Full-tensor magnetotelluric data scoring method and device, medium and program product

Abstract

The invention discloses a full-tensor magnetotelluric data scoring method and device, a medium and a program product, and the method comprises the steps: firstly obtaining the actual measurement data of a magnetotelluric measurement point, carrying out the response synthesis of the actual measurement data through a one-dimensional anisotropic model and a phase drift function, and obtaining the corresponding synthesis response data, the method comprises the following steps: constructing a robust regression model on the basis of the residual error of actually measured data and synthesized response data, then performing data fitting on the actually measured data by using the robust regression model to obtain a corresponding fitting reference line, and finally, calculating the residual error of the fitting reference line according to the actually measured data of the magnetotelluric measuring point at each frequency under different rotation azimuth angles and the residual error of the fitting reference line. And scoring the magnetotelluric measuring points to obtain a measuring point scoring result. According to the method, objective and automatic magnetotelluric measuring point data scoring is realized, and the requirement of three-dimensional inversion for data evaluation can be effectively met.

Description

Technical Field

[0001] The present invention relates to the field of magnetotelluric technology, and in particular to a full-tensor magnetotelluric data scoring method, device, medium and program product. Background Art

[0002] Magnetotellurics (MT) is a geophysical exploration method that uses naturally occurring alternating electromagnetic fields to study the Earth's electrical structure. With decades of theoretical and technological advancement, MT has found widespread application in various fields, including lithosphere exploration, seismic tectonic research, oil and gas exploration, geothermal exploration, and deep mineral exploration.

[0003] With the rapid development of instrumentation and computer technology, magnetotelluric observation projects have evolved into three-dimensional, large-scale, and high-density projects since the beginning of the 21st century. my country launched the SinoProbe national magnetotelluric standard network (approximately 1,000 measurement points) in 2009. In 2018, PetroChina Oriental Geophysical Company won the bid for the Kuwait 3D oil and gas exploration project. The first phase of the project collected data from 4,137 measurement points at a 2km grid. The "Magnetotelluric Three-Dimensional Exploration Network (Phase I)," a major national science and technology project for deep geophysical exploration, plans to establish a 50km x 50km magnetotelluric standard network across my country by 2024 (work is ongoing, with over 4,000 measurement points planned).

[0004] Magnetotellurics uses the Earth's natural electromagnetic field as a basis for data processing and inversion interpretation. For each measurement point, the data processing output is 12 curves (amplitude and phase, or real and imaginary parts) derived from six sets of parameters (Zxx, Zxy, Zyx, Zyy, Tzx, and Tzy) along the frequency axis. When working near residential areas, factories, high-voltage transmission networks, highways, and other locations with high-voltage equipment, electromagnetic interference can severely distort parts of the frequency band or even the entire measurement point, resulting in erroneous results when interpreting these six curves.

[0005] Therefore, it is crucial to conduct standardized, objective, and quantitative data quality evaluations of large quantities of magnetotelluric data collected under diverse environments. This is crucial for assessing the overall reliability of deep exploration for key national lithosphere exploration, strategic mineral resource evaluation, and deep processes in typical regional mineralization (oil and gas) systems. Furthermore, field data quality evaluations serve as a fundamental reference for inversion and interpretation of data from local oil and gas, geothermal, and mineral exploration activities.

[0006] Currently, there is no objective standard for evaluating the data quality of magnetotelluric field data. For a long time, the quality evaluation of magnetotelluric data has only been possible through the experience of different practitioners, manual visual judgment based on the curve graph, and manual evaluation. This cannot be achieved through efficient automation, and it is impossible to guarantee the objectivity and standardization of the evaluation standards.

[0007] On the other hand, existing MT standards are still based on the late 20th century, when one- and two-dimensional inversions dominated interpretation methods. They evaluate only two complex parameters, Zxy and Zyx, out of six sets of parameters. This is no longer suitable given the increasing sophistication of three-dimensional inversion. Data interpretation for the "Magnetotelluric Three-Dimensional Exploration Network (Phase I)," a major national science and technology project, requires a joint inversion of all six sets of parameters in the three-dimensional full tensor. Therefore, quality evaluation must be performed on all 12 curves at the planned 4,000+ measurement points. Research on automated MT rating methods is urgent.

[0008] In summary, the existing technologies have the following defects: the evaluation criteria for magnetotelluric observation data of specific measuring points are not objective enough, are not quantitative, rely on manual experience, and are not automated; the Dplus and Rhoplus fitting line technologies applied before inversion can only adapt to the four curves of the two components Zxy and Zxy due to their one-dimensional isotropy assumption. When the phase is out of quadrant due to strong three-dimensionality or anisotropy in the ground, the evaluation of the phase data will fail and cannot meet the data evaluation requirements of today's three-dimensional inversion; the existing technologies have not formed an automatic scoring mechanism and cannot cope with the data rating needs of large data volumes and different noise environments (such as the Deep Earth National Magnetotelluric Network Project).

[0009] Therefore, it is urgent to invent a magnetotelluric data scoring method to solve the problems that the existing magnetotelluric data rating methods are not objective enough, rely on manual experience, and cannot adapt to the data evaluation needs of three-dimensional inversion. Summary of the Invention

[0010] In view of this, embodiments of the present invention provide a full-tensor magnetotelluric data scoring method, device, medium, and program product, which at least partially solve the problems existing in the prior art.

[0011] Other features and advantages of the present invention will become apparent from the following detailed description, or may be learned in part by practice of the present invention.

[0012] In order to achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0013] According to a first aspect of an embodiment of the present invention, a full-tensor magnetotelluric data scoring method is provided, the method comprising:

[0014] Acquiring measured data of the magnetotelluric measuring point, wherein the measured data includes impedance parameters and dipole parameters;

[0015] Performing response synthesis on the measured data using a one-dimensional anisotropic model and a phase shift function to obtain corresponding synthetic response data, wherein the synthetic response data includes synthetic impedance response data and synthetic tilt response data;

[0016] Based on the residuals of the measured data and the synthetic response data, a robust regression model is constructed, wherein the robust regression model includes an impedance robust regression model and a dipole regression model;

[0017] Performing data fitting on the measured data using the robust regression model to obtain a fitting reference line corresponding to the measured data;

[0018] The magnetotelluric measuring points are scored according to the residuals between the measured data of each frequency at different rotation azimuths and the fitted reference line to obtain the measuring point scoring results.

[0019] Furthermore, the synthesis process of the synthetic impedance response data includes:

[0020] Using the first one-dimensional anisotropic model and the first phase shift function, the impedance parameter is subjected to a quasi-three-dimensional response synthesis process to obtain first impedance response data. The impedance parameter includes components Zxx, Zxy, Zyx, and Zyy in four directions. The calculation formula of the first impedance response data is: in, is the first impedance response data, f is the frequency, h1 is the thickness of each layer of the first one-dimensional anisotropic model, a1, b1 and θ1 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the first one-dimensional anisotropic model, respectively;

[0021] Wherein, c1 is the drift slope of the first phase drift function, and d1 is the drift frequency origin of the first phase drift function;

[0022] Using the second one-dimensional anisotropic model and the second phase shift function, the impedance parameter is subjected to a pseudo-three-dimensional response synthesis process to obtain second impedance response data. The calculation formula of the second impedance response data is: in, is the second impedance response data, f is the frequency, h2 is the thickness of each layer of the second one-dimensional anisotropic model, a2, b2 and θ2 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the second one-dimensional anisotropic model, respectively;

[0023] Wherein, c2 is the drift slope of the second phase drift function, and d2 is the drift frequency origin of the second phase drift function;

[0024] The first impedance response data and the second impedance response data are synthesized to obtain synthesized impedance response data. in,

[0025] Furthermore, the synthesis process of the synthetic tilter response data includes:

[0026] The impedance parameter is forward modeled using the third one-dimensional anisotropic model to obtain a first forward modeling result. The calculation formula of the first forward modeling result is: in is the first forward modeling result, f is the frequency, h3 is the thickness of each layer of the third one-dimensional anisotropic model, a3, b3 and θ3 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the third one-dimensional anisotropic model, respectively;

[0027] The impedance parameter is forward modeled using the fourth one-dimensional anisotropic model to obtain a second forward modeling result. The calculation formula for the second forward modeling result is: in is the second forward modeling result, f is the frequency, h4 is the thickness of each layer of the fourth one-dimensional anisotropic model, a4, b4 and θ4 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the fourth one-dimensional anisotropic model, respectively;

[0028] The synthetic tilt sub-response data is calculated using the first forward modeling result and the second forward modeling result. The synthetic tilt sub-response data includes two components T' zx and T' zy ,in,

[0029] Furthermore, the loss function φ of the impedance robust regression model Z for:

[0030] in, r ρ Represents the apparent resistivity residual between the measured impedance parameter and the synthetic impedance response data, ρ obv is the apparent resistivity of the measured impedance parameter, ρ fwd is the apparent resistivity of the synthetic impedance response data, It represents the apparent resistivity residual between the impedance parameter and the synthetic impedance response data when the observation azimuth is 30°;

[0031] represents the impedance phase residual between the impedance parameter and the synthetic impedance response data, is the measured impedance phase, is the composite impedance phase.

[0032] Furthermore, the loss function φ of the tilt regression model t for:

[0033] in, It represents the real part residual of the dipole parameter and the synthetic dipole response data when the observation azimuth is 0°. It represents the residual of the imaginary part of the dipole parameter and the synthetic dipole response data when the observation azimuth is 0°. r t represents the residual between the measured tilt parameter and the synthetic tilt response data, t obv are the real and imaginary parts of the measured tilt parameter, t fwd are the real and imaginary parts of the synthetic tilter response data.

[0034] Furthermore, the magnetotelluric measuring points are scored based on the residuals between the measured data of each frequency at different rotation azimuths and the fitted reference line, and the scoring results of the measuring points are obtained, including:

[0035] For each measured data curve of the magnetotelluric measuring point at different rotation azimuths, the measured data curve includes an apparent resistivity curve, an impedance phase curve, a real part curve of the dip, and an imaginary part curve of the dip;

[0036] Based on the measured data curve and the corresponding fitting reference line, the frequency point score of each frequency is calculated according to the residual of the measured data and the fitting reference data of the measured data curve at each frequency;

[0037] According to the frequency point scores of the frequencies, the frequency points having a frequency point score lower than a preset score threshold are regarded as unqualified frequency points;

[0038] In a preset high frequency band, determining whether the number of unqualified frequency points reaches a first preset ratio of the total number of frequency points in the preset high frequency band;

[0039] If the number of unqualified frequency points reaches a first preset ratio of the total number of frequency points in the preset high frequency band, the first preset score is used as the curve score of the measured data curve;

[0040] If the number of unqualified frequency points does not reach the first preset proportion of the total number of frequency points in the preset high frequency band, then determining in the preset low frequency band whether the number of unqualified frequency points reaches the first preset proportion of the total number of frequency points in the preset low frequency band;

[0041] If the number of unqualified frequency points reaches a first preset ratio of the total number of frequency points in the preset low frequency band, the first preset score is used as the curve score of the measured data curve;

[0042] If the number of unqualified frequency points does not reach the first preset proportion of the total number of frequency points in the preset low frequency band, determining whether the number of unqualified frequency points in the entire frequency band reaches the first preset proportion of the total number of frequency points in the entire frequency band;

[0043] If the number of unqualified frequency points in the entire frequency band reaches a first preset ratio of the total number of frequency points in the entire frequency band, the first preset score is used as the curve score of the measured data curve;

[0044] If the number of unqualified frequency points in the entire frequency band does not reach the first preset ratio of the total number of frequency points in the entire frequency band, all unqualified frequency points are eliminated, and the average frequency point score of the qualified frequencies is used as the average curve score of the measured data curve;

[0045] Use a sliding window of a preset length to perform sliding detection on the low-frequency direction of the measured data curve. If the number of unqualified frequency points in the window reaches a second preset proportion of the total number of frequency points in the window, deduct points from the average curve score based on the proportion of the number of unqualified frequency points in the window, and slide the sliding window forward by a preset length.

[0046] If the number of unqualified frequency points in the window does not reach a second preset ratio of the total number of frequency points in the window, sliding the sliding window forward by one frequency point;

[0047] After the sliding detection is completed, the corrected average curve score is used as the curve score corresponding to the measured data curve;

[0048] The total score of the magnetotelluric measuring point is obtained by using the curve scores corresponding to all the measured data curves.

[0049] Furthermore, the total score of the magnetotelluric measuring point is obtained by using the curve scores corresponding to all the measured data curves, including:

[0050] The impedance scores at different rotation azimuths were obtained by taking the arithmetic average of the curve scores corresponding to the apparent resistivity curve and the impedance phase curve at different rotation azimuths.

[0051] Performing a weighted average on the impedance scores at different rotation azimuths according to a first preset weight to obtain a comprehensive impedance score;

[0052] The curve scores of the real part curve and the imaginary part curve of the inclination under different rotation azimuth angles are used to perform arithmetic averaging to obtain the comprehensive score of the inclination;

[0053] According to the second preset weight, the impedance comprehensive score and the inclination comprehensive score are weighted averaged to obtain the total score of the measuring point.

[0054] According to a second aspect of an embodiment of the present invention, there is provided a device, comprising: a processor and a memory;

[0055] The memory is used to store one or more program instructions;

[0056] The processor is used to run one or more program instructions to perform the steps of a full tensor magnetotelluric data scoring method as described in any one of the above items.

[0057] According to a third aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of a full-tensor magnetotelluric data scoring method as described in any one of the above items are implemented.

[0058] According to a fourth aspect of an embodiment of the present invention, a computer program product is provided, comprising a computer program stored on a non-transitory computer-readable storage medium, wherein the computer program comprises program instructions, which, when executed by a computer, enable the computer to implement the steps of a full tensor magnetotelluric data scoring method as described in any one of the above items.

[0059] The full-tensor magnetotelluric data scoring method, device, medium, and program product provided by the embodiments of the present invention have the following advantages:

[0060] The regression model constructed based on multiple one-dimensional magnetotelluric anisotropy forward models and a hybrid phase drift function is more adaptable to common field phenomena such as three-dimensional response and phase super-quadrant compared to the existing Dplus and Rhoplus technologies.

[0061] The fitting reference line of the measured data with interference is automatically calculated through robust regression without human intervention. The designed fitting objective function can still provide reliable fitting results when the outliers reach 40% of the total sample number.

[0062] Taking into account factors such as shallow resolution, maximum detection depth, and near-field frequency band, the score of each frequency point and each curve is automatically calculated according to the residual distribution of robust regression. Finally, the purpose of automatic comprehensive evaluation of the measuring points is achieved through the weighted average of the scores of different types of data.

[0063] The impedance scoring in the embodiment of the present invention uses data in multiple directions of 0°, 30°, and 60°, taking into account both Zxx and Zyy information, and avoiding the mathematical problem of meaningless small-amplitude complex phase caused by directly scoring Zxx and Zyy.

[0064] The full-tensor magnetotelluric data scoring method provided in an embodiment of the present invention has a scoring calculation time of 5 seconds for each magnetotelluric measurement point. The whole process is automatic and does not require human intervention. It can adapt to the automatic scoring and grading of large quantities of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other implementation drawings based on the provided drawings without inventive effort.

[0066] Figure 1 A schematic flow chart of a full-tensor magnetotelluric data scoring method provided in an embodiment of the present invention;

[0067] Figure 2 A schematic diagram of the effect of a fitting reference line for a full-tensor magnetotelluric data scoring method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0069] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0070] Figure 1 The figure shows a flow chart of a full tensor magnetotelluric data scoring method according to an embodiment of the present invention.

[0071] like Figure 1 As shown, the full tensor magnetotelluric data scoring method according to an embodiment of the present invention may include step S100, step S200, step S300, step S400 and step S500.

[0072] In step S100 , measured data of magnetotelluric measuring points are obtained, where the measured data include impedance parameters (Zxx, Zxy, Zyx, Zyy) and tilt parameters (Tzx, Tzy).

[0073] Next, in step S200 , the measured data are subjected to response synthesis using a one-dimensional anisotropic model and a phase shift function to obtain corresponding synthetic response data, which includes synthetic impedance response data and synthetic tilt response data.

[0074] Specifically, the above steps include:

[0075] To construct an impedance model, we first use the one-dimensional anisotropic model F(f,h,a,b,θ) to obtain a quasi-three-dimensional, non-phase super-quadrant response. Then, we simulate the phase super-quadrant through the phase shift function G(f,c,d) to obtain the synthetic impedance response data. The specific process includes:

[0076] First, a quasi-three-dimensional response synthesis process is performed on the impedance parameter using the first one-dimensional anisotropic model and the first phase shift function to obtain first impedance response data. The impedance parameter includes components Zxx, Zxy, Zyx and Zyy in four directions.

[0077] The calculation formula for the above first impedance response data is: in, is the first impedance response data, f is the frequency, h1 is the thickness of each layer of the first one-dimensional anisotropic model, a1, b1 and θ1 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the first one-dimensional anisotropic model, respectively, and h, a, b and θ are all vectors.

[0078] Wherein, c1 is the drift slope of the first phase drift function, and d1 is the drift frequency origin of the first phase drift function.

[0079] Then, the second one-dimensional anisotropic model and the second phase shift function are used to perform a quasi-three-dimensional response synthesis process on the impedance parameter to obtain second impedance response data.

[0080] The calculation formula for the above second impedance response data is: in, is the second impedance response data, f is the frequency, h2 is the thickness of each layer of the second one-dimensional anisotropic model, a2, b2 and θ2 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the second one-dimensional anisotropic model, respectively.

[0081] Wherein, c2 is the drift slope of the second phase drift function, and d2 is the drift frequency origin of the second phase drift function.

[0082] Finally, the first impedance response data and the second impedance response data are synthesized to obtain the synthesized impedance response data at each frequency. in,

[0083] By drawing on the method of using the difference quotient to approximate the vertical magnetic field in the three-dimensional numerical forward modeling of magnetotelluric problems, for non-one-dimensional isotropic responses, the electric field curl can be used to approximate the true vertical magnetic field. Taking Tzx as an example, the inclination expression can be approximated as follows: The two terms in the numerator of the above formula are the electric fields at the two surface edges of a grid along the y direction, and the denominator D is an undetermined complex number that is related to the complex wave number, the forward modeling grid size in the y direction, and the cross-correlation power spectrum of the magnetic field.

[0084] Based on this, assuming that the surface magnetic field is normalized to 1 and that the dipon amplitude cannot exceed 1, a dipon model is constructed to obtain synthetic dipon response data. The specific process includes:

[0085] The impedance parameters are forward modeled using the third one-dimensional anisotropic model, and the first forward modeling results are obtained.

[0086] The calculation formula for the first forward modeling result is: in is the first forward modeling result, f is the frequency, h3 is the thickness of each layer of the third one-dimensional anisotropic model, a3, b3 and θ3 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the third one-dimensional anisotropic model, respectively.

[0087] The impedance parameters are forward modeled using the fourth one-dimensional anisotropic model to obtain the second forward modeling result.

[0088] The calculation formula for the second forward modeling result is: in is the second forward modeling result, f is the frequency, h4 is the thickness of each layer of the fourth one-dimensional anisotropic model, a4, b4 and θ4 are the two equivalent electrical conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the fourth one-dimensional anisotropic model, respectively.

[0089] The synthetic dipole response data are calculated using the first forward modeling results and the second forward modeling results.

[0090] The above synthetic tilt response data includes two components T' zx and T' zy ,in,

[0091] Next, in step S300, a robust regression model is constructed based on the residuals between the measured data and the synthetic response data. The robust regression model includes an impedance robust regression model and a dipole regression model.

[0092] Specifically, the impedance robust regression model is a robust regression function based on the residuals of observed data and simulated data. The formula of the impedance robust regression model is described as follows:

[0093] st h 1,2 >0,

[0094] a 1，2 >0,

[0095] 0≤θ 1,2 ≤π,

[0096] 0≤c 1,2 ≤2,

[0097] -5≤d 1,2 ≤5.

[0098] In order to evaluate the four parameters of impedance (Zxx, Zxy, Zyx, Zyy) at the same time, the residuals must be calculated at the three azimuths of 0°, 30° and 60° respectively. The loss function φ of the above impedance robust regression model is Z for:

[0099] in, r ρ Represents the apparent resistivity residual between the measured impedance parameter and the synthetic impedance response data, ρ obv is the apparent resistivity of the measured impedance parameter, ρ fwd is the apparent resistivity of the synthetic impedance response data, It represents the apparent resistivity residual of the impedance parameter and the synthetic impedance response data when the observation azimuth is 30°. The superscripts 0, 30, and 60 represent the different rotation azimuth angles of the impedance parameter, respectively.

[0100] represents the impedance phase residual between the measured impedance parameter and the synthetic impedance response data, is the measured impedance phase, is the composite impedance phase.

[0101] Specifically, the above-mentioned dipole regression model is a robust regression function based on the residuals of observed data and simulated data. The formula of the dipole regression model is described as

[0102] In order to evaluate the two parameters of the dipole at the same time, the residuals must be calculated at the three azimuths of 0°, 30° and 60° respectively. The loss function φ of the above dipole regression model is t for:

[0103] The superscript numbers are the rotation angles, and the summation symbol represents the summation of all frequencies. It represents the real part residual of the dipole parameter and the synthetic dipole response data when the observation azimuth is 0°. It represents the residual of the imaginary part of the dipole parameter and the synthetic dipole response data when the observation azimuth is 0°. r t represents the residual between the measured tilt parameter and the synthetic tilt response data, t obv are the real and imaginary parts of the measured tilt parameter, t fwd are the real and imaginary parts of the synthetic tilter response data.

[0104] Using the bound-constrained LBFGS-B optimization method (which has callable functions in both Python and Matlab software), we can directly calculate the fitting data when the loss function is minimized.

[0105] Next, in step S400, the measured data is fitted using the robust regression model to obtain a fitting reference line corresponding to the measured data.

[0106] Figure 2 A schematic diagram of the effect of the fitting reference line is shown, where the scattered points are measured data and the solid line is the fitting reference data.

[0107] Next, in step S500 , the magnetotelluric measuring points are scored according to the residuals between the measured data of each frequency at different rotation azimuths of the magnetotelluric measuring points and the fitted reference line to obtain a measuring point scoring result.

[0108] Specifically, first calculate the single frequency point score. The specific steps include:

[0109] For each measured data curve of the magnetotelluric measuring point at different rotation azimuths, the measured data curve includes an apparent resistivity curve, an impedance phase curve, a real part curve of the dip, and an imaginary part curve of the dip.

[0110] Based on the measured data curve and the corresponding fitting reference line, the frequency point score of each frequency is calculated according to the residual of the measured data and the fitting reference data at each frequency. The calculation formula is: s = |100-20×min(5, |r|)|. The corresponding scoring basis is that when the fitting residual is less than or equal to 1.5 times the expected noise, it is excellent; 1.5-2.5 times is good; 2.5-3.5 times is qualified; 3.5-5 times is unqualified.

[0111] Then calculate the score of a single curve. The specific steps include:

[0112] According to the frequency point scores of each frequency, the frequency points with scores lower than the preset score threshold are regarded as unqualified frequency points, and the preset score threshold is preferably 25 points.

[0113] In the preset high-frequency band (a frequency band of two orders of magnitude from the highest frequency to the low-frequency direction), determine whether the number of unqualified frequency points reaches the first preset proportion of the total number of frequency points in the preset high-frequency band, and the first preset proportion is preferably 75%; if the number of unqualified frequency points reaches the first preset proportion of the total number of frequency points in the preset high-frequency band, then the first preset score is used as the curve score of the measured data curve, and the first preset score is preferably 24 points.

[0114] If the number of unqualified frequency points does not reach the first preset proportion of the total number of frequency points in the preset high-frequency band, then in the preset low-frequency band (a frequency band of 2 orders of magnitude from the lowest frequency to the high-frequency direction), it is determined whether the number of unqualified frequency points reaches the first preset proportion of the total number of frequency points in the preset low-frequency band; if the number of unqualified frequency points reaches the first preset proportion of the total number of frequency points in the preset low-frequency band, the first preset score is used as the curve score of the measured data curve.

[0115] If the number of unqualified frequency points does not reach the first preset proportion of the total number of frequency points in the preset low frequency band, it is determined whether the number of unqualified frequency points in the entire frequency band reaches the first preset proportion of the total number of frequency points in the entire frequency band; if the number of unqualified frequency points in the entire frequency band reaches the first preset proportion of the total number of frequency points in the entire frequency band, the first preset score is used as the curve score of the measured data curve.

[0116] If the number of unqualified frequency points in the entire frequency band does not reach the first preset proportion of the total number of frequency points in the entire frequency band, all unqualified frequency points are eliminated, and the average frequency point score of the qualified frequencies is used as the average curve score of the measured data curve.

[0117] For the above-mentioned average curve score, a sliding window of a preset length is used to perform sliding detection on the measured data curve. If the number of unqualified frequency points in the window reaches a second preset proportion of the total number of window frequency points, and the second preset proportion is preferably 25%, then the deduction points are calculated based on the proportion of the number of unqualified frequency points in the window, and the average curve score is deducted based on the deduction points, where the deduction points = 25 × (proportion of unqualified frequency points - 0.25), and the preset length is one order of magnitude; if the number of unqualified frequency points in the window does not reach the second preset proportion of the total number of window frequency points, no deduction is performed.

[0118] The sliding window slides toward the low frequency direction. If no deduction is generated, the sliding window slides forward one frequency point; if a deduction is generated, the sliding window slides forward one window length. In addition, when the remaining frequency band is less than 1 / 5 of a window length, the new window is expanded to the last frequency point.

[0119] After the sliding detection is completed, the corrected average curve score is used as the curve score corresponding to the measured data curve.

[0120] The scoring strategy in the embodiment of the present invention takes into account the effectiveness of shallow and deep frequency bands of magnetotelluric data observations, and the deduction mechanism for dead-band data and near-field data.

[0121] Finally, the total score of the magnetotelluric measurement points is calculated. The specific steps include:

[0122] First, the curve scores corresponding to the apparent resistivity curve and the impedance phase curve at different rotation azimuths are arithmetic averaged to obtain the impedance score at different rotation azimuths. The impedance scores at different rotation azimuths are then weighted averaged according to the first preset weight to obtain the impedance comprehensive score. The curve scores of the real and imaginary parts of the dip curves at different rotation azimuths are arithmetic averaged to obtain the dip comprehensive score. Finally, the impedance comprehensive score and the dip comprehensive score are weighted averaged according to the second preset weight to obtain the total score of the measurement point.

[0123] Alternatively, the total score of the measurement point can be calculated based on the impedance score only, and the dip sub-score and rating are only used as important reference indicators for data inversion and data sharing.

[0124] For example, the total score of each impedance rotation azimuth is the arithmetic average of the scores of two apparent resistivity curves and two impedance phase curves, which are recorded as s0 and s 30 、s 60 , the impedance data is divided into s0, s 30 、s 60 The total score of the dip is the weighted average of the real and imaginary part scores of the four dip curves, with weights of 3, 1, and 1 respectively. The total score of the measuring point is the weighted average of the impedance score and the dip score, with weights of 2 and 1 respectively. In addition, when the impedance is a non-one-dimensional response (the apparent resistivity and impedance phase curves of the curve are not parallel and do not overlap), if the real and imaginary parts of the dip are both less than 0.05, the dip score is 24.

[0125] Data quality evaluation is scored on a percentage basis, with the corresponding score intervals for "excellent (A)", "good (B)", "medium (C)", and "unqualified (D)" being [75, 100], (50, 75], (25, 50], and [0, 25], respectively.

[0126] This embodiment of the present invention uses theoretical magnetotelluric curves as a reference and incorporates robust regression technology to calculate the fitted reference line, fully accounting for factors such as the three-dimensional nature of the medium, phase out-of-quadrant, and the influence of noise flying spots. The calculated total score for the measurement point reflects the reliability and overall quality of the five parameters: Zxy, Zyx, Zxx-Zyy, Tzx, and Tzy.

[0127] In addition, an embodiment of the present invention also provides a device, which includes: a processor and a memory; the memory is used to store one or more program instructions; the processor is used to run one or more program instructions to execute the steps of the full tensor magnetotelluric data scoring method as described above.

[0128] In addition, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the full tensor magnetotelluric data scoring method described above are implemented.

[0129] In addition, an embodiment of the present invention further provides a computer program product, which includes computer program instructions. When the computer program instructions are executed by a processor, the steps of the full tensor magnetotelluric data scoring method described above are implemented.

[0130] The full-tensor magnetotelluric data scoring method, device, medium and program product provided in the embodiments of the present invention adopt multiple one-dimensional anisotropic model forward modeling, add a mixed phase drift function, synthesize and simulate theoretical three-dimensional magnetotelluric impedance (Zxx, Zxy, Zyx, Zyy) and dipole (Tzx, Tzy) response data, use the regression model to perform robust regression (Robust Regression) on the observed data to calculate the fitting line of the magnetotelluric measured data, and construct an automatic scoring algorithm based on the residual value of the measured data to the fitting line and the distribution of large outliers.

[0131] The embodiment of the present invention adopts a regression model constructed based on multiple one-dimensional magnetotelluric anisotropy forward models and a hybrid phase drift function. Compared with the existing Dplus and Rhoplus technologies, it is more adaptable to common field phenomena such as three-dimensional response and phase super-quadrant.

[0132] The embodiment of the present invention automatically calculates the fitting reference line for the measured data containing interference through robust regression without human intervention. The designed fitting objective function can still provide a reliable fitting effect when the outliers reach 40% of the total sample number.

[0133] The embodiment of the present invention comprehensively considers factors such as shallow resolution, maximum detection depth, and near-field frequency band, automatically calculates the score of each frequency point and each curve according to the residual distribution of robust regression, and finally achieves the purpose of automatic comprehensive evaluation of the measuring points through the weighted average of the scores of different types of data.

[0134] The impedance scoring in the embodiment of the present invention uses data in multiple directions of 0°, 30°, and 60°, taking into account both Zxx and Zyy information, and avoiding the mathematical problem of meaningless small-amplitude complex phase caused by directly scoring Zxx and Zyy.

[0135] The full-tensor magnetotelluric data scoring method provided in an embodiment of the present invention has a scoring calculation time of 5 seconds for each magnetotelluric measurement point. The whole process is automatic and does not require human intervention. It can adapt to the automatic scoring and grading of large quantities of data.

[0136] In the embodiments of the present invention, the processor may be an integrated circuit chip with signal processing capabilities. The processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component. It may implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of the present invention may be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules within the decoding processor. The software module may be located in a storage medium well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The processor reads information from the storage medium and, in conjunction with its hardware, completes the steps of the aforementioned method. The storage medium may be a memory, for example, volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Among them, non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct rambus random access memory (DRRAM).The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory. Those skilled in the art will appreciate that in one or more of the above examples, the functions described in the present invention can be implemented using a combination of hardware and software. When software is used, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media includes any medium that facilitates the transmission of computer programs from one place to another. The storage medium can be any available medium that can be accessed by a general or special-purpose computer. Although the present invention has been described in detail above using general instructions and specific embodiments, it will be apparent to those skilled in the art that modifications or improvements can be made based on the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention are within the scope of protection claimed in the present invention.

[0137] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Those skilled in the art can make some simple modifications, equivalent changes or modifications based on the technical content disclosed above, which all fall within the scope of protection of the present invention.

Claims

1. A full tensor magnetotelluric data scoring method, characterized in that: The method comprises: Acquiring measured data of the magnetotelluric measuring point, wherein the measured data includes impedance parameters and dipole parameters; Performing response synthesis on the measured data using a one-dimensional anisotropic model and a phase shift function to obtain corresponding synthetic response data, wherein the synthetic response data includes synthetic impedance response data and synthetic tilt response data; Based on the residuals of the measured data and the synthetic response data, a robust regression model is constructed, wherein the robust regression model includes an impedance robust regression model and a dipole regression model; Performing data fitting on the measured data using the robust regression model to obtain a fitting reference line corresponding to the measured data; The magnetotelluric measuring points are scored according to the residuals between the measured data of each frequency at different rotation azimuths and the fitted reference line to obtain the measuring point scoring results.

2. A full tensor magnetotelluric data scoring method according to claim 1, characterized in that: The synthesis process of the synthetic impedance response data includes: Using the first one-dimensional anisotropic model and the first phase shift function, the impedance parameter is subjected to a quasi-three-dimensional response synthesis process to obtain first impedance response data. The impedance parameter includes components Zxx, Zxy, Zyx, and Zyy in four directions. The calculation formula of the first impedance response data is: in, is the first impedance response data, f is the frequency, h1 is the thickness of each layer of the first one-dimensional anisotropic model, a1, b1 and θ1 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the first one-dimensional anisotropic model, respectively; Wherein, c1 is the drift slope of the first phase drift function, and d1 is the drift frequency origin of the first phase drift function; Using the second one-dimensional anisotropic model and the second phase shift function, the impedance parameter is subjected to a pseudo-three-dimensional response synthesis process to obtain second impedance response data. The calculation formula of the second impedance response data is: in, is the second impedance response data, f is the frequency, h2 is the thickness of each layer of the second one-dimensional anisotropic model, a2, b2 and θ2 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the second one-dimensional anisotropic model, respectively; Wherein, c2 is the drift slope of the second phase drift function, and d2 is the drift frequency origin of the second phase drift function; The first impedance response data and the second impedance response data are synthesized to obtain synthesized impedance response data. in, 3. A full tensor magnetotelluric data scoring method according to claim 1, characterized in that: The synthesis process of the synthetic tilter response data includes: The impedance parameter is forward modeled using the third one-dimensional anisotropic model to obtain a first forward modeling result. The calculation formula of the first forward modeling result is: in is the first forward modeling result, f is the frequency, h3 is the thickness of each layer of the third one-dimensional anisotropic model, a3, b3 and θ3 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the third one-dimensional anisotropic model, respectively; The impedance parameter is forward modeled using the fourth one-dimensional anisotropic model to obtain a second forward modeling result. The calculation formula for the second forward modeling result is: in is the second forward modeling result, f is the frequency, h4 is the thickness of each layer of the fourth one-dimensional anisotropic model, a4, b4 and θ4 are the two equivalent conductivities and equivalent principal axis azimuths of the anisotropy of each layer corresponding to the fourth one-dimensional anisotropic model, respectively; The synthetic tilt sub-response data is calculated using the first forward modeling result and the second forward modeling result. The synthetic tilt sub-response data includes two components T' zx and T' zy ,in, 4. A full tensor magnetotelluric data scoring method according to claim 1, characterized in that: The loss function φ of the impedance robust regression model is Z for: in, r ρ Represents the apparent resistivity residual between the measured impedance parameter and the synthetic impedance response data, ρ obv is the apparent resistivity of the measured impedance parameter, ρ fwd is the apparent resistivity of the synthetic impedance response data, It represents the apparent resistivity residual between the impedance parameter and the synthetic impedance response data when the observation azimuth is 30°; represents the impedance phase residual between the impedance parameter and the synthetic impedance response data, is the measured impedance phase, is the composite impedance phase.

5. A full tensor magnetotelluric data scoring method according to claim 1, characterized in that: The loss function φ of the tilted regression model t for: in, It represents the real part residual of the dipole parameter and the synthetic dipole response data when the observation azimuth is 0°. It represents the residual of the imaginary part of the dipole parameter and the synthetic dipole response data when the observation azimuth is 0°. r t represents the residual between the measured tilt parameter and the synthetic tilt response data, t obv are the real and imaginary parts of the measured tilt parameter, t fwd are the real and imaginary parts of the synthetic tilter response data.

6. A full tensor magnetotelluric data scoring method according to claim 1, characterized in that: The magnetotelluric measuring points are scored based on the residuals between the measured data of each frequency at different rotation azimuths and the fitted reference line, and the scoring results are obtained, including: For each measured data curve of the magnetotelluric measuring point at different rotation azimuths, the measured data curve includes an apparent resistivity curve, an impedance phase curve, a real part curve of the dip, and an imaginary part curve of the dip; Based on the measured data curve and the corresponding fitting reference line, the frequency point score of each frequency is calculated according to the residual of the measured data and the fitting reference data of the measured data curve at each frequency; According to the frequency point scores of the frequencies, the frequency points having a frequency point score lower than a preset score threshold are regarded as unqualified frequency points; In a preset high frequency band, determining whether the number of unqualified frequency points reaches a first preset ratio of the total number of frequency points in the preset high frequency band; If the number of unqualified frequency points reaches a first preset ratio of the total number of frequency points in the preset high frequency band, the first preset score is used as the curve score of the measured data curve; If the number of unqualified frequency points does not reach the first preset proportion of the total number of frequency points in the preset high frequency band, then determining in the preset low frequency band whether the number of unqualified frequency points reaches the first preset proportion of the total number of frequency points in the preset low frequency band; If the number of unqualified frequency points reaches a first preset ratio of the total number of frequency points in the preset low frequency band, the first preset score is used as the curve score of the measured data curve; If the number of unqualified frequency points does not reach the first preset proportion of the total number of frequency points in the preset low frequency band, determining whether the number of unqualified frequency points in the entire frequency band reaches the first preset proportion of the total number of frequency points in the entire frequency band; If the number of unqualified frequency points in the entire frequency band reaches a first preset ratio of the total number of frequency points in the entire frequency band, the first preset score is used as the curve score of the measured data curve; If the number of unqualified frequency points in the entire frequency band does not reach the first preset ratio of the total number of frequency points in the entire frequency band, all unqualified frequency points are eliminated, and the average frequency point score of the qualified frequencies is used as the average curve score of the measured data curve; Use a sliding window of a preset length to perform sliding detection on the low-frequency direction of the measured data curve. If the number of unqualified frequency points in the window reaches a second preset proportion of the total number of frequency points in the window, deduct points from the average curve score based on the proportion of the number of unqualified frequency points in the window, and slide the sliding window forward by a preset length. If the number of unqualified frequency points in the window does not reach a second preset ratio of the total number of frequency points in the window, sliding the sliding window forward by one frequency point; After the sliding detection is completed, the corrected average curve score is used as the curve score corresponding to the measured data curve; The total score of the magnetotelluric measuring point is obtained by using the curve scores corresponding to all the measured data curves.

7. A full tensor magnetotelluric data scoring method according to claim 6, characterized in that: The total score of the magnetotelluric measuring point is obtained by using the curve scores corresponding to all the measured data curves, including: The impedance scores at different rotation azimuths were obtained by taking the arithmetic average of the curve scores corresponding to the apparent resistivity curve and the impedance phase curve at different rotation azimuths. Performing a weighted average on the impedance scores at different rotation azimuths according to a first preset weight to obtain a comprehensive impedance score; The curve scores of the real part curve and the imaginary part curve of the inclination under different rotation azimuth angles are used to perform arithmetic averaging to obtain the comprehensive score of the inclination; According to the second preset weight, the impedance comprehensive score and the inclination comprehensive score are weighted averaged to obtain the total score of the measuring point.

8. A full tensor magnetotelluric data scoring device, characterized in that: The device includes: a processor and a memory; The memory is used to store one or more program instructions; The processor is configured to run one or more program instructions to execute the steps of a full tensor magnetotelluric data scoring method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a full-tensor magnetotelluric data scoring method according to any one of claims 1 to 7.

10. A computer program product, characterized in that The computer program product comprises computer program instructions, which, when executed by a processor, implement the steps of a full tensor magnetotelluric data scoring method according to any one of claims 1 to 7.

Images