Adjacent well anti-collision method and device and adjacent well anti-collision auxiliary decision-making system

Abstract

The invention provides an adjacent well anti-collision method and device and an adjacent well anti-collision auxiliary decision making system, and relates to the technical field of underground detection. The method comprises the following steps: performing inversion on acquired magnetic field intensity data to obtain magnetic logging distance data between a drilling well and an adjacent well of a drilled well section in a current drilling section; inputting the well body characteristic parameters of the adjacent well into the constructed well spacing prediction model to obtain well spacing prediction data between the drilling well and the adjacent well in the current drilling section; correcting the magnetic survey trajectory of the adjacent well by using the well spacing prediction data; forming a reference trajectory of the adjacent well by the corrected magnetic measurement trajectory of the adjacent well and an unmeasured trajectory of the adjacent well determined according to the well spacing prediction data; and adjusting the drilling trajectory of the non-drilling section in the current drilling section by using the reference trajectory of the adjacent well. According to the method, the magnetic distance measurement mode and the well spacing prediction model are combined, accurate positioning of the adjacent well track is achieved, and therefore a high-precision adjacent well anti-collision decision is achieved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of downhole detection, and particularly relates to a method for preventing collision between adjacent wells, a device for preventing collision between adjacent wells, an auxiliary decision-making device for preventing collision between adjacent wells, an auxiliary decision-making system for preventing collision between adjacent wells and a machine readable storage medium. BACKGROUND

[0002] The dense well pattern can improve the efficiency of oilfield exploration and development, and can also reasonably control the drilling cost and increase the oil / gas production of the underground oil / gas block. In order to realize the efficient development of offshore and onshore oilfields, cluster dense wells are widely deployed in Shengli, Daqing and Bohai oilfield blocks, which leads to smaller and smaller well spacing when drilling in shallow layers, so that the drilling operation faces a great risk of wellbore collision. If the wellbore collision is not accurately identified in time at the drilling site, the adjacent well casing may be deformed, or the adjacent well casing may be drilled through, causing serious wellbore accidents and seriously affecting the normal production of the adjacent well.

[0003] In view of the wellbore collision problem, a relatively mature on-site wellbore collision prevention design method and measure have been formed at home and abroad, mainly including pre-drilling well trajectory optimization design, collision prevention scanning and while-drilling monitoring. The well trajectory optimization design needs detailed underground data and accurate calculation, and has a high requirement for the accuracy and integrity of the data, thereby increasing the collision risk. Moreover, due to the complex and changeable underground conditions, the missing or inaccurate inclination data of the old well, the inaccurate or uncorrected azimuth angle, and other factors will lead to the difficulty in accurately predicting the well trajectory, so that the pre-drilling collision prevention design cannot meet the collision prevention requirement. In addition, the inclination data of the collision prevention scanning and the while-drilling monitoring data are affected by factors such as data transmission delay, so that they cannot meet the on-site collision prevention requirement in some cases, leading to false positives or false negatives of the collision risk. SUMMARY

[0004] The purpose of the embodiments of the application is to provide a method for preventing collision between adjacent wells, a device for preventing collision between adjacent wells, an auxiliary decision-making device for preventing collision between adjacent wells, an auxiliary decision-making system for preventing collision between adjacent wells and a machine readable storage medium, so as to overcome one or more defects of the existing technology, such as the collision prevention scheme between adjacent wells based on the pre-drilling well trajectory optimization design, collision prevention scanning and while-drilling monitoring.

[0005] In order to achieve the above purpose, a first aspect of the embodiments of the application provides a method for preventing collision between adjacent wells, which comprises:

[0006] obtaining magnetic field intensity data, and inverting the magnetic field intensity data to obtain magnetic logging distance data between the well drilled in the current drilling section and the adjacent well;

[0007] obtaining wellbore characteristic parameters of the adjacent well, inputting the wellbore characteristic parameters into the constructed well spacing prediction model to obtain well spacing prediction data between the current drilling well and the adjacent well at the current drilling section;

[0008] obtaining the adjacent well magnetic measurement trajectory by calculation based on the magnetic well spacing data and the current drilling well trajectory at the current drilling section, and correcting the adjacent well magnetic measurement trajectory by using the well spacing prediction data;

[0009] obtaining the adjacent well unmeasured trajectory at the current drilling section by calculation based on the well spacing prediction data of the un-drilled well section in the current drilling section and the current drilling well trajectory, and composing the adjacent well reference trajectory by the corrected adjacent well magnetic measurement trajectory and the adjacent well unmeasured trajectory;

[0010] adjusting the current drilling well trajectory of the un-drilled well section in the current drilling section by using the adjacent well reference trajectory;

[0011] The well spacing prediction model is obtained by training and parameter optimization of an initial regression network, and the wellbore characteristic parameters are characteristics of the adjacent well affecting the spatial coordinates of the current drilling well at each well depth.

[0012] In specific embodiments of the present application, the wellbore characteristic parameters include one or more of lithology, well depth, hole inclination angle and dogleg severity.

[0013] In specific embodiments of the present application, the well spacing prediction model is obtained by training and parameter optimization of an initial regression network, including:

[0014] obtaining wellbore characteristic data and well spacing data of a new experimental well drilled in an external field;

[0015] taking the wellbore characteristic data of the experimental well as input data of the regression network and taking the well spacing data of the experimental well as label data of the initial regression network, constructing training samples and test samples according to the input data and the label data;

[0016] training the initial regression network by using the training samples;

[0017] evaluating the trained regression network by using the test samples, adjusting the hyperparameters of the regression network according to the evaluation results until the evaluation results meet the preset requirements, assigning the latest hyperparameters to the regression network to obtain the well spacing prediction model.

[0018] In specific embodiments of the present application, the correction of the adjacent well magnetic measurement trajectory by using the well spacing prediction data includes:

[0019] obtaining predicted well spacing between the adjacent well and the current drilling well at a plurality of well depth positions from the well spacing prediction data;

[0020] The magnetic logging distances between adjacent wells and the well being drilled at these multiple well depth locations are obtained from the magnetic logging distance data.

[0021] The magnetic ranging deviation is obtained by subtracting the predicted well distance at each well depth from the magnetic well distance at each well depth.

[0022] The maximum confidence interval of the magnetic ranging error is obtained by comparing the various magnetic ranging errors.

[0023] The spatial coordinates of adjacent wells whose magnetic ranging deviations are not within the maximum confidence interval at the corresponding well depth are excluded from the magnetic ranging trajectory of the adjacent wells.

[0024] In a specific embodiment of this application, the correction of adjacent well magnetic trajectories using the well spacing prediction data includes:

[0025] Retain the magnetic survey trajectory of adjacent wells within the preset well depth range;

[0026] Calculate the first corrected trajectory, which is the predicted trajectory of adjacent wells outside the preset well depth range of the current drilling section, calculated based on the well spacing prediction data of each well depth position outside the preset well depth range of the current drilling section and the positive drilling trajectory.

[0027] The corrected magnetic survey trajectory of the adjacent well is composed of the retained magnetic survey trajectory of the adjacent well and the first corrected trajectory.

[0028] In a specific embodiment of this application, adjusting the forward drilling trajectory of the un-drilled section in the current drilling segment using the adjacent well reference trajectory includes:

[0029] The starting and ending points of the adjacent wells at the current drilling section are determined based on the reference trajectory of the adjacent wells;

[0030] With the constraint that the drilling well does not collide with the adjacent well, the optimal path from the starting point of the un-drilled section in the current drilling segment of the drilling well to the end point of the adjacent well in the current drilling segment is determined by numerical optimization algorithm.

[0031] The trajectory composed of the coordinate points in the optimal path is used as the adjusted forward drilling trajectory for the un-drilled section in the current drilling segment.

[0032] In a specific embodiment of this application, the numerical optimization algorithm is a simulated annealing algorithm. The step of determining the optimal path from the starting point of the un-drilled section in the current drilling segment of the current drilling well to the endpoint of the adjacent well in the current drilling segment, with the constraint that the drilling does not collide with the adjacent well, using the numerical optimization algorithm includes:

[0033] Parameter initialization;

[0034] Generate an initial path from the start point of the un-drilled section in the current drilling segment to the end point of the adjacent well in the current drilling segment;

[0035] Perform the new path generation step;

[0036] Calculate the total cost of the new path generated in the current iteration. If the total cost of the new path is lower than the total cost of the path in the previous iteration, accept the new path; otherwise, proceed to execute the new path generation step.

[0037] Determine if the temperature parameter has reached the preset threshold. If so, stop the iteration and use the current latest path as the optimal path between the starting point of the un-drilled section in the current drilling segment and the end point of the adjacent well in the current drilling segment. Otherwise, jump to execute the new path generation step.

[0038] The new path generation step includes: applying a perturbation to the current path to generate a new path.

[0039] In a specific embodiment of this application, the magnetic field strength data is acquired by a magnetic ranging device lowered into the drilling well.

[0040] A second aspect of this application provides an adjacent well collision prevention device, the device comprising:

[0041] The magnetic logging distance determination module is used to acquire magnetic field strength data and invert the magnetic field strength data to obtain the magnetic logging distance data between the well being drilled in the current drilling section and the adjacent well in the current drilling section.

[0042] The well spacing prediction module is used to obtain the well characteristic parameters of adjacent wells, input the well characteristic parameters into the constructed well spacing prediction model, and obtain the well spacing prediction data between the well being drilled in the current drilling section and the adjacent wells.

[0043] The magnetic logging trajectory correction module is used to calculate the magnetic logging trajectory of adjacent wells using the magnetic logging well distance data and the positive drilling trajectory at the current drilling section, and to correct the magnetic logging trajectory of adjacent wells using the well distance prediction data.

[0044] The adjacent well reference trajectory determination module is used to calculate and obtain the unmeasured trajectory of the adjacent wells in the current drilling section based on the well spacing prediction data of the un-drilled section in the current drilling section and the current drilling trajectory, and the adjacent well reference trajectory is composed of the corrected magnetic survey trajectory of the adjacent well and the unmeasured trajectory of the adjacent well.

[0045] The trajectory adjustment module is used to adjust the forward drilling trajectory of the un-drilled section in the current drilling segment using the adjacent well reference trajectory;

[0046] The well spacing prediction model is obtained by training and optimizing the parameters of an initial regression network, and the well characteristic parameters are the characteristics of neighboring wells that affect the spatial coordinates of the drilling well at each well depth.

[0047] In specific embodiments of this application, the wellbore characteristic parameters include one or more of lithology, well depth, well inclination angle, and dogleg degree.

[0048] In a specific embodiment of this application, adjusting the forward drilling trajectory of the un-drilled section in the current drilling segment using the adjacent well reference trajectory includes:

[0049] The starting and ending points of the adjacent wells at the current drilling section are determined based on the reference trajectory of the adjacent wells;

[0050] With the constraint that the drilling well does not collide with the adjacent well, the optimal path from the starting point of the un-drilled section in the current drilling segment of the drilling well to the end point of the adjacent well in the current drilling segment is determined by numerical optimization algorithm.

[0051] The trajectory composed of the coordinate points in the optimal path is used as the adjusted forward drilling trajectory for the un-drilled section in the current drilling segment.

[0052] A third aspect of this application provides a neighboring well collision prevention auxiliary decision-making device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the neighboring well collision prevention method described in the first aspect of this application.

[0053] A fourth aspect of the present application provides a collision prevention auxiliary decision-making system for adjacent wells, the system including the collision prevention auxiliary decision-making device for adjacent wells as described in the third aspect of the present application and a magnetic ranging device, wherein the magnetic ranging device is communicatively connected to the collision prevention auxiliary decision-making device for adjacent wells via a transmission device;

[0054] During drilling, the magnetic ranging device is lowered into the well to magnetize the casing of the adjacent well through its own alternating current, and detects the magnetic field generated by the magnetized casing of the adjacent well to obtain magnetic field strength data. The magnetic field strength data is then sent to the adjacent well anti-collision auxiliary decision-making device via the transmission device.

[0055] In a specific embodiment of this application, the magnetic ranging device includes:

[0056] Electrode assembly for generating alternating current to magnetize adjacent well casing;

[0057] The sensor assembly is used to detect the magnetic field generated by the casing of the adjacent well after magnetization, obtain magnetic field strength data, and send the magnetic field strength data to the adjacent well anti-collision auxiliary decision-making device via the transmission device.

[0058] The sensor assembly includes two triaxial fluxgate sensors located on the same cross-section perpendicular to the central axis of the electrode assembly. After being lowered into the drilling well, the central axis of the electrode assembly is parallel to the central axis of the drilling well.

[0059] A fifth aspect of this application provides a machine-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the adjacent well collision prevention method described in the first aspect of this application.

[0060] The key to well collision prevention lies in determining the trajectory of adjacent wells that may collide with the main drilling well in the drilling direction. In this application, adjacent wells are also referred to as drilled wells or old wells. The above technical solution incorporates a well distance prediction model that combines magnetic field strength data, uses the wellbore characteristic parameters of adjacent wells as input data, and outputs the well distance between the main drilling well and the adjacent wells (the relative distance between the main drilling well and the adjacent wells, also known as the horizontal distance between the main drilling well and the adjacent wells). First, magnetic logging well distance data is inverted from the magnetic field strength data. Then, the well distance prediction data obtained from the well distance prediction model is used to calibrate the magnetic logging well distance data. This avoids using inclination measurement data from adjacent wells for adjacent well collision prevention design. Simultaneously, the high prediction accuracy of the well distance prediction model based on regression networks is utilized to obtain a highly accurate adjacent well reference trajectory. Finally, the main drilling well trajectory is adjusted based on the determined highly accurate adjacent well reference trajectory, achieving highly accurate adjacent well collision prevention.

[0061] Compared to existing technologies such as anti-collision scanning and monitoring while drilling, the above-mentioned technical solution avoids the risk of false alarms or missed collisions caused by data transmission delays by determining the trajectory of adjacent wells in the un-drilled section of the drilling stage before drilling and adjusting the forward drilling trajectory based on the determined trajectory of adjacent wells.

[0062] Compared to pre-drilling well trajectory optimization design, the above technical solutions use magnetic ranging and well distance prediction based on well distance prediction models, both of which incorporate real-time underground conditions, avoiding the defect of poor anti-collision effect of pre-drilling well trajectory design due to inaccurate data.

[0063] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0064] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings:

[0065] Figure 1 The schematic diagram illustrates a flow chart of an adjacent well collision prevention method according to an embodiment of this application;

[0066] Figure 2 This diagram illustrates a specific application example of a magnetic ranging device.

[0067] Figure 3The diagram illustrates the distribution of field test wells in a specific application example.

[0068] Figure 4 This diagram illustrates the network structure of the ridge regression model in a specific application example.

[0069] Figure 5 The illustration shows a schematic diagram of adjacent well trajectory correction in a specific application example;

[0070] Figure 6 This illustration shows the effect of positive drilling trajectory prediction based on simulated annealing algorithm in a specific application example.

[0071] Figure 7 This schematic diagram illustrates a component block diagram of an adjacent well anti-collision device according to an embodiment of this application;

[0072] Figure 8 The diagram illustrates a structural block diagram of an adjacent well collision prevention auxiliary decision-making device according to an embodiment of this application.

[0073] Explanation of reference numerals in the attached figures

[0074] In the figure, 1 is the sensor assembly; 2 is the electrode assembly; 3 is the right radial triaxial fluxgate sensor; and 4 is the left radial triaxial fluxgate sensor. Detailed Implementation

[0075] The specific embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the embodiments of this application.

[0076] Figure 1 The illustration schematically shows a flow diagram of an adjacent well collision prevention method according to an embodiment of this application. Figure 1 As shown in the embodiments of this application, a method for preventing collisions between adjacent wells may include the following implementation steps:

[0077] Step S100: Obtain magnetic field strength data, and invert the magnetic field strength data to obtain the magnetic logging distance data between the drilled section and the adjacent well in the current drilling section.

[0078] Step S102: Obtain the wellbore characteristic parameters of adjacent wells, and input these parameters into a pre-constructed well spacing prediction model to obtain the well spacing prediction data between the drilling well and adjacent wells. The well spacing prediction model is obtained by training and optimizing the parameters of an initial regression network. The wellbore characteristic parameters are the features of adjacent wells that influence the spatial coordinates of the drilling well at various well depths.

[0079] In this application, the well spacing prediction model can be a machine learning model that implements linear or nonlinear regression, such as a neural network regression model. This application does not specify the regression network used.

[0080] It is important to understand that, to avoid collisions between the drilling well and existing wells, the relative position of the drilling well at each depth depends on one or more key factors of the existing well at each depth, such as lithology, inclination angle, and dogleg degree. For different formations, other key factors may also be involved. This application does not provide a detailed description of these key factors in its embodiments. Therefore, the wellbore characteristic parameters of adjacent wells represent the characteristics of adjacent wells that influence the spatial coordinates of the drilling well at each depth.

[0081] For example, in one specific embodiment, the wellbore characteristic parameters include one or more of lithology, well depth, well inclination angle, and dogleg degree.

[0082] Step S104: Calculate the magnetic logging trajectory of the adjacent well using the magnetic logging distance data and the positive drilling trajectory at the current drilling section, and correct the magnetic logging trajectory of the adjacent well using the well distance prediction data obtained in step 102.

[0083] It is important to understand that before the current drilling section is started, there is an initial positive drilling trajectory corresponding to that drilling section, namely the positive drilling trajectory described in step 104, which is also the well trajectory that needs to be adjusted in subsequent steps.

[0084] As is known, in some common embodiments, there are differences in the relative position data of the drilling well and the old well measured by different measuring instruments. To eliminate the differences in relative position data, the old well trajectory obtained by measuring with multiple measuring instruments is fused, analyzed, and calibrated to determine the final old well trajectory. For example, the fusion analysis and calibration may involve the following: using the inclination angle and azimuth angle of the first type of old well trajectory as a reference, calculating the angle difference between the inclination angle of the second type of old well trajectory and the aforementioned reference inclination angle, calculating the angle difference between the azimuth angle of the second type of old well trajectory and the aforementioned reference azimuth angle, and adjusting the second type of old well trajectory according to the calculated angle difference; using the changing trend of the third type of old well trajectory within a preset well depth range as a reference, by translating the wellhead, moving the abnormal trend change points in the second type of old well trajectory to coincide with the coordinate points of the third type of old well trajectory at the same well depth, and through iterative correction, ensuring that the cumulative error between the second type of old well trajectory and the third type of old well trajectory within a preset length of well section is within the allowable range. To correct the magnetic trajectories of adjacent wells using well spacing prediction data, the process of mutual correction of multiple well trajectories as described in the above general embodiment can be combined.

[0085] Step S106: Calculate the unmeasured trajectory of adjacent wells in the current drilling section based on the predicted well spacing data of the un-drilled sections in the current drilling section and the current drilling trajectory, and form the adjacent well reference trajectory by combining the corrected magnetic survey trajectory of adjacent wells and the unmeasured trajectory of adjacent wells.

[0086] Step S108: Adjust the forward drilling trajectory of the un-drilled section in the current drilling segment using the reference trajectory of the adjacent well.

[0087] For example, in one specific embodiment, magnetic field strength data is acquired by a magnetic ranging device installed in the drilling well. In a comparative embodiment, magnetic field strength data is measured by installing a magnetic guidance device in an adjacent well. It is evident that to obtain various relative position data between the drilling well and adjacent wells, the installation of a magnetic ranging device in the drilling well is combined with a well distance prediction model. This eliminates the need to place tools or instruments in adjacent wells, thus avoiding any impact on them. This non-contact adjacent well positioning method enables precise location of distant adjacent wells from the drilling well in practical applications. Therefore, compared to adjacent well collision prevention schemes based on pre-drilling well trajectory optimization design, anti-collision scanning during drilling, and monitoring while drilling, this method has greater applicability.

[0088] It is important to understand that magnetic ranging equipment typically includes an electrode assembly for transmitting alternating current and a sensor assembly 1 for detecting alternating magnetic fields. Its working principle is as follows: after the magnetic ranging equipment is lowered into the current well, alternating current is transmitted to the formation through the electrode assembly 2. The magnetic field generated by the alternating current magnetizes the metal casing of the adjacent well. Then, the sensor assembly 1 detects the magnetic field generated by the magnetized metal casing of the adjacent well, thereby obtaining magnetic field strength data. The inversion process of the magnetic field strength data can be based on a theoretical derivation formula between magnetic field strength and well distance (the horizontal distance between the current well and the adjacent well, also known as the relative distance between the current well and the adjacent well). This theoretical derivation formula corresponds to the specific magnetic ranging equipment.

[0089] As described in the above embodiment, to obtain highly accurate adjacent well trajectories, two types of relative position data between the drilling well and adjacent wells are combined: magnetic well distance data obtained through magnetic ranging in the drilling well and well distance prediction data obtained based on regression principles using a well distance prediction model. First, leveraging the high-precision prediction capabilities of the well distance prediction model, the unmeasured trajectory of adjacent wells is deduced from the well distance prediction data obtained based on regression principles and the drilling well trajectory. Compared to deducing the unmeasured trajectory of adjacent wells using directional survey data or magnetic ranging data, the deduced trajectory is more accurate. Furthermore, factors such as the instability of the alternating current in the magnetized casing of adjacent wells and magnetic interference in the measurement environment can lead to inaccuracies in the magnetic field strength data. Therefore, some abnormal values ​​exist in the magnetic ranging trajectory of adjacent wells. By utilizing the high-precision prediction capabilities of the well distance prediction model, the well distance prediction data is used to correct the magnetic ranging trajectory of adjacent wells. This achieves highly accurate adjacent well positioning, thereby improving the accuracy of adjacent well collision avoidance decisions.

[0090] For example, in one specific embodiment, the well spacing prediction model is obtained by training and optimizing the parameters of the initial regression network, which may include the following steps:

[0091] Obtain wellbore characteristic data and well spacing data of newly drilled experimental wells in the field;

[0092] The wellbore feature data of the experimental wells were used as the input data of the regression network, and the well spacing data of the experimental wells were used as the label data of the initial regression network. Training samples and test samples were constructed based on the above input data and label data.

[0093] The initial regression network is trained using the constructed training samples;

[0094] The trained regression network is evaluated using the constructed test samples, and the hyperparameters of the regression network are adjusted according to the evaluation results until the evaluation results meet the preset requirements. Then, the latest hyperparameters are assigned to the regression network to obtain the well spacing prediction model.

[0095] For example, to illustrate the method of correcting the magnetic trajectories of adjacent wells based on well spacing prediction data, this application proposes the following two correction methods. It is understood that more feasible correction methods can be designed by combining the multi-trajectory correction process in the general embodiments.

[0096] The first correction method proposed in this application follows the procedure as follows:

[0097] The predicted well distances between adjacent wells and the well being drilled are obtained from the well distance prediction data at multiple well depth locations;

[0098] The magnetic logging distances between adjacent wells and the well being drilled at these multiple well depths are obtained from the magnetic logging distance data.

[0099] The magnetic ranging deviation is obtained by subtracting the predicted well distance at each well depth from the magnetic well distance at each well depth.

[0100] The maximum confidence interval of the magnetic ranging error is obtained by comparing the various magnetic ranging errors.

[0101] The spatial coordinates of adjacent wells whose magnetic ranging deviations are not within the maximum confidence interval at the corresponding well depth are excluded from the magnetic ranging trajectory of adjacent wells.

[0102] The second correction method proposed in this application follows the procedure as follows:

[0103] Retain the magnetic survey trajectory of adjacent wells within the preset well depth range;

[0104] Calculate the first corrected trajectory, which is the predicted trajectory of adjacent wells outside the preset well depth range of the current drilling section, calculated based on the well spacing prediction data of each well depth position outside the preset well depth range of the current drilling section and the positive drilling trajectory.

[0105] The corrected magnetic survey trajectory of the adjacent well is composed of the retained magnetic survey trajectory of the adjacent well and the first corrected trajectory.

[0106] For example, a first threshold can be set. If the well depth is greater than the first threshold, the magnetic well distance at each well depth position that is greater than the first threshold is replaced with the predicted well distance at the corresponding well depth position. If the well depth is less than or equal to the first threshold, the magnetic well distance at each well depth position that is less than or equal to the first threshold is retained.

[0107] For example, in one specific embodiment, adjusting the forward drilling trajectory of the undrilled section at the current drilling stage using the adjacent well reference trajectory includes the following steps:

[0108] The starting and ending points of the adjacent wells at the current drilling section are determined based on the reference trajectory of the adjacent wells;

[0109] With the constraint that the drilling well does not collide with the adjacent well, the optimal path from the starting point of the current drilling section of the drilling well to the end point of the adjacent well at the current drilling section is determined by numerical optimization algorithm.

[0110] The trajectory formed by the coordinates of each point in the optimal path obtained in the previous step is used as the adjusted forward drilling trajectory for the un-drilled section in the current drilling segment.

[0111] As in the above embodiment, the optimal path is automatically given through numerical optimization algorithm, thereby automatically giving the well trajectory adjustment plan for the next drilling step, which improves the decision efficiency of adjacent well collision prevention during the drilling process.

[0112] In one specific embodiment, to determine the optimal path, the numerical optimization algorithm can employ various optimization algorithms for finding the optimal solution.

[0113] For example, when using the simulated annealing algorithm to solve for the optimal path, with the constraint that the drilling well and the adjacent well do not collide, the optimal path from the starting point of the current drilling section of the drilling well to the end point of the adjacent well at the current drilling section can be determined using a numerical optimization algorithm, which may include the following steps:

[0114] Step SS1: Initialize the simulated annealing algorithm parameters;

[0115] Step SS2 generates an initial path from the start point of the un-drilled section in the current drilling segment to the end point of the adjacent well in the current drilling segment;

[0116] Step SS3: Based on a preset path perturbation function, apply a perturbation to the current path to generate a new path;

[0117] Step SS4: Calculate the total path cost of the new path generated in the current iteration round. If the total path cost of the new path is lower than the total path cost of the previous round, accept the new path; otherwise, go to step SS3.

[0118] Step SS5: Determine whether the temperature parameter has reached the preset threshold. If so, stop the iteration and find the optimal path between the starting point of the un-drilled section in the current drilling segment and the end point of the adjacent well in the current drilling segment. Otherwise, jump to step SS3.

[0119] For example, in one specific embodiment, the total path cost is the sum of the path length and the penalty term. The penalty term for a point in the path whose distance to the center point of an adjacent well at the same depth is less than or equal to the radius of the adjacent well is greater than the penalty term for a point in the path whose distance to the center point of an adjacent well at the same depth is greater than the radius of the adjacent well. For example, the total path cost can be expressed as follows:

[0120]

[0121] Where f(path) represents the total cost of the path, length(path) represents the path length, and r i W represents the radius of the adjacent well, point represents a point in the path. i This represents the center point of the adjacent well at the same depth as point. Distance represents the distance function between the two points, and penalty factor represents the penalty parameter. The penalty parameter value is greater than the penalty parameter value when the distance between a point in the path and the center point of the adjacent well at the same depth is less than or equal to the radius of the adjacent well.

[0122] The following is a specific application example of the above method. In this specific application example, the magnetic ranging device includes an electrode assembly 2 and a sensor assembly 1. The sensor assembly 1 is disposed inside a pressure-resistant cylinder and includes two radial triaxial fluxgate sensors located within the same cross-section. After the magnetic ranging device is lowered into the drilling well, the sensor assembly 1 is positioned below the electrode assembly 2. The two radial triaxial fluxgate sensors are equidistant from the center of the cross-section. The Z-axis of the triaxial fluxgate sensors is parallel to the central axis of the pressure-resistant cylinder, the X-axis is parallel to and in the same direction as the X-axis of the pressure-resistant cylinder, and the Y-axis is parallel to and in the same direction as the Y-axis of the pressure-resistant cylinder. The triaxial fluxgate sensors can be high-precision sensors, for example, sensors with an accuracy of 0.05 nT or higher. Figure 2 As shown, two radial triaxial fluxgate sensors can be placed at the left and right ends of the center of the cross section, that is, the included angle between the two radial triaxial fluxgate sensors is 180°.

[0123] The process of a magnetic ranging device acquiring magnetic field strength data at a specified depth during drilling may include the following steps:

[0124] Using conventional directional methods, when drilling to the magnetic ranging detection range, a magnetic ranging device is lowered, with the central axis of the magnetic ranging device located at the central axis of the drilling well.

[0125] After sensor assembly 1 reaches the designated well depth, electrode assembly 2 injects alternating current into the formation. Since the conductivity of the adjacent well metal casing is much greater than that of the formation, the alternating current is concentrated on the adjacent well metal casing, forming an upward concentrated current and a downward concentrated current. Among them, the downward concentrated current generates an alternating magnetic field around the adjacent well.

[0126] The triaxial fluxgate sensor inside the pressure-resistant cylinder detects the alternating magnetic field signal at the specified depth of the well.

[0127] To obtain the magnetic well distance data through inversion, the inversion equation established for the aforementioned magnetic ranging equipment is derived based on electromagnetic field theory. The specific theoretical derivation process is as follows:

[0128] Step A1, based on electromagnetic field theory, reveals the following theoretical relationship between the magnetic field strength generated by the adjacent well casing magnetized by alternating current and the horizontal distance between the drilling well and the adjacent well:

[0129]

[0130] In the above formula, H0 represents the magnetic field strength, d represents the distance between the midpoint of the central axis of electrode assembly 2 and the center point of the cross-section where the triaxial fluxgate sensor is located, and σ e σ represents the formation conductivity. c The electrical conductivity r of the adjacent well casing is expressed as r.c h represents the radius of the metal casing of the adjacent well. c The wall thickness of the casing of the adjacent well is represented by μ0, the vacuum permeability is represented by I0, the magnitude of the alternating current is represented by L, and the horizontal distance between the drilling well and the adjacent well is represented by L.

[0131] Step A2: Based on the positions of the two triaxial fluxgate sensors and the magnetic field strengths they respectively acquire, construct the following system of equations:

[0132] L A =L OA +L

[0133] L B =LL OB

[0134]

[0135]

[0136] In the above system of equations, L represents the distance from the midpoint of the cross-section at the location of the radial triaxial fluxgate sensor in the positive drilling well to the central axis of the adjacent well; L A This indicates the distance from the left radial triaxial fluxgate sensor 4 inside the pressure-resistant cylinder to the center axis of the adjacent well; L B H represents the distance from the right-side radial triaxial fluxgate sensor 3 inside the pressure-resistant cylinder to the center axis of the adjacent well; 0A H 0B σ represents the magnetic field strength measured by the left radial triaxial fluxgate sensor 4 and the right radial triaxial fluxgate sensor 3, respectively; d represents the distance between the midpoint of the central axis of the electrode assembly 2 and the midpoint of the cross-section where the triaxial fluxgate sensor is located; σ represents the magnetic field strength measured by the left radial triaxial fluxgate sensor 4 and the right radial triaxial fluxgate sensor 3, respectively ... e σ represents the formation conductivity; c Indicates the conductivity of the metal casing of the adjacent well; r c Indicates the radius of the metal casing of the adjacent well; h c The value represents the wall thickness of the metal casing in the adjacent well; μ0 represents the vacuum permeability; and I0 represents the magnitude of the alternating current.

[0137] Step A3: Solve the system of equations to obtain the formula for calculating the distance L from the midpoint of the cross-section at the location of the radial triaxial fluxgate sensor in the positive drilling well to the central axis of the adjacent well. The formula is expressed as follows:

[0138]

[0139] In the above formula, θ represents the horizontal distance between the left radial triaxial fluxgate sensor 4 and the right radial triaxial fluxgate sensor 3 inside the pressure-resistant cylinder.

[0140] To construct a well spacing prediction model, a dataset for training and testing the initial regression network was obtained by drilling new experimental wells in the field. Specifically, such as... Figure 3 As shown, four experimental wells, T, R1, R2, and R3, were drilled in the field, and relevant downhole data were collected during drilling. T is the main drilling well used to verify the above method, and R1, R2, and R3 are adjacent wells. The horizontal distance from a point on the central axis of well T to a point on the central axis of wells R1, R2, and R3 at the same depth is regarded as the well distance L between the main drilling well and the adjacent wells.

[0141] To obtain the dataset for training and testing the initial regression network, underground rock formations were explored during the drilling of the experimental well to generate lithological data. Well inclination angle data was measured and recorded using a slewing machine, dogleg degree data was calculated and recorded based on the slewing data, and well depth data was recorded while drilling. Since the lithology and drilling parameters of wells R1 and R2 are crucial for guiding the development of well R3, the data from wells R1 and R2 were integrated into a dataset for training and testing the well spacing prediction model. The dataset uses the collected lithology, well depth, well inclination angle, and dogleg degree data from wells R1 and R2 as independent variables, and the well spacing between well T and wells R1 and R2 as the target variable. Before training and testing the well spacing prediction model, the dataset was preprocessed, including removing missing data, taking the median for outliers, and normalization, to ensure data accuracy and consistency. The dataset was then split into training and testing sets in a 0.8:0.2 ratio.

[0142] In this application example, the well spacing prediction model is a ridge regression model, and the network structure of the ridge regression model is as follows: Figure 4 As shown. When training a ridge regression model, the goal is to minimize the loss function to find the optimal model parameters. The minimized loss function can be expressed as follows:

[0143]

[0144] Where L represents the loss function, and a = [a0, a1, a2, ..., a...]. p ] represents the model parameters, X i =[x i1 x i2 , ..., x ip ] represents the i-th wellbore characteristic parameter. In this application example, wellbore characteristic parameters include lithology, well depth, well inclination angle, and dogleg degree, etc. i Let represent the target value of the i-th sample, that is, the label data corresponding to the i-th wellbore feature parameter, i.e., the corresponding well spacing. λ represents the regularization parameter, N represents the number of samples, j represents the subscript number of the model parameter, and p represents the total number of model parameters. Figure 4In the diagram, x1, x2 to x3 represent the elements in the current sample, and y represents the output of the current sample.

[0145] After training, the accuracy of the trained ridge regression model is evaluated using the test set, and the model is tuned by adjusting the regularization parameter of the ridge regression model to improve accuracy. After tuning, the final well spacing prediction model is obtained.

[0146] After obtaining the final well spacing prediction model, the following verification process was performed to verify the effectiveness of the positive drilling trajectory adjustment process:

[0147] Step B1: Collect wellbore characteristic parameters such as lithology, well depth, well inclination angle, and dogleg degree of well R3 as input parameters for the well distance prediction model, and predict the well distance data between well R3 and the drilling well T.

[0148] Step B2: Since the trajectory of the drilling well T is known, the magnetic logging trajectory of well R3 is calculated using magnetic well spacing data and the trajectory of well T in the current drilling section. The magnetic logging trajectory of well R3 is then corrected using well spacing prediction data. Based on the well spacing prediction data for the un-drilled sections in the current drilling section and the trajectory of well T, the un-measured trajectory of well R3 in the current drilling section is calculated. The corrected magnetic logging trajectory of well R3 and the un-measured trajectory of well R3 form the reference trajectory of well R3. The trajectory correction effect of well R3 is as follows: Figure 5 As shown.

[0149] Step B3, with the constraint that the drilling well T does not collide with wells R1, R2, and R3, uses a simulated annealing algorithm to determine the optimal path from the starting point of the undrilled section in the current drilling segment of well T to the set target point. Figure 6 As shown. Figure 6 In the diagram, O1 to O7 represent the coordinates of the points traversed by the optimal path.

[0150] It is important to understand that, in order to avoid collisions with wells R1, R2, and R3, Figure 6 The optimal path for drilling a T-well is not fully shown in the document. For those skilled in the art, it can be found through... Figure 6 The presentation clearly illustrates the well trajectory adjustment process of the positive drilling T-well implemented based on the simulated annealing algorithm in the above embodiments.

[0151] In this application example, the optimal path from the starting point of the un-drilled section in the drilling phase of well T during active drilling is determined using a simulated annealing algorithm. This process includes the following steps:

[0152] Step C1: Initialize the starting point coordinates, ending point coordinates, and adjacent well radius. The starting point is the starting point of the un-drilled section in the drilling section of well T, and the ending point is the end point of the underground drilling of the adjacent well in that drilling section.

[0153] Step C2: Initialize path and temperature parameters;

[0154] Step C3: Based on the preset path scrambling function, randomly select a coordinate on the path, make small-range changes to the x and y coordinates, and generate a new path;

[0155] Step C4: Calculate the total cost of the new path based on the predefined distance calculation function between two coordinates and the total path cost calculation formula. If the total cost of the new path is lower than the total cost of the path in the previous iteration, accept the new path; otherwise, return to the previous step.

[0156] Step C5: If the temperature parameter that gradually decreases in each iteration is less than the preset threshold of 1×10⁻⁶, -3 If the current path is not found, the iteration stops, and the current latest path is taken as the optimal path from the starting point of the drilling section of well T to the set target point. Otherwise, return to step C3 and continue to generate a new path according to the preset path disturbance function until the temperature parameter is less than the preset threshold of 1×10. -3 The iteration stops when the time is right.

[0157] Figure 7 This schematically illustrates a block diagram of a neighboring well anti-collision device 400 according to an embodiment of this application. Figure 7 As shown, an embodiment of this application provides an adjacent well anti-collision device 400 comprising:

[0158] The magnetic logging distance determination module 410 is used to acquire magnetic field strength data and invert the magnetic field strength data to obtain the magnetic logging distance data between the drilled section and the adjacent well in the current drilling section.

[0159] The well spacing prediction module 420 is used to obtain the well characteristic parameters of adjacent wells, input the well characteristic parameters into the constructed well spacing prediction model, and obtain the well spacing prediction data between the well being drilled in the current drilling section and the adjacent wells.

[0160] The magnetic trajectory correction module 430 is used to calculate the magnetic trajectory of the adjacent well using the magnetic well spacing data and the positive drilling trajectory at the current drilling section, and to correct the magnetic trajectory of the adjacent well using the well spacing prediction data.

[0161] The adjacent well reference trajectory determination module 440 is used to calculate and obtain the unmeasured trajectory of the adjacent well in the current drilling section based on the well spacing prediction data of the un-drilled section in the current drilling section and the current drilling trajectory, and the adjacent well reference trajectory is composed of the corrected magnetic survey trajectory of the adjacent well and the unmeasured trajectory of the adjacent well.

[0162] The trajectory adjustment module 450 is used to adjust the forward drilling trajectory of the un-drilled section in the current drilling section using the adjacent well reference trajectory;

[0163] The well spacing prediction model is obtained by training and optimizing the parameters of an initial regression network, and the well characteristic parameters are the characteristics of neighboring wells that affect the spatial coordinates of the drilling well at each well depth.

[0164] In specific embodiments of this application, the wellbore characteristic parameters include one or more of lithology, well depth, well inclination angle, and dogleg degree.

[0165] In a specific embodiment of this application, the well spacing prediction model is obtained by training and optimizing the parameters of the initial regression network, including:

[0166] Obtain wellbore characteristic data and well spacing data of newly drilled experimental wells in the field;

[0167] The wellbore feature data of the experimental wells are used as the input data of the regression network, and the well spacing data of the experimental wells are used as the label data of the initial regression network. Training samples and test samples are constructed based on the input data and label data.

[0168] The initial regression network is trained using the training samples;

[0169] The trained regression network is evaluated using the test samples, and the hyperparameters of the regression network are adjusted according to the evaluation results until the evaluation results meet the preset requirements. Then, the latest hyperparameters are assigned to the regression network to obtain the well spacing prediction model.

[0170] In a specific embodiment of this application, the correction of adjacent well magnetic trajectories using the well spacing prediction data includes:

[0171] The predicted well distances between adjacent wells and the well being drilled at multiple well depth locations are obtained from the well distance prediction data.

[0172] The magnetic logging distances between adjacent wells and the well being drilled at these multiple well depth locations are obtained from the magnetic logging distance data.

[0173] The magnetic ranging deviation is obtained by subtracting the predicted well distance at each well depth from the magnetic well distance at each well depth.

[0174] The maximum confidence interval of the magnetic ranging error is obtained by comparing the various magnetic ranging errors.

[0175] The spatial coordinates of adjacent wells whose magnetic ranging deviations are not within the maximum confidence interval at the corresponding well depth are excluded from the magnetic ranging trajectory of the adjacent wells.

[0176] In a specific embodiment of this application, the correction of adjacent well magnetic trajectories using the well spacing prediction data includes:

[0177] Retain the magnetic survey trajectory of adjacent wells within the preset well depth range;

[0178] Calculate the first corrected trajectory, which is the predicted trajectory of adjacent wells outside the preset well depth range of the current drilling section, calculated based on the well spacing prediction data of each well depth position outside the preset well depth range of the current drilling section and the positive drilling trajectory.

[0179] The corrected magnetic survey trajectory of the adjacent well is composed of the retained magnetic survey trajectory of the adjacent well and the first corrected trajectory.

[0180] In a specific embodiment of this application, adjusting the forward drilling trajectory of the un-drilled section in the current drilling segment using the adjacent well reference trajectory includes:

[0181] The starting and ending points of the adjacent wells at the current drilling section are determined based on the reference trajectory of the adjacent wells;

[0182] With the constraint that the drilling well does not collide with the adjacent well, the optimal path from the starting point of the un-drilled section in the current drilling segment of the drilling well to the end point of the adjacent well in the current drilling segment is determined by numerical optimization algorithm.

[0183] The trajectory composed of the coordinate points in the optimal path is used as the adjusted forward drilling trajectory for the un-drilled section in the current drilling segment.

[0184] In a specific embodiment of this application, the numerical optimization algorithm is a simulated annealing algorithm. The step of determining the optimal path from the starting point of the un-drilled section in the current drilling segment of the current drilling well to the endpoint of the adjacent well in the current drilling segment, with the constraint that the drilling does not collide with the adjacent well, using the numerical optimization algorithm includes:

[0185] Parameter initialization;

[0186] Generate an initial path from the start point of the un-drilled section in the current drilling segment to the end point of the adjacent well in the current drilling segment;

[0187] Perform the new path generation step;

[0188] Calculate the total cost of the new path generated in the current iteration. If the total cost of the new path is lower than the total cost of the path in the previous iteration, accept the new path; otherwise, proceed to execute the new path generation step.

[0189] Determine if the temperature parameter has reached the preset threshold. If so, stop the iteration and use the current latest path as the optimal path between the starting point of the un-drilled section in the current drilling segment and the end point of the adjacent well in the current drilling segment. Otherwise, jump to execute the new path generation step.

[0190] The new path generation step includes: applying a perturbation to the current path to generate a new path.

[0191] In a specific embodiment of this application, the magnetic field strength data is acquired by a magnetic ranging device lowered into the drilling well.

[0192] In one specific embodiment, the adjacent well anti-collision device 400 provided in this application can be implemented as a computer program, and the computer program can be implemented in, for example... Figure 8 The adjacent well collision prevention auxiliary decision-making device shown operates on this device. The memory of the adjacent well collision prevention auxiliary decision-making device can store the various program modules that make up the adjacent well collision prevention device 400. The computer program composed of these program modules causes the processor to execute the steps in the adjacent well collision prevention method described in this specification.

[0193] Figure 8 This diagram schematically illustrates a structural block diagram of an adjacent well collision avoidance auxiliary decision-making device according to an embodiment of this application. Figure 8 As shown, the adjacent well collision avoidance auxiliary decision-making device provided in this application embodiment includes a processor A01, a network interface A02, a display screen A04, an input device A05, and a memory (not shown) connected via a system bus. The processor A01 of the adjacent well collision avoidance auxiliary decision-making device provides computing and control capabilities. The memory of the adjacent well collision avoidance auxiliary decision-making device includes internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01 and a computer program B02. The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A06. The network interface A02 of the adjacent well collision avoidance auxiliary decision-making device is used to communicate with external terminals via a network connection. When the computer program is executed by the processor A01, it implements any of the adjacent well collision avoidance methods in the above embodiments. The display screen A04 of the adjacent well collision prevention auxiliary decision-making device can be an LCD screen or an e-ink screen. The input device A05 of the adjacent well collision prevention auxiliary decision-making device can be a touch layer covering the display screen, or a button, trackball, or touchpad set on the outer shell of the adjacent well collision prevention auxiliary decision-making device, or an external keyboard, touchpad, or mouse, etc.

[0194] On the other hand, this application also provides an adjacent well collision avoidance auxiliary decision-making system, including an adjacent well collision avoidance auxiliary decision-making device and a magnetic ranging device. The magnetic ranging device is communicatively connected to the adjacent well collision avoidance auxiliary decision-making device via a transmission device. During drilling, the magnetic ranging device is lowered into the well to magnetize the adjacent well casing through its own generated alternating current, and detects the magnetic field generated by the magnetized adjacent well casing to obtain magnetic field strength data. The magnetic field strength data is then transmitted to the adjacent well collision avoidance auxiliary decision-making device via the transmission device.

[0195] In one specific embodiment, the magnetic ranging device includes an electrode assembly and a sensor assembly, wherein:

[0196] Electrode assembly for generating alternating current to magnetize adjacent well casing;

[0197] The sensor assembly is used to detect the magnetic field generated by the casing of the adjacent well after magnetization, obtain magnetic field strength data, and send the magnetic field strength data to the adjacent well anti-collision auxiliary decision-making equipment via a transmission device.

[0198] In one specific embodiment, the magnetic ranging device may be: Figure 2 The magnetic ranging device shown.

[0199] For each drilling section, the process of making adjacent well collision avoidance decisions using the aforementioned adjacent well collision avoidance auxiliary decision-making system may include:

[0200] Ground personnel lowered the magnetic ranging equipment into the drilling well and controlled the electrode assembly to inject low-frequency alternating current into the surrounding formation to magnetize the casing of the adjacent well.

[0201] Ground personnel acquire magnetic field strength data collected by sensor components, and use geometric distance calibration calculation methods to invert and calculate magnetic logging distance data;

[0202] The wellbore characteristic parameters detected during the drilling of adjacent wells are input into the constructed ridge regression model to predict the well spacing. The well spacing prediction data is used to correct the outliers in the magnetic logging well spacing data of the drilled sections in the current drilling segment, and the corrected magnetic logging trajectory of the adjacent wells is obtained. Based on the well spacing prediction data and the positive drilling trajectory of the current drilling segment, the undetected trajectory of the adjacent wells in the un-drilled sections of the current drilling segment is deduced. Finally, the undetected trajectory of the adjacent wells and the corrected magnetic logging trajectory of the adjacent wells are combined to form the final reference trajectory of the adjacent wells.

[0203] Using the simulated annealing algorithm, the optimal path between the starting point of the un-drilled section in the current drilling segment of the well and the end point of the underground drilling of the adjacent well in the current drilling segment of the well is found, and then the well trajectory adjustment scheme of the well at the starting point of the well is automatically generated.

[0204] On the other hand, embodiments of this application also provide a machine-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the aforementioned methods for preventing collisions between adjacent wells.

[0205] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0206] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0207] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0208] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for preventing collisions between adjacent wells, characterized in that, The method includes: Obtain magnetic field strength data, and invert the magnetic field strength data to obtain the magnetic logging distance data between the drilled section and the adjacent well in the current drilling section; Obtain the wellbore characteristic parameters of the adjacent wells, input the wellbore characteristic parameters into the constructed well distance prediction model, and obtain the well distance prediction data between the well being drilled in the current drilling section and the adjacent wells; The magnetic logging trajectory of the adjacent well is calculated using the magnetic logging distance data and the positive drilling trajectory at the current drilling section, and the magnetic logging trajectory of the adjacent well is corrected using the well distance prediction data. The unmeasured trajectory of adjacent wells in the current drilling section is calculated based on the predicted well spacing data of the un-drilled section in the current drilling section and the current drilling trajectory. The reference trajectory of adjacent wells is composed of the corrected magnetic survey trajectory of adjacent wells and the unmeasured trajectory of adjacent wells. The forward drilling trajectory of the un-drilled section in the current drilling segment is adjusted using the adjacent well reference trajectory; The well spacing prediction model is obtained by training and optimizing the parameters of an initial regression network, and the well characteristic parameters are the characteristics of neighboring wells that affect the spatial coordinates of the drilling well at each well depth.

2. The method according to claim 1, characterized in that, The wellbore characteristic parameters include one or more of the following: lithology, well depth, well inclination angle, and dogleg degree.

3. The method according to claim 1 or 2, characterized in that, The well spacing prediction model is obtained by training and optimizing the parameters of the initial regression network, including: Obtain wellbore characteristic data and well spacing data of newly drilled experimental wells in the field; The wellbore feature data of the experimental wells are used as the input data of the regression network, and the well spacing data of the experimental wells are used as the label data of the initial regression network. Training samples and test samples are constructed based on the input data and label data. The initial regression network is trained using the training samples; The trained regression network is evaluated using the test samples, and the hyperparameters of the regression network are adjusted according to the evaluation results until the evaluation results meet the preset requirements. Then, the latest hyperparameters are assigned to the regression network to obtain the well spacing prediction model.

4. The method according to claim 1, characterized in that, Correcting the magnetic trajectories of adjacent wells using the well spacing prediction data includes: The predicted well distances between adjacent wells and the well being drilled at multiple well depth locations are obtained from the well distance prediction data. The magnetic logging distances between adjacent wells and the well being drilled at these multiple well depth locations are obtained from the magnetic logging distance data. The magnetic ranging deviation is obtained by subtracting the predicted well distance at each well depth from the magnetic well distance at each well depth. The maximum confidence interval of the magnetic ranging error is obtained by comparing the various magnetic ranging errors. The spatial coordinates of adjacent wells whose magnetic ranging deviations are not within the maximum confidence interval at the corresponding well depth are excluded from the magnetic ranging trajectory of the adjacent wells.

5. The method according to claim 1, characterized in that, Correcting the magnetic trajectories of adjacent wells using the well spacing prediction data includes: Retain the magnetic survey trajectory of adjacent wells within the preset well depth range; Calculate the first corrected trajectory, which is the predicted trajectory of adjacent wells outside the preset well depth range of the current drilling section, calculated based on the well spacing prediction data of each well depth position outside the preset well depth range of the current drilling section and the positive drilling trajectory. The corrected magnetic survey trajectory of the adjacent well is composed of the retained magnetic survey trajectory of the adjacent well and the first corrected trajectory.

6. The method according to claim 1, characterized in that, The adjustment of the forward drilling trajectory of the un-drilled section in the current drilling segment using the adjacent well reference trajectory includes: The starting and ending points of the adjacent wells at the current drilling section are determined based on the reference trajectory of the adjacent wells; With the constraint that the drilling well does not collide with the adjacent well, the optimal path from the starting point of the un-drilled section in the current drilling segment of the drilling well to the end point of the adjacent well in the current drilling segment is determined by numerical optimization algorithm. The trajectory composed of the coordinate points in the optimal path is used as the adjusted forward drilling trajectory for the un-drilled section in the current drilling segment.

7. The method according to claim 6, characterized in that, The numerical optimization algorithm is a simulated annealing algorithm. The process involves using the constraint that the drilling well and adjacent wells do not collide, and determining the optimal path from the starting point of the un-drilled section in the current drilling segment to the endpoint of the adjacent well in the current drilling segment using the numerical optimization algorithm. This includes: Parameter initialization; Generate an initial path from the start point of the un-drilled section in the current drilling segment to the end point of the adjacent well in the current drilling segment; Perform the new path generation step; Calculate the total cost of the new path generated in the current iteration. If the total cost of the new path is lower than the total cost of the path in the previous iteration, accept the new path; otherwise, proceed to execute the new path generation step. Determine if the temperature parameter has reached the preset threshold. If so, stop the iteration and use the current latest path as the optimal path between the starting point of the un-drilled section in the current drilling segment and the end point of the adjacent well in the current drilling segment. Otherwise, jump to execute the new path generation step. The new path generation step includes: applying a perturbation to the current path to generate a new path.

8. The method according to claim 1, characterized in that, The magnetic field strength data was acquired by a magnetic ranging device installed in the drilling well.

9. A device for preventing collisions between adjacent wells, characterized in that, The device includes: The magnetic logging distance determination module is used to acquire magnetic field strength data and invert the magnetic field strength data to obtain the magnetic logging distance data between the well being drilled in the current drilling section and the adjacent well in the current drilling section. The well spacing prediction module is used to obtain the well characteristic parameters of adjacent wells, input the well characteristic parameters into the constructed well spacing prediction model, and obtain the well spacing prediction data between the well being drilled in the current drilling section and the adjacent wells. The magnetic logging trajectory correction module is used to calculate the magnetic logging trajectory of adjacent wells using the magnetic logging well distance data and the positive drilling trajectory at the current drilling section, and to correct the magnetic logging trajectory of adjacent wells using the well distance prediction data. The adjacent well reference trajectory determination module is used to calculate and obtain the unmeasured trajectory of the adjacent wells in the current drilling section based on the well spacing prediction data of the un-drilled section in the current drilling section and the current drilling trajectory, and the adjacent well reference trajectory is composed of the corrected magnetic survey trajectory of the adjacent well and the unmeasured trajectory of the adjacent well. The trajectory adjustment module is used to adjust the forward drilling trajectory of the un-drilled section in the current drilling segment using the adjacent well reference trajectory; The well spacing prediction model is obtained by training and optimizing the parameters of an initial regression network, and the well characteristic parameters are the characteristics of neighboring wells that affect the spatial coordinates of the drilling well at each well depth.

10. The apparatus according to claim 9, characterized in that, The wellbore characteristic parameters include one or more of the following: lithology, well depth, well inclination angle, and dogleg degree.

11. The apparatus according to claim 9, characterized in that, The adjustment of the forward drilling trajectory of the un-drilled section in the current drilling segment using the adjacent well reference trajectory includes: The starting and ending points of the adjacent wells at the current drilling section are determined based on the reference trajectory of the adjacent wells; With the constraint that the drilling well does not collide with the adjacent well, the optimal path from the starting point of the un-drilled section in the current drilling segment of the drilling well to the end point of the adjacent well in the current drilling segment is determined by numerical optimization algorithm. The trajectory composed of the coordinate points in the optimal path is used as the adjusted forward drilling trajectory for the un-drilled section in the current drilling segment.

12. A collision avoidance auxiliary decision-making device for adjacent wells, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the adjacent well collision prevention method as described in any one of claims 1 to 8.

13. A collision avoidance auxiliary decision-making system for adjacent wells, characterized in that, The system includes the adjacent well collision prevention auxiliary decision-making device and the magnetic ranging device according to claim 12, wherein the magnetic ranging device is communicatively connected to the adjacent well collision prevention auxiliary decision-making device via a transmission device; During drilling, the magnetic ranging device is lowered into the well to magnetize the casing of the adjacent well through its own alternating current, and detects the magnetic field generated by the magnetized casing of the adjacent well to obtain magnetic field strength data. The magnetic field strength data is then sent to the adjacent well anti-collision auxiliary decision-making device via the transmission device.

14. The system according to claim 13, characterized in that, The magnetic ranging device includes: Electrode assembly for generating alternating current to magnetize adjacent well casing; The sensor assembly is used to detect the magnetic field generated by the casing of the adjacent well after magnetization, obtain magnetic field strength data, and send the magnetic field strength data to the adjacent well anti-collision auxiliary decision-making device via the transmission device. The sensor assembly includes two triaxial fluxgate sensors located on the same cross-section perpendicular to the central axis of the electrode assembly. After being lowered into the drilling well, the central axis of the electrode assembly is parallel to the central axis of the drilling well.

15. A machine-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the adjacent well collision prevention method as described in any one of claims 1 to 8.

Images