the second network is The second network weight parameter ψ 2 The differential of

[0144] Continue iterating until the maximum number of iterations is reached.

[0145] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0146] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0147] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0148] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a coalbed methane well productivity evaluation device based on artificial intelligence at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the aforementioned coalbed methane well productivity evaluation methods based on artificial intelligence.

[0149] The above application Figure 1 The embodiment shown discloses a method for evaluating the productivity of coalbed methane wells based on artificial intelligence, which can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or instructions in the form of software. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiment of the present application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0150] The electronic device may also perform Figure 1 A coalbed methane well productivity evaluation method based on artificial intelligence is proposed and implemented Figure 1 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0151] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple application programs, any one of the aforementioned artificial intelligence-based coalbed methane well productivity evaluation methods is executed.

[0152] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0153] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0154] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0155] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0156] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0157] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0158] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0159] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0160] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0161] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A coalbed methane well productivity evaluation method based on artificial intelligence, characterized in that: The following steps are involved: Collect production data and basic data of developed coalbed methane wells, and pre-process the production data and basic data; the basic data include the average thickness, occurrence, burial depth, porosity, permeability, gas content, caprock thickness, lithology, development well type, fracturing conditions, etc. of coal seams; the production data include actual annual gas production, maximum daily gas production, 20-year gas production, etc.; Inputting the production data into a coalbed methane production prediction model to obtain coalbed methane well predicted production data, and comparing the predicted production data with actual production data to obtain an actual-prediction ratio; Performing an impact analysis on the basic data according to the change correlation to obtain a production data change coefficient, and constructing a production capacity function according to the actual-to-predict ratio and the production data change coefficient; Constructing a coalbed methane well productivity evaluation model according to the production capacity function, and optimizing the coalbed methane well productivity evaluation model; The data to be evaluated is input into the coalbed methane well productivity evaluation model, and the evaluation result is output.

2. According to the method for evaluating coalbed methane well productivity based on artificial intelligence in claim 1, it is characterized in that: The method of inputting the basic data and production data into a coalbed methane well productivity evaluation model to obtain predicted production data comprises: The gas layer well productivity evaluation model is a prediction model based on the time series analysis algorithm and the BP neural network algorithm; The decision tree is used to extract the characteristics of production data, obtain the production capacity characteristics, and input the production capacity characteristics into the gas layer production capacity prediction model; The time series analysis algorithm learns the time change law of production characteristics, simulates the time change law to predict production data and obtains the first prediction value; the BP neural network algorithm uses back propagation to supervise the production capacity characteristics and obtain the second prediction value; the Gaussian mixture is used to fuse the first prediction value and the second prediction value to obtain the predicted production data.

3. The method for evaluating coalbed methane well productivity based on artificial intelligence according to claim 1, characterized in that: The method of performing impact analysis on the impact data according to the change correlation to obtain the production change coefficient includes: Calculate the sensitivity of changes affecting the data: The impact data The sensitivity of the change is The fuzzy coefficient is The fuzzy neighborhood particle that affects the data attribute τ is Affects the data sensitive attribute set The fuzzy neighborhood particle is Impact data The adjustment constant is q; Taking the influencing data with change sensitivity greater than 0 as the key data, a fuzzy similarity relationship matrix is ​​established, and the expression is: The key data in row v and column j is r zj , the key data center of the p-th feature in the v-th row is The number of features is N, and the genetic factor in the zth row and jth column is The change sensitivity of the key data in row v and column j is To solve the fuzzy equivalent matrix of the fuzzy similarity relationship matrix, the expression is: W * ={h vj * } The fuzzy equivalence matrix is ​​W * , the fuzzy equivalent element is h vj * , the fuzzy equivalent elements are ranked in descending order, and the key data whose elements in the matrix are 1 are taken as category 1; Calculate the change correlation of key data in category 1: The relevant data are Related data The number of related data is n. The correlation degree with the production data change w is Related data The control factor is The correlation constant is L; According to the change correlation, the relationship between the basic data and production data of coalbed methane wells is captured through machine learning algorithm to obtain the production data change coefficient.

4. The method for evaluating coalbed methane well productivity based on artificial intelligence according to claim 1, characterized in that: The method for constructing a production capacity function according to the prediction-actual ratio and the production data variation coefficient comprises: The basic parameters of the i-th coalbed methane well are: The actual prediction ratio of the i-th coalbed methane well is The coefficient of variation of production data of the i-th coalbed methane well is The i+1th The satisfactory deviation value is The δ-1th The satisfactory deviation value is No. The maximum satisfactory deviation is No. The minimum satisfactory deviation is The contribution weight of the basic parameters of the i-th coalbed methane well is The mass multiplier is The proportional control coefficient is The constant of change is 5. The method for evaluating coalbed methane well productivity based on artificial intelligence according to claim 1, characterized in that: The method for constructing a coalbed methane well productivity evaluation model according to the production capacity function comprises: The objective function is constructed based on the production capacity function, and the expression is: The predicted production capacity of the i-th coalbed methane well is The first linear weight is ζ1, the second linear weight is ζ2, and the i-th production capacity evaluation factor is g i , production capacity evaluation factor g i The maximum information coefficient is f i , the L2 norm function is ||·||2, production capacity and predicting production capacity The loss function is The coalbed methane well productivity evaluation model includes random forest algorithm, fuzzy C-means clustering algorithm, sensitivity screening algorithm, neural network fusion algorithm, and recurrent neural network algorithm; The random forest algorithm divides the data to be evaluated into a training set and a test set in a ratio of 4:1; The fuzzy C-means clustering algorithm calculates the membership of the training set data, clusters the training set data according to the membership, and obtains the classified data; The sensitivity screening algorithm calculates the sensitivity of the categorical data and outputs the categorical data with a sensitivity greater than 0.416 as important factors; The neural network fusion algorithm fuses the important factors of the same cluster through the connection between neurons to obtain the cluster index; The recurrent neural network algorithm compares the cluster index and the actual evaluation result, obtains the corresponding relationship between the cluster index and the actual evaluation result, and predicts the evaluation value according to the corresponding relationship to obtain the predicted evaluation value; Calculate the mean square error between the predicted evaluation value and the actual evaluation value of the test set data, and adjust the hyperparameters of the recurrent neural network algorithm based on the mean square error.

6. The method for evaluating coalbed methane well productivity based on artificial intelligence according to claim 1, characterized in that: The method for optimizing the coalbed methane well productivity evaluation model comprises: Create an experience pool, collect actions based on the policy network, and calculate the probability of the action occurring: The probability of action c occurring is The first network weight parameter is ψ 1 , the state is y, the action c is in state y and the first network weight parameter ψ 1 The policy network under 1 ), the logarithmic function with base 2 is log2(·), and the natural constant is e; Execute the action, get the new state and immediate reward. If the new state exceeds the boundary, the task is terminated and the experience data is stored in the experience pool. If the task termination is 1, all data is obtained from the experience pool, input into the evaluation network to obtain the state value function, and calculate the advantage function: The time step of the tth iteration is The second network weight parameter for the t+1th iteration is The second network weight parameter for the tth iteration is The time step of the t+1th iteration is The state value of the second network weight parameter in the tth iteration state y is The state value of the second network weight parameter in the t+1th iteration state y is The decay rate is η, and the maximum number of iterations is t max , the reward function for the tth iteration is s t , t max The step size of the -1 iteration is No. max The decay rate of -t+1 iterations is Calculate the value: The value of the tth iteration is The task progress value is The value of task completion is The task penalty value is The decay rate of the at-th evaluation is η a-t , the control factor of the ath evaluation is b a ; Sample small batches of data, calculate the probability of the action occurring in the time step, and calculate the objective function value: The entropy function is B(·), the entropy weight coefficient is ρ, the clipping factor is x, and the evaluation error of the tth iteration is The number of evaluation indicators is u, and the first network weight parameter of the tth iteration is The round number function of the tth iteration is Z t [·]; Update the first network weight parameter, the expression is: The updated first network weight parameter is The first network weight parameter ψ 1 The differential of Calculate the loss function: The loss function is The loss constant is υ; Update the second network weight parameter, the expression is: The updated weight parameter of the second network is The second network weight parameter ψ 2 The differential of Continue iterating until the maximum number of iterations is reached.

7. An electronic device comprising: processor; as well as A memory arranged to store computer executable instructions, wherein when the executable instructions are executed, the processor executes the method according to any one of claims 1 to 6.

8. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, enables the electronic device to execute the method according to any one of claims 1 to 6.

Images