A Superconducting Electromagnetic 3D Modeling Method Based on Measured Emission Waveform Reconstruction

By separating and reconstructing the distortion characteristics of the emission current, a measured emission waveform model is established, which solves the problem that existing three-dimensional electromagnetic modeling methods cannot reflect the actual emission current distortion, and realizes accurate simulation and data interpretation of superconducting electromagnetic response.

CN123118976APending Publication Date: 2026-11-17JILIN UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611619137.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-20
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing three-dimensional electromagnetic modeling methods are difficult to reflect the distortion characteristics of actual transmission current, such as reverse overshoot, tail current, and damped oscillation. Furthermore, directly using measured waveforms can easily introduce noise or weaken the physical meaning of overshoot, tail current, and oscillation components.

Method used

By separating and reconstructing the overshoot tail component and damped oscillation component of the transmission current, a model representing the measured transmission waveform is established. Measurement noise is removed, and three-dimensional electromagnetic calculations are performed using the finite difference time-domain method with recursive convolution and non-uniform time steps, thus achieving accurate simulation of the actual transmission waveform.

Benefits of technology

It improves the simulation accuracy and data interpretation reliability of superconducting transient electromagnetic response, distinguishes the emission current distortion response from the polarization response of the subsurface medium, and provides a modeling basis for identifying anomalous attenuation and polarity reversal.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123118976A_ABST
    Figure CN123118976A_ABST
Patent Text Reader

Abstract

This application belongs to the field of transient electromagnetic detection technology, specifically a superconducting electromagnetic three-dimensional modeling method based on the reconstruction of measured transmission waveforms. The method includes: obtaining the measured transmission current of the transmission system; separating the overshoot tail component and damped oscillation component from the residual current after turn-off during the turn-off transition and zeroing phases; reconstructing the transmission waveform; mapping the reconstructed transmission waveform to the transmission source region of the three-dimensional electromagnetic calculation model, converting it into an equivalent source current density that varies with time; establishing a three-dimensional time-domain electromagnetic forward model; iteratively solving the electric field using the finite difference time-domain method based on the electric field diffusion equation containing source and polarization convolution terms and the recursive convolution update format; and calculating the magnetic induction intensity based on the electric field to obtain the three-dimensional electromagnetic response under actual transmission waveform conditions. This application achieves accurate simulation of the three-dimensional electromagnetic response of superconducting transient electromagnetic systems under actual transmission waveform conditions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application belongs to the field of transient electromagnetic detection technology, specifically a superconducting electromagnetic three-dimensional modeling method based on the reconstruction of measured transmission waveforms. Background Technology

[0002] Superconducting quantum interference device transient electromagnetic systems (SQUID-TEM) possess high magnetic field sensitivity, a large dynamic range, and the ability to directly measure transient magnetic induction intensity responses, making them widely used in deep resource exploration and subsurface anomaly detection. During actual detection, factors such as parasitic inductance and capacitance in the transmitting circuit, grounding load, and impedance matching networks often lead to distortions after the transmitting current is turned off, including reverse overshoot, tail current, and damped oscillations. These distortions generate additional source-related magnetic fields, which superimpose with the secondary electromagnetic responses generated by the subsurface medium, resulting in abnormal attenuation or polarity reversal. Such responses may be confused with subsurface induced polarization responses, affecting the accuracy of SQUID-TEM data interpretation. Existing three-dimensional electromagnetic modeling methods typically use ideal step waveforms, ramp waveforms, or preset analytical waveforms as excitations, which are insufficient to accurately reflect the subtle distortion characteristics in the measured transmitting current. Directly using noisy measured waveforms easily introduces random noise into numerical calculations; while conventional filtering methods may weaken physically meaningful overshoot, tail current, and oscillation components. Furthermore, the measured distortion waveform exhibits significant variations in its rate of change across different time intervals, making it difficult to balance waveform description accuracy and computational efficiency using a uniform time step. Therefore, a modeling method is needed that can extract and reconstruct distortion features from the measured transmission current and load the reconstructed waveform as a time-varying excitation source into a three-dimensional electromagnetic model. This aims to improve the reliability of SQUID-TEM electromagnetic response simulation and data interpretation under actual transmission conditions. Summary of the Invention

[0003] This application provides a superconducting electromagnetic three-dimensional modeling method based on the reconstruction of measured transmission waveforms, which solves the problem that existing three-dimensional electromagnetic modeling methods are unable to reflect the distortion characteristics such as reverse overshoot, tail current and damped oscillation in actual transmission current.

[0004] This application provides a method for superconducting electromagnetic three-dimensional modeling based on reconstructed measured emission waveforms, including: The measured transmission current of the transmission system is obtained, and the measured transmission current is divided into a steady-state stage, a turn-off transition stage and a zero-return stage according to the time change characteristics of the transmission current after the turn-off. The overshoot tail component and the damped oscillation component are separated from the residual current after the turn-off in the turn-off transition stage and the zero-return stage. A measured waveform representation model is established, consisting of an ideal transmit current, overshoot tail component, damped oscillation component, and measurement noise. Deterministic distortion components are extracted through parameter fitting, and measurement noise is removed to obtain the reconstructed transmit waveform; including: The measured emission current is considered as the superposition of the ideal emission current, overshoot tail component, damped oscillation component, and measurement noise. Using the characteristic parameters of the ideal transmit current, overshoot tail component and damped oscillation component as the fitting parameters, the measured transmit current is fitted with parameters, the measurement noise is separated and removed, and the deterministic distortion characteristics generated by the transient dynamic process of the transmit system are retained, so as to obtain the reconstructed transmit waveform composed of the superposition of the ideal transmit current, overshoot tail component and damped oscillation component. The reconstructed emission waveform is mapped to the emission source region of the three-dimensional electromagnetic calculation model and converted into an equivalent source current density that varies with time. A three-dimensional time-domain electromagnetic forward model is established under actual transmitted waveform excitation and polarized medium to obtain the electric field diffusion equation and recursive convolution update format containing source and polarization convolution terms. The source term in the source-containing part is the equivalent source current density. Based on the electric field diffusion equation containing source and polarization convolution terms and the recursive convolution update scheme, the electric field is solved iteratively using the finite difference time-domain method, and the magnetic induction intensity is calculated based on the electric field to obtain the three-dimensional electromagnetic response under the actual emission waveform conditions.

[0005] Furthermore, the residual current after the measured transmit current is turned off is expressed as the sum of the overshoot tail component and the damped oscillation component; whereby the overshoot tail component is used to characterize the low-frequency reverse overshoot and slow return to zero process after the transmit current is turned off, and the damped oscillation component is used to characterize the high-frequency decaying oscillation process that occurs in a short time after the transmit current is turned off.

[0006] Further, the reconstructed emission waveform is mapped to the emission source region of the three-dimensional electromagnetic calculation model and converted into an equivalent source current density that varies with time, including: The equivalent source current density is calculated based on the reconstructed emission current value, the two grid dimensions perpendicular to the emission source direction, and the unit vector along the emission source direction; the positive and negative signs of the equivalent source current density correspond to the opposite current directions at both ends of the grounding conductor; the equivalent source current density is applied to the grid edge region where the emission source is located.

[0007] Furthermore, a three-dimensional time-domain forward model is established under actual transmitted waveform excitation and containing a polarized medium, including: In the three-dimensional electromagnetic calculation model, the source-containing time-domain Maxwell equations are used to describe the electromagnetic diffusion process in the underground medium. For the computational domain containing polarized media, the frequency-dependent polarization characteristics of the polarized media are described using the Cole-Cole conductivity model, and the frequency-dependent polarization characteristics are converted into time-domain conductivity operators. The polarization convolution term of the time-domain conductivity operator and the electric field intensity is updated step by step using a recursive convolution method, and the historical response of the polarization medium is introduced into the electric field diffusion equation. By substituting the polarization convolution term and the equivalent source current density formed by the reconstructed emission waveform into the electric field diffusion equation, and using the finite difference method for time discretization, the electric field intensity at each time node is obtained.

[0008] Furthermore, the electric field diffusion equation is expressed as: , in, Indicates electric field strength. Represents the permeability of free space. Represents the time-domain conductivity operator. This represents the time-domain polarizability operator for a polarized medium, with the symbol * indicating the time convolution operation. To reconstruct the equivalent source current density formed by the transmitted waveform, the time-domain conductivity operator is used to describe the inductive effect of the medium, the time-domain polarizability operator is used to describe the time-domain polarization response characteristics of the polarized medium, the time convolution operation represents the historical response of the polarized medium, and the equivalent source current density is used as the source term of the source-containing time-domain Maxwell equation.

[0009] Furthermore, it also includes setting a non-uniform time step based on the rate of change of the reconstructed transmission waveform in different time intervals: a first time step is used in the time intervals of the turn-off transition phase and the damped oscillation characteristic set, and a second time step is used in the steady phase and the decay phase. The change of adjacent time steps is controlled by a smoothing function, and the first time step is smaller than the second time step.

[0010] Furthermore, the iterative solution of the electric field using the finite difference time-domain method includes: iteratively solving the electric field strength and the recursive convolution memory variable of the polarization medium; calculating the time rate of change of the magnetic field strength according to Faraday's law of electromagnetic induction; and using the trapezoidal integral method to calculate the magnetic field strength at the next time node based on the time rate of change of the magnetic field strength at the current time node and the next time node, as well as the time step.

[0011] Furthermore, the characteristic parameters of the ideal transmit current, the overshoot tail component, and the damped oscillation component used as fitting parameters include: the fitting parameters of the ideal transmit current include at least one of the transmit current amplitude, the transmit current steady-state duration, the transmit current turn-off start time, and the transmit current turn-off time; the fitting parameters of the overshoot tail component include at least one of the overshoot tail component amplitude, the start time of the distortion component after turn-off, the relaxation mode time constant, and the number of relaxation modes; the fitting parameters of the damped oscillation component include at least one of the number of damped oscillation modes, the amplitude of each oscillation mode, the damping coefficient, the oscillation frequency, and the initial phase.

[0012] Furthermore, the parameter fitting of the measured transmission current includes: using the waveform parameters of the ideal transmission current, the characteristic parameters of the overshoot tail component, and the characteristic parameters of the damped oscillation component as the parameters to be fitted; using the deviation between the measured transmission current and the measured waveform expression model composed of the ideal transmission current, the overshoot tail component, the damped oscillation component, and the measurement noise as the objective function; iteratively adjusting the parameters to be fitted through an optimization algorithm to make the objective function reach the preset convergence condition; separating and removing the measurement noise; retaining the deterministic distortion characteristics generated by the transient dynamic process of the transmission system; and obtaining the reconstructed transmission waveform composed of the ideal transmission current, the overshoot tail component, and the damped oscillation component.

[0013] Compared with the prior art, the advantages of this application are as follows: Based on measured emission current, this application introduces the actual emission waveform into a three-dimensional time-domain electromagnetic forward model containing polarized media through distortion feature separation, physical constraint parameterization reconstruction, and source term transformation. Recursive convolution and non-uniform time steps are used to improve computational accuracy and efficiency, ultimately achieving accurate simulation of the three-dimensional electromagnetic response of superconducting transient electromagnetic under actual emission waveform conditions. This provides a modeling basis for distinguishing the emission current distortion response from the polarization response of the underground medium and improving the reliability of superconducting transient electromagnetic data interpretation.

[0014] This application overcomes the problem that existing modeling methods mainly use ideal or preset transmission waveforms, which are difficult to reflect the actual transmission current distortion characteristics. While suppressing measurement noise, it retains deterministic distortion characteristics such as overshoot tail and damped oscillation, and realizes direct connection between measured transmission waveforms and three-dimensional electromagnetic calculations, providing a modeling basis for identifying abnormal attenuation and polarity reversal response caused by transmission current distortion. Attached Figure Description

[0015] Figure 1 A flowchart illustrating a superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction, provided in this application embodiment; Figure 2 The diagram below shows the separation and reconstruction of the emission current distortion features provided in the embodiments of this application. (a) is the ideal emission current, (b) is the overshoot tail component, (c) is the damped oscillation component, and (d) is the reconstruction result of the fully distorted waveform. Figure 3 The numerical simulation results and semi-analytical solution responses and errors provided in the embodiments of this application are shown in the following figures: (a) shows the comparison curve between the magnetic induction intensity response calculated by the method of this application and the analytical solution response under different turn-off times; (b) shows the relative error curve. Figure 4The following are schematic diagrams of different distortion waveform response analysis provided in the embodiments of this application: (a) is the magnetic field response result diagram under the damped oscillation component; (b) is the magnetic field response result diagram under the overshoot tail component. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0017] This application provides a superconducting electromagnetic three-dimensional modeling method based on the reconstruction of measured transmission waveforms. This method solves the problem that existing three-dimensional electromagnetic modeling methods mainly use ideal or preset transmission waveforms, which are difficult to reflect the distortion characteristics such as reverse overshoot, tail current and damped oscillation in the actual transmission current. It also avoids the computational interference caused by directly loading noisy measured waveforms, and realizes three-dimensional numerical simulation of the SQUID-TEM electromagnetic response under actual transmission conditions.

[0018] See Figure 1 As shown, a superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction includes: S101, obtain the measured transmission current of the transmission system, and divide the measured transmission current into a steady-state stage, a turn-off transition stage and a zero-return stage according to the time change characteristics of the transmission current after the turn-off. Separate the overshoot tail component and the damped oscillation component from the remaining current after the turn-off in the turn-off transition stage and the zero-return stage. S102, establish a measured waveform representation model consisting of ideal transmission current, overshoot tail component, damped oscillation component and measurement noise, extract deterministic distortion component by parameter fitting and remove measurement noise to obtain reconstructed transmission waveform; S103, the reconstructed emission waveform is mapped to the emission source region of the three-dimensional electromagnetic calculation model and converted into an equivalent source current density that varies with time. S104. Establish a three-dimensional time-domain electromagnetic forward model with polarized medium under actual transmitted waveform excitation, and obtain the electric field diffusion equation with source and polarization convolution term and the recursive convolution update format. The source term in the source-containing part is the equivalent source current density. S105. Based on the electric field diffusion equation containing source and polarization convolution terms and the recursive convolution update scheme, the electric field is iteratively solved using the finite difference time-domain method, and the magnetic induction intensity is calculated based on the electric field to obtain the three-dimensional electromagnetic response under the actual emission waveform conditions.

[0019] The transmission system in this application refers to a superconducting quantum interference device transient electromagnetic system (SQUID-TEM system), which simultaneously acquires the transmission current waveform and transient magnetic flux density response. The transmission current is measured using a Hall current sensor with a response time in the microsecond range; the transient magnetic flux density response is measured using a SQUID sensor. In this embodiment, the transmission current amplitude is set to 21A, the transmission current turn-off time is 145μs, the transmission current sampling rate is 96kHz, the SQUID system sampling rate is 47kHz, and the SQUID system dynamic range is 132dB.

[0020] Multiple continuously acquired transmission cycles are sequentially divided into periods, aligned to a time reference, and averaged to reduce the impact of random measurement noise and obtain representative measured transmission current waveforms. Simultaneously, the magnetic flux density response acquired by the SQUID sensor undergoes the same period division, time alignment, and averaging process.

[0021] Based on the measured time-varying characteristics of the transmit current after shutdown, the transmit current is divided into three stages: steady-state, shutdown transition, and zeroing-out. The steady-state stage is characterized by a relatively constant transmit current; the shutdown transition stage is characterized by a rapid decrease in transmit current from its steady-state value with a reverse overshoot; and the zeroing-out stage is characterized by a slow return of the transmit current to zero accompanied by damped oscillations. Overshoot tail components and damped oscillation components are separated from the residual current after shutdown in the shutdown transition and zeroing-out stages. The steady-state stage is characterized by a relatively constant transmit current. During this stage, the transmit current amplitude remains essentially constant with time, the transmit system is in a stable operating state, and the current in the transmit circuit is mainly determined by the transmitter output. The current waveform in the steady-state stage can serve as a reference for the ideal transmit current, used to determine parameters such as the transmit current amplitude, steady-state duration, and shutdown start time. The shutdown transition stage is characterized by a rapid decrease in transmit current from its steady-state value with a reverse overshoot. During this phase, the transmit current rapidly decreases under the turn-off command. Due to factors such as parasitic inductance, parasitic capacitance, grounding load, and impedance matching network in the transmit circuit, the current decrease is not an ideal step, but rather exhibits rapid decay accompanied by reverse overshoot. The turn-off transition phase has a short timescale and a large rate of current change, making it one of the intervals with the most concentrated transmit current distortion characteristics. The current change during this phase mainly reflects the rapid dynamic response characteristics of the transmit system, including the amplitude of the reverse overshoot, the rise and fall rates, and the initial decay characteristics. The zeroing phase is the stage where the transmit current slowly returns to zero accompanied by damped oscillations. During this phase, after the turn-off transition phase, the transmit current enters a slow decay process, with the current amplitude gradually decreasing and eventually returning to zero. Due to the potential energy exchange between parasitic capacitance and parasitic inductance in the transmit circuit, the current during the zeroing phase is often accompanied by high-frequency damped oscillations. This phase has a relatively long timescale and a relatively small rate of current change, but the damped oscillation characteristics are more obvious, making it the main interval for extracting damped oscillation components and overshoot tail components.

[0022] The overshoot tail component and the damped oscillation component are separated from the post-turn-off residual current during the turn-off transition phase and the zeroing phase. The post-turn-off residual current refers to the current remaining in the transmitter circuit after the transmitter current is turned off, including the reverse overshoot current during the turn-off transition phase and the slowly decaying current during the zeroing phase. The post-turn-off residual current is expressed as the sum of the overshoot tail component and the damped oscillation component.

[0023] The overshoot tail component characterizes the low-frequency reverse overshoot and slow return to zero process after the transmit current is turned off. This component mainly reflects the low-frequency relaxation characteristics determined by the parasitic inductance and equivalent damping resistance in the transmit circuit. Its change is relatively slow and its duration is relatively long. It manifests as a reverse overshoot appearing first after the current is turned off, followed by a slow decay over time and a gradual return to zero. The time scale and amplitude of the overshoot tail component mainly depend on factors such as the equivalent inductance and equivalent damping resistance of the transmit circuit, as well as the grounding load and impedance matching network.

[0024] The damped oscillation component characterizes the high-frequency decaying oscillation process that occurs shortly after the transmitter current is turned off. This component mainly reflects the energy exchange between parasitic inductance and parasitic capacitance in the transmitter circuit. It has a high frequency of change and decays rapidly, manifesting as a high-frequency oscillation that appears shortly after turn-off and gradually decays. The frequency, amplitude, and decay rate of the damped oscillation component mainly depend on factors such as the equivalent inductance, equivalent parasitic capacitance, and equivalent damping resistance of the transmitter circuit.

[0025] Because the overshoot tail component and the damped oscillation component have different time scales and variation characteristics, they are separated and parameterized separately. The overshoot tail component is mainly concentrated in the low-frequency, slowly changing range of the zero-return phase, while the damped oscillation component is mainly concentrated in the high-frequency decay range of the turn-off transition phase and the early zero-return phase. By dividing the phases, the main time intervals of the two types of distortion components can be determined. Therefore, the overshoot tail component is parameterized using an exponential decay form, and the damped oscillation component is parameterized using a damped sinusoidal form. Their respective characteristic parameters are then fitted into the measured waveform representation model to retain deterministic distortion characteristics and remove measurement noise.

[0026] The residual current after the transmitting current is turned off is denoted as... The residual current includes an overshoot tail component. and damped oscillation components The sum is expressed as: , The two types of distortion components in the measured waveform have different time scales and variation characteristics, and therefore are characterized by parameterization.

[0027] The overshoot tail component primarily reflects the low-frequency reverse overshoot and slow return to zero process after the transmit current is turned off. This component arises from the low-frequency relaxation characteristics determined by the parasitic inductance, equivalent damping resistance, ground load, and impedance matching network in the transmit circuit. Its time scale is relatively long, typically covering the entire time interval from the turn-off transition phase to the later stages of the return-to-zero phase. In terms of its variation characteristics, the overshoot tail component exhibits an initial reverse overshoot after current turn-off, followed by a slow decay over time, gradually returning to zero. The overall change is gentle, with a small rate of change of current per unit time and no obvious high-frequency oscillations. Since this component is mainly determined by the equivalent inductance and equivalent damping resistance, its decay process can be described by exponential decay. When a single relaxation mode exists, it exhibits single exponential decay; when multiple relaxation modes are introduced by the impedance matching network, it exhibits multi-exponential decay.

[0028] The damped oscillation component primarily reflects the high-frequency decaying oscillation process that occurs shortly after the transmitter current is turned off. This component arises from the energy exchange between parasitic inductance and capacitance in the transmitter circuit. Its timescale is relatively short, mainly concentrated in the turn-off transition phase and the early return-to-zero phase, appearing rapidly and decaying quickly shortly after the transmitter current is turned off. In terms of its variation characteristics, the damped oscillation component exhibits high-frequency oscillation with a high oscillation frequency, a rapid decay rate, and a large rate of change of current per unit time, showing obvious periodicity and alternating polarity. Since this component is mainly determined by the equivalent inductance, equivalent parasitic capacitance, and equivalent damping resistance, its oscillation process can be described using a damped sinusoidal form, including characteristic parameters such as the number of oscillation modes, the amplitude of each oscillation mode, the damping coefficient, the oscillation frequency, and the initial phase.

[0029] The two types of distortion components differ significantly in the following aspects: The time scales are different. The overshoot tail component lasts for a longer period, covering most of the turn-off transition and zeroing phases; the damped oscillation component lasts for a shorter period, mainly concentrated in the short time after turn-off.

[0030] The rates of change are different. The overshoot tail component changes slowly, with a smaller rate of change of current per unit time; the damped oscillation component changes rapidly, with a larger rate of change of current per unit time.

[0031] The frequency characteristics are different. The overshoot tail component is a low-frequency component and does not exhibit obvious periodic oscillations; the damped oscillation component is a high-frequency component and has obvious periodic oscillation characteristics.

[0032] The physical mechanisms are different. The overshoot tail component is mainly generated by the low-frequency relaxation process determined by the equivalent inductance and equivalent damping resistance of the transmitting circuit; the damped oscillation component is mainly generated by the energy exchange between the parasitic inductance and parasitic capacitance of the transmitting circuit.

[0033] The mathematical representations differ. Overshoot tail components are best represented using an exponential decay form; damped oscillation components are best represented using a damped sine form.

[0034] Due to the aforementioned differences, if a unified functional form is used to characterize the two types of distortion components, it is difficult to accurately describe the two distinctly different processes of slow decay at low frequencies and rapid oscillation at high frequencies simultaneously. If only the exponential decay form is used, the periodicity and alternating polarity characteristics of the damped oscillation component cannot be described; if only the damped sine form is used, the slow return to zero process of the overshoot tail component cannot be described.

[0035] Based on the physical mechanisms and temporal variation characteristics of the two types of distortion components, this application employs different parameterization forms to characterize them: the overshoot tail component adopts a single exponential decay form or a multi-exponential decay form, and its fitting parameters include the amplitude of the overshoot tail component, the start time of the distortion component after turn-off, the relaxation mode time constant, and the number of relaxation modes, etc.; the damped oscillation component adopts a damped sine form, and its fitting parameters include the number of damped oscillation modes, the amplitude of each oscillation mode, the damping coefficient, the oscillation frequency, and the initial phase, etc.

[0036] After separate parameterization, the characteristic parameters of the two types of distortion components can be fitted separately in the measured waveform representation model. This allows the parameter fitting process to better adapt to the variation patterns of different components, improving the accuracy of parameter extraction and physical interpretability. Furthermore, separate parameterization facilitates the construction of transmit waveforms containing only overshoot and tail components, transmit waveforms containing only damped oscillation components, and transmit waveforms containing both types of distortion features, as needed in subsequent modeling. This provides a foundation for analyzing the influence of different distortion components on the three-dimensional electromagnetic response.

[0037] Furthermore, since the time scales of the two types of distortion components differ significantly, a smaller time step can be used during the time discretization process in the turn-off transition phase and the time interval where damped oscillation characteristics are concentrated, while a larger time step can be used in the steady phase and the decay phase, so as to balance the waveform discretization accuracy and the computational efficiency of three-dimensional electromagnetic forward modeling.

[0038] A measured waveform representation model consisting of ideal transmit current, overshoot tail component, damped oscillation component, and measurement noise is established. The deterministic distortion component is extracted by parameter fitting and the measurement noise is removed to obtain the reconstructed transmit waveform.

[0039] The measured transmit current includes the ideal transmit current component reflecting the normal operation of the transmit system, as well as the deterministic distortion component caused by factors such as parasitic parameters of the transmit circuit and impedance matching network, and the random noise component introduced by the measurement system. Therefore, the measured transmit current can be expressed as the superposition of the ideal transmit current, overshoot tail component, damped oscillation component, and measurement noise.

[0040] The ideal transmit current is a current that does not contain reverse overshoot, tail current and damped oscillation characteristics after turn-off. It is used to characterize the transmit current waveform of the transmit system under ideal conditions. Its fitting parameters include at least one of the transmit current amplitude, transmit current steady-state duration, transmit current turn-off start time and transmit current turn-off time.

[0041] The overshoot tail component is the low-frequency reverse overshoot and slow return to zero component after the transmit current is turned off. The parameters to be fitted include at least one of the following: the amplitude of the overshoot tail component, the start time of the distortion component after turn-off, the relaxation mode time constant, and the number of relaxation modes. When a multi-exponential decay form is used, the amplitude and relaxation time corresponding to each relaxation mode are also included.

[0042] The damped oscillation component is a high-frequency decaying oscillation component that appears shortly after the emission current is turned off. Its fitting parameters include at least one of the following: the number of damped oscillation modes, the amplitude of each oscillation mode, the damping coefficient, the oscillation frequency, and the initial phase.

[0043] Measurement noise is random noise generated by the measurement system. It does not have a deterministic physical change law and does not belong to the deterministic distortion characteristics generated by the transient dynamic process of the launch system.

[0044] The overshoot tail component and the damped oscillation component are generated by the transient dynamic process after the transmitter system is turned off. They have clear physical mechanisms and deterministic change patterns, and belong to deterministic distortion components. Among them, the overshoot tail component is generated by the low-frequency relaxation process determined by the equivalent inductance and equivalent damping resistance of the transmitter circuit, while the damped oscillation component is generated by the energy exchange between the parasitic inductance and parasitic capacitance of the transmitter circuit.

[0045] Measurement noise is generated by factors such as the measurement system and environmental interference, and does not have a deterministic physical change law, belonging to random components. If the noisy measured waveform is directly loaded into the three-dimensional electromagnetic model, the measurement noise will be introduced into the numerical calculation as a spurious source term, causing computational interference; if conventional filtering methods are used, it may be possible to remove noise while weakening physically meaningful overshoot, tailing, and oscillation components.

[0046] In this embodiment, the transient process after the transmission system is turned off is equivalent to a second-order damped network, and the residual current satisfies: , in, The equivalent inductance is formed by the combined action of the transmitting circuit, grounding load, and connecting lines. This is the equivalent damping resistance during the transient process after turn-off. This is the equivalent parasitic capacitance in the transmission system.

[0047] The eigenvalues ​​of the second-order damped network are: , in, This represents the characteristic root of the equivalent damped network. When no impedance matching branch is provided or the residual current is primarily controlled by a single relaxation mode, the overshoot tail component... Represented as: , in, This represents the amplitude of the overshoot trailing component. This indicates the start time of the distortion component after it is turned off. This represents the time constant corresponding to the relaxation mode. Represents a unit step function, used to limit the distortion component from... From that moment on, Indicates time, This represents the characteristic root of the first equivalent damped network. When an impedance matching circuit is added, causing the residual current after turn-off to contain multiple relaxation modes, the overshoot tail component... Represented as: , in, Indicates the number of relaxation modes. Indicates the first A relaxed mode, , and They represent the first The amplitude, eigenvalues, and relaxation time of each relaxation mode can be used to describe the overshoot tail evolution process formed by multiple relaxation processes under impedance matching conditions.

[0048] Damped oscillation component Represented as: , In the formula, This indicates the number of damped oscillation modes. Indicates the first The amplitude of each oscillation mode, Indicates the damping coefficient. Indicates the oscillation frequency. Indicates the initial phase.

[0049] Among them, a measured waveform representation model composed of ideal emission current, overshoot tail component, damped oscillation component, and measurement noise is established. : , In the formula, This represents the ideal emission current that does not exhibit post-turn-off distortion characteristics. This represents the random noise generated by the measurement system. During waveform reconstruction, the parameters of the overshoot tail component and the damped oscillation component are fitted, and the random noise term is removed to obtain the reconstructed emission current: , In the formula, This represents the reconstructed transmit waveform after preserving the deterministic transmit current distortion characteristics and suppressing random noise. Depending on the requirements, ideal transmit waveforms, transmit waveforms containing only overshoot tail components, transmit waveforms containing only damped oscillation components, and transmit waveforms containing all distortion components can be constructed.

[0050] Therefore, in this embodiment of the application, by establishing a measured waveform representation model, the deterministic distortion component and random noise are separated in the representation model, thereby extracting the deterministic distortion component and removing measurement noise during the parameter fitting process.

[0051] The measured emission current is fitted using the waveform parameters of the ideal emission current, the characteristic parameters of the overshoot tail component, and the characteristic parameters of the damped oscillation component as the parameters to be fitted.

[0052] During the parameter fitting process, the deviation between the measured emission current and the measured waveform expression model composed of the ideal emission current, overshoot tail component, damped oscillation component and measurement noise is used as the objective function. The parameters to be fitted are iteratively adjusted through optimization algorithm so that the objective function reaches the preset convergence condition.

[0053] Specifically, parameter fitting includes the following process: First, the amplitude and steady-state duration of the ideal transmitting current are determined based on the steady-state phase, and the turn-off time of the ideal transmitting current is determined based on the turn-off start time; Second, based on the residual current after turn-off in the turn-off transition phase and the zeroing phase, the amplitude, start time, and time constant of the overshoot tail component, as well as the amplitude, damping coefficient, oscillation frequency, and initial phase of the damped oscillation component are initially determined; Then, using the initially determined parameters as initial values, and the deviation between the measured transmitting current and the measured waveform expression model as the objective function, the waveform parameters of the ideal transmitting current, the characteristic parameters of the overshoot tail component, and the characteristic parameters of the damped oscillation component are iteratively adjusted through an optimization algorithm until the objective function reaches the preset convergence condition.

[0054] Through the above parameter fitting, the characteristic parameters of the ideal emission current, overshoot tail component, and damped oscillation component are gradually determined. However, the measurement noise, due to its lack of deterministic variation law, cannot be fitted by the parameterized forms of the ideal emission current, overshoot tail component, and damped oscillation component, and is thus separated during the parameter fitting process.

[0055] After parameter fitting, the measurement noise term in the measured waveform representation model is removed, and the reconstructed transmission waveform, composed of the ideal transmission current, overshoot tail component, and damped oscillation component, is retained. The reconstructed transmission waveform retains the deterministic distortion characteristics generated by the transient dynamic process of the transmission system, including overshoot tail and damped oscillation characteristics, while suppressing the random noise in the original measured waveform.

[0056] Specifically, the reconstructed transmit waveform is composed of the superposition of the ideal transmit current, the overshoot tail component, and the damped oscillation component. The ideal transmit current is determined based on the fitted waveform parameters; the overshoot tail component is determined based on the fitted amplitude, start time, time constant, and number of relaxation modes; and the damped oscillation component is determined based on the fitted number of oscillation modes, amplitude, damping coefficient, oscillation frequency, and initial phase.

[0057] Depending on the modeling requirements, different transmission waveforms can be constructed based on the parameter fitting results, including: an ideal transmission waveform containing only the ideal transmission current; a transmission waveform containing only the overshoot tail component; a transmission waveform containing only the damped oscillation component; and a transmission waveform that simultaneously contains all distortion characteristics of both the overshoot tail component and the damped oscillation component.

[0058] By constructing the different transmission waveforms mentioned above, the computational domain, underground medium parameters, boundary conditions, mesh generation, and receiver position can be kept unchanged in the three-dimensional electromagnetic calculation. Only the source current density time function of the transmission source region can be changed, thereby comparing and analyzing the influence of different transmission current distortion components on the three-dimensional electromagnetic response.

[0059] By establishing a measured waveform representation model composed of ideal emission current, overshoot tail component, damped oscillation component, and measurement noise, and extracting deterministic distortion components and removing measurement noise through parameter fitting, it is possible to suppress random measurement noise while retaining physically meaningful deterministic distortion features such as overshoot tail and damped oscillation. This avoids computational interference caused by directly loading noisy measured waveforms and also avoids weakening of the true distortion components by conventional filtering methods. Thus, it provides an accurate, reliable, and physically meaningful time-varying excitation source for three-dimensional electromagnetic calculations.

[0060] The reconstructed transmission waveform is a continuously varying current waveform over time. In three-dimensional time-domain electromagnetic calculations, the reconstructed transmission waveform needs to be discretized according to a predetermined time step to obtain the reconstructed transmission current value at each time node. The time step can be set to a non-uniform time step based on the rate of change of the reconstructed transmission waveform in different time intervals. A smaller time step is used in the time intervals of the turn-off transition phase and the damped oscillation characteristic concentration, while a larger time step is used in the steady phase and the decay phase. Through time discretization, the reconstructed transmission waveform is converted into a series of reconstructed transmission current values ​​arranged according to time nodes.

[0061] In the three-dimensional electromagnetic computational model, the source region is determined based on the spatial location, orientation, and geometry of the transmitting loop or grounding conductor. The source region is mapped to the edge region of the three-dimensional mesh, i.e., the edge region of the mesh where the source is located. The spatial location and geometry of the source remain unchanged during the modeling process to allow for comparison of the three-dimensional electromagnetic responses corresponding to different transmitted waveforms under the same computational domain, underground medium parameters, boundary conditions, mesh generation, and receiver location.

[0062] For each time node, the equivalent source current density is calculated based on the reconstructed emission current value at that time node, the two grid dimensions perpendicular to the emission source direction, and the unit vector along the emission source direction.

[0063] Mapping the reconstructed emission current to the emission source region in the three-dimensional computational model, the corresponding equivalent source current density is expressed as: , In the formula, Indicates spatial location At the The equivalent source current density at each time point This indicates the grid edge region where the emission source is located. and This represents the dimensions of two grids perpendicular to the direction of the emission source. This represents a unit vector along the direction of the emission source, with positive and negative signs corresponding to opposite current directions at the two ends of the grounding conductor. Indicates the first There are discrete time points. The above formula keeps the spatial position and geometry of the emission source unchanged, and loads all the distortion changes in the measured emission current into the time function of the source term.

[0064] Specifically, the reconstructed emission current is a line current flowing along the emission source direction. In three-dimensional finite-difference time-domain calculations, the line current needs to be converted into an equivalent source current density applied to the grid edge region. During this conversion, the reconstructed emission current value is divided by two grid dimensions perpendicular to the emission source direction to obtain the equivalent source current density value on that grid edge region. The two grid dimensions are the grid dimensions in two spatial directions perpendicular to the emission source direction, and their product represents the equivalent cross-sectional area of ​​the grid edge region where the emission source is located. The unit vector along the emission source direction is used to determine the direction of the equivalent source current density.

[0065] The signs of the equivalent source current density correspond to opposite current directions at both ends of the grounding conductor. For a grounded conductor-based transmitter, current flows into the ground from one grounding electrode, diffuses through the underground medium, and flows out from the other grounding electrode; therefore, the current directions at both ends of the grounding conductor are opposite. When mapping the reconstructed transmitter current to the edge region of the three-dimensional mesh, the signs of the equivalent source current density are determined based on the respective current directions at both ends of the grounding conductor, thus ensuring that the current distribution within the transmitter region is consistent with the current direction in the actual transmitter loop.

[0066] The calculated equivalent source current density is applied to the edge region of the grid where the transmitter is located. During the application process, the spatial position and geometry of the transmitter remain unchanged, and only the time function of the source current density is changed. That is, the position, orientation, and geometry of the transmitter in the three-dimensional grid are determined by the actual layout of the transmission loop or grounding conductor and do not change with the transmitted waveform; while the magnitude and sign of the source current density change with time nodes and are determined by the reconstructed transmitted waveform.

[0067] The equivalent source current density applied to the edge region of the grid containing the transmitter is used as a source term in the source-domain Maxwell's equations in the time domain for three-dimensional time-domain electromagnetic calculations. The equivalent source current density includes deterministic distortion features such as overshoot tails and damped oscillations determined by the reconstructed transmitter waveform. Through this mapping, all distortion changes in the actual transmitter waveform are loaded into the time function of the source term, while the spatial position and geometry of the transmitter remain unchanged. Thus, the time-iterative solution of the electric field diffusion equation can reflect the non-ideal turn-off process of the actual transmitter waveform in the source term, thereby obtaining the three-dimensional electromagnetic response under the actual transmitter waveform conditions.

[0068] By mapping the reconstructed transmitted waveform to the source region of a three-dimensional electromagnetic calculation model and converting it into an equivalent source current density that varies with time, this application achieves a direct connection between the measured reconstructed waveform and three-dimensional time-domain electromagnetic calculation. The spatial position and geometry of the source remain unchanged; only the time function of the source current density changes with the reconstructed transmitted waveform. This facilitates comparison of the three-dimensional electromagnetic responses corresponding to the ideal waveform, the waveform containing only overshoot tails, the waveform containing only damped oscillations, and the waveform containing all distortion features under the same calculation conditions. This provides a basis for analyzing the influence of different transmitted current distortion components on the three-dimensional electromagnetic response.

[0069] A three-dimensional time-domain electromagnetic forward model is established under actual transmitted waveform excitation and polarized medium, yielding the electric field diffusion equation containing source and polarization convolution terms, and the recursive convolution update scheme, including: In a three-dimensional electromagnetic computational model, the source-containing time-domain Maxwell's equations are used to describe the electromagnetic diffusion process in the subsurface medium. Based on the constructed three-dimensional electromagnetic computational model, the source-containing time-domain Maxwell's equations are used as the fundamental governing equations describing the electromagnetic diffusion process in the subsurface medium. The source-containing time-domain Maxwell's equations include the equivalent source current density formed by the reconstructed transmitted waveform as a source term. The equivalent source current density has been mapped to the transmission source region of the three-dimensional electromagnetic computational model according to the reconstructed transmitted waveform and loaded onto the grid edge region where the transmission source is located. Therefore, the source terms in the source-containing time-domain Maxwell's equations no longer use ideal step waveforms, ramp waveforms, or preset analytical waveforms, but instead use time-varying source terms corresponding to the actual transmitted waveforms that include deterministic distortion characteristics such as overshoot tails and damped oscillations.

[0070] In transient electromagnetic methods, a quasi-static condition is typically satisfied, meaning that the displacement current is negligible compared to the conduction current. Based on the quasi-static approximation, Maxwell's equations in the time domain with sources are rearranged to eliminate magnetic induction, resulting in an electric field diffusion equation with electric field intensity as the solution. The electric field diffusion equation describes the diffusion process of the electric field in the subsurface medium, including the equivalent source current density formed by the reconstructed emission waveform as an excitation source term.

[0071] For computational domains containing polarized media, the Cole-Cole conductivity model is used to describe the frequency-dependent polarization characteristics of the polarized media, converting these characteristics into a time-domain conductivity operator. For computational domains containing polarized media, the conductivity of the subsurface medium is no longer a constant dependent only on spatial location, but rather a complex conductivity that varies with frequency. The Cole-Cole conductivity model describes the frequency-dependent polarization characteristics of the polarized media. This model includes parameters such as zero-frequency conductivity, infinite-frequency conductivity, polarizability, time constant, and frequency correlation coefficient, enabling it to describe the conductivity variations of the medium at different frequencies and the excitation of polarization effects.

[0072] Since three-dimensional time-domain electromagnetic forward modeling requires iterative solutions in the time domain, the Cole-Cole conductivity model is transformed from the frequency domain to the time domain, resulting in a time-domain conductivity operator. This time-domain conductivity operator describes the conductivity response characteristics of polarized media in the time domain. Furthermore, based on the Cole-Cole conductivity model, a time-domain polarizability operator can also be obtained to describe the time-domain polarization response characteristics of polarized media. Through this transformation, the description of polarized media in the frequency domain is converted into an operator form that can be used in time-domain finite-difference calculations.

[0073] In the frequency domain, the complex conductivity of a polarized medium is described using the Cole-Cole model: , Where: σ(ω) represents the complex conductivity in the frequency domain; ω represents the angular frequency; Represents the imaginary unit; This represents the zero-frequency conductivity, that is, the conductivity when the frequency approaches zero. This represents the conductivity at infinite frequency, that is, the conductivity when the frequency approaches infinity. Indicates the polarization time constant; This represents the frequency correlation coefficient, and its value range is typically [value range missing]. .

[0074] Among them, polarizability It can be represented as: ; Therefore, the Cole-Cole conductivity model can also be written as: ; In the time domain, the electric field strength With current density They satisfy a convolutional relationship: , in, This represents the time-domain conductivity operator, and the symbol * represents the time convolution operation.

[0075] Time-domain conductivity operator It can be obtained from the complex conductivity σ(ω) in the frequency domain through the inverse Fourier transform: , According to the Cole-Cole conductivity model, the time-domain polarizability operator of a polarized medium can be expressed as: , in, The time-domain polarizability operator represents the polarization of a polarized medium. This represents the relaxation function of the polarized medium.

[0076] The polarization convolution term of the time-domain conductivity operator and the electric field intensity is updated step-by-step using a recursive convolution method, incorporating the historical response of the polarized medium into the electric field diffusion equation. In the time domain, the relationship between the time-domain conductivity operator and the electric field intensity is no longer a simple product but a time convolution. This convolution term constitutes the polarization convolution term, used to characterize the historical response of the polarized medium. That is, the electric field intensity at the current time node depends not only on the medium parameters and source term at the current time node but also on the historical response accumulated in the polarized medium by the electric field intensity at previous time nodes. Directly calculating the convolution term of the time-domain conductivity operator and the electric field intensity would require storing the electric field intensity of all historical time nodes, resulting in significant computational and storage requirements. Therefore, a recursive convolution method is used to calculate the polarization convolution term. This method utilizes the polarization memory variable from the previous time node and the electric field intensity at the current time node to recursively derive the polarization convolution term for the current time node. The polarization memory variable is used to store the historical response information of the polarization medium and is updated at each time step.

[0077] Specifically, the initial electric field, initial magnetic induction, and initial terms of the recursive convolution are all set to zero. During the time-step iteration, the polarization convolution term for the current time step is calculated using a recursive convolution update format based on the electric field intensity of the previous time step, the polarization memory variable of the previous time step, and the electric field intensity of the current time step. Subsequently, the polarization convolution term for the current time step is stored as a polarization memory variable for recursive calculation in the next time step. Through the above recursive convolution method, the historical response of the polarized medium is introduced into the electric field diffusion equation step by step.

[0078] The polarization convolution term and the equivalent source current density formed by the reconstructed transmitted waveform are substituted into the electric field diffusion equation. A finite difference method is used for time discretization to obtain the electric field intensity at each time node. The polarization convolution term obtained from recursive convolution calculation and the equivalent source current density formed by the reconstructed transmitted waveform are then substituted into the electric field diffusion equation. At this point, the electric field diffusion equation simultaneously contains two key terms: first, the equivalent source current density formed by the reconstructed transmitted waveform, used to reflect deterministic distortion characteristics such as overshoot tailing and damped oscillations in the actual transmitted waveform; second, the time-domain polarization convolution term obtained from the Cole-Cole conductivity model, used to reflect the historical response of the polarized medium. Therefore, the electric field diffusion equation simultaneously possesses the ability to describe both the actual transmitted waveform excitation and the historical response of the polarized medium. The finite difference method is used to perform time discretization on the electric field diffusion equation containing the source and polarization convolution terms. During the time discretization process, the time derivative term and spatial derivative term in the electric field diffusion equation are approximated using a finite difference scheme, the polarization convolution term is calculated using a recursive convolution update scheme, and the source term uses the equivalent source current density at the current time node. Through iterative time-step calculations, starting from the initial conditions, the electric field intensity at each time node is solved sequentially.

[0079] During the time discretization process, the time step can be set to a non-uniform time step according to the rate of change of the reconstructed transmission waveform in different time intervals. A smaller time step is used in the turn-off transition phase and the time intervals where damped oscillation characteristics are concentrated, to accurately describe the rapid changes in the transmission current and the damped oscillation characteristics; a larger time step is used in the steady phase and the decay phase to reduce the computational burden of three-dimensional time-domain electromagnetic forward modeling. A smoothing function controls the change of adjacent time steps to avoid numerical oscillations or iterative instability caused by abrupt changes in time step size.

[0080] The non-uniform time step is set according to the rate of change of the reconstructed transmit current over time: , In the formula, Indicates the first Each time step This indicates the preset maximum time step. A proportionality coefficient representing the sensitivity of the control time step to waveform changes. This prevents positive numbers with a denominator of zero. Indicates the first The rate of change of the emitter current with respect to time is reconstructed at each time point. The time step is reduced when the emitter current changes rapidly during the turn-off transition phase; the time step is increased when the emitter current is in the steady-state phase or the slow decay phase, in order to balance the discretization accuracy of the turn-off distortion waveform and the efficiency of 3D modeling.

[0081] Through the above process, a three-dimensional time-domain forward model of the polarized medium under actual transmitted waveform excitation is finally obtained. The core of this model is the electric field diffusion equation containing source and polarization convolution terms, and the recursive convolution update scheme. The electric field intensity at each time point is obtained by solving under actual transmitted waveform excitation and considering the historical response of the underground polarized medium. This provides a foundation for calculating the time-varying rate of magnetic flux density according to Faraday's law of electromagnetic induction and obtaining the magnetic flux density at each time point using the trapezoidal integral method. Therefore, the three-dimensional electromagnetic response of the polarized medium under actual transmitted waveform conditions can be obtained.

[0082] The above electric field diffusion equation is expressed as: , in, Indicates electric field strength. Represents the permeability of free space. Represents the time-domain conductivity operator. This represents the time-domain polarizability operator for a polarized medium, with the symbol * indicating the time convolution operation. To reconstruct the equivalent source current density formed by the transmitted waveform, the time-domain conductivity operator is used to describe the inductive effect of the medium, the time-domain polarizability operator is used to describe the time-domain polarization response characteristics of the polarized medium, the time convolution operation represents the historical response of the polarized medium, and the equivalent source current density is used as the source term of the source-containing time-domain Maxwell equation.

[0083] The curl, representing the electric field intensity, reflects the diffusion, spatial variation, and boundary constraints of the electric field in three-dimensional space. This term is derived from the spatial derivative term in Maxwell's equations with sources in the time domain and is used to describe the spatial distribution characteristics of the electric field in the subsurface medium. In finite-difference time-domain calculations, this term is discretized using a spatial finite-difference scheme.

[0084] This represents the medium-sensing effect term. Among them, This represents the time-domain conductivity operator, used to describe the inductive effect of a medium, i.e., the contribution of conduction current to the diffusion of electromagnetic fields. This represents the rate of change of the electric field strength over time. This term reflects the diffusion and attenuation process of the electromagnetic field in the underground medium and is a fundamental conduction term in the electric field diffusion equation.

[0085] This is the history response term for the polarized medium, where γ(t) represents the time-domain polarizability operator of the polarized medium, used to describe the time-domain polarization response characteristics of the polarized medium; the symbol * represents the time convolution operation, indicating the history response of the polarized medium. This term shows that the electric field at the current time node depends not only on the medium parameters and source terms at the current time node, but also on the history response accumulated in the polarized medium by the electric field at previous time nodes.

[0086] After transforming the Cole-Cole conductivity model from the frequency domain to the time domain, we obtain the time-domain conductivity operator and the time-domain polarizability operator. The convolution term of the time-domain conductivity operator with the electric field intensity constitutes the polarization convolution term. For ease of expression, the polarization-related term can be denoted as... ,but This represents the polarization convolution term formed by the combined action of the time-domain conductivity operator and the time-domain polarizability operator. The polarization convolution term is used to incorporate the historical response of the polarized medium into the electric field diffusion equation.

[0087] This is the source term formed by reconstructing the transmitted waveform. Among them, To reconstruct the equivalent source current density formed by the transmitted waveform, it is used as a source term in the source-containing time-domain Maxwell equations; This represents the time rate of change of the equivalent source current density. The equivalent source current density has been mapped to the source region of the three-dimensional electromagnetic calculation model based on the reconstructed transmit waveform and loaded onto the grid edge region where the source is located. Since the reconstructed transmit waveform contains deterministic distortion features such as overshoot tails and damped oscillations, this term can introduce the non-ideal turn-off process in the actual transmit waveform into the three-dimensional time-domain electromagnetic forward model.

[0088] If calculated directly This requires storing the electric field time derivatives for all historical time points, resulting in significant computational and storage requirements. Therefore, a recursive convolution method is employed to update the polarization convolution term step-by-step.

[0089] Specifically, the initial electric field, initial magnetic induction, and initial terms of the recursive convolution are all set to zero. At the nth time step, the polarization convolution term for the current time step is recursively derived based on the polarization memory variable from the previous time step and the electric field time derivative at the current time step. Subsequently, the polarization convolution term for the current time step is saved as a polarization memory variable for recursive calculation in the next time step. Through this recursive convolution update format, the historical response of the polarized medium is introduced into the electric field diffusion equation step by step, without needing to store all historical electric field data.

[0090] The electric field diffusion equation containing source and polarization convolution terms is discretized in time using the finite difference method. During the time discretization process, the curl term on the left side of the equation is discretized using a spatial finite difference scheme, the electric field time derivative term on the right side of the equation is discretized using a time finite difference scheme, the polarization convolution term is calculated using a recursive convolution update scheme, and the source term uses the equivalent source current density and its time rate of change at the current time node.

[0091] When the polarization convolution term is computed using a recursive convolution update format, its recursive convolution update format can be expressed as: , The polarization convolution term is obtained by weighted summation of the polarization medium recursively convolved memory variables of each relaxation mode: , in, Indicates the first The relaxation mode is in the first The polarization medium recursive convolution memory variable at each time point Indicates the first The relaxation mode is in the first A recursive convolution memory variable of the polarization medium at a single time point. Indicates the first Each time step Indicates the first The amplitude coefficients of each relaxation mode, Indicates the first The attenuation coefficient of a relaxation mode, Indicates the first The electric field intensity at each time point.

[0092] This yields the iterative equations for the electric field intensity at each time node. Starting from the initial conditions, the electric field intensity at each time node is solved step-by-step, and the recursive convolution memory variables of the polarization medium are updated synchronously. During the time discretization process, the time step size can be set to a non-uniform time step size according to the rate of change of the reconstructed emission waveform in different time intervals: a smaller time step size is used in the time intervals of the turn-off transition phase and the damped oscillation characteristic set, and a larger time step size is used in the steady phase and the decay phase. The change of adjacent time steps is controlled by a smoothing function to balance the waveform discretization accuracy and the computational efficiency of three-dimensional time-domain electromagnetic forward modeling.

[0093] The iterative solution of the electric field using the finite difference time-domain method includes: iteratively solving the electric field strength and the recursive convolution memory variable of the polarization medium; calculating the time rate of change of the magnetic field strength according to Faraday's law of electromagnetic induction; and using the trapezoidal integral method to calculate the magnetic field strength at the next time node based on the time rate of change of the magnetic field strength at the current time node and the next time node, as well as the time step.

[0094] When using the finite difference time-domain method to iteratively solve the electric field diffusion equation containing source and polarization convolution terms, the objects of the iterative solution include two types of variables: First, the electric field intensity at each time point. The electric field intensity is the direct solution object of the electric field diffusion equation, which is solved step by step at each time point.

[0095] Secondly, the polarization medium recursive convolution memory variable. This polarization medium recursive convolution memory variable is used to store historical response information of the polarization medium and is updated at each time step based on the electric field strength at the current time node, so as to calculate the polarization convolution term in the next time step.

[0096] Therefore, the electric field strength and the recursive convolution memory variable of the polarization medium are updated synchronously and coupled with each other during the time iteration. Solving for the electric field strength requires the polarization memory variable from the previous time step, while updating the recursive convolution memory variable of the polarization medium requires the electric field strength obtained in the current time step.

[0097] The initial electric field, initial magnetic field strength, and initial terms of the recursive convolution are all set to zero. Starting from the initial time step, the process proceeds sequentially according to the time step.

[0098] In the Given the electric field intensity at the previous time step, the recursive convolution memory variable of the polarization medium at the previous time step, the equivalent source current density at the current time step, and the current time step size, solve for the electric field intensity at the current time step using the finite difference discretization scheme of the electric field diffusion equation containing source and polarization convolution terms.

[0099] After obtaining the electric field strength at the current time node, the polarization convolution term for the current time node is calculated using a recursive convolution update format based on the electric field strength at the current time node and the polarization medium recursive convolution memory variable from the previous time node. This polarization convolution term is then stored as the polarization medium recursive convolution memory variable for the current time node. The polarization medium recursive convolution memory variable is used for the recursive convolution calculation in the next time step.

[0100] Through the above process, the electric field strength and the recursive convolution memory variables of the polarization medium are updated iteratively step by step until the preset calculation termination time is reached.

[0101] After obtaining the electric field strength at each time point, the rate of change of the magnetic field strength over time is calculated according to Faraday's law of electromagnetic induction. Faraday's law of electromagnetic induction states that the rate of change of the magnetic field strength over time is equal to the negative of the curl of the electric field strength. Therefore, for each time point, the curl of the electric field strength is calculated based on the electric field strength at that time point, and its negative value is taken to obtain the rate of change of the magnetic field strength over time at that time point.

[0102] The time-varying rate of magnetic flux density is used to calculate the magnetic flux density at each time point using the trapezoidal integral method. Since the electric field strength has already been obtained under the actual transmitted waveform excitation and after considering the historical response of the polarization medium, the time-varying rate of magnetic flux density calculated from the electric field strength also includes the contributions of the actual transmitted waveform distortion and the polarization medium response.

[0103] After obtaining the rate of change of magnetic flux density over time at the current and next time points, the magnetic flux density at the next time point is calculated using the trapezoidal integral method.

[0104] The trapezoidal integral method approximates the change in magnetic field strength between the current and next time points as the product of the average rate of change of magnetic field strength over time for both points and the time step. Therefore, the magnetic field strength at the next time point is equal to the magnetic field strength at the current time point plus the product of the average rate of change of magnetic field strength over time for both points and the time step.

[0105] Specifically, when calculating the magnetic flux density at the next time point, it is necessary to first calculate the rate of change of the magnetic flux density over time based on the electric field intensity at the next time point. Then, by combining the magnetic flux density at the current time point, the rate of change of the magnetic flux density over time at the current time point, the rate of change of the magnetic flux density over time at the next time point, and the current time step, the trapezoidal integral formula is used to calculate the magnetic flux density at the next time point.

[0106] The rate of change of magnetic flux density over time is determined according to Faraday's law of electromagnetic induction, and the magnetic flux density at the next time node is calculated using the trapezoidal integral: , In the formula, and These represent the magnetic flux density at two adjacent time points. The magnetic flux density response is obtained through trapezoidal integration and can be compared with transient magnetic flux density data directly acquired by the SQUID sensor.

[0107] Since the time step can be set to a non-uniform time step according to the rate of change of the reconstructed transmission waveform in different time intervals, the time step used in the trapezoidal integral calculation is the non-uniform time step corresponding to the current time node. In the turn-off transition phase and the time interval where damped oscillation characteristics are concentrated, the time step is smaller, and the trapezoidal integral can describe the rapid changes in magnetic induction intensity more precisely; in the steady phase and decay phase, the time step is larger, which can reduce the number of integration steps and reduce the computational load.

[0108] The electric field strength, magnetic induction intensity, and recursive convolution memory variables of the polarization medium at the current time node are updated to the corresponding values ​​at the next time node. The electric field strength iterative solution, polarization memory variable update, magnetic induction intensity time change rate calculation, and trapezoidal integral calculation are repeated until the preset calculation termination time is reached.

[0109] Through the above time-step iterations, the magnetic flux density at each time node is finally obtained. The magnetic flux density at each time node is the three-dimensional electromagnetic response of the polarized medium under the actual transmission waveform conditions. Since the equivalent source current density formed by the reconstructed transmission waveform and the polarization convolution term obtained by the Cole-Cole conductivity model are considered simultaneously during the iteration process, the obtained magnetic flux density response includes both the influence of deterministic distortion characteristics such as overshoot tail and damped oscillations in the actual transmission waveform and the influence of the historical response of the underground polarized medium.

[0110] See Figure 2 As shown, this application provides a schematic diagram of the separation and reconstruction of transmit current distortion features, wherein... Figure 2 In the figure, (a) represents the ideal emission current. Figure 2 (b) in the figure represents the overshoot tail component. Figure 2 In the equation (c), the damped oscillation component and... Figure 2 (d) in the figure represents the reconstructed waveform of the fully distorted waveform. Each figure shows the original measured transmit current waveform, the corresponding reconstructed waveform, and the zero baseline. A magnified view is also provided in the distorted region after turn-off to show the subtle changes in the overshoot tail component and the damped oscillation component. Wherein: Figure 2 In the diagram, (a) represents a waveform that retains only the ideal emission current and removes all distortion features after turn-off; Figure 2 (b) in the figure represents the waveform that retains only the overshoot tail component after the turn-off, which is characterized by low-frequency reverse overshoot and a slow return to zero process; Figure 2 (c) in the figure represents the waveform that retains only the damped oscillation component after the turn-off, which is a high-frequency decaying oscillation process; Figure 2 (d) in the figure represents a waveform that simultaneously contains overshoot tail components and damped oscillation components, which matches well with the original measured emission current waveform.

[0111] Figure 3 The numerical simulation results and semi-analytical solution responses and errors for different turn-off times are compared. This is used to verify the computational accuracy of the three-dimensional time-domain electromagnetic forward model under different emission current turn-off times. Figure 3 Figure (a) shows a comparison curve between the magnetic flux density response calculated using the method of this application and the analytical solution response under different turn-off times. Figure 3 (b) in the figure represents the relative error curve. It can be concluded that... Figure 3Figure (a) presents the analytical solution, the numerical solution calculated using the method of this application under turn-off times of 10 μs, 50 μs, and 100 μs, and the convolutional analytical solution under the same turn-off times. Under different turn-off times, the numerical solution curve, the convolutional analytical solution curve, and the analytical solution curve show the same overall trend, all gradually decreasing with time, and the decreasing pattern matches that of the analytical solution.

[0112] Figure 3 (b) in the figure gives the relative error between the numerical solution and the analytical solution of the convolution under the conditions of 10μs, 50μs and 100μs respectively.

[0113] Depend on Figure 3 As shown in (b), the relative errors under the three shutdown time conditions are at a low level for most of the time intervals, and the relative errors are below or close to the 3% reference line for most time periods. This indicates that the method of this application has high computational accuracy under different shutdown time conditions.

[0114] Figure 4 This is a schematic diagram illustrating the response analysis of different distortion waveforms. Figure 4 (a) in the figure shows the magnetic field response results under the damped oscillation component; Figure 4 (b) in the figure shows the magnetic field response results under the overshoot tail component, from Figure 4 As can be seen in (a), when the waveform containing only the damped oscillation component is used as the excitation source, the overall magnetic induction intensity response still decays with time, but there is a significant difference in the early time period compared with the response corresponding to the ideal waveform. This difference manifests as high-frequency disturbances in the early magnetic field response, indicating that the damped oscillation component mainly affects the early magnetic field response. As time goes by, the damped oscillation component decays rapidly, and the magnetic field responses corresponding to the two waveforms gradually become consistent. Therefore, the influence of the damped oscillation component is mainly concentrated in the short time after the turn-off, and has little impact on the middle and late magnetic field response.

[0115] Figure 4 As shown in (b), when the waveform containing only the overshoot tail component is used as the excitation source: the magnetic induction intensity response differs from the response corresponding to the ideal waveform in the early stage; as time progresses, the magnetic field response corresponding to the overshoot tail component becomes significantly separated from the response corresponding to the ideal waveform; the magnetic field response corresponding to the overshoot tail component shows a change in the middle and late stages that is significantly different from the monotonic decay trend of the ideal waveform, and even shows local minima and rebound phenomena; this indicates that the overshoot tail component forms an additional source-related magnetic field with a long duration and changes its superposition relationship with the underground secondary magnetic field; this additional source-related magnetic field may cause abnormal decay or even polarity reversal, thereby affecting the accuracy of superconducting transient electromagnetic data interpretation.

[0116] This indicates that the damped oscillation component mainly causes high-frequency disturbances in the early magnetic field response; the overshoot tail component forms an additional source-related magnetic field with a longer duration and changes its superposition relationship with the underground secondary magnetic field, thus providing a basis for identifying abnormal attenuation and polarity reversal caused by emission current distortion, and improving the reliability of superconducting transient electromagnetic data interpretation.

[0117] Figures 2-4 Together, they demonstrate that the method of this application can realize three-dimensional time-domain electromagnetic forward modeling under actual emission waveform excitation and polarized media, and provide a reliable basis for the interpretation of superconducting transient electromagnetic data.

[0118] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for superconducting electromagnetic three-dimensional modeling based on reconstructed measured emission waveforms, characterized in that, include: The measured transmission current of the transmission system is obtained, and the measured transmission current is divided into a steady-state stage, a turn-off transition stage and a zero-return stage according to the time change characteristics of the transmission current after the turn-off. The overshoot tail component and the damped oscillation component are separated from the residual current after the turn-off in the turn-off transition stage and the zero-return stage. A measured waveform representation model is established, consisting of an ideal transmit current, overshoot tail component, damped oscillation component, and measurement noise. Deterministic distortion components are extracted through parameter fitting, and measurement noise is removed to obtain the reconstructed transmit waveform; including: The measured emission current is considered as the superposition of the ideal emission current, overshoot tail component, damped oscillation component, and measurement noise. Using the characteristic parameters of the ideal transmit current, overshoot tail component and damped oscillation component as the fitting parameters, the measured transmit current is fitted with parameters, the measurement noise is separated and removed, and the deterministic distortion characteristics generated by the transient dynamic process of the transmit system are retained, so as to obtain the reconstructed transmit waveform composed of the superposition of the ideal transmit current, overshoot tail component and damped oscillation component. The reconstructed emission waveform is mapped to the emission source region of the three-dimensional electromagnetic calculation model and converted into an equivalent source current density that varies with time. A three-dimensional time-domain electromagnetic forward model is established under actual transmitted waveform excitation and polarized medium to obtain the electric field diffusion equation and recursive convolution update format containing source and polarization convolution terms. The source term in the source-containing part is the equivalent source current density. Based on the electric field diffusion equation containing source and polarization convolution terms and the recursive convolution update scheme, the electric field is solved iteratively using the finite difference time-domain method, and the magnetic induction intensity is calculated based on the electric field to obtain the three-dimensional electromagnetic response under the actual emission waveform conditions.

2. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 1, characterized in that, The residual current after the measured transmit current is turned off is expressed as the sum of the overshoot tail component and the damped oscillation component; where the overshoot tail component is used to characterize the low-frequency reverse overshoot and slow return to zero process after the transmit current is turned off, and the damped oscillation component is used to characterize the high-frequency decay oscillation process that occurs in a short time after the transmit current is turned off.

3. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 1, characterized in that, Mapping the reconstructed emission waveform to the emission source region of the three-dimensional electromagnetic calculation model, and converting it into an equivalent source current density that varies with time, includes: The equivalent source current density is calculated based on the reconstructed emission current value, the two grid dimensions perpendicular to the emission source direction, and the unit vector along the emission source direction; the positive and negative signs of the equivalent source current density correspond to the opposite current directions at both ends of the grounding conductor; the equivalent source current density is applied to the grid edge region where the emission source is located.

4. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 1, characterized in that, A three-dimensional time-domain forward model is established under actual transmitted waveform excitation and containing a polarized medium, including: In the three-dimensional electromagnetic calculation model, the source-containing time-domain Maxwell equations are used to describe the electromagnetic diffusion process in the underground medium. For the computational domain containing polarized media, the frequency-dependent polarization characteristics of the polarized media are described using the Cole-Cole conductivity model, and the frequency-dependent polarization characteristics are converted into time-domain conductivity operators. The polarization convolution term of the time-domain conductivity operator and the electric field intensity is updated step by step using a recursive convolution method, and the historical response of the polarization medium is introduced into the electric field diffusion equation. By substituting the polarization convolution term and the equivalent source current density formed by the reconstructed emission waveform into the electric field diffusion equation, and using the finite difference method for time discretization, the electric field intensity at each time node is obtained.

5. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 4, characterized in that the electric field The diffusion equation is expressed as: , in, Indicates electric field strength. Represents the permeability of free space. Represents the time-domain conductivity operator. This represents the time-domain polarizability operator for a polarized medium, with the symbol * indicating the time convolution operation. To reconstruct the equivalent source current density formed by the transmitted waveform, the time-domain conductivity operator is used to describe the inductive effect of the medium, the time-domain polarizability operator is used to describe the time-domain polarization response characteristics of the polarized medium, the time convolution operation represents the historical response of the polarized medium, and the equivalent source current density is used as the source term of the source-containing time-domain Maxwell equation.

6. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 1, characterized in that, It also includes setting a non-uniform time step based on the rate of change of the reconstructed transmission waveform in different time intervals: a first time step is used in the time intervals of the turn-off transition phase and the damped oscillation characteristic set, and a second time step is used in the steady phase and the decay phase. The change of adjacent time steps is controlled by a smoothing function, and the first time step is smaller than the second time step.

7. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 6, characterized in that, The iterative solution of the electric field using the finite difference time-domain method includes: iteratively solving the electric field strength and the recursive convolution memory variable of the polarization medium; calculating the time rate of change of the magnetic field strength according to Faraday's law of electromagnetic induction; and using the trapezoidal integral method to calculate the magnetic field strength at the next time node based on the time rate of change of the magnetic field strength at the current time node and the next time node, as well as the time step.

8. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 1, characterized in that, The method of using characteristic parameters of the ideal transmit current, overshoot tail component, and damped oscillation component as fitting parameters includes: the fitting parameters of the ideal transmit current include at least one of the transmit current amplitude, transmit current steady-state duration, transmit current turn-off start time, and transmit current turn-off time; the fitting parameters of the overshoot tail component include at least one of the overshoot tail component amplitude, the start time of the distortion component after turn-off, the relaxation mode time constant, and the number of relaxation modes; the fitting parameters of the damped oscillation component include at least one of the number of damped oscillation modes, the amplitude of each oscillation mode, the damping coefficient, the oscillation frequency, and the initial phase.

9. The superconducting electromagnetic three-dimensional modeling method based on measured emission waveform reconstruction according to claim 1, characterized in that, The parameter fitting of the measured transmission current includes: using the waveform parameters of the ideal transmission current, the characteristic parameters of the overshoot tail component, and the characteristic parameters of the damped oscillation component as the parameters to be fitted; using the deviation between the measured transmission current and the measured waveform expression model composed of the ideal transmission current, the overshoot tail component, the damped oscillation component, and the measurement noise as the objective function; iteratively adjusting the parameters to be fitted through an optimization algorithm to make the objective function reach the preset convergence condition; separating and removing the measurement noise; retaining the deterministic distortion characteristics generated by the transient dynamic process of the transmission system; and obtaining the reconstructed transmission waveform composed of the ideal transmission current, the overshoot tail component, and the damped oscillation component.