Drill bit optimization method, device, equipment and medium
By integrating multi-source data through machine learning algorithms, the performance of drill bits in specific formations can be predicted, solving the problem of existing drill bit selection relying on experience. This achieves standardization and accuracy in drill bit selection, improving drilling efficiency and cost control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-27
- Publication Date
- 2026-04-07
AI Technical Summary
Existing drill bit selection methods rely on engineers' experience, lack data integration, cannot predict the actual performance of drill bits in specific formations, and fail to model the dynamic interaction between formation and drill bit as a whole, making it difficult to optimize drilling efficiency and cost.
A drill bit rock-breaking performance prediction model is established based on machine learning algorithms. Geological, engineering and drill bit product data are integrated. The model is trained to predict mechanical drilling speed and drill bit footage. A weighted score is performed using preset scoring rules, and the optimal drill bit result is output.
It has achieved standardization and precision in drill bit selection, reduced the uncertainty of subjective judgment in new blocks or complex formations, improved the accuracy of predicting mechanical drilling speed and drill bit footage, and met diverse engineering needs.
Smart Images
Figure CN121808457A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of drilling optimization design technology, specifically to a drill bit selection method, apparatus, equipment, and medium. Background Technology
[0002] In oil and gas drilling operations, the drill bit is a crucial tool that directly interacts with the formation and breaks up the rock. The quality of drill bit selection directly affects drilling efficiency, drilling costs, and operational safety. Traditional drill bit selection methods mainly rely on engineers' experience, simple rock mechanics parameter analysis, and records of drill bit usage in adjacent wells. However, these methods have significant limitations. They heavily rely on engineers' personal experience, making standardization and knowledge transfer difficult, and introducing high uncertainty when facing new blocks or complex formations. Data utilization is insufficient, typically considering only a limited number of parameters and failing to fully integrate massive amounts of logging data, formation lithology information, and drill bit rock-breaking structural parameters. Furthermore, existing methods often compare the formation and the drill bit, lacking a holistic model of the dynamic interaction system of formation, drilling parameters, and drill bit structure, making it impossible to predict the true performance of the drill bit under specific formation conditions. Existing methods often consider mechanical drilling rate or per-trip footage as the sole or isolated evaluation criterion, ignoring the constraints of mechanical drilling rate and per-trip footage. Summary of the Invention
[0003] This application provides a method, apparatus, equipment, and medium for selecting drill bits, which can solve the technical problem of inaccurate drill bit selection in the prior art.
[0004] In a first aspect, embodiments of this application provide a method for selecting the best drill bit, the method comprising: A drill bit rock breaking performance prediction model was established based on machine learning algorithms. A historical dataset including geological data, engineering data and drill bit product data was constructed to train the drill bit rock breaking performance prediction model. Input the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock breaking performance prediction model to obtain the predicted engineering parameters of each candidate drill bit, including the predicted mechanical drilling rate and the predicted drill bit footage. Based on preset scoring rules, the predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit are weighted and scored to obtain the predicted comprehensive score of each candidate drill bit, so as to output the drill bit selection result.
[0005] In conjunction with the first aspect, in one implementation, the step of establishing a drill bit rock-breaking performance prediction model based on machine learning algorithms, and training the drill bit rock-breaking performance prediction model using a historical dataset including geological data, engineering data, and drill bit product data, includes: A drill bit rock-breaking performance prediction model was established based on the AutoML framework; Acquire historical datasets including geological data, engineering data, and drill bit product data, and align the geological data, engineering data, and drill bit product data by well depth to form a sample set; Based on the sample set, geological data and drill bit product data are constructed as input feature vectors, and mechanical drilling speed and drill bit footage in engineering data are used as output labels. Construct a configuration space that includes feature selection space, hyperparameter space, and model architecture space; Using the input feature vector as training input and mechanical drilling rate and drill bit footage as prediction targets, an optimizer searches for the optimal model configuration combination in the configuration space. Based on the optimal model configuration combination, the drill bit rock breaking performance prediction model is trained to obtain a drill bit rock breaking performance prediction model that can map from the input feature vector to the mechanical drilling rate and drilling footage. The drill bit rock breaking performance prediction model is evaluated using a validation set, and the trained drill bit rock breaking performance prediction model is output.
[0006] In conjunction with the first aspect, in one implementation, acquiring a historical dataset including geological data, engineering data, and drill bit product data, and aligning the geological data, engineering data, and drill bit product data by well depth to form a sample set, includes: Geological data, engineering data, and drill bit product data are aligned and correlated according to well depth sequence to form a historical dataset with depth points as sample units; The historical dataset is subjected to outlier detection and processing, missing value imputation, and data noise reduction in sequence. Categorical variables in the processed historical dataset are encoded, and numerical variables are standardized to form a sample set.
[0007] In conjunction with the first aspect, in one implementation, before inputting the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock-breaking performance prediction model, the following steps are included: Based on the wellhead coordinates of the well to be drilled, the nearest neighboring well in terms of plane distance to the well to be drilled is selected from the historical dataset as the reference well; Obtain the stratigraphic lithology sequence and geological parameters distributed with well depth of the reference well; The formation lithology distribution and geological parameters of the reference well are used as the geological parameters input for the well to be drilled.
[0008] In conjunction with the first aspect, in one implementation, the weighted scoring of the predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit based on preset scoring rules to obtain a comprehensive predicted score for each candidate drill bit includes: The predicted mechanical drilling rate and the predicted drill bit footage are normalized respectively. Weighting coefficients are set for the predicted mechanical drilling rate and the predicted drill bit footage according to the engineering requirements. The measured mechanical drilling rate and measured drill bit footage of each candidate drill bit were obtained based on historical datasets. Based on a preset scoring rule, the predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit are weighted and scored to obtain a comprehensive predicted score for each candidate drill bit. The preset scoring rule is as follows:
[0009] in, This indicates the predicted overall score. and All represent weighting coefficients. , , , This indicates the measured mechanical drilling speed. This indicates the measured drill bit advance. This indicates the predicted mechanical drilling speed. This indicates the predicted drill bit advance.
[0010] In conjunction with the first aspect, in one implementation, obtaining the predicted comprehensive score of each candidate drill bit to output the drill bit selection result includes: Multiple preset depth points are set in the well to be drilled. For each preset depth point, the candidate drill bit with the highest predicted comprehensive score is selected, and the corresponding drill bit model is determined. According to the engineering requirements, the well to be drilled is divided into multiple openings. For each opening, the drill bit models selected for all preset depth points within the current opening are counted, and the drill bit model with the highest frequency is selected as the preferred drill bit model for the current opening. Based on the preferred drill bit models for all drill passes, generate a drill bit selection report that includes at least the preferred drill bit models for each drill pass and a sorted list of candidate drill bits.
[0011] In conjunction with the first aspect, in one implementation method: Geological data should include at least: well depth, stratigraphic sequence, lithological sequence, well logging data, and geomechanical parameters; Engineering data should include at least: drilling pressure, rotation speed, displacement, torque, pump pressure, mechanical drilling speed, mud density, drill bit diameter, drill bit model, drill bit size, and drill bit serial number; Drill bit product data should include at least: number of blades, tooth density, profile aggression, main cutting tooth diameter, main cutting tooth back slope, main cutting tooth exposure height, cutting tooth type - nose and shoulder, and nozzle combination.
[0012] Secondly, embodiments of this application provide a drill bit selection device, the drill bit selection device comprising: The training module is used to build a drill bit rock breaking performance prediction model based on machine learning algorithms. It constructs a historical dataset including geological data, engineering data and drill bit product data to train the drill bit rock breaking performance prediction model. The execution module is used to input the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock breaking performance prediction model to obtain the predicted engineering parameters of each candidate drill bit, including the predicted mechanical drilling rate and the predicted drill bit footage. The optimization module is used to perform weighted scoring on the predicted mechanical drilling speed and predicted drill bit footage of each candidate drill bit based on preset scoring rules, so as to obtain the predicted comprehensive score of each candidate drill bit and output the drill bit optimization result.
[0013] Thirdly, embodiments of this application provide a drill bit selection device, which includes a processor, a memory, and a drill bit selection program stored in the memory and executable by the processor, wherein when the drill bit selection program is executed by the processor, it implements the steps of the drill bit selection method as described in any of the above embodiments.
[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing a drill bit selection program, wherein when the drill bit selection program is executed by a processor, it implements the steps of the drill bit selection method as described in any of the above embodiments.
[0015] The beneficial effects of the technical solutions provided in this application include: This application employs machine learning algorithms to train a drill bit rock-breaking performance prediction model, replacing engineers' personal experience with data-driven approaches to standardize the drill bit selection process and avoid subjective judgments in selecting drill bits for new blocks or complex formations. This application integrates multi-source data, including geological, engineering, and drill bit product parameters, enabling drill bit selection based on more comprehensive information. This application predicts and outputs two drill bit performance indicators: mechanical drilling rate and drill bit footage. A weighted comprehensive score is applied based on preset scoring rules, while also considering the constraint relationship between drilling speed and drill bit life, resulting in more accurate drill bit selection. Attached Figure Description
[0016] Figure 1 This is a flowchart illustrating the preferred drill bit method of this application; Figure 2 A flowchart of the AutoML framework provided in this application; Figure 3 Geological parameter profile of a reference well provided for embodiments of this application; Figure 4 A flowchart illustrating the architecture of the drill bit optimization method provided in this application embodiment; Figure 5 This is a comparison chart of the preferred results between human experience and the embodiments of this application; Figure 6 A comparison chart of the predicted mechanical drilling rate and drill bit footage; Figure 7 This is a schematic diagram of the functional modules of the preferred drill bit device of this application; Figure 8 This is a schematic diagram of the hardware structure of the preferred drill bit device involved in the embodiments of this application. Detailed Implementation
[0017] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.
[0018] This application provides a method, apparatus, equipment, and medium for selecting drill bits, which can solve the technical problem of inaccurate drill bit selection in the prior art.
[0019] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.
[0020] In a first aspect, embodiments of this application provide a preferred method for drill bits.
[0021] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating the drill bit selection method of this application. Figure 1 As shown, the drill bit optimization method specifically includes the following steps: Step S1: Establish a drill bit rock breaking performance prediction model based on machine learning algorithms, and construct a historical dataset including geological data, engineering data and drill bit product data to train the drill bit rock breaking performance prediction model.
[0022] In this embodiment of the application, step S1 specifically includes the following steps: Step S11: Establish a drill bit rock breaking performance prediction model based on the AutoML framework.
[0023] Specifically, the AutoML framework is a type of machine learning algorithm. It can obtain models with higher prediction accuracy within a limited time, avoiding the limitations of manual trial and error. In some other embodiments of this application, other machine learning algorithms can also be used to establish drill bit rock-breaking performance prediction models, which are not limited here.
[0024] Step S12: Obtain a historical dataset including geological data, engineering data, and drill bit product data, and align the geological data, engineering data, and drill bit product data by well depth to form a sample set.
[0025] In this embodiment of the application, step S12 specifically includes the following steps: Step S121: Align and correlate geological data, engineering data, and drill bit product data according to well depth sequence to form a historical dataset with depth points as sample units.
[0026] Specifically, geological data should include at least: well depth, stratigraphic sequence, lithological sequence, well logging data, and geomechanical parameters. The lithological sequence should at least include sandstone, mudstone, and limestone; the well logging data should at least include gamma rays, density, and sonic transit time; and the geomechanical parameters should at least include uniaxial compressive strength, internal friction angle, elastic modulus, Poisson's ratio, and rock drillability. Geological data can be derived from well logging data, core experiments, seismic data, etc.
[0027] Engineering data should include at least: drilling pressure, rotational speed, displacement, torque, pump pressure, mechanical drilling speed, mud density, drill bit diameter, drill bit model, drill bit size, and drill bit serial number. Engineering data reflects the dynamic inputs and drill bit responses during the drilling process. Engineering data can be derived from real-time logging data, drilling log information, etc.
[0028] Drill bit product data should include at least: drill body material, number of blades, tooth density, profile aggression, main cutting tooth diameter, main cutting tooth rake angle, main cutting tooth exposure height, cutting tooth type - nose and shoulder, and nozzle assembly. Drill bit product data can be obtained from the drill bit manufacturer's product design documents.
[0029] The above data are aligned and correlated according to the well depth sequence to form a dataset with depth points as sample units. Geological data and engineering data are directly aligned through well depth, while drill bit product data are secondary correlated with drill bit usage records in the engineering data through drill bit model or drill bit number. Therefore, by using a dataset with depth points as sample units, the drill bit rock-breaking performance prediction model can learn and predict the variation of drill bit rock-breaking performance with formation changes.
[0030] Step S122: Perform outlier monitoring and processing, missing value imputation, and data noise reduction on the historical dataset in sequence.
[0031] Specifically, outlier monitoring and handling involves calculating the mean and standard deviation of any parameter, using the 3σ (three sigma) principle to identify data points deviating from the mean by more than three times the standard deviation. Business rules are then applied to the parameters, and identified outliers are marked as missing or directly removed, thus enabling statistical testing of numerical parameters. Missing value imputation is performed using methods selected based on data characteristics and correlations. For parameters that continuously vary along depth, spline interpolation is used, fitting smooth curves with neighboring data points to fill gaps in the data sequence caused by measurement omissions, transmission interruptions, etc., ensuring the continuity and integrity of the dataset. Data denoising employs time-series filtering algorithms such as Kalman filtering, using recursive calculations to optimally estimate the true values of parameters while considering process and observation noise, thereby eliminating high-frequency noise and preserving the true trend. The processed high-quality data is then structured and stored, establishing primary and foreign keys between databases, and used as the database for drill bit optimization tasks.
[0032] Step S123: Encode the categorical variables in the processed historical dataset and standardize the numerical variables to form a sample set.
[0033] Specifically, feature encoding and standardization primarily involve converting categorical variables such as lithology and cutting tooth type into binary columns using one-hot encoding to eliminate potential order misleading effects from categorical numbers. For numerical variables such as compressive strength and drill pressure, Z-score standardization is typically used to eliminate differences in dimensions and orders of magnitude, ensuring that each feature has a mean of 0 and a variance of 1. The final output is a purely numerical feature matrix, where each row represents a sample and each column corresponds to a processed feature, which can be directly used as input for training the model. This unifies the data scale, accelerates the convergence process of model training, and improves computational stability.
[0034] Step S13: Based on the sample set, construct the geological data and drill bit product data into an input feature vector, and use the mechanical drilling rate and drill bit footage from the engineering data as output labels.
[0035] Specifically, for each well section corresponding to the drill bit, the mean and variance of eight main geomechanical parameters are calculated to form a geological feature vector. These geomechanical parameters include uniaxial compressive strength, drillability, hardness, cohesion, internal friction angle, plasticity coefficient, abrasiveness, and impact resistance. A structural feature vector is also extracted from the drill bit, including the number of cutter wings, tooth density, profile aggression, main cutting tooth diameter, main cutting tooth rake angle, main cutting tooth exposure height, cutting tooth type (nose and shoulder), and nozzle combination. The geological and structural feature vectors are concatenated to form the input feature vector. The mechanical drilling rate and drilling footage recorded for the same drilling run are used as the corresponding predicted target values.
[0036] Step S14: Construct a configuration space that includes feature selection space, hyperparameter space, and model architecture space.
[0037] Specifically, Figure 2 A flowchart of the AutoML framework provided for this application. (See attached diagram.) Figure 2 As shown, the dashed box represents the configuration space, including the feature selection space, hyperparameter space, and model architecture space. The left side is used to input training data, and through the connected optimizer, the optimal configuration is automatically found based on the defined evaluation metrics. Finally, the trained model is output, and the test data is then run on the model to achieve the prediction goal.
[0038] The feature selection space determines which input parameters are most effective for prediction. The model architecture space defines a set of optional machine learning algorithms from which the model can choose the most suitable one. The hyperparameter space sets the internal control parameters for the selected algorithm.
[0039] Step S15: Using the input feature vector as training input and the mechanical drilling rate and drill bit footage as prediction targets, the optimizer searches for the optimal model configuration combination in the configuration space.
[0040] Specifically, the input feature vector is fed into the framework, and the optimizer automatically performs numerous trials and comparisons within a defined configuration space based on preset performance evaluation metrics to search for the optimal configuration combination. The preset performance evaluation metrics can be used to minimize the prediction error.
[0041] Step S16: Based on the optimal model configuration combination, train the drill bit rock breaking performance prediction model to obtain a drill bit rock breaking performance prediction model that can map from the input feature vector to the mechanical drilling rate and drilling footage.
[0042] Specifically, a sample set is used to train the drill bit rock-breaking performance prediction model.
[0043] Step S17: Use the validation set to evaluate the drill bit rock breaking performance prediction model and output the trained drill bit rock breaking performance prediction model.
[0044] Specifically, a reserved validation set is used as input to the final model for prediction. The predicted results are compared with the actual values, and indices such as root mean square error and coefficient of determination are calculated to evaluate the model's predictive accuracy and confirm that the completed drill bit rock-breaking performance prediction model meets the application requirements.
[0045] Step S2: Input the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock breaking performance prediction model to obtain the predicted engineering parameters of each candidate drill bit. The predicted engineering parameters include the predicted mechanical drilling rate and the predicted drill bit footage.
[0046] In this embodiment of the application, before step S2, the following steps are also included: Step A1: Select the nearest neighboring well from the historical dataset based on the wellhead coordinates of the well to be drilled as the reference well.
[0047] Specifically, since information about the well to be drilled is difficult to obtain or is usually limited, and geological features exhibit lateral similarity and coherence of stratigraphic sequences, the nearest neighbor well is used as a reference well to characterize the distribution of geological features of the well to be drilled. Based on the wellhead coordinates in the drilling engineering design, the planar Euclidean distances between the well to be drilled and multiple existing wells near the well are calculated, and the nearest existing well is determined as the reference well.
[0048] Step A2: Obtain the stratigraphic lithology sequence of the reference well and the geological parameters distributed with well depth.
[0049] Specifically, the entire well section data of the reference well was retrieved from the geomechanical database, including: stratigraphic sequence, lithological sequence, logging curves, and eight geomechanical parameters. The entire well section data of the reference well was sampled at 1-meter depth intervals to form a vertically continuous geological parameter profile.
[0050] Figure 3 A geological parameter profile of a reference well provided for an embodiment of this application. For example... Figure 3 As shown, the left vertical axis represents the well depth, ranging from approximately 2500 meters to 5000 meters, covering the target section to be drilled. The formation and lithology column shows the formation sequence and lithological distribution as the depth changes. The geomechanical parameter curves show the changes in key mechanical parameters such as compressive strength, hardness, drillability, and internal friction angle as the depth changes. The drill bit structural parameter matching diagram shows the recommended key structural parameters of the drill bit for the corresponding depth range, such as the number of blades (6), tooth density, main cutting tooth diameter (16mm), back rake angle, and profile aggression (medium / strong).
[0051] Step A3: Input the formation lithology distribution and geological parameters of the reference well as the geological parameters of the well to be drilled.
[0052] Specifically, the geological parameter profile of the reference well is used to obtain the distribution of geological features of the well to be drilled, and the auxiliary drill bit is selected in the best way.
[0053] Step S3: Based on the preset scoring rules, the predicted mechanical drilling speed and predicted drill bit footage of each candidate drill bit are weighted and scored to obtain the predicted comprehensive score of each candidate drill bit, so as to output the drill bit selection result.
[0054] In this embodiment of the application, step S3 involves weighting the predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit based on a preset scoring rule to obtain a comprehensive predicted score for each candidate drill bit. Specifically, this includes the following steps: Step S311: Normalize the predicted mechanical drilling speed and the predicted drill bit footage respectively.
[0055] Specifically, before scoring, a drill bit population is constructed by obtaining the geomechanical parameters per meter of the reference well and the structural parameters of all drill bits used in the same block as a search space for parameter combinations, thereby expanding the number of candidate drill bits. The predicted mechanical rate of penetration and predicted drill bit footage of all candidate drill bits in the drill bit population are normalized to convert the predicted mechanical rate of penetration and predicted drill bit footage to a unified scoring scale.
[0056] Step S312: Set weighting coefficients for the predicted mechanical drilling speed and the predicted drill bit footage according to the project requirements.
[0057] Specifically, weighting coefficients are set according to the requirements of different well sections for mechanical drilling rate and bit footage. The weighting coefficient for predicting mechanical drilling rate indicates the importance of mechanical drilling rate, and the weighting coefficient for predicting bit footage indicates the importance of bit footage. In a specific embodiment, if the requirement is to increase drilling speed, such as needing to drill quickly in a complex formation to avoid formation collapse, then the weighting coefficient for mechanical drilling rate is set to be greater than the weighting coefficient for bit footage. If the requirement is to increase drilling footage per trip, such as avoiding increased non-productive time due to frequent tripping, then the weighting coefficient for mechanical drilling rate is set to be less than the weighting coefficient for bit footage.
[0058] Step S313: Obtain the measured mechanical drilling speed and measured drill bit footage for each candidate drill bit based on the historical dataset.
[0059] Step S314: Based on the preset scoring rules, weight the predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit to obtain the predicted comprehensive score of each candidate drill bit. The preset scoring rules are as follows:
[0060] in, This indicates the predicted overall score. and All represent weighting coefficients. , , , This indicates the measured mechanical drilling speed. This indicates the measured drill bit advance. This indicates the predicted mechanical drilling speed. This indicates the predicted drill bit advance.
[0061] A higher value indicates better overall performance of the candidate drill bit in the current well section. This application's embodiments utilize configurable weighting coefficients to allow drill bit selection to flexibly adapt to diverse field engineering needs.
[0062] In this embodiment of the application, step S3 obtains the predicted comprehensive score of each candidate drill bit to output the drill bit selection result, specifically including the following steps: Step S321: Set multiple preset depth points in the well to be drilled. For each preset depth point, select the candidate drill bit with the highest predicted comprehensive score and determine the drill bit model corresponding to the candidate drill bit.
[0063] Step S322: According to the engineering requirements, the well to be drilled is divided into multiple sessions. For each session, the drill bit models selected for all preset depth points within the current session are counted, and the drill bit model with the highest frequency is selected as the preferred drill bit model for the current session.
[0064] Step S323: Based on the preferred drill bit models for all drill passes, generate a drill bit selection report that includes at least the preferred drill bit models for each drill pass and a sorted list of candidate drill bits.
[0065] Specifically, in this embodiment, a series of preset depth points are set at certain intervals. The interval can be every 1 meter or every 10 meters. By evaluating the candidate drill bits point by point, a drill bit selection scheme that conforms to the actual engineering is generated. By optimizing based on the preset depth points, the optimal drill bit model corresponding to each point is obtained. Subsequently, according to the drilling design's division into openings, the recommended results of all depth points within each opening are statistically analyzed. The majority decision principle is used to determine the final recommended drill bit model for that opening, and a drill bit selection report is generated.
[0066] In one specific embodiment Figure 4 This is a flowchart illustrating the architecture of the drill bit optimization method provided in this application embodiment. Figure 4 As shown, the formation, lithology, and geomechanical parameters of the entire well section are extracted, along with the drill bit type, size, and rock-breaking structure, to form the input of the training samples. Then, the mechanical drilling rate and drilling footage of the corresponding well section are extracted as dual output labels. Two independent drilling rate prediction models and footage prediction models are trained separately. The two models share the input feature space but the objective functions are decoupled to avoid gradient conflicts in multi-task learning.
[0067] Based on the location of the well to be drilled, a geological profile from a reference well is selected as the input for the expected geological conditions of the well to be drilled. A candidate drill bit population is generated through parametric grid search or evolutionary algorithms. The geological parameters of the well to be drilled are combined with the structural parameters of each candidate drill bit and input into the trained rate of drilling (RDR) and footage prediction models to obtain the predicted mechanical rate of drilling (MRD) and predicted footage per trip for each candidate drill bit. Then, the comprehensive score of each candidate drill bit is calculated according to a preset scoring rule to identify the theoretically optimal combination of drill bit structural parameters. The optimal parameter combination or the drill bit with the highest score is matched with existing drill bit models to find physical drill bits that can be directly purchased or used. The optimization results are further aggregated according to the drilling design sequence to form a drill bit selection scheme.
[0068] Figure 5 This is a comparison chart showing the preferred results between human experience and the embodiments of this application. For example... Figure 5 As shown, the "Manual" column on the left displays five drill bit models (ae) used in six sub-sections, with structural parameters for each model selected based on manual experience. The "This Method" column on the right displays five model models (AE) recommended by the algorithm, with eight parameters, including the number of blades, tooth density, and cutting tooth diameter, showing systematic differences from the manual selection. For example, in the 2300-2378m section, manual selection used model a with 6 blades and moderate profile aggression, while this method recommends model A with 5 blades and strong profile aggression. In this embodiment, the predictive model identifies a low drillability rating for this formation; reducing the number of blades and increasing aggression improves the mechanical drilling rate, while optimizing tooth density extends the drilling footage. Figure 5 The comparative study verified the magnitude and effect of data-driven selection on the structural parameters of experience-based selection.
[0069] Figure 6 This is a comparison chart of the predicted mechanical drilling rate and drill bit footage. (Example) Figure 6 As shown, the horizontal axis represents drill bit footage (0m - 2500m), and the vertical axis represents mechanical drilling speed (0m / h - 35m / h). The left figure shows the selection results based on traditional manual experience, while the right figure shows the selection results of the embodiments of this application. The scattered points in the left figure are dispersed, with a large number of inefficient samples having mechanical drilling speeds <20m / h and drill bit footage <1500m. The scattered points in the right figure generally shift to the upper right, with mechanical drilling speeds concentrated in the 25m / h-30m / h range and footage concentrated in the 1800-2200m range. The preferred drill bits in the embodiments of this application have improved average mechanical drilling speed and average footage, and the scattered points are more concentrated, exhibiting better performance consistency and repeatability.
[0070] This application employs machine learning algorithms to train a drill bit rock-breaking performance prediction model, replacing engineers' personal experience with data-driven approaches. This standardizes the drill bit selection process and solidifies knowledge, reducing the uncertainty of subjective judgments in new blocks or complex formations. This application integrates multi-source data, including geological, engineering, and drill bit product parameters, enabling drill bit selection based on more comprehensive information. This application predicts and outputs two drill bit performance indicators: mechanical drilling rate and drill bit footage. A weighted comprehensive score is applied based on preset scoring rules, while also considering the constraint relationship between drilling speed and drill bit life, resulting in more accurate drill bit selection.
[0071] Secondly, embodiments of this application also provide a drill bit selection device.
[0072] In one embodiment, reference is made to Figure 7 , Figure 7 This is a schematic diagram of the functional modules of the preferred drill bit device of this application. Figure 7 As shown, the preferred drill bit assembly includes: The training module is used to build a drill bit rock breaking performance prediction model based on machine learning algorithms. It constructs a historical dataset including geological data, engineering data and drill bit product data to train the drill bit rock breaking performance prediction model. The execution module is used to input the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock breaking performance prediction model to obtain the predicted engineering parameters of each candidate drill bit, including the predicted mechanical drilling rate and the predicted drill bit footage. The optimization module is used to perform weighted scoring on the predicted mechanical drilling speed and predicted drill bit footage of each candidate drill bit based on preset scoring rules, so as to obtain the predicted comprehensive score of each candidate drill bit and output the drill bit optimization result.
[0073] The functions of each module in the drill bit selection device correspond to the steps in the drill bit selection method embodiment, and their functions and implementation processes will not be described in detail here.
[0074] Thirdly, embodiments of this application provide a drill bit selection device, which includes a processor, a memory, and a drill bit selection program stored in the memory and executable by the processor, wherein when the drill bit selection program is executed by the processor, it implements the steps of the drill bit selection method as described in any of the above embodiments.
[0075] The preferred equipment for drilling bits can be personal computers (PCs), laptops, servers, or other devices with data processing capabilities.
[0076] Reference Figure 8 , Figure 8This is a schematic diagram of the hardware structure of the drill bit optimization device involved in the embodiments of this application. In the embodiments of this application, the drill bit optimization device may include a processor, a memory, a communication interface, and a communication bus.
[0077] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.
[0078] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces used for interconnecting devices within the drill bit selection device, as well as interfaces used for interconnecting the drill bit selection device with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.
[0079] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.
[0080] The processor can be a general-purpose processor, which can call the drill bit selection program stored in memory and execute the drill bit selection method provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the drill bit selection program is called can be referred to in the various embodiments of the drill bit selection method of this application, and will not be repeated here.
[0081] Those skilled in the art will understand that Figure 8 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.
[0082] Fourthly, embodiments of this application provide a computer-readable storage medium storing a drill bit selection program, wherein when the drill bit selection program is executed by a processor, it implements the steps of the drill bit selection method as described in any of the above embodiments.
[0083] The present application has a drill bit selection program stored on a computer-readable storage medium, wherein when the drill bit selection program is executed by a processor, it implements the steps of the drill bit selection method as described above.
[0084] The method implemented when the drill bit selection procedure is executed can be referred to in various embodiments of the drill bit selection method of this application, and will not be repeated here.
[0085] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0086] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.
[0087] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.
[0088] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.
[0089] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.
[0090] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.
[0091] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. A method for selecting the best drill bit, characterized in that, The preferred method for the drill bit includes: A drill bit rock breaking performance prediction model was established based on machine learning algorithms. A historical dataset including geological data, engineering data and drill bit product data was constructed to train the drill bit rock breaking performance prediction model. Input the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock breaking performance prediction model to obtain the predicted engineering parameters of each candidate drill bit, including the predicted mechanical drilling rate and the predicted drill bit footage. Based on preset scoring rules, the predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit are weighted and scored to obtain the predicted comprehensive score of each candidate drill bit, so as to output the drill bit selection result.
2. The preferred drill bit method according to claim 1, characterized in that, The drill bit rock-breaking performance prediction model based on machine learning algorithms is trained by constructing a historical dataset including geological data, engineering data, and drill bit product data. A drill bit rock-breaking performance prediction model was established based on the AutoML framework; Acquire historical datasets including geological data, engineering data, and drill bit product data, and align the geological data, engineering data, and drill bit product data by well depth to form a sample set; Based on the sample set, geological data and drill bit product data are constructed as input feature vectors, and mechanical drilling speed and drill bit footage in engineering data are used as output labels. Construct a configuration space that includes feature selection space, hyperparameter space, and model architecture space; Using the input feature vector as training input and mechanical drilling rate and drill bit footage as prediction targets, an optimizer searches for the optimal model configuration combination in the configuration space. Based on the optimal model configuration combination, the drill bit rock breaking performance prediction model is trained to obtain a drill bit rock breaking performance prediction model that can map from the input feature vector to the mechanical drilling rate and drilling footage. The drill bit rock breaking performance prediction model is evaluated using a validation set, and the trained drill bit rock breaking performance prediction model is output.
3. The preferred drill bit method according to claim 2, characterized in that, The process involves acquiring historical datasets including geological data, engineering data, and drill bit product data, and aligning these datasets by well depth to form a sample set, including: Geological data, engineering data, and drill bit product data are aligned and correlated according to well depth sequence to form a historical dataset with depth points as sample units; The historical dataset is subjected to outlier detection and processing, missing value imputation, and data noise reduction in sequence. Categorical variables in the processed historical dataset are encoded, and numerical variables are standardized to form a sample set.
4. The preferred drill bit method according to claim 2, characterized in that, Before inputting the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock-breaking performance prediction model, the following steps are included: Based on the wellhead coordinates of the well to be drilled, the nearest neighboring well in terms of plane distance to the well to be drilled is selected from the historical dataset as the reference well; Obtain the stratigraphic lithology sequence and geological parameters distributed with well depth of the reference well; The formation lithology distribution and geological parameters of the reference well are used as the geological parameters input for the well to be drilled.
5. The preferred method for drill bits according to claim 1, characterized in that, The predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit are weighted and scored based on a preset scoring rule to obtain a comprehensive predicted score for each candidate drill bit, including: The predicted mechanical drilling rate and the predicted drill bit footage are normalized respectively. Weighting coefficients are set for the predicted mechanical drilling rate and the predicted drill bit footage according to the engineering requirements. The measured mechanical drilling rate and measured drill bit footage of each candidate drill bit were obtained based on historical datasets. Based on a preset scoring rule, the predicted mechanical drilling rate and predicted drill bit footage of each candidate drill bit are weighted and scored to obtain a comprehensive predicted score for each candidate drill bit. The preset scoring rule is as follows: in, This indicates the predicted overall score. and All represent weighting coefficients. , , , This indicates the measured mechanical drilling speed. This indicates the measured drill bit advance. This indicates the predicted mechanical drilling speed. This indicates the predicted drill bit advance.
6. The preferred drill bit method according to claim 5, characterized in that, The process of obtaining the predicted comprehensive score for each candidate drill bit to output the drill bit selection result includes: Multiple preset depth points are set in the well to be drilled. For each preset depth point, the candidate drill bit with the highest predicted comprehensive score is selected, and the corresponding drill bit model is determined. According to the engineering requirements, the well to be drilled is divided into multiple openings. For each opening, the drill bit models selected for all preset depth points within the current opening are counted, and the drill bit model with the highest frequency is selected as the preferred drill bit model for the current opening. Based on the preferred drill bit models for all drill passes, generate a drill bit selection report that includes at least the preferred drill bit models for each drill pass and a sorted list of candidate drill bits.
7. The preferred method for drill bits according to claim 1, characterized in that: Geological data should include at least: well depth, stratigraphic sequence, lithological sequence, well logging data, and geomechanical parameters; Engineering data should include at least: drilling pressure, rotation speed, displacement, torque, pump pressure, mechanical drilling speed, mud density, drill bit diameter, drill bit model, drill bit size, and drill bit serial number; Drill bit product data should include at least: number of blades, tooth density, profile aggression, main cutting tooth diameter, main cutting tooth back slope, main cutting tooth exposure height, cutting tooth type - nose and shoulder, and nozzle combination.
8. A drill bit selection device, characterized in that, The preferred drill bit device includes: The training module is used to build a drill bit rock breaking performance prediction model based on machine learning algorithms. It constructs a historical dataset including geological data, engineering data and drill bit product data to train the drill bit rock breaking performance prediction model. The execution module is used to input the geological parameters of the well to be drilled and the structural parameters of each candidate drill bit into the trained drill bit rock breaking performance prediction model to obtain the predicted engineering parameters of each candidate drill bit, including the predicted mechanical drilling rate and the predicted drill bit footage. The optimization module is used to perform weighted scoring on the predicted mechanical drilling speed and predicted drill bit footage of each candidate drill bit based on preset scoring rules, so as to obtain the predicted comprehensive score of each candidate drill bit and output the drill bit optimization result.
9. A drill bit optimization device, characterized in that, The drill bit selection device includes a processor, a memory, and a drill bit selection program stored in the memory and executable by the processor, wherein when the drill bit selection program is executed by the processor, it implements the steps of the drill bit selection method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a drill bit selection program, wherein when the drill bit selection program is executed by a processor, it implements the steps of the drill bit selection method as described in any one of claims 1 to 7.