An adaptive wellbore correction method based on a wellbore response database

Abstract

This invention discloses an adaptive wellbore correction method based on a wellbore response database, belonging to the field of geophysical logging technology. This adaptive wellbore correction method, based on a wellbore response database, is applied to actual logging data from the small-diameter array induction logging instrument HACRT. For wellbore impacts with intrusion, a wellbore correction model with intrusion is introduced. The intrusion impact is calculated using geometric factors, and the wellbore response is calculated using the wellbore response database to obtain a logging response estimate. This invention solves problems such as skin effect correction of the wellbore model response signal, inconsistency between the wellbore model and the actual logging environment, and multiple solutions in wellbore model parameter inversion. It considers engineering problems in actual logging data processing and is adaptable to harsh well conditions such as larger well diameters and low-resistivity salt-cement slurries.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of geophysical well logging, and particularly relates to an adaptive wellbore correction method based on a wellbore response database. BACKGROUND

[0002] Array induction logging tool provides multiple resolution and multiple detection depth curves, compared with traditional induction logging, has the advantages of high resolution, deep detection depth, obvious invasion indication, accurate measurement of formation resistivity, etc., and has become an important resistivity logging tool for sand shale formation. Array induction logging tool collects downhole rich formation information through design of multiple sub-arrays from short to long and multiple frequencies, and the ground obtains 5 kinds (0.254m, 0.508m, 0.762m, 1.524m and 2.286m) or 6 kinds (increased by 3.048m) of detection depth curves with 3 kinds of resolution (0.305m, 0.610m and 1.219m) through software processing. The short sub-array has high resolution while the wellbore effect is serious, therefore, wellbore effect correction is the key link of array induction logging data processing, especially in the case of large wellbore and low resistance mud. There are four kinds of commercial array induction logging tools: Schlumberger's AIT, Baker Atlas's HDIL, Halliburton's HACRT and China Petroleum Group Logging Company's MIT, each company adopts different wellbore correction methods.

[0003] In AIT of Schlumberger, the borehole response database is established based on numerical calculation and water tank simulation, and the forward calculation of borehole response with skin effect is fitted by double functions. The first weight is to fit the nonlinear formation conductivity function with borehole radius and eccentricity as polynomial fitting coefficients, and the second weight is to fit the two-dimensional weighted linear borehole response function varying with mud and formation conductivity with formation conductivity, borehole radius and eccentricity as fitting function coefficients. The measured and simulated response error square minimum objective function is solved by Levenberg-Marquardt inversion method to obtain borehole radius, eccentricity, mud conductivity and formation conductivity, and then the borehole influence correction is realized by using model response and average response. Corresponding to multiple original signals (AIT patent is 18 real parts, and commercial instrument is 14 real parts) with skin effect influence, the data amount is large, the process is complex, the actual logging data exists measurement error, and the multi-solution is difficult to solve, and real-time processing result is difficult to provide in actual logging. In HDIL of Baker Atlas, the borehole response database based on numerical calculation is established, and then the skin effect correction is carried out, and 56 real part signals are converted into 7 sub-array borehole response libraries without skin effect influence. The polynomial fitting function of formation conductivity with borehole radius, eccentricity and mud conductivity as coefficients is designed to calculate the forward response, and the adaptive borehole influence correction is carried out on the 7 sub-array logging data after skin effect correction. Mud, hole diameter parameters and eccentricity can only be adaptively corrected one by one, and it is required that the measured data has multiple frequency signals which can realize multi-frequency skin effect correction. In HACRT of Halliburton, the borehole correction based on eccentric geometric factor is carried out, and in the process of realizing the present application, the inventor finds that the analytical form of the eccentric geometric factor does not exist, and must be calculated by numerical calculation method. This method has certain limitations. It is not suitable for the case of low mud resistivity and high mud and formation resistivity contrast. In MIT of China Petroleum Group Logging Company, the borehole response database with skin effect of all arrays and all frequencies is established based on numerical calculation, and the logging response of given borehole model parameters is quickly calculated by multiple linear interpolation. This method has large data amount, large storage space requirement and slow calculation speed, the borehole correction considers skin effect, and multi-solution is easy to appear, and the adaptive range is narrow. SUMMARY

[0004] The present application aims at overcoming the shortcomings of the prior art and providing an adaptive borehole correction method based on borehole response database.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] An adaptive borehole correction method based on borehole response database is used for the correction of borehole with invasion effect:

[0007] The invaded borehole model is composed of borehole, invasion, and formation. The logging response described by the geometric factors is:

[0008]

[0009] Without invasion, the logging response is:

[0010]

[0011] Without invasion and borehole, the logging response of the degenerated uniform formation is:

[0012]

[0013] In Equations (3) to (5), and are the borehole, invasion, and formation contribution geometric factors of the jthsubarray, respectively; σ m , σ xo , and σ t correspond to the conductivities of mud, invasion, and formation, respectively; and are the logging responses of the invasion model, borehole model, and uniform formation, respectively;

[0014] According to Equations (4) and (5), the borehole effect without invasion is:

[0015]

[0016] Subtracting from the logging response of the borehole model realizes the borehole correction without invasion, i.e.,

[0017]

[0018] According to Equations (3) and (4), the invasion effect with invasion is equal to the logging response of the invasion model minus the logging response of the borehole model, i.e.,

[0019]

[0020] Therefore, the logging response with invasion is represented by the logging response without borehole and the invasion effect, i.e.,

[0021]

[0022] is the logging response estimation of the invasion model considering invasion;

[0023] The borehole correction formula with borehole filling invasion is

[0024]

[0025] Equations (6) to (10) are respectively the borehole effect without invasion, the borehole correction without invasion, the invasion effect with invasion, the logging response with invasion, and the borehole correction with invasion of the wellbore filling. and respectively the borehole effect without invasion, the borehole correction without invasion, the invasion effect with invasion, the logging response with invasion, and the borehole correction with invasion of the wellbore filling.

[0026] Further, the logging response of the invasion model the logging response of the borehole model and the logging response of the homogeneous formation have been subjected to skin effect correction, which includes the following steps:

[0027] (101) Designing the borehole model parameter range and sectioning data

[0028] The borehole model parameters are the borehole radius, the eccentric distance, the mud conductivity and the formation conductivity, and the value range of each of the four borehole model parameters and the corresponding sectioning data are determined;

[0029] (102) Establishing the borehole model response database

[0030] (103) Establishing the borehole response database without skin effect

[0031] The borehole responses of the six subarrays at three frequencies in the borehole model are subjected to three-frequency and normalized skin effect correction, and in the three-frequency skin effect correction, the borehole responses of subarrays 1 to 5 are first-order derivative correction:

[0032]

[0033] In Equation (1), σ aR is the real part of the subarray logging response; y is the square root of the frequency; is the first-order derivative of the real part of the logging response with respect to the square root of the frequency; σ sc1 is the result after first-order derivative skin effect correction;

[0034] The longest subarray 6 simultaneously uses first-order and second-order derivative correction

[0035]

[0036] In Equation (2), σ is the second-order derivative of the logging response with respect to the square root of the frequency, and σ sc2 is the result after second-order derivative skin effect correction.

[0037] Further, in step (103), when the low-frequency logging response is positive, if the first-order derivative σ in Equation (1) is positive, it is not corrected;

[0038] When the second-order derivative σ If the second derivative is negative, then the corresponding term of the second derivative does not participate in the correction.

[0039] Further, the method further comprises: inverting the borehole model parameters based on the borehole model logging response.

[0040] Further, the specific steps of inverting the borehole model parameters based on the borehole model logging response are:

[0041] (501) determining an inversion objective function

[0042] The invasion model parameter inversion objective function is:

[0043]

[0044] In formula (11), the first term is the error square sum of the resolution-matched logging response of the jth subarray and the invasion model response formula (9); the second term is a borehole correction non-negative inequality constraint penalty function, where α is a penalty factor, β j is a non-negative minimum value of the jth subarray borehole correction; the non-negative constraint is:

[0045]

[0046] In formula (12), the first term is the error square sum of the resolution-matched logging response of the jth subarray and the unprocessed skin effect correction logging value

[0047] The invasion borehole model parameters, borehole radius, instrument position, mud conductivity, formation conductivity, invasion conductivity, and invasion depth are found to minimize the objective function formula (11);

[0048] (502) determining an inversion method

[0049] The invasion parameter inversion is combined with the borehole model parameter inversion, and when the borehole model parameters are inverted, a set of borehole model parameters is given, the invasion parameters are first inverted to determine the invasion depth and the invasion conductivity, and then the borehole model four parameters are inverted by formula (11);

[0050] (503) solving the multi-extremum problem of the objective function formula (11)

[0051] All regions in the search range are searched one by one according to the grid regions of the borehole response database, and finally the parameters at which the global minimum of the objective function is found, and the response in each region is calculated by linear interpolation;

[0052] (504) analyzing the function characteristics of the search interval

[0053] The change characteristics of the objective function with the parameter change are analyzed, the function change characteristics are determined based on a 5-point method of a grid interval, and different parameters adopt different segmentation points;

[0054] (505) determining convergence of golden section extreme value seeking

[0055] The golden point is continuously used in the variable change interval until the relative error of the variable change meets a predetermined requirement, and the search is stopped;

[0056] When the absolute values of two adjacent golden points of the variable are less than a preset value, the iteration is terminated;

[0057] When the absolute value of the average error of the objective function of two adjacent times is less than 0.001, the iteration is terminated, and the average value of the two adjacent variables is taken.

[0058] Further, in the 5-point function characteristic analysis method in step (504), if the intermediate minimum value is missed, then:

[0059] The midpoint of the previous interval and the midpoint of the next interval are taken as the search range, the golden section method is used to determine the minimum value and the corresponding value, and the next interval is continuously searched.

[0060] Further, the invasion parameter inversion in step (502) is specifically:

[0061] The invasion depth parameter inversion objective function is determined by formula (8):

[0062]

[0063] Formula (12) shows that the invasion depth parameter inversion is reduced to finding the invasion radius, calculating the invasion geometric factor and the invasion influence, and minimizing formula (12); the best invasion radius is found, and the invasion conductivity is calculated by formula (13):

[0064]

[0065] In formula (13), N is less than J;

[0066] The four parameters of the borehole model in step (502) are specifically inverted by formula (11):

[0067] ① The instrument is centered, the well radius and the mud conductivity are given, and the formation conductivity is inverted;

[0068] ② The instrument is centered, the well radius is given, and the mud conductivity and the formation conductivity are inverted;

[0069] ③ The instrument is centered, the mud conductivity is given, and the well radius and the formation conductivity are inverted;

[0070] ④ The instrument is centered, and the well radius, the mud conductivity and the formation conductivity are inverted;

[0071] ⑥Instrument eccentricity, given well radius, mud conductivity, inversion instrument eccentricity and formation conductivity;

[0072] ⑦Instrument eccentricity, given well radius, inversion mud conductivity, instrument eccentricity and formation conductivity;

[0073] ⑧Instrument eccentricity, given mud conductivity, inversion well radius, instrument eccentricity and formation conductivity;

[0074] ⑨Instrument eccentricity, inversion well radius, mud conductivity, instrument eccentricity and formation conductivity.

[0075] Further, the specific implementation steps of step (503) are:

[0076] ①Determine the interval number contained in the parameter search range;

[0077] ②Analyze whether the first term characteristics and the non-negative constraint of the second term in the objective function formula (11) are met;

[0078] ③If the objective function formula (11) has a minimum value and meets the constraint, use the golden section to find the minimum value, compare the error with the error obtained last time, and keep the minimum error and the parameter;

[0079] ④If the objective function formula (11) is monotonic and partially meets the constraint, compare the error of the minimum error end point with the error of the previous time, and keep the minimum error and the corresponding end point parameter;

[0080] ⑤If the objective function formula (11) does not meet the constraint, only keep the error minimum end point and the error;

[0081] ⑥After traversing all intervals in the search range, judge whether the interval number of the objective function formula (11) that does not meet the constraint is equal to the total interval number; if the two are equal, the parameters in the search range cannot meet the constraint condition, and the parameter with the minimum error is taken; if the two are not equal, the parameter with the minimum error that meets the constraint condition is taken.

[0082] Further, the parameter segmentation method in step (504) is specifically:

[0083] ①For the formation and mud, compare the errors of the midpoint, left and right golden section points and the two end points;

[0084] ②For well radius and eccentricity, compare the errors of the midpoint, 5% of the starting end point, 95% of the terminal end point and the two end points;

[0085] ③Firstly, the interval end point error and whether the borehole influence correction non-negative constraint is satisfied are calculated; if both end points satisfy the non-negative constraint, the error of three points in the interval is calculated; if the error of one point in the three points is less than the minimum error of the two end points, it is considered that there is a minimum value in the interval; otherwise, it is considered that the function in the interval is monotonic;

[0086] ④If one of the two end points does not satisfy the non-negative requirement, the error of the three points in the middle is not calculated, the function is monotonic, and the two end point errors and the corresponding parameters are output.

[0087] Further, the borehole model parameters are low-pass filtered.

[0088] Further, it further comprises:

[0089] The different sub-arrays are converted into the same resolution by using a low-pass filtering and resolution matching method.

[0090] Further, the logging data are low-pass filtered.

[0091] Compared with the prior art, the present application has the following beneficial effects:

[0092] An adaptive borehole correction method based on a borehole response database is used for the actual logging data of a small-diameter array induction logging instrument HACRT, and the borehole influence correction when there is invasion influence is introduced, the borehole correction model with invasion is introduced, the invasion influence is calculated by using a geometric factor, the borehole response is calculated by using a borehole response database, and the logging response estimation is obtained. The present application solves the problems of skin effect correction of borehole model response signals, inconsistency between the borehole model and the actual logging environment, and multi-solution of borehole model parameter inversion. The engineering problems in the actual logging data processing are considered, and the present application is suitable for harsh well conditions such as large well diameter and low-resistivity salt mud. Mud loss in deep wells and ultra-deep wells is a difficulty in deep fracture sandstone oil and gas evaluation, and the borehole correction with invasion of the present application solves the borehole influence correction problem in mud loss reservoirs, meets the market demand of array induction logging, and can improve the reliability of oil and gas interpretation and evaluation of logging data.

[0093] Further, the skin effect correction of the model library logging response with both borehole and skin effect is performed, the skin effect correction is performed first, then the borehole influence correction is performed, and in the polynomial fitting and normalization correction, the first derivative non-negative and second derivative non-positive control are introduced.

[0094] Further, for the multi-solution problem of borehole model parameter inversion, a method of searching all regions in the search range one by one according to the grid regions of the database is proposed, and finally the parameters at the global minimum of the objective function are found out, so that the multi-solution problem is solved.

[0095] Further, the borehole model parameters are low-pass filtered, and the jumping problem of the model parameters is eliminated.

[0096] Further, low-pass filtering and resolution matching method is used to realize the same resolution of different sub-arrays and eliminate the influence of surrounding rock effect.

[0097] Further, low-pass filtering is performed on the logging data to eliminate high-frequency noise. BRIEF DESCRIPTION OF DRAWINGS

[0098] Figure 1 For the coil system structure of the small-diameter array induction logging instrument HCRT;

[0099] Figure 2 For the wellbore model response library (wellbore radius 0.28m), wherein, Figure 2 (a) is the sub-array 2 wellbore response when the instrument is centered in the wellbore with a radius of 0.28m, Figure 2 (b) is the sub-array 5 wellbore response when the instrument is centered in the wellbore with a radius of 0.28m, Figure 2 (c) is the sub-array 2 wellbore response when the instrument is eccentric with a distance of 0.238m in the wellbore with a radius of 0.28m, Figure 2 (d) is the sub-array 5 wellbore response when the instrument is eccentric with a distance of 0.238m in the wellbore with a radius of 0.28m;

[0100] Figure 3 For the wellbore model containing intrusion;

[0101] Figure 4 For the optional case of four-parameter inversion of the wellbore model;

[0102] Figure 5 For the schematic diagram of missing minimum points in function characteristic analysis;

[0103] Figure 6 For the wellbore correction simulation results, wherein, Figure 6 (a) is the skin effect correction result, Figure 6 (b) is the wellbore correction result, Figure 6 (c) is the synthetic focusing processing result without wellbore correction, Figure 6 (d) is the synthetic focusing processing result after wellbore correction;

[0104] Figure 7 For the wellbore correction results of actual logging data, wherein, Figure 7 (a) is the skin effect correction result, Figure 7 (b) is the wellbore correction result, Figure 7 (c) is the synthetic focusing processing result without wellbore correction, Figure 7 (d) is the synthetic focusing processing result after wellbore correction. DETAILED DESCRIPTION

[0105] In the following, the technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application, so that those skilled in the art can better understand the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work should fall within the scope of the present application.

[0106] It should be noted that the terms "first", "second" and the like in the description and claims of the present application and the above drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units need not be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0107] The borehole response original database of HACRT is the logging response of 6 sub-arrays at 3 frequencies about 4 borehole model parameters (borehole radius, instrument eccentricity, borehole mud conductivity and formation conductivity). In addition to the influence of 4 parameters, the skin effect is also included in the logging influence. The general borehole influence correction is to eliminate the influence of borehole radius, instrument eccentricity and borehole mud conductivity in the 6 sub-array logging signals. However, the actual situation is very complex, the skin effect correction is performed first, and then the borehole influence correction; few formations are regular boreholes, no invasion and no surrounding rock, most formations have invasion and surrounding rock; the borehole environment is harsh: the borehole is irregular, and the high temperature of deep well affects the measurement accuracy; the instrument is required to adapt to salt mud and large borehole environment. There are the following problems in the actual logging data borehole correction.

[0108] (1) The skin effect correction of the skin effect correction of the model library logging response of the borehole and the skin effect. The borehole correction first needs to perform the skin effect correction, and convert into the database without the skin effect influence. Generally, the skin effect influence refers to the influence in the uniform formation, the high frequency measurement value is low, but due to the complex function of the four parameters of the borehole model, the low frequency measurement signal can be smaller than the high frequency measurement signal.

[0109] (2) Borehole influence correction when there is surrounding rock effect. In the limited thickness of non-permeable mudstone, the logging data has both surrounding rock and borehole influence, and the forward borehole model response only has borehole influence.

[0110] (3) Correction for wellbore effects when intrusion occurs. In thick permeable formations, mud intrudes into the formation, and logging data simultaneously show both intrusion and wellbore effects. The forward model response only shows the wellbore effect.

[0111] (4) Correction for wellbore effects when both intrusion and surrounding rock effects are present. In finite thickness permeable formations, logging data simultaneously contain the effects of surrounding rock, intrusion, and wellbore, while the forward model response only contains the wellbore effect.

[0112] (5) The problem of multiple solutions in wellbore model parameter inversion. The wellbore model logging response is a function of four model parameters. With high conductivity mud and large eccentricity, the response is non-monotonic. Given the logging response, there are multiple solutions in inverting the wellbore model parameters. The actual logging response includes invasion. How can we achieve the inversion of the wellbore model parameters corresponding to the actual logging response?

[0113] (6) Noise issues in logging data. Due to well depth, temperature, high pressure, vibration and other factors, the actual logging signal may contain errors.

[0114] (7) Problem of irregular wellbore. In reality, the wellbore is rough, irregular, and elliptical, which leads to inconsistencies between the wellbore model and the theoretical model.

[0115] The present invention will now be described in further detail with reference to the accompanying drawings:

[0116] See Figure 1 , Figure 1 This describes the coil system structure of the HCRT (High-Cut Array Induction Logging) instrument. The HCRT consists of six sub-arrays arranged on one side, operating at three frequencies (12kHz, 36kHz, and 72kHz) with main receiver spacing of 0.152m, 0.254m, 0.432m, 0.737m, 1.270m, and 2.032m, respectively. Each sub-array comprises a transmitting coil, a receiving coil, and a shielding coil. The instrument diameter is 0.079m. After skin effect correction, borehole effect correction, and synthetic focusing processing, the logging data provides curves at three vertical resolutions (0.305m, 0.610m, and 1.219m) for five different detection depths (0.254m, 0.508m, 0.762m, 1.524m, and 2.286m) at five different depths.

[0117] This invention proposes an adaptive wellbore correction method based on a wellbore response database for actual logging data from the small-diameter array induction logging instrument HACRT. This method addresses issues such as skin effect correction of wellbore model response signals, inconsistencies between the wellbore model and the actual logging environment, and multiple solutions in wellbore model parameter inversion. The invention also considers engineering challenges in processing actual logging data and is adaptable to harsh well conditions such as larger well diameters and low-resistivity salt-cement slurries.

[0118] An adaptive wellbore correction method based on a wellbore response database specifically includes the following steps:

[0119] I. Skin effect correction for logging responses of models with both wellbore and skin effects.

[0120] (101) Design the wellbore model parameter range and profile data

[0121] The wellbore model has four parameters: wellbore radius, eccentricity, mud conductivity, and formation conductivity. The value ranges and corresponding meshing data for each parameter are determined. The parameter meshing is determined while ensuring that the response between connected grids is approximately linear, specifically including the following parameters:

[0122] Borehole radius. Based on the instrument radius of 0.04m and the maximum possible borehole radius of 0.30m, the borehole radius range was determined to be 0.05-0.3m, with 14 data points: 0.05, 0.06, 0.08, 0.10, 0.12, 0.14, 0.16, 0.18, 0.20, 0.22, 0.24, 0.26, 0.28, 0.30.

[0123] Eccentricity. Eccentricity is the maximum distance between the instrument housing surface and the wellbore wall. Since the diameter of the instrument housing is a fixed value, a smaller wellbore results in a smaller eccentricity and a smaller eccentricity effect, while a larger wellbore results in a larger eccentricity and a larger eccentricity effect. To reduce computational load, the eccentricity distribution varies depending on the wellbore diameter. The determined eccentricity range and distribution for the 14 wellbore diameters are shown in Table 1.

[0124] Table 1. Eccentricity range and profile for different well diameters (unit: m)

[0125]

[0126] In the table: the first row is the wellbore radius number; the second row is the wellbore radius; the third row is the maximum instrument eccentricity for each wellbore radius; the fourth row is the eccentricity fraction for each wellbore radius; the first column of rows 5 to 17 is the eccentricity fraction data number; the second to 15th columns of rows 5 to 17 are the specific eccentricity fraction data.

[0127] Mud conductivity. The conductivity of oil-based mud is set at 0.01 S / m. Below this value, the wellbore impact is negligible. The conductivity of saturated brine mud can reach 100 S / m. Above this value, the wellbore impact is extremely severe, and logging data becomes invalid. Therefore, the mud conductivity range is determined to be 0.01-100 S / m. Considering the skin effect, the data is divided into 22 points: 0.01, 0.1, 0.2, 0.5, 0.8, 1.0, 2.0, 4.0, 6.0, 8.0, 10.0, 13.0, 16.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, 100.0.

[0128] Formation electrical conductivity. The instrument's resistivity measurement range is 0.2-2000 Ω·m, corresponding to a conductivity of 0.0005-5 S / m. For resistivity greater than 1000 Ω·m (conductivity less than 0.001 S / m), the skin effect is very small, and the signal is less than the instrument's measurement accuracy. In low-resistivity oil layers, the resistivity may reach 0.1 Ω·m (conductivity less than 0.001 S / m). Therefore, the formation electrical conductivity range is taken as 0.001-10 S / m. Considering the skin effect, the data is divided into 23 points: 0.001, 0.005, 0.01, 0.02, 0.05, 0.08, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.5, 2.0, 3.0, 4.0, 5.0, 8.0, 10.0.

[0129] (102) Establish a wellbore model response database

[0130] The well logging response on four parameter grid nodes was calculated in three dimensions, including six subarrays, three operating frequencies, and the real and imaginary parts of the response. Since the real part is not used in actual well logging signal processing, only the real part is ultimately retained. The total number of real part data points is 1,211,364. Figure 2 This is an example of a wellbore model response database. Figure 2 In the middle, the wellbore radius is 0.28m. Figure 2 (a) and Figure 2 (b) shows the instrument centering and the responses of subarrays 2 and 5 at different formations and mud conductivity. Figure 2 (c) and Figure 2 (d) is the response of subarrays 2 and 5 under different formations and mud conductivity with instrument eccentricity (eccentricity distance of 0.238m). Figure 2 This indicates that: ① When centered, the wellbore response changes approximately linearly with the changes in mud and formation conductivity; ② When eccentric, the short subarray exhibits nonlinear changes with the changes in mud and formation conductivity, and the response becomes negative when the mud conductivity is significantly greater than the formation conductivity.

[0131] (103) Establish a database of wellbore responses without skin effect

[0132] Three-frequency and normalized skin effect corrections were performed on the wellbore responses of the six subarrays at three frequencies in the wellbore model. In the three-frequency skin effect correction, the wellbore responses of subarrays 1 to 5 were corrected using the first derivative.

[0133]

[0134] In equation (1), σ aR y represents the real part of the subarray logging response; y is the square root of the frequency. The first derivative of the real part of the logging response with respect to frequency is obtained by fitting three frequency responses with a polynomial and then calculating the first derivative. σ sc1 This is the result after correction for the skin effect of the first derivative.

[0135] The longest subarray 6 is corrected using both first and second derivatives.

[0136]

[0137] In equation (2), σ is the second derivative of the logging response with respect to the square root of the frequency. sc2 This is the result after correction for the second derivative skin effect.

[0138] As can be seen from equation (1), the first derivative must be negative to correct for the skin effect in the logging response. Therefore, if the first derivative is positive, it must be controlled. Based on extensive analysis of measurement signals, this invention finds that when the low-frequency logging response is positive, the first derivative must be controlled to prevent it from being positive; otherwise, the correction will be unreasonable. Therefore, during correction, if the first derivative is positive, no correction is performed, and the low-frequency signal is used directly.

[0139] As can be seen from equation (2), the second derivative must be positive to correct for the skin effect in the logging response. Therefore, when the second derivative is negative, no correction is needed, and the first derivative is used directly for correction.

[0140] II. Wellbore Influence Correction in the Presence of Surrounding Rock Effects

[0141] Low-pass filtering and resolution matching methods are used to ensure that different subarrays have the same resolution, thus eliminating the influence of the surrounding rock effect. Specifically:

[0142] Design a low-pass filter library based on a Gaussian function with 6 subarrays;

[0143] Design a library of low-pass filters for two connected subarrays with different background conductivity.

[0144] Design and implement a low-pass filter program with 6 subarrays based on a low-pass filter library;

[0145] Design and implement a resolution matching program for six subarrays based on a low-pass filter library.

[0146] III. Wellbore Impact Correction When Intrusion Involves

[0147] See Figure 3 , Figure 3 This is a wellbore model with intrusion, consisting of the wellbore, intrusion, and formation. The logging response described by the geometric factor is as follows:

[0148]

[0149] When there is no intrusion into the wellbore, it is simplified to:

[0150]

[0151] Log logging response degenerating into a homogeneous formation without intrusion or wellbore.

[0152]

[0153] In equations (3) to (5), and These are the geometry factors for the wellbore, invasion, and formation contributions of the j-th subarray, respectively; σ m σ xo and σ t These correspond to the electrical conductivity of mud, intrusion, and formation, respectively. and The logging responses for the invasion model, wellbore model, and homogeneous formation, respectively, have all undergone skin effect correction. The purpose of wellbore correction is to eliminate the influence of well diameter, eccentricity, and mud conductivity in the wellbore model and restore it to the mean formation condition, since subsequent signal processing is based on the homogeneous formation geometric factor theory.

[0154] According to equations (4) and (5), the wellbore effect without intrusion is:

[0155]

[0156] Subtract from the wellbore model logging response This allows for non-invasive wellbore correction, i.e.

[0157]

[0158] According to equations (3) and (4), the impact of invasion when there is invasion is equal to the logging response of the invasion model minus the logging response of the wellbore model, that is...

[0159]

[0160] Therefore, the logging response under invasion is represented by the non-invaded wellbore response and the invasion effect, i.e.

[0161]

[0162] Equation (9) is the logging response estimate of the invasion model considering invasion.

[0163] The wellbore correction formula for wellbore filling invasion is as follows:

[0164]

[0165] IV. Correction for the simultaneous presence of surrounding rock effects and intrusion influences

[0166] The second and third scenarios are combined to correct for the effects of surrounding rock and intrusion.

[0167] V. Problem of multiple solutions in wellbore model parameter inversion.

[0168] (501) Determine the inversion objective function

[0169] This invention proposes an intrusive model parameter inversion objective function.

[0170]

[0171] In equation (11), the first term is the logging response of the j-th subarray after resolution matching processing. The sum of squared errors of the intrusion model response (9); the second term is the wellbore correction nonnegative inequality constraint penalty function, where α is the penalty factor, β j The non-negative minimum value is set for the correction of the j-th subarray wellbore. The non-negativity constraint is...

[0172]

[0173] In formula (12) Correcting logging values ​​using untreated skin effect

[0174] Equation (11) shows that the problem of inverting parameters of an invading wellbore model is reduced to finding parameters of an invading wellbore model (well radius, instrument location, mud conductivity, formation conductivity, invading conductivity and invading depth) that minimize the objective function (11).

[0175] (502) Determine the inversion method

[0176] An inversion method combining invasion parameter inversion and wellbore model parameter inversion is proposed. When inverting the wellbore model parameters, for each set of wellbore model parameters, the invasion parameters are inverted first to determine the invasion depth and invasion conductivity, and then the four parameters of the wellbore model are inverted using Equation (11).

[0177] (5021) Intrusion Parameter Inversion Method

[0178] Equation (8) determines the objective function for inverting the intrusion depth parameters.

[0179]

[0180] Equation (12) shows that the inversion of the penetration depth parameter boils down to finding the penetration radius, calculating the penetration geometry factor and penetration effect, and minimizing Equation (12). The optimal penetration radius is then determined, and the penetration conductivity is calculated using Equation (13).

[0181]

[0182] In equation (13), N is less than J.

[0183] (5022) Wellbore Model Four-Parameter Inversion

[0184] The inversion of wellbore model parameters is based on the determination of four wellbore model parameters (wellbore radius, eccentricity, mud conductivity, and formation conductivity) from six subarray logging signals. Based on possible scenarios in actual logging, the inversion is divided into the following nine cases. The instrument's position in the wellbore is categorized into two types: centered and eccentric. Centered positioning has four cases, and eccentric positioning has five cases, namely:

[0185] ① With the instrument centered, given the well radius and mud conductivity, invert the formation conductivity;

[0186] ② With the instrument centered and the well radius given, invert the mud conductivity and formation conductivity;

[0187] ③ With the instrument centered, the mud conductivity, inversion well radius, and formation conductivity are given;

[0188] ④ With the instrument centered, invert the well radius, mud conductivity, and formation conductivity;

[0189] ⑤ Instrument eccentricity: Given the well radius, mud conductivity, and instrument eccentricity, invert the formation conductivity;

[0190] ⑥ Instrument eccentricity: Given the well radius and mud conductivity, invert the instrument eccentricity and formation conductivity;

[0191] ⑦ Instrument eccentricity: Given the well radius, invert mud conductivity, instrument eccentricity, and formation conductivity;

[0192] ⑧ Instrument eccentricity, given mud conductivity, invert well radius, instrument eccentricity and formation conductivity;

[0193] ⑨ Instrument eccentricity, inversion well radius, mud conductivity, instrument eccentricity and formation conductivity.

[0194] For all nine scenarios, formation conductivity is inverted. When considering the invasion model, the invasion depth and invasion conductivity must be inverted. When the instrument is eccentric, the instrument eccentricity in the well is a variable and cannot remain constant. Therefore, scenario 5 is impossible, or in other words, the instrument's condition in the wellbore can only be one of the other eight scenarios. Figure 4 This is a possible scenario in four-parameter inversion.

[0195] (503) Solve the problem of multiple extrema.

[0196] The objective function (11) has multiple variables. Even with given wellbore diameter, mud, and eccentricity, three parameters—formation conductivity, invasion conductivity, and invasion depth—still need to be determined. Therefore, equation (11) is prone to multiple extrema. Analysis of the reasons for these extrema reveals two aspects: firstly, the function's variation is non-monotonic; secondly, the search area is too large. To address this, this invention proposes a method that searches through all regions within the search range according to the grid regions of the database, finally finding the parameters at the global minimum of the objective function. The response in each region is calculated using linear interpolation, eliminating the problem of multiple extrema. Specific implementation steps:

[0197] ① Determine the number of intervals included in the parameter search range.

[0198] ②Analyze the first characteristic and the second non-negativity constraint of the objective function (11) on an interval-by-interval basis to see if they are satisfied.

[0199] ③ If the function has a minimum value and satisfies the constraints, use the golden ratio to find the minimum value, compare it with the error obtained in the previous step, and retain the minimum error and the parameter.

[0200] ④ If the function is monotonic and partially satisfies the constraints, compare the error at the minimum error endpoint with the previous error, and retain the minimum error and the corresponding endpoint parameters.

[0201] ⑤ If the function does not satisfy the constraints, retain only the endpoint with the smallest error and the error itself.

[0202] ⑥ After traversing all intervals within the search range, determine whether the number of intervals where the function does not satisfy the constraints equals the total number of intervals. If they are equal, it means that the parameter within the search range does not satisfy the constraints, and the parameter with the smallest error is selected. If they are not equal, the parameter with the smallest error that satisfies the constraints is selected.

[0203] (504) Quickly and robustly analyze the function characteristics of the search interval.

[0204] There are various methods for analyzing the characteristics of interval functions, such as the equal division method, the midpoint method, the golden section method, and the random point placement method. This invention analyzes the changing characteristics of the objective function when parameters change, and finds that different parameters have different effects on the objective function. The eccentricity has the greatest and nonlinear effect, while the mud effect has an approximately linear impact on the three short subarrays. A five-point grid interval method is proposed to determine the changing characteristics of the function, using different division points for different parameters.

[0205] ① For strata and mud, calculate and compare the errors of the midpoint, left and right golden section points with the two endpoints.

[0206] ② For well radius and eccentricity, calculate the error at the midpoint, 5% of the starting endpoint, 95% of the ending endpoint, and compare it with the error at both endpoints.

[0207] ③ First, calculate the interval endpoint errors and whether the wellbore influence correction non-negativity constraint is satisfied. If both endpoints satisfy the non-negativity constraint, calculate the errors at three points in the interval. If the error at any of the three points is less than the minimum error at the endpoints, then a local minimum exists within the interval. Otherwise, the function is considered monotonic within the interval, and the maximum value problem is not considered.

[0208] ④ If one of the two endpoints does not meet the non-negativity requirement, the error of the three intermediate points is no longer calculated. The function is considered to be monotonic, and the errors of the two endpoints and their corresponding parameters are output.

[0209] (507) Determine the convergence of finding the extrema of the golden ratio.

[0210] The idea behind the golden ratio is to continuously use the golden point within the range of the independent variable until the relative error of the independent variable meets a predetermined requirement before stopping the search. However, in practical application, the following problems have been observed:

[0211] ① When the variable is very small, such as when the instrument is centered, the eccentricity is equal to 0, and the relative error is difficult to meet.

[0212] ② The error changes very slowly, and it takes a long time to make the relative error of the variable meet the requirements, and sometimes it cannot meet the requirements.

[0213] To address these two issues, two additional checks were added. The first check is to terminate the iteration when the absolute value of two adjacent golden points of a variable is less than a certain number (0.005 for eccentricity and well radius, and 0.001 for mud and formation).

[0214] The second criterion is to terminate the iteration when the absolute value of the average error of the objective function between two consecutive iterations is less than 0.001, and then take the average value of the two adjacent variables.

[0215] (505) Solving anomalies in function property analysis

[0216] During testing of the inversion algorithm, it was found that the 5-point function characteristic analysis method misses the minimum value of the objective function, such as... Figure 4 As shown. The first interval [x] a ,x b The function within the second interval [x] is monotonically decreasing. d ,x e The value increases monotonically within the range, missing the intermediate minimum value x. c .

[0217] Let ku = 1 and kd = 1 represent the curves monotonically increasing and monotonically decreasing, respectively. If kd = 1 appears for the first time, and ku = 1 appears for the second time immediately afterward, then... Figure 5 If the situation is such that a minimum value exists in the middle, then we take the midpoint of the previous interval and the midpoint of the next interval as the search range, use the golden section method to determine the minimum value and its corresponding value, and assign ku and kd to 0, and continue searching for the next interval.

[0218] VI. Noise issues in well logging data

[0219] In deep wells, the first step is to determine if the shortest subarray 1 is faulty; if so, it is not used. A low-pass filter is designed, and the mud and formation conductivity low-pass filters are inverted from the ground. Then, the wellbore model logging response database is quickly calculated, and the logging response is calculated. Finally, the influence of mud conductivity is eliminated.

[0220] VII. Problems with irregular wellbore

[0221] The design incorporates a low-pass filter to filter the inverted wellbore radius and eccentricity, and then rapidly calculates the wellbore model logging response database to eliminate the influence of wellbore radius and eccentricity.

[0222] Example 1

[0223] Wellbore impact correction simulation. A 7-layer formation model with wellbore was established. The well diameter is 0.2032m, with two layers each of 1m and 2m thickness, one layer of 3m thickness, and upper and lower surrounding rock layers. The resistivity of the target layer is 20Ω·m, the resistivity of the surrounding rock is 2Ω·m, and the resistivity of the wellbore is 0.1Ω·m. Figure 6 These are the simulation results of wellbore calibration. During wellbore calibration, both the wellbore radius and mud conductivity are simultaneously adaptive. Figure 6 (a), (b), (c), and (d) in the figure represent the skin effect correction result, the wellbore correction result, the composite focusing result without wellbore correction, and the composite focusing result after wellbore correction, respectively. Figure 6 The following characteristics are displayed.

[0224] (1) In the skin effect correction results ( Figure 6a) Subarrays 1 and 2 have high resolution and can distinguish 2m and 3m thick strata. In the surrounding rock, the readings are 0.46Ω·m and 1.07Ω·m, respectively, which are significantly deviated from the surrounding rock resistivity of 2Ω·m due to the influence of low-resistivity mud. In the target layer, the readings are 0.54Ω·m and 1.85Ω·m, respectively, which are completely deviated from the target layer resistivity of 20Ω·m due to the influence of low-resistivity mud. The remaining subarrays have low resolution and cannot distinguish 3m thick strata. The response is recorded using the transmitting coil as the recording point. Due to the inconsistency of the recording points of each subarray, the position of the maximum response value is inconsistent.

[0225] (2) In the wellbore correction results ( Figure 6 (b) The wellbore response of the surrounding rock and the target layer is effectively corrected. In the surrounding rock above 5m from the formation, the wellbore influence is completely corrected, and the subarrays almost overlap. In the target layer, for a 1m thick formation, subarray 1 is larger than subarray 2; for 2m and 3m thick formations, subarray 1 and subarray 2 almost overlap. The curve fluctuations are caused by fluctuations in calculation accuracy, which correspond to measurement errors in actual well logging.

[0226] (3) Synthetic focusing processing results without wellbore correction ( Figure 6 c) At the shallowest depth of 0.254 m, the readings for the surrounding rock and the target layer are 0.99 Ω·m and 1.68 Ω·m, respectively, significantly deviating from the true values ​​of 2 Ω·m and 20 Ω·m. Within the target layer, the reading at the 0.508 m depth is 6.23 Ω·m, also significantly deviating from the true value of 20 Ω·m. The other depth curves largely overlap, but due to the combined influence of the wellbore and the surrounding rock, the maximum reading for the target layer of 12.58 Ω·m still deviates from the true value of 20 Ω·m.

[0227] (4) Synthetic focusing processing results after wellbore correction Figure 6 d) The wellbore and surrounding rock were significantly and effectively corrected, the five depth curves almost overlapped, and the true value was read from the formation with a thickness of more than 2m.

[0228] Example 2

[0229] Correction for wellbore impact based on actual logging data. Figure 7 This is the wellbore correction effect based on actual logging data from a certain oilfield. Figure 7 (a), (b), (c), and (d) in the figures represent the skin effect correction result, the caliper curve, the composite focusing result without wellbore correction, and the composite focusing result after wellbore correction, respectively. This is a saline mud well with a measured mud resistivity of approximately 0.037 Ω·m and a wellbore diameter of 0.168 m. The well has low mud resistivity, but exhibits enlargement and eccentricity. Therefore, during wellbore correction, a given caliper curve is selected, and the mud conductivity, eccentricity, and formation conductivity are adaptively searched. Figure 7 The following characteristics are displayed.

[0230] (1) Skin effect correction results Figure 7 In (a), short subarrays 1 and 2 are severely affected by the wellbore, deviating significantly from other curves; near 7216m, due to wellbore enlargement (… Figure 7 b) Subarray 2 shows a negative value anomaly, and subarrays 3 to 6 show fluctuations.

[0231] (2) The well diameter measurement curve shows that the well diameter value fluctuates between 0.174m and 0.276m, which is greater than the drill bit diameter of 0.168m, and shows obvious enlargement between 7210m and 7212m.

[0232] (3) Synthetic focusing processing results without wellbore correction ( Figure 7 In c), the deepest depth curve of 2.286m and the second deepest depth curve of 1.524m basically overlap, while the shallower depth curves of 0.254m, 0.508m and 0.762m are separated from them. It is impossible to distinguish whether the wellbore influence or the intrusion influence is present, and it is impossible to distinguish between the permeable layer and the non-permeable layer.

[0233] (4) Synthetic focusing processing results after wellbore correction Figure 7 In section d), the wellbore influence was significantly eliminated, clearly showing the permeable and non-permeable layers. In the 7191m-7202m interval, the curves largely overlapped, indicating a non-permeable layer. In the 7203m-7216m interval, among the five curves at different detection depths, three shallow detection depth curves showed different separation degrees, indicating different invasion depths. The processed and reasonable curves provide a guarantee for accurate oil and gas interpretation and evaluation.

[0234] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. An adaptive wellbore correction method based on a wellbore response database, characterized in that, The correction for wellbore with intrusion effects is as follows: A wellbore model containing invasion is obtained, the wellbore model consisting of the wellbore, invasion, and formation. The logging response of the wellbore model, described using geometric factors, is as follows: When there is no intrusion into the wellbore, the logging response is: When there is no intrusion or wellbore, the logging response for a formation degenerating into a homogeneous formation is: In equations (3) to (5), and These are the geometry factors for the wellbore, invasion, and formation contributions of the j-th subarray, respectively; σ m σ xo and σ t These correspond to the electrical conductivity of mud, intrusion, and formation, respectively. and These correspond to the logging responses of the intrusion model, the wellbore model, and the homogeneous formation, respectively. According to equations (4) and (5), the wellbore effect without intrusion is: Log logging response from the wellbore model Subtracting wellbore impact Achieve wellbore correction without intrusion, i.e. According to equations (3) and (4), the impact of invasion when there is invasion is equal to the logging response of the invasion model minus the logging response of the wellbore model, that is: Therefore, the logging response and invasion effect during invasion are expressed as follows: Equation (9) considers the logging response estimation of the invasion model; The wellbore is corrected during wellbore invasion filling; the correction formula is as follows: In equations (6) to (10), and These are the wellbore influence when the j-th subarray has no invasion, the wellbore correction when there is no invasion, the invasion influence when there is invasion, the logging response when there is invasion, and the wellbore correction when the wellbore is filled with invasion. Wellbore correction during wellbore filling and invasion Evaluate the hydrocarbon properties of deep fractured sandstone.

2. The adaptive wellbore correction method based on a wellbore response database according to claim 1, characterized in that, Well logging response of the intrusion model Logging response of wellbore model Logging response of uniform formation All have undergone skin effect correction, which includes the following steps: (101) Design the wellbore model parameter range and profile data The wellbore model parameters are wellbore radius, eccentricity, mud conductivity, and formation conductivity. The value ranges and corresponding subdivision data of the four wellbore model parameters are determined respectively. (102) Establish a wellbore model response database (103) Establish a database of wellbore responses without skin effect Three-frequency and normalized skin effect corrections were performed on the wellbore responses of the six subarrays at three frequencies in the wellbore model. In the three-frequency skin effect correction, the wellbore responses of subarrays 1 to 5 were corrected using the first derivative. In equation (1), σ aR y represents the real part of the subarray logging response; y is the square root of the frequency. σ is the first derivative of the real part of the logging response with respect to frequency; sc1 This is the result after correction for the skin effect of the first derivative; The longest subarray 6 utilizes both first-order and second-order derivative corrections. In equation (2), σ is the second derivative of the logging response with respect to the square root of the frequency. sc2 This is the result after correction for the second derivative skin effect.

3. The adaptive wellbore correction method based on a wellbore response database according to claim 2, characterized in that, In step (103), when the low-frequency logging response is positive, if the first derivative in equation (1) If a positive value is found, no correction is performed; When the second derivative of equation (2) When it is negative, the second derivative is... The corresponding item is not included in the correction.

4. The adaptive wellbore correction method based on a wellbore response database according to claim 1, characterized in that, Also includes: Wellbore model parameters are inverted based on well logging response from the wellbore model.

5. The adaptive wellbore correction method based on a wellbore response database according to claim 4, characterized in that, The specific steps for inverting wellbore model parameters based on wellbore model logging response are as follows: (501) Determine the inversion objective function The objective function for inverting intrusive model parameters is: In equation (11), the first term is the logging response of the j-th subarray after resolution matching processing. The sum of squared errors of the intrusion model response (9); the second term is the wellbore correction nonnegative inequality constraint penalty function, where α is the penalty factor, β j The minimum non-negative value for the correction of the j-th subarray wellbore; the non-negativity constraint is: In formula (12) Correcting logging values ​​using untreated skin effect Find the well radius, instrument location, mud conductivity, formation conductivity, invasion conductivity and invasion depth of the invading wellbore model to minimize the objective function (11); (502) Determine the inversion method The inversion of invasion parameters is combined with the inversion of wellbore model parameters. When inverting wellbore model parameters, for each set of wellbore model parameters, the invasion parameters are inverted first to determine the invasion depth and invasion conductivity, and then the four parameters of the wellbore model are inverted by equation (11). (503) Solve the multi-extremum problem of objective function (11). The search is performed by iterating through all regions in the grid region of the well response database, and finally the parameters that are minimized globally in the objective function are found. The response in each region is calculated by linear interpolation. (504) Analyze the functional characteristics of the search interval. The characteristics of the objective function change when the parameters change are analyzed. The characteristics of the function change are determined based on the 5-point grid interval method, and different division points are used for different parameters. (505) Determine the convergence of finding the extrema of the golden ratio. The search continues to use the golden point within the range of the independent variable until the relative error of the independent variable meets the predetermined requirements. The iteration terminates when the absolute value of two adjacent golden points of a variable is less than a preset value. The iteration terminates when the absolute value of the average error of the objective function between two consecutive iterations is less than 0.001, and the average value of the two adjacent variables is taken.

6. The adaptive wellbore correction method based on a wellbore response database according to claim 5, characterized in that, If the intermediate minimum value is missed when using the 5-point function characteristic analysis method in step (504), then: Take the midpoint of the previous interval and the midpoint of the next interval as the search range, use the golden section method to determine the minimum value and its corresponding value, and continue searching the next interval.

7. The adaptive wellbore correction method based on a wellbore response database according to claim 5, characterized in that, The intrusion parameter inversion in step (502) specifically involves: The objective function for inverting the intrusion depth parameter, determined by equation (8), is as follows: Equation (12) shows that the inversion of the penetration depth parameter boils down to finding the penetration radius, calculating the penetration geometry factor and penetration effect, and minimizing Equation (12); the optimal penetration radius is then determined, and the penetration conductivity is calculated using Equation (13): In equation (13), N is less than J; In step (502), the four parameters of the wellbore model are specifically obtained by inverting equation (11): ① With the instrument centered, given the well radius and mud conductivity, invert the formation conductivity; ② With the instrument centered and the well radius given, invert the mud conductivity and formation conductivity; ③ With the instrument centered, the mud conductivity, inversion well radius, and formation conductivity are given; ④ With the instrument centered, invert the well radius, mud conductivity, and formation conductivity; ⑥ Instrument eccentricity: Given the well radius and mud conductivity, invert the instrument eccentricity and formation conductivity; ⑦ Instrument eccentricity: Given the well radius, invert mud conductivity, instrument eccentricity, and formation conductivity; ⑧ Instrument eccentricity, given mud conductivity, invert well radius, instrument eccentricity and formation conductivity; ⑨ Instrument eccentricity, inversion well radius, mud conductivity, instrument eccentricity and formation conductivity.

8. The adaptive wellbore correction method based on a wellbore response database according to claim 5, characterized in that, The specific implementation steps of step (503) are as follows: ① Determine the number of intervals included in the parameter search range; ②Analyze the first characteristic and the second non-negativity constraint in the objective function (11) on an interval-by-interval basis to see if they are satisfied; ③ If the objective function (11) has a minimum value and satisfies the constraints, use the golden section to find the minimum value, compare it with the error obtained in the previous step, and retain the minimum error and parameters; ④ If the objective function (11) is monotonic and partially satisfies the constraints, compare the error of the minimum error endpoint with the previous error, and retain the minimum error and the corresponding endpoint parameters; ⑤ If the objective function (11) does not satisfy the constraints, retain only the endpoint with the smallest error and the error; ⑥ After traversing all intervals within the search range, determine whether the number of intervals in which the objective function (11) does not satisfy the constraints is equal to the total number of intervals; if the two are equal, then the parameters within the search range cannot satisfy the constraints, and the parameter with the smallest error is selected; if the two are not equal, the parameter with the smallest error that satisfies the constraints is selected.

9. The adaptive wellbore correction method based on a wellbore response database according to claim 5, characterized in that, The specific method for segmenting the parameters in step (504) is as follows: ① For strata and mud, calculate and compare the errors of the midpoint, left and right golden section points with the two endpoints; ② For well radius and eccentricity, calculate the errors at the midpoint, 5% of the starting endpoint, and 95% of the ending endpoint and compare them with the errors at the two endpoints; ③ First, calculate the interval endpoint error and whether the wellbore influence correction non-negative constraint is satisfied; if both endpoints satisfy the non-negative constraint, calculate the error of 3 points in the interval; if the error of one of the 3 points is less than the minimum error of the two endpoints, then it is considered that there is a local minimum in the interval; otherwise, it is considered that the function in the interval is monotonic. ④ If one of the two endpoints does not meet the non-negativity requirement, the error of the three intermediate points will not be calculated. The function is monotonic, and the error of the two endpoints and the corresponding parameters will be output.

10. The adaptive wellbore correction method based on a wellbore response database according to claim 1, characterized in that, Low-pass filtering is applied to the parameters of the inverted wellbore model.

11. The adaptive wellbore correction method based on a wellbore response database according to claim 1, characterized in that, Also includes: Low-pass filtering and resolution matching methods are used to convert different subarrays to the same resolution.

12. The adaptive wellbore correction method based on a wellbore response database according to claim 1, characterized in that, Low-pass filtering is applied to the well logging data.

Images