Well drilling overflow early warning method and device
The drilling well recording data is processed through neural network models, and the drilling overflow probability is automatically identified, which solves the problem of difficulty in identifying drilling overflow signals in the existing technology, and improves the accuracy of overflow warning and drilling safety.
Patent Information
- Application Number
- CN202310587752.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-23
- Publication Date
- 2024-11-26
AI Technical Summary
In the prior art, drilling overflow monitoring methods are difficult to accurately identify overflow signals under complex geological conditions, which can easily cause false alarms or missed reports, seriously affecting drilling safety.
The neural network model is adopted, combining long and short-term memory neural networks and fully convolutional neural networks, and connected in series with the full connection layer and the output layer. By obtaining drilling and recording data, the drilling overflow probability is output. If the preset is exceeded, The threshold will issue an overflow warning to achieve automatic overflow warning.
No sensors are required to identify overflow signals, which reduces the number of early warning false alarms and missed reports, improves the accuracy of drilling overflow warning, reduces the risk of drilling accidents, and ensures drilling safety.
Smart Images

Figure CN119027258A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of oil drilling, and in particular to a drilling overflow early warning method and device. Background Art
[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No admission is made that the description herein is prior art by inclusion in this section.
[0003] Oil resources are an important energy source for the rapid development of industry. With the continuous deepening of oil resource exploration and development, drilling projects are facing many challenges such as drilling safety. Drilling overflow is a common and serious accident in drilling projects, so the monitoring and early warning of drilling overflow is very important. The drilling overflow monitoring method in the existing technology mainly collects the overflow signal of the drilling site in real time through sensors, and relies on the observation and experience of on-site operators to analyze and warn of drilling overflow. However, under complex geological conditions, it is difficult for sensors to accurately identify drilling overflow signals, which is easy to cause false alarms or missed alarms, seriously affecting drilling safety. Summary of the invention
[0004] In an embodiment of the present invention, a drilling overflow warning method is proposed to improve the accuracy of drilling overflow warning, help reduce the risk of drilling accidents, and ensure drilling safety, including:
[0005] Obtain drilling logging data;
[0006] The drilling logging data is input into the drilling overflow warning model, and the drilling overflow probability is output; the drilling overflow warning model is obtained by training a neural network model based on historical drilling logging data and historical drilling overflow data, and the neural network model is a structure in which a long short-term memory neural network is connected in parallel with a fully convolutional neural network, and then connected in series with a fully connected layer and an output layer;
[0007] If the drilling overflow probability exceeds the preset overflow threshold, a drilling overflow warning is issued.
[0008] In an embodiment of the present invention, a drilling overflow warning device is proposed to improve the accuracy of drilling overflow warning, help reduce the risk of drilling accidents, and ensure drilling safety, including:
[0009] A data acquisition module, used to acquire drilling and logging data;
[0010] A probability output module is used to input the drilling logging data into the drilling overflow warning model and output the drilling overflow probability; the drilling overflow warning model is obtained by training a neural network model based on historical drilling logging data and historical drilling overflow data, and the neural network model is a structure in which a long short-term memory neural network is connected in parallel with a fully convolutional neural network, and then connected in series with a fully connected layer and an output layer;
[0011] The overflow warning module is used to issue a drilling overflow warning if the drilling overflow probability exceeds a preset overflow threshold.
[0012] In an embodiment of the present invention, a computer device is proposed, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a drilling overflow early warning method when executing the computer program.
[0013] In an embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, a drilling overflow early warning method is implemented.
[0014] A computer program product is provided in an embodiment of the present invention. The computer program product includes a computer program. When the computer program is executed by a processor, a drilling overflow early warning method is implemented.
[0015] The drilling overflow warning method and device proposed in the embodiment of the present invention can solve the problem that the sensor in the prior art is difficult to accurately identify the drilling overflow signal, and relies on the observation and experience of the on-site operators, which is easy to cause false alarms or missed alarms, seriously affecting the drilling safety; the embodiment of the present invention obtains the drilling logging data; inputs the drilling logging data into the drilling overflow warning model, and outputs the drilling overflow probability; the drilling overflow warning model is obtained by training the neural network model according to the historical drilling logging data and the historical drilling overflow data, and the neural network model is a structure in which the long short-term memory neural network is connected in parallel with the full convolution neural network, and then connected in series with the full connection layer and the output layer; if the drilling overflow probability exceeds the preset overflow threshold, a drilling overflow warning is issued. The embodiment of the present invention does not need to use a sensor to identify the drilling overflow signal, realizes automatic drilling overflow warning, does not need to manually judge the drilling overflow situation, can reduce the number of warning false alarms and missed alarms, improves the accuracy of drilling overflow warning, helps to reduce the risk of drilling accidents, and ensures drilling safety. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0017] Figure 1 is a schematic flow chart of a drilling overflow early warning method in an embodiment of the present invention;
[0018] Figure 2 is a specific example diagram of a drilling overflow early warning method in an embodiment of the present invention;
[0019] Figure 3 is a specific example diagram of a drilling overflow early warning method in an embodiment of the present invention;
[0020] Figure 4 is a specific example diagram of a drilling overflow early warning method in an embodiment of the present invention;
[0021] Figure 5 is a specific example diagram of a drilling overflow early warning method in an embodiment of the present invention;
[0022] Figure 6 is a specific example diagram of a drilling overflow early warning method in an embodiment of the present invention;
[0023] Figure 7 is a schematic diagram of a drilling overflow warning device according to an embodiment of the present invention;
[0024] Figure 8 is a specific example diagram of a drilling overflow warning device in an embodiment of the present invention;
[0025] Fig. 9 Schematic diagram of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION
[0026] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.
[0027] The term "and / or" herein only describes an association relationship, indicating that three relationships may exist. For example, A and / or B may represent the following three situations: A exists alone, A and B exist at the same time, and B exists alone. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C may represent including any one or more elements selected from the set consisting of A, B, and C.
[0028] In the description of this specification, the terms "include", "including", "have", "contain", etc. are all open terms, which mean including but not limited to. The descriptions with reference to the terms "one embodiment", "a specific embodiment", "some embodiments", "for example", etc. mean that the specific features, structures or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. The order of steps involved in each embodiment is used to schematically illustrate the implementation of the present application, and the order of steps is not limited and can be appropriately adjusted as needed.
[0029] The principle and spirit of the present invention are explained in detail below with reference to several representative embodiments of the present invention.
[0030] Figure 1 FIG. 1 is a flow chart of a drilling overflow warning method according to an embodiment of the present invention. Figure 1 As shown, the method includes:
[0031] Step 101, obtaining drilling logging data;
[0032] Step 102, input the drilling logging data into the drilling overflow warning model, and output the drilling overflow probability; the drilling overflow warning model is obtained by training a neural network model based on historical drilling logging data and historical drilling overflow data, and the neural network model is a structure in which a long short-term memory neural network is connected in parallel with a fully convolutional neural network, and then connected in series with a fully connected layer and an output layer;
[0033] Step 103: If the drilling overflow probability exceeds a preset overflow threshold, a drilling overflow warning is issued.
[0034] Depend on Figure 1 It can be seen from the process shown that the embodiment of the present invention obtains drilling logging data; inputs the drilling logging data into the drilling overflow warning model, and outputs the drilling overflow probability; the drilling overflow warning model is obtained by training the neural network model based on historical drilling logging data and historical drilling overflow data, and the neural network model is a structure in which a long short-term memory neural network is connected in parallel with a full convolutional neural network, and then connected in series with a full connection layer and an output layer; if the drilling overflow probability exceeds a preset overflow threshold, a drilling overflow warning is issued. The embodiment of the present invention does not need to use a sensor to identify a drilling overflow signal, and realizes an automatic drilling overflow warning. There is no need to manually judge the drilling overflow situation, which can reduce the number of false alarms and missed alarms, improve the accuracy of drilling overflow warnings, and help reduce the risk of drilling accidents and ensure drilling safety.
[0035] In order to explain the above drilling overflow early warning method more clearly, each step is described in detail below.
[0036] Figure 2 It is a specific example diagram of the drilling overflow early warning method in an embodiment of the present invention.
[0037] In one embodiment of the present invention, reference Figure 2 , the drilling overflow warning model is trained and tested as follows:
[0038] Step 201, preprocessing historical drilling logging data;
[0039] Step 202, using the historical drilling logging data before and after preprocessing and the historical drilling overflow data as sample data to construct a training data set, a verification data set, and a test data set;
[0040] Step 203, training the neural network model using the training data set and the validation data set; wherein the historical drilling logging data before preprocessing is input into the long short-term memory neural network, and the historical drilling logging data after preprocessing is input into the full convolutional neural network;
[0041] Step 204, using the test data set to evaluate the trained neural network model, and determining the neural network model that passes the evaluation as the drilling overflow warning model.
[0042] Figure 3 , Figure 4 It is a specific example diagram of the drilling overflow early warning method in an embodiment of the present invention.
[0043] In one embodiment of the present invention, reference Figure 3 , a neural network model is established. The neural network model is a structure that first connects the long short-term memory neural network (LSTM) in parallel with the fully convolutional neural network (FCN), and then connects the fully connected layer and the output layer in series; among them, the FCN consists of three FCN modules, namely FCN module 1, FCN module 2, and FCN module 3. Figure 4 Each FCN module consists of two network layers: 1D convolution and batch normalization. After obtaining the drilling logging data, the drilling logging data is preprocessed, the drilling logging data before preprocessing is input into LSTM, and the drilling logging data after preprocessing is input into FCN. The outputs of LSTM and FCN are both multi-dimensional vectors. After the output results of LSTM and FCN are spliced, they enter the fully connected layer and the output layer to obtain the drilling overflow probability. If the drilling overflow probability exceeds the preset overflow threshold, a drilling overflow warning is issued.
[0044] In the specific implementation, the drilling logging data includes: inlet flow, outlet flow, inlet temperature, outlet temperature, inlet conductivity, outlet conductivity, inlet density, outlet density, total pool volume, total gas hydrocarbon value, and standpipe pressure, a total of 11 drilling logging parameters; in the process of building the neural network model, the LSTM contains 1 network layer, the input feature dimension is 11, and the output feature dimension is 64; the FCN is composed of three FCN modules, namely FCN module 1, FCN module 2, and FCN module 3. In the three FCN modules, the hidden layer output feature dimensions of the 1D convolution are 64, 128, and 64, respectively, and the 1D convolution kernel sizes are 8, 5, and 3, respectively; the input feature dimension of the output layer of the neural network model is 128, and the output feature dimension is 2. The activation function of the output layer can select the Softmax function or the Sigmoid function to obtain the drilling overflow probability.
[0045] In one embodiment of the present invention, the detailed method for preprocessing the historical drilling logging data is as follows:
[0046] Determine the Pearson product-moment correlation coefficient between historical drilling logging data with different characteristics; if the Pearson product-moment correlation coefficient is greater than or equal to a preset correlation threshold, determine the change between historical drilling logging data with different characteristics at the same time; if the Pearson product-moment correlation coefficient is less than the preset correlation threshold, determine the change of historical drilling logging data with the same characteristic at different times; for the pressure characteristic data in the historical drilling logging data, determine the relative change of the pressure characteristic data at different times.
[0047] In one embodiment of the present invention, the Pearson product-moment correlation coefficient between historical drilling logging data with different characteristics is determined, and the correlation between the historical drilling logging data with different characteristics is obtained through the Pearson product-moment correlation coefficient. The Pearson product-moment correlation coefficient between two historical drilling logging data with different characteristics is calculated according to the following formula:
[0048]
[0049] Where r is the Pearson product-moment correlation coefficient between two historical drilling logging data with different characteristics; T is the number of historical drilling logging data with different characteristics; x t ,y t They are the historical drilling logging data with different characteristics at time t; They are the average values of historical drilling logging data with different characteristics.
[0050] In one embodiment of the present invention, the historical drilling logging data includes: inlet flow rate, outlet flow rate, inlet temperature, outlet temperature, inlet conductivity, outlet conductivity, inlet density, outlet density, total pool volume, total gas hydrocarbon value, and riser pressure; the correlation between two historical drilling logging data with different characteristics is determined by the Pearson product-moment correlation coefficient; if the Pearson product-moment correlation coefficient is greater than or equal to a preset correlation threshold, the change between the historical drilling logging data with different characteristics at the same time is determined; if the Pearson product-moment correlation coefficient is less than the preset correlation threshold, the change of the historical drilling logging data with the same characteristic at different times is determined; for the pressure characteristic data in the historical drilling logging data, the relative change of the pressure characteristic data at different times is determined to improve the generalization ability of the neural network model.
[0051] In specific implementation, if the Pearson product-moment correlation coefficient is greater than or equal to the preset correlation threshold, it indicates that the correlation between the data is strong, and the variation between the historical drilling logging data with different characteristics at the same time is determined, that is, the characteristic axis data variation processing is adopted, and the historical drilling logging data processed by the characteristic axis data variation processing includes: inlet flow rate, outlet flow rate, inlet temperature, outlet temperature, inlet conductivity, outlet conductivity, inlet density, and outlet density; the variation between the historical drilling logging data with different characteristics at the same time is determined according to the following formula:
[0052] x new1 =x t -y t ;
[0053] Among them, x new1 Indicates the changes between historical drilling logging data with different characteristics at the same time, such as the changes between inlet flow and outlet flow, the changes between inlet temperature and outlet temperature, the changes between inlet conductivity and outlet conductivity, and the changes between inlet density and outlet density; x t ,y t They are the input data and output data in the historical drilling logging data at time t (current time) respectively;
[0054] If the Pearson product-moment correlation coefficient is less than the preset correlation threshold, the change amount of the historical drilling logging data with the same feature at different times is determined, that is, the time axis data change amount processing method is adopted. The historical drilling logging data processed by the time axis data change amount includes: total pool volume, total gas hydrocarbon value; the change amount of the historical drilling logging data with the same feature at different times is determined according to the following formula:
[0055] x new2 =x t -x t-1 ;
[0056] Among them, x new2Indicates the changes in historical drilling logging data of the same feature at different times, such as the changes in total pool volume at different times, and the changes in total gas hydrocarbon values at different times; x t-1 、x t The historical drilling logging data at time t-1 (previous time) and time t (current time) include: total pool volume, total gas hydrocarbon value;
[0057] For the pressure characteristic data in the historical drilling logging data, the relative changes of the pressure characteristic data at different times are determined, that is, the relative changes of the time axis data are processed. The historical drilling logging data processed by the relative changes of the time axis data include: standpipe pressure; the relative changes of the pressure characteristic data at different times are determined according to the following formula:
[0058]
[0059] Among them, x new3 Indicates the relative change of pressure characteristic data at different times, for example: the relative change of standpipe pressure at different times; x t-1 、x t They are the historical drilling logging data at time t-1 (previous moment) and time t (current moment), including: standpipe pressure.
[0060] In one embodiment of the present invention, after preprocessing the historical drilling logging data, the method further includes:
[0061] Determine the median, first quantile, and third quantile of the preprocessed historical drilling logging data; perform standardization on the preprocessed historical drilling logging data according to the median, first quantile, and third quantile of the preprocessed historical drilling logging data; wherein the standardization is to determine the ratio of the difference between the preprocessed historical drilling logging data and the median to the difference between the third quantile and the first quantile.
[0062] In specific implementation, the pre-processed historical drilling logging data is standardized according to the median, first quantile, and third quantile of the pre-processed historical drilling logging data according to the following formula:
[0063]
[0064] Among them, x scale is the historical drilling logging data after standardization; x is the historical drilling logging data after preprocessing; X median is the median of the preprocessed historical drilling logging data; X Q1 is the first quantile of the preprocessed historical drilling logging data; X Q3is the third quantile of the pre-processed historical drilling logging data. Optionally, the above formula is used to standardize the pre-processed historical drilling logging data.
[0065] Figure 5 It is a specific example diagram of the drilling overflow early warning method in an embodiment of the present invention.
[0066] In one embodiment of the present invention, reference Figure 5 , preprocess the historical drilling logging data, and then perform standardization processing, construct a training data set, a verification data set, and a test data set based on the historical drilling logging data, the standardized historical drilling data, and the historical drilling overflow data; use the training data set and the verification data set to train the neural network model, input the training data set into the initialized neural network model, train the parameters of the neural network model, and use the verification data set to verify the trained neural network model. When the verification result on the verification data set does not reach the optimal, continue to train the parameters of the neural network model; when the verification result on the verification data set is optimal, the training is completed; and use the test data set to evaluate the trained neural network model.
[0067] In one embodiment of the present invention, a trained neural network model is evaluated using a test data set, and the detailed process of determining the neural network model that passes the evaluation as a drilling overflow warning model is as follows: inputting the test data set into the trained neural network model to determine the accuracy and recall of the test data set; determining a weighted harmonic score of the test data set based on the accuracy and recall of the test data set; when the weighted harmonic score of the test data set is greater than a preset score threshold, determining the trained neural network model as a drilling overflow warning model.
[0068] In the specific implementation, the trained neural network model is evaluated using the test data set. The model evaluation indicators include: precision rate (PPV), recall rate (TPR), weighted harmonic score F2, receiver operating characteristic curve (ROC), area under the ROC curve (AUC) and prediction lead (FB); among them, the precision rate is determined according to the following formula:
[0069]
[0070] The recall rate is determined as follows:
[0071]
[0072] The weighted harmonic score is determined as follows:
[0073]
[0074] The forecast lead time is determined by the following formula:
[0075] FB=F Predict -F Label ;
[0076] Among them, TP is the number of positive labels (drilling overflows) predicted by the model; FP is the number of positive labels (drilling does not overflow) predicted by the model; FN is the number of negative labels (drilling overflows) predicted by the model; F Label and F Predict are the label frame number of the drilling overflow and the model prediction frame number, respectively, where F Label It is defined as the first predicted frame number when the model predicts the positive class for multiple consecutive times; the ROC curve is drawn by selecting different thresholds based on the precision rate PPV and the recall rate TPR; AUC is the area enclosed by the ROC curve and the two coordinate axes; for drilling overflow warning, the recall rate is more important than the precision rate, so F2 is used as the main evaluation indicator. When the weighted harmonic score F2 of the test data set is greater than the preset score threshold, the trained neural network model is determined as the drilling overflow warning model, and the prediction lead time and the area under the ROC curve are used as secondary evaluation indicators.
[0077] Figure 6 It is a specific example diagram of the drilling overflow early warning method in an embodiment of the present invention.
[0078] In one embodiment of the present invention, reference Figure 6 In order to improve the training efficiency and prediction efficiency of the neural network model, the historical drilling logging data is sliced in the form of a sliding window to ensure that the data X input into the neural network model each time is slide The same size, Among them, X slide Represents the number of data samples input into the neural network model in batches, represents the historical drilling logging data of the i-th feature at time t; W represents the size of the sliding window; M in the figure represents the number of features of the historical drilling logging data; T is the number of historical drilling logging data with different features, which represents the number of values of historical drilling logging data with different features at different times.
[0079] In one embodiment of the present invention, a data-driven method is used to train a neural network model. When the training data set and the validation data set are used to train the neural network model, the goal is to minimize the cross entropy loss function so that the cross entropy loss function converges. The cross entropy loss function is defined as:
[0080]
[0081] Where L is the cross entropy loss; N is the number of samples of historical drilling logging data; y i is the label of the sample data in the historical drilling logging data;i It represents the predicted probability of sample data in historical drilling logging data; the trained neural network model is verified by using the validation data set, the weighted harmonic score F2 of the validation data set is obtained, and the neural network model parameters with the largest weighted harmonic score F2 are recorded. When the weighted harmonic score F2 continues to decrease for 10 consecutive times, the training process is terminated, and the neural network model parameters with the largest weighted harmonic score F2 are exported and saved.
[0082] In one embodiment of the present invention, according to the neural network model completed after 50 experimental trainings, the average weighted harmonic score F2 is 0.9443, the average recall rate is 98.94%, the average precision rate is 79.89%, the average prediction advance is 226.13 frames, and the AUC average value is 0.92. Therefore, the drilling overflow warning model has high performance.
[0083] It should be noted that, although the operations of the method of the present invention are described in a specific order in the above embodiments and the accompanying drawings, this does not require or imply that the operations must be performed in the specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.
[0084] The implementation of the drilling overflow warning device can refer to the implementation of the above method, and the repeated parts will not be repeated. The term "module" or "unit" used below can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0085] Based on the same inventive concept, the present invention also proposes a drilling overflow warning device, such as Figure 7 As shown, the device comprises:
[0086] Data acquisition module 701, used to acquire drilling logging data;
[0087] The probability output module 702 is used to input the drilling logging data into the drilling overflow warning model and output the drilling overflow probability; the drilling overflow warning model is obtained by training a neural network model based on historical drilling logging data and historical drilling overflow data, and the neural network model is a structure in which a long short-term memory neural network is connected in parallel with a fully convolutional neural network, and then connected in series with a fully connected layer and an output layer;
[0088] The overflow warning module 703 is used to issue a drilling overflow warning if the drilling overflow probability exceeds a preset overflow threshold.
[0089] Figure 81 is a specific example diagram of the drilling overflow warning device in the embodiment of the present invention. Figure 8 As shown, in one embodiment of the present invention, Figure 7 The drilling overflow warning device also includes:
[0090] The model training module 801 is used to:
[0091] Preprocess historical drilling logging data;
[0092] The historical drilling logging data before and after preprocessing and the historical drilling overflow data are used as sample data to construct a training data set, a validation data set, and a test data set;
[0093] The neural network model is trained using the training data set and the validation data set; the historical drilling logging data before preprocessing is input into the long short-term memory neural network, and the historical drilling logging data after preprocessing is input into the full convolutional neural network;
[0094] The trained neural network model is evaluated using the test data set, and the neural network model that passes the evaluation is determined as the drilling overflow warning model.
[0095] In one embodiment of the present invention, the model training module 801 is specifically used to pre-process the historical drilling logging data in the following manner:
[0096] Determine the Pearson product-moment correlation coefficient between historical drilling logging data of different characteristics;
[0097] If the Pearson product-moment correlation coefficient is greater than or equal to a preset correlation threshold, the amount of change between historical drilling logging data with different characteristics at the same time is determined;
[0098] If the Pearson product-moment correlation coefficient is less than a preset correlation threshold, the change amount of the historical drilling logging data of the same feature at different times is determined;
[0099] For the pressure characteristic data in the historical drilling logging data, the relative changes of the pressure characteristic data at different times are determined.
[0100] In one embodiment of the present invention, the model training module 801 is further used for:
[0101] After preprocessing the historical drilling logging data, the median, first quantile and third quantile of the preprocessed historical drilling logging data are determined; according to the median, first quantile and third quantile of the preprocessed historical drilling logging data, the preprocessed historical drilling logging data is standardized; wherein the standardized processing is to determine the ratio of the difference between the preprocessed historical drilling logging data and the median to the difference between the third quantile and the first quantile.
[0102] In one embodiment of the present invention, the model training module 801 is specifically used for:
[0103] Input the test data set into the trained neural network model to determine the accuracy and recall of the test data set;
[0104] Determine the weighted reconciliation score of the test data set according to the precision and recall of the test data set;
[0105] When the weighted harmonic score of the test data set is greater than a preset score threshold, the trained neural network model is determined as a drilling overflow warning model.
[0106] It should be noted that although several modules of the drilling overflow warning device are mentioned in the above detailed description, this division is only exemplary and not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more modules described above can be embodied in one module. Conversely, the features and functions of one module described above can be further divided into multiple modules for embodiment.
[0107] Based on the above invention concept, Fig. 9 As shown, the present invention also proposes a computer device 900, including a memory 901, a processor 902, and a computer program 903 stored in the memory 901 and executable on the processor 902, wherein the processor 902 implements the aforementioned drilling overflow warning method when executing the computer program 903.
[0108] Based on the aforementioned inventive concept, the present invention proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the aforementioned drilling overflow warning method is implemented.
[0109] Based on the aforementioned inventive concept, the present invention proposes a computer program product, wherein the computer program product includes a computer program, and when the computer program is executed by a processor, a drilling overflow early warning method is implemented.
[0110] The drilling overflow warning method and device proposed in the embodiment of the present invention can solve the problem that the sensor in the prior art is difficult to accurately identify the drilling overflow signal, and relies on the observation and experience of the on-site operators, which is easy to cause false alarms or missed alarms, seriously affecting the drilling safety; the embodiment of the present invention obtains the drilling logging data; inputs the drilling logging data into the drilling overflow warning model, and outputs the drilling overflow probability; the drilling overflow warning model is obtained by training the neural network model according to the historical drilling logging data and the historical drilling overflow data, and the neural network model is a structure in which the long short-term memory neural network is connected in parallel with the full convolution neural network, and then connected in series with the full connection layer and the output layer; if the drilling overflow probability exceeds the preset overflow threshold, a drilling overflow warning is issued. The embodiment of the present invention does not need to use a sensor to identify the drilling overflow signal, realizes automatic drilling overflow warning, does not need to manually judge the drilling overflow situation, can reduce the number of warning false alarms and missed alarms, improves the accuracy of drilling overflow warning, helps to reduce the risk of drilling accidents, and ensures drilling safety.
[0111] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0112] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0113] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0115] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A drilling overflow early warning method, characterized in that: include: Obtain drilling logging data; The drilling logging data is input into the drilling overflow warning model to output the drilling overflow probability; The drilling overflow warning model is obtained by training a neural network model based on historical drilling logging data and historical drilling overflow data. The neural network model is a structure in which a long short-term memory neural network is connected in parallel with a fully convolutional neural network, and then connected in series with a fully connected layer and an output layer. If the drilling overflow probability exceeds the preset overflow threshold, a drilling overflow warning is issued.
2. The method according to claim 1, characterized in that The drilling overflow warning model is trained and tested as follows: Preprocess historical drilling logging data; The historical drilling logging data before and after preprocessing and the historical drilling overflow data are used as sample data to construct a training data set, a validation data set, and a test data set; The neural network model is trained using the training data set and the validation data set; the historical drilling logging data before preprocessing is input into the long short-term memory neural network, and the historical drilling logging data after preprocessing is input into the full convolutional neural network; The trained neural network model is evaluated using the test data set, and the neural network model that passes the evaluation is determined as the drilling overflow warning model.
3. The method according to claim 2, characterized in that Preprocess the historical drilling logging data, including: Determine the Pearson product-moment correlation coefficient between historical drilling logging data of different characteristics; If the Pearson product-moment correlation coefficient is greater than or equal to a preset correlation threshold, the amount of change between historical drilling logging data with different characteristics at the same time is determined; If the Pearson product-moment correlation coefficient is less than a preset correlation threshold, the change amount of the historical drilling logging data of the same feature at different times is determined; For the pressure characteristic data in the historical drilling logging data, the relative changes of the pressure characteristic data at different times are determined.
4. The method according to claim 2, characterized in that: After preprocessing the historical drilling logging data, it also includes: Determine the median, first quantile, and third quantile of the preprocessed historical drilling logging data; The preprocessed historical drilling logging data is standardized according to the median, the first quantile and the third quantile of the preprocessed historical drilling logging data; wherein the standardized processing is to determine the ratio of the difference between the preprocessed historical drilling logging data and the median to the difference between the third quantile and the first quantile.
5. The method according to claim 2, characterized in that: The trained neural network model is evaluated using the test data set, and the neural network model that passes the evaluation is determined as the drilling overflow warning model, including: Input the test data set into the trained neural network model to determine the accuracy and recall of the test data set; Determine the weighted reconciliation score of the test data set according to the precision and recall of the test data set; When the weighted harmonic score of the test data set is greater than a preset score threshold, the trained neural network model is determined as a drilling overflow warning model.
6. A drilling overflow warning device, characterized in that: include: A data acquisition module, used to acquire drilling and logging data; A probability output module is used to input the drilling logging data into the drilling overflow warning model and output the drilling overflow probability; the drilling overflow warning model is obtained by training a neural network model based on historical drilling logging data and historical drilling overflow data, and the neural network model is a structure in which a long short-term memory neural network is connected in parallel with a fully convolutional neural network, and then connected in series with a fully connected layer and an output layer; The overflow warning module is used to issue a drilling overflow warning if the drilling overflow probability exceeds a preset overflow threshold.
7. The device according to claim 6, characterized in that Also includes: Model training module, used to: Preprocess historical drilling logging data; The historical drilling logging data before and after preprocessing and the historical drilling overflow data are used as sample data to construct training data sets, verification data sets, and test data sets; The neural network model is trained using the training data set and the validation data set; the historical drilling logging data before preprocessing is input into the long short-term memory neural network, and the historical drilling logging data after preprocessing is input into the full convolutional neural network; The trained neural network model is evaluated using the test data set, and the neural network model that passes the evaluation is determined as the drilling overflow warning model.
8. The device according to claim 7, characterized in that The model training module is specifically used to preprocess historical drilling and logging data as follows: Determine the Pearson product-moment correlation coefficient between historical drilling logging data of different characteristics; If the Pearson product-moment correlation coefficient is greater than or equal to a preset correlation threshold, the amount of change between historical drilling logging data with different characteristics at the same time is determined; If the Pearson product-moment correlation coefficient is less than a preset correlation threshold, the change amount of the historical drilling logging data of the same feature at different times is determined; For the pressure characteristic data in the historical drilling logging data, the relative changes of the pressure characteristic data at different times are determined.
9. The device according to claim 7, characterized in that The model training module is also used to: After preprocessing the historical drilling logging data, the median, first quantile and third quantile of the preprocessed historical drilling logging data are determined; according to the median, first quantile and third quantile of the preprocessed historical drilling logging data, the preprocessed historical drilling logging data is standardized; wherein the standardized processing is to determine the ratio of the difference between the preprocessed historical drilling logging data and the median to the difference between the third quantile and the first quantile.
10. The device according to claim 7, characterized in that The model training module is specifically used for: Input the test data set into the trained neural network model to determine the accuracy and recall of the test data set; Determine the weighted reconciliation score of the test data set according to the precision and recall of the test data set; When the weighted harmonic score of the test data set is greater than a preset score threshold, the trained neural network model is determined as a drilling overflow warning model.
11. A computer 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, the method according to any one of claims 1 to 5 is implemented.
12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.
13. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.
