Reservoir parameter prediction method and device based on logging information and storage medium

Abstract

The invention discloses a reservoir parameter prediction method and device based on logging information and a storage medium, and relates to the technical field of petroleum earth exploration, and the key points of the technical scheme are that the method comprises the following steps: obtaining existing reservoir parameters and logging curve data of a work area; constructing a first sample set according to the existing reservoir parameters and the logging curve data, and performing normalization processing on the first sample set; constructing a convolutional neural network model, and training the convolutional neural network model to obtain an optimal intelligent model; constructing a second sample set for the unknown well, and predicting reservoir parameters of the unknown well through the intelligent model based on the second sample set; and performing interpolation and reverse normalization processing on the unknown well reservoir parameters, and outputting the processed unknown well reservoir parameters. By means of the prediction method, logging longitudinal correlation information can be reserved, the calculation error evaluation problem caused by the fact that the core reservoir parameters do not correspond to the logging curve depth is solved, and therefore the calculation requirement of the reservoir parameters in logging reservoir evaluation is met.

Description

Technical Field

[0001] The present invention relates to the technical field of petroleum geoexploration, and in particular to a reservoir parameter prediction method, device and storage medium based on well logging data. Background Art

[0002] This section is intended to provide a background or context to the embodiments recited in the claims. No admission is made that the description herein is prior art by inclusion in this section.

[0003] With the great success of artificial intelligence in the fields of image recognition and natural language processing, some scholars have gradually introduced artificial intelligence technology into the field of logging reservoir evaluation to perform quantitative calculation of reservoir parameters based on logging data. In the prior art, for the prediction of reservoir parameters, artificial intelligence reservoir parameter prediction based on logging curves usually adopts point samples, that is, one sample per depth point, and interpolates the logging curve with the depth of the reservoir parameter (as the sample set label) obtained by core analysis as the index. This method makes the input feature (logging curve) correspond to the output feature (label, reservoir parameter) at each sample depth point. When artificial intelligence modeling is performed, each sample is predicted, and then compared with the label, the training cost is calculated according to the loss function, and the weight is optimized. This method meets the general needs of artificial intelligence reservoir parameter modeling.

[0004] However, core analysis test data are non-continuous. This method of generating samples abandons the vertical correlation information between logging data. According to past studies, the vertical correlation of logging data contains important geological information such as sedimentary facies. This information may be helpful in obtaining more accurate reservoir parameters, affecting the accuracy of reservoir parameter calculation. It cannot solve the problem of calculation error evaluation caused by the mismatch between core reservoir parameters and logging curve depth, and cannot meet the needs of artificial intelligence modeling to calculate reservoir parameters. Summary of the invention

[0005] In response to the above technical problems, the present invention proposes a reservoir parameter prediction method, device and storage medium based on well logging data, which can retain the vertical correlation information of well logging and solve the calculation error evaluation problem caused by the mismatch between core reservoir parameters and logging curve depth. It can meet the needs of reservoir parameter calculation in well logging reservoir evaluation and improve the accuracy of predicted reservoir friction numbers.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention includes four aspects.

[0007] In a first aspect, a reservoir parameter prediction method based on well logging data is provided, comprising the following steps:

[0008] Based on the logging data of multiple wells in the work area, the existing reservoir parameters and logging curve data of the work area are obtained;

[0009] Constructing a first sample set according to existing reservoir parameters and logging curve data, and performing normalization processing on the first sample set;

[0010] Constructing a convolutional neural network model, and using the first sample set to train the convolutional neural network model to obtain an optimal intelligent model;

[0011] constructing a second sample set for the unknown well, and predicting reservoir parameters of the unknown well through an intelligent model based on the second sample set;

[0012] The unknown well reservoir parameters are interpolated and denormalized, and the processed unknown well reservoir parameters are output.

[0013] In some embodiments, the existing reservoir parameters include well number, core sampling depth, and core reservoir parameters obtained by lithology analysis; the first sample set is constructed based on the existing reservoir parameters and logging curve data, including:

[0014] Extract the logging curve data corresponding to the well number of the core reservoir parameter;

[0015] Determine the logging curve data of several logging sampling points before and after the core sample depth from the extracted logging curve data;

[0016] Each core reservoir parameter is associated with the well logging curve data of several well logging sampling points to form a first sample set.

[0017] In some embodiments, the well logging data of each sample is normalized and the reservoir parameters of each sample are normalized, including:

[0018] Performing linear normalization processing on the well logging curve and reservoir parameters using a linear scale using a preset first formula;

[0019] Performing logarithmic normalization processing on the well logging curve and reservoir parameters using a logarithmic scale using a preset second formula;

[0020] Among them, the first formula is:

[0021]

[0022] The second formula is:

[0023]

[0024] In the formula, gvalue i is a well logging curve corresponding to the i-th depth point after normalization, value iis the logging curve corresponding to the i-th depth point, LSCA and RSCA are the default left and right scales of the curve in the work area.

[0025] In some embodiments, the constructing of the convolutional neural network model and training the convolutional neural network model using the first sample set to obtain an optimal intelligent model includes:

[0026] The number of logging curves is set as the input attribute dimension and channel number of the convolutional neural network model, and the convolution direction is set as the depth direction, and the output attribute is set as the reservoir parameter to be predicted;

[0027] The plurality of samples in the first sample set are respectively set as a training sample set and a test sample set;

[0028] The convolutional neural network model is trained using the training sample set, and its effect is evaluated in real time using the test sample set, so that the training cost and test error are reduced to a minimum and the optimal intelligent model is obtained.

[0029] In some embodiments, constructing a second sample set for the unknown well, and predicting the reservoir parameters of the unknown well by using an intelligent model based on the second sample set includes:

[0030] Obtain logging curve data for testing multiple logging sampling points in unknown wells;

[0031] Combining the well logging curve data of each well logging sampling point of the unknown well with a number of well logging sampling points thereafter to construct a second sample;

[0032] Circularly construct a plurality of second samples according to the number of well logging sampling depths in the well logging curve data to obtain a second sample set;

[0033] Normalizing the second sample set to obtain a processed second sample set;

[0034] The processed second sample set is input into the intelligent model to predict the reservoir parameters corresponding to each second sample, and the reservoir parameters of the unknown well corresponding to each depth are obtained.

[0035] In some embodiments, the process of normalizing the second sample set to obtain the processed second sample set is performed by using a method of normalizing the first sample set.

[0036] In some embodiments, the interpolation and denormalization of the unknown well reservoir parameters and outputting the processed unknown well reservoir parameters include:

[0037] Interpolating the unknown well reservoir parameters using the preset third formula and fourth formula to obtain the unknown well reservoir parameters after difference;

[0038] The unknown well reservoir parameters after difference are denormalized using the preset fifth and sixth formulas to obtain the unknown well reservoir parameters consistent with the logging curve depth, which are output together with the logging curve.

[0039] In a second aspect, a reservoir parameter prediction device is provided, comprising:

[0040] The data collection module is used to obtain the existing reservoir parameters and logging curve data of the work area based on the logging data of multiple wells in the work area;

[0041] A data processing module, used to construct a first sample set according to existing reservoir parameters and logging curve data, and perform normalization processing on the first sample set;

[0042] A model processing module, used for constructing a convolutional neural network model, and training the convolutional neural network model using the first sample set to obtain an optimal intelligent model;

[0043] A parameter prediction module, used for constructing a second sample set for the unknown well, and predicting the reservoir parameters of the unknown well through an intelligent model based on the second sample set;

[0044] The prediction data processing module is used to interpolate and denormalize the unknown well reservoir parameters, and output the processed unknown well reservoir parameters.

[0045] In a third aspect, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the reservoir parameter prediction method as described above is performed.

[0046] In a fourth aspect, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and when the storage medium is executed by one or more processors, the reservoir parameter prediction method as described above is implemented.

[0047] Compared with the prior art, one or more embodiments of the above scheme may have the following advantages or beneficial effects:

[0048] The present application provides a reservoir parameter prediction method, device and storage medium based on logging data. The prediction method includes the following steps: obtaining the existing reservoir parameters and logging curve data of the work area based on the logging data of multiple wells in the work area; constructing a first sample set according to the existing reservoir parameters and logging curve data, and normalizing the first sample set; constructing a convolutional neural network model, and using the first sample set to train the convolutional neural network model to obtain the optimal intelligent model; constructing a second sample set for unknown wells, and predicting the reservoir parameters of the unknown wells through the intelligent model based on the second sample set; interpolating and denormalizing the reservoir parameters of the unknown wells, and outputting the processed reservoir parameters of the unknown wells. Through this method, it is possible to construct and train an intelligent model based on a layer sample set associated with continuous logging curve data and reservoir parameters, so as to predict the reservoir parameters of unknown wells using the intelligent model, which not only retains the vertical correlation information of logging, but also solves the problem of calculation error evaluation caused by the mismatch between the core reservoir parameters and the logging curve depth, thereby meeting the calculation requirements of reservoir parameters in logging reservoir evaluation and improving the accuracy and stability of reservoir parameter prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Hereinafter, the present application will be described in more detail based on embodiments and with reference to the accompanying drawings;

[0050] Figure 1 An exemplary flow chart of a reservoir parameter prediction method based on well logging data provided in Example 1 of the present invention;

[0051] Figure 2 In the embodiment of the present invention, Figure 1 An exemplary flow chart of step S2 shown in;

[0052] Figure 3 In the embodiment of the present invention, Figure 1 An exemplary flow chart of steps S24 and S25 shown in FIG.

[0053] Figure 4 In the embodiment of the present invention, Figure 1 Step S3 shown in the flowchart is specifically implemented;

[0054] Figure 5 In the embodiment of the present invention, Figure 1 Step S4 shown in the flowchart is specifically implemented;

[0055] Figure 6 In the embodiment of the present invention, Figure 1 Step S5 shown in the flowchart is specifically implemented;

[0056] Figure 7 A comparative analysis graph of the porosity of an unknown well and the porosity of a core predicted by the prediction method provided in Example 1 of the present invention;

[0057] Figure 8 A schematic block diagram of a reservoir parameter prediction device provided in Embodiment 2 of the present invention;

[0058] Fig. 9 A schematic block diagram of an electronic device provided in Embodiment 3 of the present invention;

[0059] Fig.10 A schematic diagram of a computer-readable storage medium provided for Embodiment 4 of the present invention.

[0060] In the drawings, the same reference numerals are used for the same components, and the drawings are not drawn to scale. DETAILED DESCRIPTION

[0061] The present disclosure is further described below in conjunction with the embodiments shown in the accompanying drawings.

[0062] Embodiment 1:

[0063] The present application embodiment discloses a reservoir parameter prediction method based on well logging data, such as Figure 1 As shown, the following steps are included: based on the logging data of multiple wells in the work area, the existing reservoir parameters and logging curve data of the work area are obtained; a first sample set is constructed according to the existing reservoir parameters and logging curve data, and the first sample set is normalized; a convolutional neural network model is constructed, and the convolutional neural network model is trained using the first sample set to obtain the optimal intelligent model; a second sample set is constructed for unknown wells, and the reservoir parameters of the unknown wells are predicted through the intelligent model based on the second sample set; the reservoir parameters of the unknown wells are interpolated and denormalized, and the processed reservoir parameters of the unknown wells are output. Through this prediction method, the vertical correlation information of the logging can be retained, and the calculation error evaluation problem caused by the mismatch between the core reservoir parameters and the logging curve depth can be solved, thereby meeting the calculation requirements of the reservoir parameters in the logging reservoir evaluation and improving the accuracy and stability of the reservoir parameter prediction.

[0064] At least one embodiment of the present disclosure provides a reservoir parameter prediction method based on well logging data. The prediction method can be implemented in the form of software, hardware, firmware or any combination thereof, and loaded and executed by a processor in a device such as a mobile phone, a tablet computer, a laptop computer, a desktop computer, a network server, etc., so as to retain the vertical correlation information of well logging and solve the calculation error evaluation problem caused by the mismatch between core reservoir parameters and logging curve depth, thereby meeting the calculation requirements of reservoir parameters in well logging reservoir evaluation.

[0065] Reference below Figure 1 As shown, a reservoir parameter prediction method based on logging data provided by at least one embodiment of the present disclosure is described, and the reservoir parameter prediction method includes steps S1 to S5.

[0066] S1. Based on the logging data of multiple wells in the work area, the existing reservoir parameters and logging curve data of the work area are obtained.

[0067] In some embodiments, the existing reservoir parameters include well number, core sampling depth, and core reservoir parameters obtained by analyzing and testing the core samples; the core sampling depth is the core sampling depth after depth homing; the core reservoir parameters include porosity and permeability, etc. At the same time, the logging curve data must correspond to the well number of the core sample in the logging reservoir parameters to ensure the correlation between the two, which is conducive to subsequent prediction.

[0068] S2. Construct a first sample set according to existing reservoir parameters and logging curve data, and perform normalization processing on the first sample set.

[0069] In some embodiments, step S2, such as Figure 2 As shown, steps S21-S25 are included, specifically:

[0070] S21, extracting well logging curve data corresponding to the well number of the core reservoir parameter;

[0071] S22. Determine the logging curve data of several logging sampling points before and after the core sample depth based on the extracted logging curve data.

[0072] In some embodiments, the number of logging sampling points before and after the core sample depth is recorded as N, that is, there are 2*N continuous logging sampling points in depth, so as to effectively use the vertical correlation information of the logging data for prediction and ensure the integrity of the prediction results.

[0073] S23. Associating each core reservoir parameter with the well logging curve data of several well logging sampling points to form a first sample set.

[0074] S24, normalizing the logging curve data of each sample in the first sample set.

[0075] S25. Normalize the reservoir parameters of each sample in the first sample set.

[0076] In some embodiments, specifically, Figure 3 As shown, step S24 and step S25 include the following steps:

[0077] S201, performing linear normalization processing on the well logging curve and reservoir parameters using a linear scale using a preset first formula;

[0078] S202, performing logarithmic normalization processing on the well logging curve and reservoir parameters using a logarithmic scale using a preset second formula.

[0079] Among them, the first formula is:

[0080]

[0081] The second formula is:

[0082]

[0083] In the formula, gvalue i is a well logging curve corresponding to the i-th depth point after normalization, value i is the logging curve corresponding to the i-th depth point, LSCA and RSCA are the default left and right scales of the curve in the work area.

[0084] S3. Construct a convolutional neural network model, and use the first sample set to train the convolutional neural network model to obtain an optimal intelligent model.

[0085] In some embodiments, step S3, such as Figure 4 As shown, including:

[0086] S31, setting the number of logging curves as the input attribute dimension and the number of channels of the convolutional neural network model, setting the convolution direction to the depth direction, and the output attribute to the reservoir parameter to be predicted;

[0087] S32, setting the multiple samples in the first sample set as a training sample set and a test sample set respectively;

[0088] S33. Use the training sample set to train the convolutional neural network model, and use the test sample set to evaluate its effect in real time, so as to reduce the training cost and test error to a minimum and obtain the optimal intelligent model.

[0089] The training cost is the sum of the prediction errors of the training sample set, and the test error is the sum of the prediction errors of the test sample set.

[0090] S4. Construct a second sample set for the unknown well, and predict reservoir parameters of the unknown well through an intelligent model based on the second sample set.

[0091] In some embodiments, step S4, such as Figure 5 As shown, including S41-S45, specifically:

[0092] S41, obtaining logging curve data of an unknown well for testing multiple logging sampling points.

[0093] In some embodiments, unknown wells are wells where core samples have not been collected or core sample analysis and testing has not been carried out, and the logging curve data is measured by the work area for each well during the logging stage. Among them, the depths of each logging sampling point are different, and the data are collected at different logging sampling depths.

[0094] S42, combining the well logging curve data of each well logging sampling point of the unknown well with several well logging sampling points thereafter to construct a second sample.

[0095] S43. Circularly construct a plurality of second samples according to the number of well logging sampling depths in the well logging curve data to obtain a second sample set.

[0096] In some embodiments, assuming that there are M logging sampling depths in the logging curve data, the second samples are constructed in a loop according to step S42, and a total of (M-2*N+1) second samples can be constructed to constitute the second sample set.

[0097] S44: normalize the second sample set to obtain a processed second sample set.

[0098] In some embodiments, the process of normalizing the second sample set in step S44 is performed using the method of normalizing the first sample set described in step S44, that is, the normalization process in step S44 is performed using the normalization method in step S2, and the normalized logging curve data and reservoir parameters need to correspond to step S2. The dimension of each second sample is (2*N, the number of logging curves), which is consistent with the input attribute dimension of the one-dimensional convolutional neural network model in step S3, so as to facilitate the input of the second sample set into the intelligent model for prediction.

[0099] S45. Input the processed second sample set into the intelligent model, predict the reservoir parameters corresponding to each second sample, and obtain the unknown well reservoir parameters corresponding to each depth.

[0100] In some embodiments, the second sample set of the unknown well is input into the trained intelligent model to predict the reservoir parameters corresponding to each second sample. Each second sample corresponds to a layer with a different depth, and the reservoir parameters of each layer of the unknown well can be obtained. The depth of the predicted reservoir parameter is set to the average depth of the Nth logging sampling point and the N+1th logging sampling point of the second sample, and the predicted reservoir parameter data is obtained as (depi, vali), where vali is the reservoir parameter predicted by the i-th layer sample of the unknown well, and depi is the depth corresponding to the predicted reservoir parameter of the i-th layer sample of the unknown well, so that the vertical correlation information of the well logging can be retained, the core reservoir parameters are guaranteed to correspond to the depth of the logging curve, the calculation error evaluation is reduced, and the accuracy of the reservoir parameter calculation in the logging reservoir evaluation is improved.

[0101] S5. Interpolate and denormalize the unknown well reservoir parameters, and output the processed unknown well reservoir parameters.

[0102] In some embodiments, step S5, such as Figure 6As shown, including:

[0103] S51, interpolating the unknown well reservoir parameters using the preset third formula and the fourth formula to obtain the unknown well reservoir parameters after difference;

[0104] S52, performing inverse normalization processing on the unknown well reservoir parameters after difference using the preset fifth formula and sixth formula to obtain the unknown well reservoir parameters consistent with the depth of the logging curve, and outputting them together with the logging curve.

[0105] Specifically, the third formula for difference calculation is:

[0106]

[0107] The fourth formula is:

[0108]

[0109] Among them, depc j The depth is the same as that of the N+1th logging sampling point of the i-th layer sample, that is, the interpolated reservoir parameters.

[0110] The fifth formula for the denormalization process is:

[0111] fvalc j =valc j ×RSCA+(1-valc j )×LSCA

[0112] The sixth formula is:

[0113]

[0114] Wherein, LSCA and RSCA are the default left and right scales of the curve in the working area, respectively.

[0115] Among them, the fifth formula is used for linear inverse normalization processing, and the sixth formula is used for logarithmic inverse normalization processing. The reservoir parameter data after interpolation and inverse normalization is consistent with the depth of the well logging curve data, and the dimension and physical meaning of the reservoir parameter data obtained are consistent with those obtained by the traditional method, and can be output together with the well logging curve data.

[0116] In some specific embodiments, according to step S1 and step S2 of the present invention, a first sample set is constructed for 10 wells with core reservoir parameters in the work area and well logging is normalized. Specifically, the reservoir parameters are sorted, there are 616 core porosity data obtained by core sample analysis and testing, the well logging curve data files of the 10 wells are sorted, and the well logging curve data are read according to the well name and depth of the core porosity data in the reservoir parameter file, and N=16 is set, that is, the well logging curve data of 16 well logging sampling points before and after the core depth corresponding to the core porosity data are read, and each core porosity data is associated with the well logging curve data of the 32 well logging sampling points read to obtain a layer sample set. There are 5 well logging curves in the layer sample set, which are natural gamma (G R), neutron porosity (CNL), density (DEN), acoustic time difference (AC), deep induction resistivity (ILD). Usually, GR, CNL, DEN, AC adopt linear scale, ILD adopts logarithmic scale, and the default left and right scales in the work area are: GR: 0, 250, CNL: 37, -13, DEN: 1.975, 2.775, AC: 365, 145, ILD: 2, 200; the logging data of the layer sample are normalized according to formula (1) and formula (2). The core porosity (core-POR) in the layer sample adopts linear scale. The default left and right scales in the work area are: POR: 0, 30. The logging normalization of the core porosity data of the layer sample is carried out according to formula (1).

[0117] Furthermore, since each first sample logging data has 32 logging sampling points and each logging sampling point has 5 logging curves, the input attribute dimension of the designed convolutional neural network model is (32,5), that is, the number of input channels is 5, the convolution direction is the depth direction, and the first hidden layer of the designed network model is a one-dimensional convolution layer: the parameters of this layer are: 48 one-dimensional convolution kernels, the convolution kernel size is 3, the step is 1, and the activation function is the RELU function; the second hidden layer of the designed network model is a pooling layer: the parameters of this layer are: the pooling method is maximum pooling, the filter size is 2, The step is 2; the third hidden layer of the network model is designed to be a one-dimensional convolution layer: the parameters of this layer are: 96 one-dimensional convolution kernels, the convolution kernel size is 3, the step is 1, and the activation function is the RELU function; the fourth hidden layer of the network model is designed to be a pooling layer: the parameters of this layer are: the pooling method is maximum pooling, the filter size is 2, and the step is 2; the fifth hidden layer of the network model is designed to be a flat layer: the output of the fourth hidden layer is flattened and output as a one-dimensional vector; the output layer of the network model is designed to be a fully connected layer with 1 neuron, that is, the porosity data predicted by the output network model is output.

[0118] 504 first samples from 8 wells in the first sample set were used as training sets, and 112 second samples from the other two wells were used as test sets. The keras tool of the tensorflow machine learning platform was used to construct the aforementioned one-dimensional convolutional neural network model, and the training set was input for iterative training. The network model weights were updated once in each iteration, and the training set was predicted by the network model after the updated weights, which was compared with the training set labels (core porosity data). The training cost was calculated and output. At the same time, the test set was predicted by the network model after the updated weights, which was compared with the test set labels, the error was calculated and the model with the smallest error was saved. After 1,000 training times, it was found that the training cost continued to decrease during the training process, while the trend of the prediction error for the test set was first reduced and then increased. The error was the smallest after 40 training times, and the network model at this time could be considered the optimal model.

[0119] Furthermore, the porosity data of 8 sections of a certain 12 wells in the area are predicted, the logging data of 8 sections of a certain 12 wells are processed depth point by depth point, each depth point and the logging curve data of the subsequent 31 (2*N-1=31) depth points constitute a second sample, the logging curve data in the second sample are normalized, and a total of 258 layer samples are obtained. The logging curve data of each layer sample has a dimension of (32,5), and the layer sample set is input into the network model to obtain the predicted porosity data, and the depth of the porosity data is set according to the method in step S4.

[0120] Finally, according to step S5, the porosity data of the 8th section of a certain 12-well predicted by the intelligent model are interpolated and denormalized, and the output is compared with the porosity obtained by core sample analysis. A certain 12-well did not participate in the training of the intelligent model and can be considered as a blind well. Figure 7 As shown in the figure, it can be seen that the graphical comparison analysis of the predicted porosity (AIPOR) and the core porosity (core-POR) shows good consistency.

[0121] The reservoir parameter prediction method based on well logging data provided by the embodiments of the present disclosure can build and train an intelligent model based on a layer sample set that associates continuous well logging curve data with reservoir parameters, thereby using the intelligent model to predict reservoir parameters of unknown wells. This not only retains the vertical correlation information of well logging, but also solves the problem of calculation error evaluation caused by the mismatch between core reservoir parameters and logging curve depth, thereby meeting the calculation requirements of reservoir parameters in well logging reservoir evaluation and improving the accuracy and stability of reservoir parameter prediction.

[0122] Embodiment 2:

[0123] At least one embodiment of the present disclosure further provides a reservoir parameter prediction device, wherein each module in the device corresponds to the steps in the method of the first embodiment above, and the effects achieved and the problems solved have been described above and will not be repeated here. Figure 8 As shown, including:

[0124] The data collection module 11 is used to obtain the existing reservoir parameters and logging curve data of the work area based on the logging data of multiple wells in the work area;

[0125] The data processing module 12 is used to construct a first sample set according to the existing reservoir parameters and logging curve data, and perform normalization processing on the first sample set;

[0126] The model processing module 13 is used to construct a convolutional neural network model and train the convolutional neural network model using the first sample set to obtain an optimal intelligent model;

[0127] A parameter prediction module 14 is used to construct a second sample set for the unknown well, and predict the reservoir parameters of the unknown well through an intelligent model based on the second sample set;

[0128] The prediction data processing module 15 is used to interpolate and denormalize the unknown well reservoir parameters, and output the processed unknown well reservoir parameters.

[0129] In some embodiments, the output end of the data collection module 11 is connected to the input end of the data processing module 12, the output end of the data processing module 12 is connected to the input end of the model processing module 13, the output end of the model processing module 13 is connected to the input end of the parameter prediction module 14, and the output end of the parameter prediction module 14 is connected to the prediction data processing module 15.

[0130] In some embodiments, the device may be a part of the software, or may be implemented in combination with corresponding hardware, which will not be described in detail here.

[0131] Embodiment three:

[0132] At least some embodiments of the present disclosure also provide an electronic device, such as Fig. 9 As shown, the electronic device includes a memory 21 and a processor 22, wherein the memory 21 stores a computer program, and when the computer program is executed by the processor, the steps of the reservoir parameter prediction method provided in any embodiment of the present disclosure are performed.

[0133] In some embodiments, the processor 22 is used to execute all or part of the steps in the reservoir parameter prediction method of any embodiment of the present disclosure. The memory 21 is used to store various types of data, which may include instructions of any application or method in the electronic device, and data related to the application.

[0134] The processor 22 may be an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components, and is used to execute the reservoir parameter prediction method in the above-mentioned embodiment 1.

[0135] The memory 21 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0136] Embodiment 4:

[0137] At least some embodiments of the present disclosure also provide a computer-readable storage medium, such as Fig.10 As shown, the readable storage medium stores a computer program 31, and when the computer program 31 is executed by a processor, the steps of the reservoir parameter prediction method provided by any embodiment of the present disclosure are implemented.

[0138] In some embodiments, the storage medium can be a tangible medium that can contain or store a program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the above. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0139] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0140] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0141] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0142] An embodiment of the present invention further provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the reservoir parameter prediction method provided in any embodiment of the present disclosure.

[0143] The various embodiments in the present disclosure are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0144] The protection scope of the present disclosure is not limited to the above-mentioned embodiments. Obviously, those skilled in the art can make various changes and modifications to the present disclosure without departing from the scope and spirit of the present disclosure. If these changes and modifications fall within the scope of the claims of the present disclosure and their equivalents, the intention of the present disclosure also includes these changes and modifications.

Claims

1. A reservoir parameter prediction method based on well logging data, characterized in that: The following steps are involved: Based on the logging data of multiple wells in the work area, the existing reservoir parameters and logging curve data of the work area are obtained; Constructing a first sample set according to existing reservoir parameters and logging curve data, and performing normalization processing on the first sample set; Constructing a convolutional neural network model, and using the first sample set to train the convolutional neural network model to obtain an optimal intelligent model; constructing a second sample set for the unknown well, and predicting reservoir parameters of the unknown well through an intelligent model based on the second sample set; The unknown well reservoir parameters are interpolated and denormalized, and the processed unknown well reservoir parameters are output.

2. The reservoir parameter prediction method based on well logging data according to claim 1, characterized in that: The existing reservoir parameters include well number, core sampling depth and core reservoir parameters obtained by lithology analysis; the first sample set is constructed based on the existing reservoir parameters and logging curve data, including: Extract the logging curve data corresponding to the well number of the core reservoir parameter; Determine the logging curve data of several logging sampling points before and after the core sample depth from the extracted logging curve data; Each core reservoir parameter is associated with the well logging curve data of several well logging sampling points to form a first sample set.

3. The reservoir parameter prediction method based on well logging data according to claim 1, characterized in that: The normalizing process of the first sample set includes: Normalizing the well logging curve data of each sample in the first sample set; The reservoir parameters of each sample in the first sample set are normalized.

4. The reservoir parameter prediction method based on well logging data according to claim 3 is characterized in that: The well logging curve data of each sample is normalized and the reservoir parameters of each sample are normalized, including: Performing linear normalization processing on the well logging curve and reservoir parameters using a linear scale using a preset first formula; Performing logarithmic normalization processing on the well logging curve and reservoir parameters using a logarithmic scale using a preset second formula; Among them, the first formula is: The second formula is: In the formula, gvalue i is a well logging curve corresponding to the i-th depth point after normalization, value i is the logging curve corresponding to the i-th depth point, LSCA and RSCA are the default left and right scales of the curve in the work area.

5. The reservoir parameter prediction method based on well logging data according to claim 1, characterized in that: The method of constructing a convolutional neural network model and training the convolutional neural network model using the first sample set to obtain an optimal intelligent model includes: The number of logging curves is set as the input attribute dimension and channel number of the convolutional neural network model, and the convolution direction is set as the depth direction, and the output attribute is set as the reservoir parameter to be predicted; The plurality of samples in the first sample set are respectively set as a training sample set and a test sample set; The convolutional neural network model is trained using the training sample set, and its effect is evaluated in real time using the test sample set, so that the training cost and test error are reduced to a minimum and the optimal intelligent model is obtained.

6. The reservoir parameter prediction method based on well logging data according to claim 1, characterized in that: The step of constructing a second sample set for the unknown well and predicting reservoir parameters of the unknown well by using an intelligent model based on the second sample set includes: Obtain logging curve data for testing multiple logging sampling points in unknown wells; Combining the well logging curve data of each well logging sampling point of the unknown well with a number of well logging sampling points thereafter to construct a second sample; Circularly construct a plurality of second samples according to the number of well logging sampling depths in the well logging curve data to obtain a second sample set; Normalizing the second sample set to obtain a processed second sample set; The processed second sample set is input into the intelligent model to predict the reservoir parameters corresponding to each second sample, and the reservoir parameters of the unknown well corresponding to each depth are obtained.

7. A reservoir parameter prediction method based on well logging data according to claim 6, characterized in that: The process of normalizing the second sample set to obtain the processed second sample set is performed by using the method of normalizing the first sample set.

8. The reservoir parameter prediction method based on well logging data according to claim 1, characterized in that: The interpolation and denormalization processing of the unknown well reservoir parameters and outputting the processed unknown well reservoir parameters include: Interpolating the unknown well reservoir parameters using the preset third formula and fourth formula to obtain the unknown well reservoir parameters after difference; The unknown well reservoir parameters after difference are denormalized using the preset fifth and sixth formulas to obtain the unknown well reservoir parameters consistent with the logging curve depth, which are output together with the logging curve.

9. A reservoir parameter prediction device, characterized in that: include: The data collection module is used to obtain the existing reservoir parameters and logging curve data of the work area based on the logging data of multiple wells in the work area; A data processing module, used to construct a first sample set according to existing reservoir parameters and logging curve data, and perform normalization processing on the first sample set; A model processing module, used for constructing a convolutional neural network model, and training the convolutional neural network model using the first sample set to obtain an optimal intelligent model; A parameter prediction module, used for constructing a second sample set for the unknown well, and predicting the reservoir parameters of the unknown well through an intelligent model based on the second sample set; The prediction data processing module is used to interpolate and denormalize the unknown well reservoir parameters, and output the processed unknown well reservoir parameters.

10. An electronic device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the reservoir parameter prediction method according to any one of claims 1 to 8 is executed.

11. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the storage medium is executed by one or more processors, the reservoir parameter prediction method according to any one of claims 1 to 8 is implemented.

Images