Drilling fluid leakage estimation method and device, computer equipment and storage medium
Through seismic data analysis and machine learning algorithms, a drilling leakage prediction model is constructed, which solves the problem of inaccurate prediction of drilling fluid leakage in the existing technology, achieves more accurate estimation of leakage volume and risk degree, and improves the targetedness of leakage prevention measures. .
Patent Information
- Application Number
- CN202311014381.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-08-11
- Publication Date
- 2025-02-18
AI Technical Summary
The prior art is difficult to accurately predict and quantify the amount and risk of drilling fluid leakage, resulting in uncertainty in the implementation of leak prevention measures.
By collecting seismic data, multiple fault characterization attributes are extracted, and correlation analysis is performed with leakage information, the optimal fault characterization attributes are selected, and the prediction parameters are regressed and integrated with machine learning algorithms to construct a drilling leakage prediction model.
It improves the accuracy of drilling fluid leakage prediction, reduces the prediction result error due to regional differences, and can more accurately estimate the leakage amount and risk level, thereby helping drillers take leak prevention measures in advance.
Smart Images
Figure CN119475949A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of petroleum engineering, and in particular to a method, a device, a computer device and a storage medium for estimating the amount of drilling fluid loss. Background Art
[0002] Well leakage is a common underground complex situation during drilling operations. Severe leakage will cause wellbore instability and may be accompanied by a series of drilling problems such as pipe sticking and overflow, resulting in increased drilling costs and huge safety hazards. For some work areas with complex geological structures, the main cause of serious well leakage accidents is the high degree of formation fracture development. The existence of fracture zones changes the mechanical state of the rock around the well and provides a natural leakage channel for underground liquid. Compared with passive treatment of well leakage accidents, the most effective way to prevent and control leakage is to predict and warn of leakage accidents in advance, take targeted preventive measures in advance, and reduce the impact of accidents.
[0003] At present, the advance prediction methods for leakage accidents are all based on seismic data. By calculating the seismic fracture attributes such as ant bodies, the location of the drilled cracks is located to warn of leakage risks. However, the correlation between seismic attributes and leakage risks and the occurrence of complex situations is not clear, and there are regional differences. The reference for leakage warning based solely on the abnormal location of attributes is not strong. At the same time, this method only provides potential risk locations, but cannot quantify the amount of leakage and the severity of leakage risks, which hinders the implementation of targeted leakage prevention measures. Summary of the invention
[0004] Based on this, it is necessary to provide a drilling fluid loss estimation method, device, computer equipment and storage medium to address the above technical problems.
[0005] A method for estimating drilling fluid loss, comprising:
[0006] Collecting seismic data, extracting multiple fault characterization attributes from the seismic data, and collecting missing information of the work area;
[0007] Detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute;
[0008] Calculating multiple effective factor data leading to lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters;
[0009] The prediction parameters are regressed and integrated using a machine learning algorithm to construct a drilling loss prediction model.
[0010] In one embodiment, the step of detecting the correlation between each of the fracture characterization attributes and the missing information and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute comprises:
[0011] Calculate the loss probability when each of the fracture characterization attributes is abnormal according to the loss information to obtain a plurality of first probabilities, and calculate the loss probability when each of the fracture characterization attributes is normal according to the loss information to obtain a plurality of second probabilities, wherein the loss probability reflects the correlation between the fracture characterization attribute and the loss information;
[0012] The fracture characterization attribute with the largest first probability and the smallest second probability is selected as the optimal fracture characterization attribute.
[0013] In one embodiment, the step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model includes:
[0014] The prediction parameters are regressed and integrated using a machine learning algorithm, a mapping relationship between the leakage volume and the prediction parameters is established, and the drilling leakage prediction model is constructed.
[0015] In one embodiment, the step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model further includes:
[0016] Extract the values of multiple influencing factors of drilling loss along the trajectory of the well to be drilled;
[0017] The numerical values of the influencing factors are introduced into the drilling loss prediction model for calculation to obtain the loss prediction result.
[0018] In one embodiment, the prediction parameters include relative fracture strength, apparent fracture length and pressure difference.
[0019] In one embodiment, the step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model includes:
[0020] Using multiple machine learning algorithms, respectively regress and integrate the prediction parameters to construct candidate prediction models;
[0021] Calculating the missing prediction results respectively using each of the candidate prediction models;
[0022] The average error of the leakage prediction results of each of the candidate prediction models is detected, and the candidate prediction model with the smallest average error is selected as the drilling leakage prediction model.
[0023] In one embodiment, the machine learning algorithm includes a support vector machine algorithm, a neural network algorithm, an Xgboost algorithm and a multi-element linear regression algorithm.
[0024] A drilling fluid loss estimation device, comprising:
[0025] A characterization attribute extraction module, used to collect seismic data, extract multiple fracture characterization attributes from the seismic data, and collect missing information of the work area;
[0026] An optimal characterization attribute selection module, used for detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute;
[0027] A prediction parameter acquisition module, used to calculate multiple effective factor data causing lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters;
[0028] The model building module is used to utilize a machine learning algorithm to perform regression integration on the prediction parameters and build a drilling loss prediction model.
[0029] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:
[0030] Collecting seismic data, extracting multiple fault characterization attributes from the seismic data, and collecting missing information of the work area;
[0031] Detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute;
[0032] Calculating multiple effective factor data leading to lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters;
[0033] The prediction parameters are regressed and integrated using a machine learning algorithm to construct a drilling loss prediction model.
[0034] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:
[0035] Collecting seismic data, extracting multiple fault characterization attributes from the seismic data, and collecting missing information of the work area;
[0036] Detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute;
[0037] Calculating multiple effective factor data leading to lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters;
[0038] The prediction parameters are regressed and integrated using a machine learning algorithm to construct a drilling loss prediction model.
[0039] The above-mentioned drilling fluid loss estimation method, device, computer equipment and storage medium comprehensively consider the fracture characterization attributes that are highly correlated with the loss information for calculation, thereby obtaining prediction parameters, and using machine learning and data mining ideas to establish a drilling loss prediction model, thereby improving the prediction accuracy and avoiding prediction result errors caused by regional differences. This allows drillers to take measures in advance to actively prevent the occurrence of well leakage accidents based on the prediction results of the formation to be drilled, thereby reducing the complex cycle and ensuring safe well completion. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 A schematic diagram of a flow chart of a method for estimating drilling fluid loss in one embodiment;
[0041] Figure 2 is a structural block diagram of a drilling fluid loss estimation device in one embodiment;
[0042] Figure 3 is an internal structure diagram of a computer device in one embodiment;
[0043] Figure 4 A schematic diagram of earthquake fracture properties of an ant body passing through two leakage points in an embodiment;
[0044] Figure 5 It is a schematic diagram of the estimation result of the formation leakage of the test well A to be drilled in one embodiment. DETAILED DESCRIPTION
[0045] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0046] Embodiment 1
[0047] In this embodiment, Figure 1 As shown, a method for estimating the amount of drilling fluid loss is provided, which includes:
[0048] Step 110 , collect seismic data, extract multiple fault characterization attributes from the seismic data, and collect missing information of the work area.
[0049] In this embodiment, the fracture characterization attribute may also be referred to as the fracture characterization seismic attribute. The leakage information includes leakage location information and leakage amount information. In this embodiment, the post-stack seismic data of the completed wells in the work area are collected, and a variety of fracture characterization seismic attributes are extracted, and the leakage location and leakage amount information of the work area are collected.
[0050] Step 120 , detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute.
[0051] In this embodiment, the correlation is used to indicate the influence of each of the fracture characterization attributes on the leakage. The greater the correlation, the greater the influence of the fracture characterization attribute on the leakage, and the smaller the correlation, the smaller the influence of the fracture characterization attribute on the leakage. In this embodiment, the correlations between multiple fracture characterization seismic attributes and the leakage information are compared, and the optimal fracture characterization attribute is selected using the correlation.
[0052] In one embodiment, the fracture characterization attributes include seismic coherence volume characterization seismic attributes, ant volume characterization seismic attributes and curvature characterization seismic attributes.
[0053] Step 130: Calculate data of various effective factors causing lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters.
[0054] In this embodiment, the effective factor data is the data that has the main influence on the drilling loss. The prediction parameter is the specific value of the effective factor data.
[0055] In this embodiment, the optimal fracture characterization attribute is used to calculate various effective factor data corresponding to the missing events in the missing information, thereby obtaining prediction parameters.
[0056] In one embodiment, the prediction parameters include relative fracture strength, apparent fracture length and pressure difference.
[0057] It is worth mentioning that the main factors affecting drilling losses include: drilling through holes or porous formations; drilling through fracture zones or faults; the difference between the equivalent drilling fluid density ECD of the annulus and the pore pressure, that is, the pressure difference is too large. Therefore, in this embodiment, the relative strength of the fracture, the apparent length of the fracture and the pressure difference are selected as the prediction parameters for estimating the amount of drilling fluid loss based on these factors. In this embodiment, the multiple effective factor data include the relative strength of the fracture, the apparent length of the fracture and the pressure difference, and the prediction parameters are the specific values of the relative strength of the fracture, the apparent length of the fracture and the pressure difference.
[0058] Step 140, using a machine learning algorithm to perform regression integration on the prediction parameters to construct a drilling loss prediction model.
[0059] In this embodiment, a machine learning algorithm is used to regress and integrate the prediction parameters that cause well leakage, so as to construct a drilling leakage prediction model that can predict the leakage amount of the formation to be drilled.
[0060] In one embodiment, the prediction parameters are regressed and integrated using a machine learning algorithm to establish a mapping relationship between the loss volume and the prediction parameters, and to construct the drilling loss prediction model. In this embodiment, the prediction parameters are regressed and integrated to establish a mapping relationship between the drilling loss volume and the prediction parameters. The mapping relationship reflects the relationship between various prediction parameters and the drilling loss volume, and reflects the degree of influence of various prediction parameters on the drilling loss volume. By establishing the mapping relationship, a drilling loss prediction model is constructed.
[0061] In one embodiment, the machine learning algorithm includes a support vector machine algorithm, a neural network algorithm, an Xgboost algorithm, and a multi-element linear regression algorithm.
[0062] In the above embodiment, the fracture characterization attributes with a greater correlation with the leakage information are comprehensively considered for calculation to obtain prediction parameters, and a drilling leakage prediction model is established by using machine learning and data mining ideas, thereby improving the prediction accuracy and avoiding prediction result errors caused by regional differences.
[0063] In one embodiment, the step of detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the largest correlation as the optimal fracture characterization attribute includes: calculating the loss probability when each of the fracture characterization attributes is abnormal according to the missing information, and obtaining multiple first probabilities; calculating the loss probability when each of the fracture characterization attributes is normal according to the missing information, and obtaining multiple second probabilities, wherein the loss probability reflects the correlation between the fracture characterization attribute and the missing information; and selecting the fracture characterization attribute with the largest first probability and the smallest second probability as the optimal fracture characterization attribute.
[0064] In this embodiment, when a fracture characterization attribute is abnormal, the probability of leakage is relatively high, and when the fracture characterization attribute is normal (no abnormality), the probability of leakage is relatively low, indicating that the fracture characterization attribute has a greater impact on the occurrence of leakage, and the fracture characterization attribute has a high correlation with the leakage information. Therefore, in this embodiment, the fracture characterization attribute with the highest probability of leakage when abnormal and the lowest probability of leakage when normal is selected as the optimal fracture characterization attribute, which has the highest correlation with the occurrence of leakage and the highest degree of influence on leakage.
[0065] In one embodiment, three fault characterization attributes, namely, the seismic coherence body, ant body and curvature of the lost point, are calculated, and the ability of the three fault characterization attributes to indicate lost events is judged. It is calculated that the ant body attribute shows better correlation than the other two attributes. When the ant body attribute is abnormal, the probability of lost is 72%, and when the ant body attribute is normal, the probability of lost is 18%. Therefore, the ant body attribute is selected as the optimal fracture characterization seismic attribute.
[0066] In one embodiment, the step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model also includes: extracting multiple values of influencing factors of drilling loss along the trajectory to be drilled; importing the values of each of the influencing factors into the drilling loss prediction model for calculation to obtain a loss prediction result.
[0067] In this embodiment, the values of multiple influencing factors of drilling loss are extracted along the trajectory of the well to be drilled, and substituted into the mapping relationship of the constructed drilling loss prediction model, so as to calculate the predicted value of the loss amount of the formation to be drilled. This allows the drilling personnel to take measures in advance to actively prevent the occurrence of well leakage accidents based on the prediction results of the formation to be drilled, thereby reducing the complex cycle and ensuring safe well completion.
[0068] In one embodiment, the step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model includes: using multiple machine learning algorithms to regress and integrate the prediction parameters respectively to construct candidate prediction models respectively; using each of the candidate prediction models to calculate the loss prediction results respectively; detecting the average error of the loss prediction results of each of the candidate prediction models, and selecting the candidate prediction model with the smallest average error as the drilling loss prediction model.
[0069] In this embodiment, different drilling loss prediction models have different prediction accuracies. By comparing the average prediction errors of the candidate prediction models, a prediction model with higher accuracy can be selected.
[0070] In one embodiment, the support vector machine algorithm, neural network algorithm, Xgboost algorithm and multi-element linear regression algorithm are used to regress and integrate the three effective factors leading to well leakage to build a drilling loss prediction model for the loss amount. The average errors of the four drilling loss prediction models are 34m 3 , 51m 3 , 58m 3 , 59m 3 Among the average prediction errors, the average error of the drilling loss prediction model corresponding to the support vector machine algorithm is the smallest. Therefore, the support vector machine algorithm is selected as the optimal prediction algorithm, and the drilling loss prediction model corresponding to the support vector machine algorithm is selected as the drilling loss prediction model.
[0071] Embodiment 2
[0072] In this embodiment, in order to solve the problem that the existing leakage risk prediction method based on seismic data only provides potential risk locations and cannot quantitatively evaluate the leakage volume and risk level, by effectively mining and utilizing a large amount of extremely valuable prior information on completed drilling risks, the main factors of well leakage risk are fully considered. On the basis of a series of methods for predicting leakage risk based on seismic attributes, the influence of pressure difference factors is focused on, and an advance estimation method for leakage volume is established, which guides drillers to take measures in advance to actively prevent the occurrence of well leakage accidents, thereby reducing complex cycles and ensuring safe well completion.
[0073] Therefore, this embodiment provides a method for calculating the amount of drilling fluid loss. Compared with the prior art, this embodiment optimizes the main control factors of the loss, makes full use of seismic and engineering data, and realizes the advance estimation of the amount of drilling fluid loss. This method includes the following steps:
[0074] (1) The main factors affecting drilling losses include:
[0075] ① Drilling through holes or porous formations;
[0076] ②Drilling through fracture zones or faults;
[0077] ③ The difference between the annular equivalent drilling fluid density ECD and the pore pressure and the pressure difference are too large
[0078] Based on these factors, the present invention selects the relative strength of the fracture, the apparent length of the fracture and the pressure difference as the prediction parameters for the advance estimation of the drilling fluid loss.
[0079] (2) Select the optimal seismic attribute representing the fault of the leakage event in the work area to estimate the relative strength and apparent length of the fracture along the drilling trajectory: collect the post-stack seismic data around the completed wells in the work area and extract several types of seismic attributes representing the fault. Collect the information on the leakage location and leakage amount in the work area, compare them, and use the prediction accuracy to select the optimal seismic attribute representing the fault.
[0080] (3) Prepare forecasting technical data:
[0081] Using the optimal fracture characterization seismic attribute optimization result in step (2), calculate the relative strength, apparent length and pressure difference of the fracture corresponding to the leakage event in the selected work area.
[0082] (4) Establish a mapping relationship between the amount of loss and the predictive factors:
[0083] Different machine learning algorithms are optimized to regress and integrate the data of three effective factors that cause lost circulation, so as to construct an effective prediction relationship for the amount of lost circulation.
[0084] (5) Prediction of leakage in new drilling:
[0085] The values of the main influencing factors of drilling loss are extracted along the trajectory of the well to be drilled, and are brought into the loss mapping relationship established in step (4) to obtain the predicted value of the loss amount of the formation to be drilled.
[0086] Embodiment 3
[0087] In this embodiment, a block in the northwest is used as the test area, and 11 well data and related seismic data are obtained as the test basis, including 32 lost points. In this test case, most of the lost points come from the same stratum. The three fault characterization attributes of the lost point seismic coherence body, ant body and curvature are calculated, and their ability to indicate lost events is judged. Finally, the ant body attribute shows better correlation than the other attributes. The probability of loss when the attribute is abnormal is 72%, and the probability of loss when the attribute is normal is 18%. The ant body is selected as the optimal fault characterization seismic attribute.
[0088] Based on the calculation results of the ant body properties at the leakage point, the relative strength and apparent length of the crack are estimated. Figure 4 As shown in the figure, the seismic attribute map of the ant body passing through two leakage points is estimated to be 0.38 and 0.42, and the apparent length of the fracture is 55 meters and 60 meters respectively. Then the pressure difference information of all leakage points is calculated, that is, the difference between the annular equivalent drilling fluid density ECD and the pore pressure of the corresponding leakage point.
[0089] The support vector machine algorithm, neural network algorithm, Xgboost algorithm and multi-prime linear regression algorithm were used to regress and integrate the three effective factors leading to well leakage to build an effective prediction relationship for the leakage amount. The average prediction errors of these four machine learning algorithms were 34m 3 , 51m 3 , 58m 3 , 59m 3 , select the support vector machine algorithm as the optimal prediction algorithm.
[0090] A new well A in the work area was tested. The three effective factor data around the well were brought into the machine learning model obtained in the previous step to predict the formation leakage of the well. Figure 5 shown.
[0091] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.
[0092] Embodiment 4
[0093] In this embodiment, Figure 2 As shown, a drilling fluid loss estimation device is provided, comprising:
[0094] Characterization attribute extraction module 210, used to collect seismic data, extract multiple fault characterization attributes from the seismic data, and collect missing information of the work area;
[0095] An optimal characterization attribute selection module 220 is used to detect the correlation between each of the fracture characterization attributes and the missing information, and select the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute;
[0096] A prediction parameter acquisition module 230 is used to calculate multiple effective factor data causing lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters;
[0097] The model building module 240 is used to utilize a machine learning algorithm to perform regression integration on the prediction parameters to build a drilling loss prediction model.
[0098] In one embodiment, the optimal characterization attribute selection module includes:
[0099] a loss probability calculation unit, configured to calculate the loss probability when each of the fracture characterization attributes is abnormal according to the loss information to obtain a plurality of first probabilities, and to calculate the loss probability when each of the fracture characterization attributes is normal according to the loss information to obtain a plurality of second probabilities, wherein the loss probability reflects the correlation between the fracture characterization attribute and the loss information;
[0100] The optimal characterization attribute selection unit is used to select the fracture characterization attribute with the maximum first probability and the minimum second probability as the optimal fracture characterization attribute.
[0101] In one embodiment, the model building module is further used to utilize a machine learning algorithm to perform regression integration on the prediction parameters, establish a mapping relationship between the leakage volume and the prediction parameters, and construct the drilling leakage prediction model.
[0102] In one embodiment, the drilling fluid loss estimation device further comprises:
[0103] An influencing factor value acquisition module is used to extract multiple influencing factor values of drilling loss along the trajectory to be drilled;
[0104] The leakage prediction result calculation module is used to import the numerical values of each of the influencing factors into the drilling leakage prediction model for calculation to obtain the leakage prediction result.
[0105] In one embodiment, the prediction parameters include relative fracture strength, apparent fracture length and pressure difference.
[0106] In one embodiment, the model building module includes:
[0107] The prediction model building units are used to use multiple machine learning algorithms to regress and integrate the prediction parameters and build the prediction models to be selected respectively;
[0108] A leak prediction result calculation unit, used to calculate the leak prediction result using each of the candidate prediction models;
[0109] The leakage prediction model selection unit is used to detect the average error of the leakage prediction results of each of the candidate prediction models, and select the candidate prediction model with the smallest average error as the drilling leakage prediction model.
[0110] In one embodiment, the machine learning algorithm includes a support vector machine algorithm, a neural network algorithm, an Xgboost algorithm, and a multi-element linear regression algorithm.
[0111] The specific definition of the drilling fluid loss estimation device can be found in the definition of the drilling fluid loss estimation method mentioned above, which will not be repeated here. Each unit in the above-mentioned drilling fluid loss estimation device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned units can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned units.
[0112] Embodiment 5
[0113] In this embodiment, a computer device is provided. Its internal structure diagram can be shown as follows: Figure 3As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program, and the non-volatile storage medium is deployed with a database, which is used for seismic data, data of a prediction model, etc. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with other computer devices that deploy application software. When the computer program is executed by the processor, a method for estimating the amount of drilling fluid loss is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.
[0114] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0115] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:
[0116] Collecting seismic data, extracting multiple fault characterization attributes from the seismic data, and collecting missing information of the work area;
[0117] Detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute;
[0118] Calculating multiple effective factor data leading to lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters;
[0119] The prediction parameters are regressed and integrated using a machine learning algorithm to construct a drilling loss prediction model.
[0120] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0121] Calculate the loss probability when each of the fracture characterization attributes is abnormal according to the loss information to obtain a plurality of first probabilities, and calculate the loss probability when each of the fracture characterization attributes is normal according to the loss information to obtain a plurality of second probabilities, wherein the loss probability reflects the correlation between the fracture characterization attribute and the loss information;
[0122] The fracture characterization attribute with the largest first probability and the smallest second probability is selected as the optimal fracture characterization attribute.
[0123] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0124] The prediction parameters are regressed and integrated using a machine learning algorithm, a mapping relationship between the leakage volume and the prediction parameters is established, and the drilling leakage prediction model is constructed.
[0125] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0126] Extract the values of multiple influencing factors of drilling loss along the trajectory of the well to be drilled;
[0127] The numerical values of the influencing factors are introduced into the drilling loss prediction model for calculation to obtain the loss prediction result.
[0128] In one embodiment, the prediction parameters include relative fracture strength, apparent fracture length and pressure difference.
[0129] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0130] Using multiple machine learning algorithms, respectively regress and integrate the prediction parameters to construct candidate prediction models;
[0131] Calculating the missing prediction results respectively using each of the candidate prediction models;
[0132] The average error of the leakage prediction results of each of the candidate prediction models is detected, and the candidate prediction model with the smallest average error is selected as the drilling leakage prediction model.
[0133] In one embodiment, the machine learning algorithm includes a support vector machine algorithm, a neural network algorithm, an Xgboost algorithm, and a multi-element linear regression algorithm.
[0134] Embodiment 6
[0135] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:
[0136] Collecting seismic data, extracting multiple fault characterization attributes from the seismic data, and collecting missing information of the work area;
[0137] Detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute;
[0138] Calculating multiple effective factor data leading to lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters;
[0139] The prediction parameters are regressed and integrated using a machine learning algorithm to construct a drilling loss prediction model.
[0140] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0141] Calculate the loss probability when each of the fracture characterization attributes is abnormal according to the loss information to obtain a plurality of first probabilities, and calculate the loss probability when each of the fracture characterization attributes is normal according to the loss information to obtain a plurality of second probabilities, wherein the loss probability reflects the correlation between the fracture characterization attribute and the loss information;
[0142] The fracture characterization attribute with the largest first probability and the smallest second probability is selected as the optimal fracture characterization attribute.
[0143] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0144] The prediction parameters are regressed and integrated using a machine learning algorithm, a mapping relationship between the leakage volume and the prediction parameters is established, and the drilling leakage prediction model is constructed.
[0145] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0146] Extract the values of multiple influencing factors of drilling loss along the trajectory of the well to be drilled;
[0147] The numerical values of the influencing factors are introduced into the drilling loss prediction model for calculation to obtain the loss prediction result.
[0148] In one embodiment, the prediction parameters include relative fracture strength, apparent fracture length and pressure difference.
[0149] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0150] Using multiple machine learning algorithms, respectively regress and integrate the prediction parameters to construct candidate prediction models;
[0151] Calculating the missing prediction results respectively using each of the candidate prediction models;
[0152] The average error of the leakage prediction results of each of the candidate prediction models is detected, and the candidate prediction model with the smallest average error is selected as the drilling leakage prediction model.
[0153] In one embodiment, the machine learning algorithm includes a support vector machine algorithm, a neural network algorithm, an Xgboost algorithm and a multi-element linear regression algorithm.
[0154] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0155] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0156] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.
Claims
1. A method for estimating drilling fluid loss, characterized in that: include: Collecting seismic data, extracting multiple fault characterization attributes from the seismic data, and collecting missing information of the work area; Detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute; Calculating multiple effective factor data leading to lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters; The prediction parameters are regressed and integrated using a machine learning algorithm to construct a drilling loss prediction model.
2. The method according to claim 1, characterized in that The step of detecting the correlation between each of the fracture characterization attributes and the missing information and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute comprises: Calculate the loss probability when each of the fracture characterization attributes is abnormal according to the loss information to obtain a plurality of first probabilities, and calculate the loss probability when each of the fracture characterization attributes is normal according to the loss information to obtain a plurality of second probabilities, wherein the loss probability reflects the correlation between the fracture characterization attribute and the loss information; The fracture characterization attribute with the largest first probability and the smallest second probability is selected as the optimal fracture characterization attribute.
3. The method according to claim 1, characterized in that The step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model includes: The prediction parameters are regressed and integrated using a machine learning algorithm, a mapping relationship between the leakage volume and the prediction parameters is established, and the drilling leakage prediction model is constructed.
4. The method according to claim 1, characterized in that: After the step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model, the following steps are further included: Extract the values of multiple influencing factors of drilling loss along the trajectory of the well to be drilled; The numerical values of the influencing factors are introduced into the drilling loss prediction model for calculation to obtain the loss prediction result.
5. The method according to claim 1, characterized in that The prediction parameters include relative crack strength, apparent crack length and pressure difference.
6. The method according to any one of claims 1 to 5, characterized in that: The step of using a machine learning algorithm to regress and integrate the prediction parameters to construct a drilling loss prediction model includes: Using multiple machine learning algorithms, respectively regress and integrate the prediction parameters to construct candidate prediction models; Calculating the missing prediction results respectively using each of the candidate prediction models; The average error of the leakage prediction results of each of the candidate prediction models is detected, and the candidate prediction model with the smallest average error is selected as the drilling leakage prediction model.
7. The method according to claim 6, characterized in that The machine learning algorithms include support vector machine algorithm, neural network algorithm, Xgboost algorithm and multi-primitive linear regression algorithm.
8. A drilling fluid loss estimation device, characterized in that: include: A characterization attribute extraction module, used to collect seismic data, extract multiple fracture characterization attributes from the seismic data, and collect missing information of the work area; An optimal characterization attribute selection module, used for detecting the correlation between each of the fracture characterization attributes and the missing information, and selecting the fracture characterization attribute with the greatest correlation as the optimal fracture characterization attribute; A prediction parameter acquisition module, used to calculate multiple effective factor data causing lost circulation based on the optimal fracture characterization attribute to obtain prediction parameters; The model building module is used to utilize a machine learning algorithm to perform regression integration on the prediction parameters and build a drilling loss prediction model.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.