Real-time intelligent judgment method, system and device for operation state of downhole instrument and medium

Abstract

The invention discloses a real-time intelligent judgment method, system and device for the operation state of a downhole instrument and a medium, and the method comprises the steps: carrying out the data expansion through employing a CGAN, building a plurality of downhole instrument operation algorithm parameter libraries according to the classification of different well types, avoiding the limitation of a model for discriminating the operation state of the downhole instrument through a single parameter, and improving the recognition precision; and the hyper-parameters of the DNN-GRU network are optimized, so that the hyper-parameters of the machine learning model are accurately obtained, time and labor cost are saved, and the overall performance of the model is improved. According to the method, the operation state of the underground instrument can be timely and accurately judged, and the abnormal working condition can be judged, so that an important basis and a monitoring means are provided for safe and efficient conveying of the underground instrument, the burden of field engineers is relieved, the operation safety degree is improved, and automation and intelligentization of well logging service are promoted.

Description

Technical Field

[0001] This invention belongs to the field of petroleum logging technology and relates to a method, system, device and medium for real-time intelligent determination of the operating status of downhole instruments. Background Technology

[0002] The low oil price environment places new demands on enhancing the technical service capabilities for exploration and development of different types of oil and gas reservoirs, requiring safe and efficient logging operations. There is an urgent need to comprehensively improve logging service capabilities through technological innovation. Currently, downhole instrument delivery mainly relies on logging winches, and the operational status of downhole instruments is determined manually. However, logging operations are demanding, often requiring continuous operation for over ten hours, leading to fatigue for field engineers. In complex well conditions, this can affect operational efficiency and safety. Therefore, there is an urgent need to research intelligent identification methods for downhole instrument operational status to assist in identifying and handling abnormal situations, thereby reducing the operational burden on field engineers and improving the efficiency and safety of logging operations. Current research mainly focuses on single algorithms or threshold parameters for identification, which can easily lead to misjudgments of operational status. Research on the application of intelligent algorithms in real-time determination of downhole instrument operational status is also limited. Summary of the Invention

[0003] The purpose of this invention is to address the problem that existing technologies for studying the operating status of downhole instruments mainly rely on single algorithms or threshold parameters, which can easily lead to misjudgments of the operating status. Furthermore, there is a lack of application of intelligent algorithms in the real-time determination of the operating status of downhole instruments. This invention provides a method, system, device, and medium for real-time intelligent determination of the operating status of downhole instruments.

[0004] To achieve the above objectives, the present invention employs the following technical solution:

[0005] A method for real-time intelligent determination of the operating status of downhole instruments, comprising:

[0006] Collect existing logging data and operational parameters for each well, and use Conditional Generative Adversarial Network (CGAN) to expand the data. At the same time, establish a database of various downhole instrument operation algorithm parameters based on different well types.

[0007] Preprocess the datasets contained in the downhole instrument operation algorithm parameter library;

[0008] The predicted parameters are obtained by reconstructing the multi-source data contained in the downhole instrument operation algorithm parameter library;

[0009] The preprocessed downhole instrument operation algorithm parameter library dataset was divided into training set, validation set and test set, and a DNN-GRU network was built.

[0010] The hyperparameters of the DNN-GRU network are optimized, and the optimized hyperparameters are fused with the training set and validation set. The DNN-GRU network is then trained based on the test set and the fused training set and validation set to obtain the optimized DNN-GRU network model.

[0011] The reconstructed prediction parameters are used as input variables, and the downhole instrument operating status is used as output variables. The input variables are then fed into the optimized DNN-GRU network model to obtain the output variables, which are the real-time operating status of the downhole instrument.

[0012] A further improvement of the present invention is that:

[0013] Furthermore, existing logging data and operational parameters for each well were collected, and the data was augmented using a conditional generative adversarial network (CGAN). Specifically:

[0014] A generator is constructed using cable tension data and Gaussian distributed random noise as input datasets, and the output is the generated three-parameter tension data. A discriminator is also constructed, with the inputs being the real three-parameter tension data and the generated three-parameter tension data, respectively. The discriminator is used to determine the authenticity of the three-parameter tension data. Then, a conditional generative adversarial network (CGAN) is trained, and the trained CGAN network is used to generate new three-parameter tension data to achieve data augmentation.

[0015] Furthermore, the dataset contained in the downhole instrument operation algorithm parameter library is preprocessed, specifically by performing depth correction, data filtering, data completion, and normalization on the dataset.

[0016] The depth correction involves performing depth correction on the logging curves of each well to ensure that the depths of each parameter curve are aligned.

[0017] The data filtering involves removing outliers from the algorithm parameter library. If the parameter depth DEP is less than 0m or greater than the well depth, all data at that depth point are deleted.

[0018] The data completion refers to filling in missing parameters using the mean method. The calculation formula is as follows:

[0019] Parm i =Parm i-2 +Parm i-1 +Parm i+1 +Parm i+2 (1)

[0020] Where i is the current data point position, and Parm is the algorithm library parameter;

[0021] The normalization process involves applying Z-Score standardization to all parameters, calculated using the following formula:

[0022]

[0023] Where x is the original parameter, μ is the average value of the input parameter, and σ is the standard deviation of the input parameter.

[0024] Furthermore, the multi-source data contained in the downhole instrument operation algorithm parameter library are reconstructed to obtain the prediction parameters, specifically:

[0025] Based on the cable tension and three-parameter tension from multi-source data, the cable differential tension and three-parameter differential tension are obtained.

[0026] The cable tension, three-parameter tension, cable differential tension, and three-parameter differential tension are recalculated to obtain the fused tension FTEN and the fused differential tension DFTEN.

[0027] By integrating other parameters from the multi-source data, the input variables for the downhole instrument operating status are obtained; among these, other parameters from the multi-source data include logging depth (DEP), logging operation speed (SPD), logging instrument weight (TW), mud density (MD), mud viscosity (MV), and well deviation (DEV).

[0028] Furthermore, based on the cable tension and three-parameter tension from the multi-source data, the cable differential tension and three-parameter differential tension are obtained, specifically:

[0029] DTENC i =TENC i+1 -TENC i (3)

[0030] DTENT i =TENT i+1 -TENT i (4)

[0031] Where i is the current data point position, DTENC is the cable differential tension, DTENT is the three-parameter differential tension, TENC is the cable tension, and TENT is the three-parameter tension;

[0032] The process involves recalculating the cable tension, three-parameter tension, cable differential tension, and three-parameter differential tension to obtain the fused tension FTEN and the fused differential tension DFTEN, specifically as follows:

[0033]

[0034]

[0035] Among them, TENC min TENC is the minimum cable tension. max This represents the maximum cable tension; TENTmin TENT is the minimum value of the three-parameter tension. max This represents the maximum tension value of the three parameters; DTENC min DTENC is the minimum differential tension of the cable. max This represents the maximum differential tension of the cable; DTENT min DTENT is the minimum value of the three-parameter differential tension. max This represents the maximum value of the three-parameter differential tension.

[0036] The other parameters from the integrated multi-source data are used to obtain the input variables for the downhole instrument operating status. Specifically, the parameters in the final parameter library include: logging depth (DEP), logging operation speed (SPD), logging instrument weight (TW), fusion tension (TENF), fusion differential tension (DTENF), mud density (MD), mud viscosity (MV), and well deviation (DEV).

[0037] Furthermore, the DNN-GRU network includes two DNN layers, two GRU layers, and two Dropout layers. The first DNN layer is used to capture the variable features of parameters in the input parameter library. After the data passes through the first DNN layer, it enters the first GRU layer to determine the time dependency in each parameter sequence data. Then it enters the first Dropout layer to prevent overfitting of the training data. It continues to the second GRU layer to capture the time dependency of the data. Then it enters the second Dropout layer to improve the generalization of the model. Finally, it enters the DNN layer to integrate the feature information extracted by all the previous layers to generate the downhole instrument operating status.

[0038] Furthermore, the hyperparameters of the DNN-GRU network are optimized based on the ASHA algorithm, specifically as follows:

[0039] The model hyperparameters were tuned using the ASHA algorithm. The hyperparameters to be optimized included the learning rate (lr), dropout rate, and batch size. The optimized hyperparameter combination was then used to fuse the training and validation sets to retrain the DNN-GRU network model. Training was stopped when the maximum number of iterations was reached. The optimizer used RMSprop, the activation function for the GRU layer was tanh, the activation function for the last DNN layer was softmax, and the loss function used was cross-entropy loss.

[0040] The loss function is calculated as follows:

[0041]

[0042] Where C is the number of predicted categories, x i For the i-th element of the real label, y i To predict the probability that x belongs to the i-th category.

[0043] A real-time intelligent determination system for the operating status of downhole instruments, comprising:

[0044] The expansion module collects existing logging data and operational parameters for each well, and uses a conditional generative adversarial network (CGAN) to expand the data. At the same time, it establishes a database of various downhole instrument operation algorithm parameters based on different well types.

[0045] The preprocessing module preprocesses the dataset contained in the downhole instrument operation algorithm parameter library;

[0046] The reconstruction module reconstructs multi-source data contained in the downhole instrument operation algorithm parameter library to obtain prediction parameters.

[0047] The partitioning module divides the preprocessed downhole instrument operation algorithm parameter library dataset into a training set, a validation set, and a test set, and builds a DNN-GRU network.

[0048] The training module optimizes the hyperparameters of the DNN-GRU network, fuses the optimized hyperparameters with the training set and validation set, and trains the DNN-GRU network based on the test set and the fused training set and validation set to obtain the optimized DNN-GRU network model.

[0049] The acquisition module takes the reconstructed prediction parameters as input variables and the downhole instrument operating status as output variables, and inputs the input variables into the optimized DNN-GRU network model to obtain the output variables, namely the real-time operating status of the downhole instrument.

[0050] A terminal device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the steps of the method described above.

[0051] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] This invention uses CGAN to generate missing three-parameter tension curves, obtaining fused tension and fused differential tension. By comprehensively considering multiple measurement parameters, it avoids the limitations of single-parameter modeling in determining the operating status of downhole instruments, thus improving recognition accuracy. Furthermore, it optimizes the hyperparameters of the DNN-GRU network to accurately obtain the hyperparameters of the machine learning model, saving time and manpower costs and improving the overall performance of the model. This invention can promptly and accurately determine the operating status of downhole instruments and identify abnormal conditions, thereby providing important evidence and monitoring methods for the safe and efficient delivery of downhole instruments, reducing the burden on on-site engineers, and improving operational safety, thereby promoting the automation and intelligence of logging services. Attached Figure Description

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart illustrating the real-time intelligent determination method for the operating status of downhole instruments according to the present invention.

[0056] Figure 2 This is a schematic diagram of the structure of the real-time intelligent determination system for the operating status of downhole instruments according to the present invention;

[0057] Figure 3 Flowchart for generating tension data for CGAN;

[0058] Figure 4 The resistance tension and velocity curves of well A1 are shown.

[0059] Figure 5 This is a flowchart of a DNN-GRU model optimized based on ASHA. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0061] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0062] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0063] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper," "lower," "horizontal," or "inner" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of the invention is in use, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Furthermore, terms such as "first" and "second" are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0064] Furthermore, the use of the term "horizontal" does not imply that the component must be absolutely horizontal, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0065] In the description of the embodiments of the present invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention according to the specific circumstances.

[0066] The present invention will now be described in further detail with reference to the accompanying drawings:

[0067] See Figure 1 This invention discloses a real-time intelligent determination method for the operating status of downhole instruments, comprising:

[0068] S101 collects existing logging data and operational parameters for each well, and uses Conditional Generative Adversarial Network (CGAN) to expand the data. At the same time, it establishes a database of various downhole instrument operation algorithm parameters based on different well types.

[0069] Based on the collected logging data and operational parameters, a Conditional Generative Adversarial Network (CGAN) was used to augment the data for the missing three-parameter tension in certain wells. Specifically, a generator was constructed using cable tension data and Gaussian distributed random noise as input datasets, and the output was the generated three-parameter tension data. A discriminator was also constructed, with the inputs being the real three-parameter tension data and the generated three-parameter tension data. The discriminator was used to determine the authenticity of the three-parameter tension data. Then, the CGAN was trained, and the trained CGAN network was used to generate new three-parameter tension data, thus achieving data augmentation.

[0070] S102, preprocess the dataset contained in the downhole instrument operation algorithm parameter library;

[0071] The downhole instrument operation algorithm parameter database dataset is subjected to depth correction, data filtering, data completion and normalization.

[0072] Depth correction involves performing depth correction on the logging curves of each well to ensure that the depths of each parameter curve are aligned.

[0073] The data filtering involves removing outliers from the algorithm parameter library. If the parameter depth DEP is less than 0m or greater than the well depth, all data at that depth point are deleted.

[0074] The data completion refers to filling in missing parameters using the mean method. The calculation formula is as follows:

[0075] Parm i =Parm i-2 +Parm i-1 +Parm i+1 +Parm i+2 (1)

[0076] Where i is the current data point position, and Parm is the algorithm library parameter;

[0077] The normalization process involves applying Z-Score standardization to all parameters, calculated using the following formula:

[0078]

[0079] Where x is the original parameter, μ is the average value of the input parameter, and σ is the standard deviation of the input parameter.

[0080] S103, Reconstruct the multi-source data contained in the downhole instrument operation algorithm parameter library to obtain the prediction parameters;

[0081] Based on the cable tension and three-parameter tension from multi-source data, the cable differential tension and three-parameter differential tension are obtained.

[0082] The cable tension, three-parameter tension, cable differential tension, and three-parameter differential tension are recalculated to obtain the fused tension FTEN and the fused differential tension DFTEN.

[0083] By integrating other parameters from the multi-source data, the input variables for the downhole instrument operating status are obtained; among these, other parameters from the multi-source data include logging depth (DEP), logging operation speed (SPD), logging instrument weight (TW), mud density (MD), mud viscosity (MV), and well deviation (DEV).

[0084] Based on the cable tension and three-parameter tension from multi-source data, the cable differential tension and three-parameter differential tension are obtained, specifically:

[0085] DTENC i =TENC i+1 -TENC i (3)

[0086] DTENT i =TENT i+1 -TENT i (4)

[0087] Where i is the current data point position, DTENC is the cable differential tension, DTENT is the three-parameter differential tension, TENC is the cable tension, and TENT is the three-parameter tension.

[0088] The process involves recalculating the cable tension, three-parameter tension, cable differential tension, and three-parameter differential tension to obtain the fused tension FTEN and the fused differential tension DFTEN, specifically as follows:

[0089]

[0090]

[0091] Among them, TENC min TENC is the minimum cable tension. max This represents the maximum cable tension; TENT min TENT is the minimum value of the three-parameter tension. max This represents the maximum tension value of the three parameters; DTENC min DTENC is the minimum differential tension of the cable. max This represents the maximum differential tension of the cable; DTENT min DTENT is the minimum value of the three-parameter differential tension. max This represents the maximum value of the three-parameter differential tension.

[0092] The other parameters from the integrated multi-source data are used to obtain the input variables for the downhole instrument operating status. Specifically, the parameters in the final parameter library include: logging depth (DEP), logging operation speed (SPD), logging instrument weight (TW), fusion tension (TENF), fusion differential tension (DTENF), mud density (MD), mud viscosity (MV), and well deviation (DEV).

[0093] S104, the preprocessed downhole instrument operation algorithm parameter library dataset is divided into training set, validation set and test set, and a DNN-GRU network is built;

[0094] The DNN-GRU network consists of two DNN layers, two GRU layers, and two Dropout layers. The first DNN layer captures the variable features of the parameters in the input parameter library. After the data passes through the first DNN layer, it enters the first GRU layer to determine the time dependencies in the parameter sequence data. Then it enters the first Dropout layer to prevent overfitting of the training data. It continues to the second GRU layer to capture the time dependencies of the data. Then it enters the second Dropout layer to improve the generalization of the model. Finally, it enters the DNN layer to integrate the feature information extracted by all the previous layers to generate the downhole instrument operating status.

[0095] S105, optimize the hyperparameters of the DNN-GRU network, fuse the optimized hyperparameters with the training set and validation set, train the DNN-GRU network based on the test set and the fused training set and validation set, and obtain the optimized DNN-GRU network model.

[0096] The model hyperparameters were tuned using the ASHA algorithm. The hyperparameters to be optimized included the learning rate (lr), dropout rate, and batch size. The optimized hyperparameter combination was then used to fuse the training and validation sets to retrain the DNN-GRU network model. Training was stopped when the maximum number of iterations was reached. The optimizer used RMSprop, the activation function for the GRU layer was tanh, the activation function for the last DNN layer was softmax, and the loss function used was cross-entropy loss.

[0097] The loss function is calculated as follows:

[0098]

[0099] Where C is the number of predicted categories, x i For the i-th element of the real label, y i To predict the probability that x belongs to the i-th category.

[0100] S106 uses the reconstructed prediction parameters as input variables and the downhole instrument operating status as output variables. The input variables are then fed into the optimized DNN-GRU network model to obtain the output variables, which are the real-time operating status of the downhole instrument.

[0101] See Figure 2 This invention discloses a real-time intelligent determination system for the operating status of downhole instruments, comprising:

[0102] The expansion module collects existing logging data and operational parameters for each well, and uses a conditional generative adversarial network (CGAN) to expand the data. At the same time, it establishes a database of various downhole instrument operation algorithm parameters based on different well types.

[0103] The preprocessing module preprocesses the dataset contained in the downhole instrument operation algorithm parameter library;

[0104] The reconstruction module reconstructs multi-source data contained in the downhole instrument operation algorithm parameter library to obtain prediction parameters.

[0105] The partitioning module divides the preprocessed downhole instrument operation algorithm parameter library dataset into a training set, a validation set, and a test set, and builds a DNN-GRU network.

[0106] The training module optimizes the hyperparameters of the DNN-GRU network, fuses the optimized hyperparameters with the training set and validation set, and trains the DNN-GRU network based on the test set and the fused training set and validation set to obtain the optimized DNN-GRU network model.

[0107] The acquisition module takes the reconstructed prediction parameters as input variables and the downhole instrument operating status as output variables, and inputs the input variables into the optimized DNN-GRU network model to obtain the output variables, namely the real-time operating status of the downhole instrument.

[0108] Example:

[0109] (1) Logging data and field operation parameters of 10 casing vertical wells in Changqing Oilfield were collected, denoted as wells A1 to A10. The logging data sampling interval was 0.1m. The main downhole instruments were: remote gamma ray TELC_C, three-parameter TPMS, magnetic signature CCL, and acoustic variable density VDL. A downhole instrument operation algorithm parameter library was established. For the problem of missing three-parameter tension in wells A5, A6, and A7, conditional generative adversarial network (CGAN) was used to expand the three-parameter tension data and establish a casing vertical well operation parameter library.

[0110] (2) Preprocess the algorithm parameter library data. First, perform depth correction on the logging data of each well to ensure that the depth of each parameter is aligned. Then, remove outliers. If the logging depth is less than 0m or greater than the well depth, delete all data at that depth point. For missing parameters, fill them with the average of the five points before and after the current point. Finally, use Z-Score standardization to process all parameters to eliminate the differences caused by different data units and ranges.

[0111] (3) The differential tension of the cable and the differential tension of the three parameters are calculated using cable tension and three-parameter tension. Then, the cable tension, the three-parameter tension, the differential tension of the cable, and the differential tension of the three parameters are recalculated to obtain two parameters: fused tension FTEN and fused differential tension DFTEN. Based on this, other operational data are integrated to obtain the input variable of the downhole instrument operating status. At the same time, the downhole instrument operating status is used as the output variable. Since all 10 wells in this case were logging, the logging status is divided into two types: normal logging and logging encountering resistance, with output labels of 1 and 0 respectively. The algorithm parameter library is divided into training set, validation set and test set according to an 8:1:1 ratio. A DNN-GRU network is built, including 2 DNN layers, 2 GRU layers and 2 Dropout layers.

[0112] (4) The hyperparameters were tuned using the ASHA algorithm. The searched hyperparameters included the learning rate (lr), dropout rate, and batch size. The final model parameters were as follows: 32 neurons in the DNN layer, 16 neurons in each of the two GRU layers, a batch size of 64, a dropout rate of 0.5, 80 iterations, a learning rate of 0.0012, a time step of 20, the tanh activation function, and the RMSprop optimizer.

[0113] (5) Using the hyperparameter combination obtained from ASHA, the training and validation sets were fused to retrain the DNN-GRU network model. The trained model was then used to make real-time judgments on the operating status of downhole instruments using the test set data. The prediction results are shown in the table below. It can be seen that the algorithm's accuracy reached 92.7%.

[0114]

[0115] An embodiment of the present invention provides a terminal device. This terminal device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the various method embodiments described above. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the various device embodiments described above.

[0116] The computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention.

[0117] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0118] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0119] The memory can be used to store the computer program and / or module. The processor implements various functions of the terminal device by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0120] If the modules / units integrated into the terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0121] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for real-time intelligent determination of the operating status of downhole instruments, characterized in that, include: Collect existing logging data and operational parameters for each well, and use Conditional Generative Adversarial Network (CGAN) to expand the data. At the same time, establish a database of various downhole instrument operation algorithm parameters based on different well types. Preprocess the datasets contained in the downhole instrument operation algorithm parameter library; The predicted parameters are obtained by reconstructing the multi-source data contained in the downhole instrument operation algorithm parameter library; The preprocessed downhole instrument operation algorithm parameter library dataset was divided into training set, validation set and test set, and a DNN-GRU network was built. The hyperparameters of the DNN-GRU network are optimized, and the optimized hyperparameters are fused with the training set and validation set. The DNN-GRU network is then trained based on the test set and the fused training set and validation set to obtain the optimized DNN-GRU network model. The reconstructed prediction parameters are used as input variables, and the downhole instrument operating status is used as output variables. The input variables are then fed into the optimized DNN-GRU network model to obtain the output variables, which are the real-time operating status of the downhole instrument.

2. The real-time intelligent determination method for the operating status of downhole instruments according to claim 1, characterized in that, The process involves collecting existing logging data and operational parameters for each well, and then augmenting the data using a conditional generative adversarial network (CGAN). Specifically: A generator is constructed using cable tension data and Gaussian distributed random noise as input datasets, and the output is the generated three-parameter tension data. A discriminator is also constructed, with the inputs being the real three-parameter tension data and the generated three-parameter tension data, respectively. The discriminator is used to determine the authenticity of the three-parameter tension data. Then, a conditional generative adversarial network (CGAN) is trained, and the trained CGAN network is used to generate new three-parameter tension data to achieve data augmentation.

3. The real-time intelligent determination method for the operating status of downhole instruments according to claim 1, characterized in that, The preprocessing of the dataset contained in the downhole instrument operation algorithm parameter library specifically involves: performing depth correction, data filtering, data completion, and normalization on the dataset of the downhole instrument operation algorithm parameter library. The depth correction involves performing depth correction on the logging curves of each well to ensure that the depths of each parameter curve are aligned. The data filtering involves removing outliers from the algorithm parameter library. If the parameter depth DEP is less than 0m or greater than the well depth, all data at that depth point are deleted. The data completion refers to filling in missing parameters using the mean method. The calculation formula is as follows: Parma i =Parm i-2 +Parm i-1 +Parm i+1 +Parm i+2 (1) Where i is the current data point position, and Parm is the algorithm library parameter; The normalization process involves applying Z-Score standardization to all parameters, calculated using the following formula: Where x is the original parameter, μ is the average value of the input parameter, and σ is the standard deviation of the input parameter.

4. The real-time intelligent determination method for the operating status of downhole instruments according to claim 1, characterized in that, The process of reconstructing the multi-source data contained in the downhole instrument operation algorithm parameter library to obtain prediction parameters is as follows: Based on the cable tension and three-parameter tension from multi-source data, the cable differential tension and three-parameter differential tension are obtained. The cable tension, three-parameter tension, cable differential tension, and three-parameter differential tension are recalculated to obtain the fused tension FTEN and the fused differential tension DFTEN. By integrating other parameters from the multi-source data, the input variables for the downhole instrument operating status are obtained; among these, other parameters from the multi-source data include logging depth (DEP), logging operation speed (SPD), logging instrument weight (TW), mud density (MD), mud viscosity (MV), and well deviation (DEV).

5. The real-time intelligent determination method for the operating status of downhole instruments according to claim 4, characterized in that, The cable differential tension and three-parameter differential tension are obtained based on the cable tension and three-parameter tension from multi-source data, specifically as follows: DTENC i =I HAVE i+1 -I HAVE i (3) DTENT i =TENT i+1 -TENT i (4) Where i is the current data point position, DTENC is the cable differential tension, DTENT is the three-parameter differential tension, TENC is the cable tension, and TENT is the three-parameter tension; The process involves recalculating the cable tension, three-parameter tension, cable differential tension, and three-parameter differential tension to obtain the fused tension FTEN and the fused differential tension DFTEN, specifically as follows: Among them, TENC min TENC is the minimum cable tension. max This represents the maximum cable tension; TENT min TENT is the minimum value of the three-parameter tension. max This represents the maximum tension value of the three parameters; DTENC min DTENC is the minimum differential tension of the cable. max This represents the maximum differential tension of the cable; DTENT min DTENT is the minimum value of the three-parameter differential tension. max This represents the maximum value of the three-parameter differential tension. The other parameters from the integrated multi-source data are used to obtain the input variables for the downhole instrument operating status. Specifically, the parameters in the final parameter library include: logging depth (DEP), logging operation speed (SPD), logging instrument weight (TW), fusion tension (TENF), fusion differential tension (DTENF), mud density (MD), mud viscosity (MV), and well deviation (DEV).

6. The real-time intelligent determination method for the operating status of downhole instruments according to claim 5, characterized in that, The DNN-GRU network comprises two DNN layers, two GRU layers, and two Dropout layers. The first DNN layer captures the variable features of parameters in the input parameter library. After passing through the first DNN layer, the data enters the first GRU layer to determine the time dependencies in each parameter sequence data. Then, it enters the first Dropout layer to prevent overfitting of the training data. It continues to the second GRU layer to capture the time dependencies of the data. Then, it enters the second Dropout layer to improve the generalization of the model. Finally, it enters the DNN layer to integrate the feature information extracted by all the previous layers to generate the downhole instrument operating status.

7. The real-time intelligent determination method for the operating status of downhole instruments according to claim 6, characterized in that, The optimization of hyperparameters of the DNN-GRU network based on the ASHA algorithm is specifically as follows: The model hyperparameters were tuned using the ASHA algorithm. The hyperparameters to be optimized included the learning rate (lr), dropout rate, and batch size. The optimized hyperparameter combination was then used to fuse the training and validation sets to retrain the DNN-GRU network model. Training was stopped when the maximum number of iterations was reached. The optimizer used RMSprop, the activation function for the GRU layer was tanh, the activation function for the last DNN layer was softmax, and the loss function used was cross-entropy loss. The loss function is calculated as follows: Where C is the number of predicted categories, x i For the i-th element of the real label, y i To predict the probability that x belongs to the i-th category.

8. A real-time intelligent determination system for the operating status of downhole instruments, characterized in that, include: The expansion module collects existing logging data and operational parameters for each well, and uses a conditional generative adversarial network (CGAN) to expand the data. At the same time, it establishes a database of various downhole instrument operation algorithm parameters based on different well types. The preprocessing module preprocesses the dataset contained in the downhole instrument operation algorithm parameter library; The reconstruction module reconstructs multi-source data contained in the downhole instrument operation algorithm parameter library to obtain prediction parameters. The partitioning module divides the preprocessed downhole instrument operation algorithm parameter library dataset into a training set, a validation set, and a test set, and builds a DNN-GRU network. The training module optimizes the hyperparameters of the DNN-GRU network, fuses the optimized hyperparameters with the training set and validation set, and trains the DNN-GRU network based on the test set and the fused training set and validation set to obtain the optimized DNN-GRU network model. The acquisition module takes the reconstructed prediction parameters as input variables and the downhole instrument operating status as output variables, and inputs the input variables into the optimized DNN-GRU network model to obtain the output variables, namely the real-time operating status of the downhole instrument.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-7.

Images