Method and system for removing data linear drift for time domain induced polarization method

Abstract

The application discloses a data linear drift removal method and system for a time domain induced polarization method, and relates to the technical field of geophysical exploration.The method comprises the following steps: acquiring original time sequence data through the time domain induced polarization method, and performing low-pass digital filtering on the original time sequence data with a cutoff frequency of 50 Hz; calculating the number of sampling points of each de-energization time base according to a transmission period and a sampling rate of the time domain induced polarization method; extracting a plurality of sampling points in each de-energization time base based on the number of sampling points, so as to obtain a sampling point set; obtaining a plurality of sampling point sets based on a preset number of transmission periods; performing least square fitting on the plurality of sampling point sets, so as to obtain a linear drift trend; and subtracting the linear drift trend from the original time sequence, so as to complete data repair.The application solves the problem that a traditional time domain induced polarization method decay curve crosses a 0 value line, and improves the data linear drift removal effect.

Description

Technical Field

[0001] This application relates to the field of geophysical exploration technology, and in particular to a method and system for removing linear data drift in time-domain induced polarization. Background Technology

[0002] The time-domain induced polarization method is a geophysical exploration method that uses the induced polarization effect of rocks and ores to find metal mines, groundwater, and solve engineering geological problems. Its core is to apply direct current underground and observe the decay process of the secondary field potential difference after the power is cut off, and analyze the difference in polarization effect of the double electric layer at the interface between rock particles and solution.

[0003] Currently, in time-domain induced polarization (TNP) exploration, linear drift caused by factors such as geodetic current, industrial stray current, and electrode range difference can significantly impact the results, particularly in the extraction of secondary voltage and the calculation of apparent polarizability. Linear drift interference is a core issue affecting data accuracy. Figure 1 As shown, tiny linear drifts in the original time series (difficult to detect with the naked eye, but having a significant impact on the decay curve due to the extremely weak secondary voltage after power failure) can cause the decay curve calculated from the secondary voltage to abnormally cross zero. For example... Figure 2 As shown, the early values ​​are positive while the later values ​​turn negative. Although the shape of the attenuation curve calculated using the secondary voltage is normal, it is clearly erroneous because it crosses the zero line, exhibiting a situation where the early values ​​are positive and the later values ​​are negative. The linear drift of the attenuation dashed line needs to be corrected. This severely violates the physical laws governing the excitation polarization effect.

[0004] Existing mainstream solutions have significant drawbacks. For example, the least squares global fitting method removes drift by fitting a straight line across the entire sequence. However, it suffers from the problem of exponential decay (non-linearity) of the secondary voltage in the initial stage of power failure, which, combined with linear drift, leads to fitting distortion. While it can remove linear drift to some extent, it cannot completely correct the secondary voltage. Empirical mode decomposition (EMD) methods require extracting the maxima and minima of the time series. If there are sudden disturbances, the maxima and minima may deviate significantly from the linear drift trend, resulting in excessively large errors in the calculated envelope.

[0005] In summary, there is an urgent need to propose a new method that is more suitable for removing the linear drift of time-domain excited polarization data and to solve the problem of the decay curve of time-domain excited polarization crossing the zero line. Summary of the Invention

[0006] To propose a new method more suitable for removing linear drift in time-domain excited polarization (TNP) data, address the issue of the TNP decay curve crossing the zero line, and improve the data linear drift removal effect, this application provides a method and system for removing linear drift in time-domain excited polarization (TNP).

[0007] Firstly, the objective of this invention is achieved through the following technical solution:

[0008] Data linear drift removal methods for time-domain excited polarization include:

[0009] The original time series data was obtained by time-domain excited polarization method, and the original time series data was subjected to low-pass digital filtering with a cutoff frequency of 50Hz.

[0010] Based on the emission period and sampling rate of the time-domain excitation polarization method, calculate the number of sampling points for each power-off time base;

[0011] Based on the number of sampling points, several sampling points are extracted in each power outage time base to obtain a sampling point set;

[0012] Multiple sampling point sets are obtained based on a preset number of emission cycles; the multiple sampling point sets are fitted using the least squares method to obtain a linear drift trend;

[0013] The data repair is completed by subtracting the linear drift trend from the original time series.

[0014] By adopting the above technical solution, this invention improves the traditional least squares method for fitting the linear drift trend of time series, and proposes a new method more suitable for removing linear drift of time-domain excited polarization (ETP) data. It solves the problem of the attenuation curve of the TEP method crossing the zero line, thus improving the linear drift removal effect. The original time series data obtained by TEP is first subjected to a 50Hz low-pass digital filter to remove the influence of the 50Hz industrial frequency sine wave on the least squares, making the linear drift trend more obvious. Then, based on the transmission period and sampling rate, the number of sampling points for each power outage time base is calculated. Because in some cases (e.g., strong anisotropy of underground resistivity), the received time series amplitude may differ between positive and negative power supply, and is not completely inversely symmetrical; furthermore, changes in the transmission current also affect the received time series amplitude during power supply. Therefore, this invention does not use sampling points during power supply periods, but rather uses sampling points during power outages throughout. In the early stages of a power outage, the secondary voltage exhibits an exponential decay trend over time, which is not a completely linear drift trend. Furthermore, in the late stages of a power outage, due to the imminent resumption of power supply and the influence of the transmission switch activation, significant pulse interference often occurs, although the duration of this pulse interference is very short. Therefore, the selection of sampling points is also crucial. This application reasonably avoids these data interference factors based on the start and end points of the sampling points, obtaining multiple sampling point sets based on a preset number of transmission cycles. Least square fitting is performed on these multiple sampling point sets to obtain a relatively obvious linear drift trend function. Finally, the linear drift trend is subtracted from the original time series to complete data repair. The linear drift in the repaired time series of this invention almost completely disappears, the attenuation curve has a normal shape, and all values ​​are positive, solving the problem of the attenuation curve crossing the zero line. The data linear drift removal method is more accurate and reliable.

[0015] In a preferred embodiment of this application: the starting sampling point of the sampling point set is set in the early stage of power outage based on the acquired attenuation trend data of the secondary voltage, where the secondary voltage at the starting sampling point has attenuated to close to the background value; the ending sampling point of the sampling point set is set in the late stage of power outage, before the power supply pulse interference is generated at the power supply switching moment, avoiding the switching pulse interference range.

[0016] By adopting the above technical solution, the starting sampling point is set before the secondary voltage decays to near the background value to avoid early noise interference, and the ending sampling point is set before the power supply pulse to avoid pulse interference at the power supply switching time, so as to eliminate the disturbance of the non-steady-state signal at both ends of the power outage time base to the drift region.

[0017] In a preferred example, this application further includes, after obtaining the set of sampling points:

[0018] The power outage process is divided into M consecutive windows, and the average value of the secondary field potential in each window is calculated. , where k ;

[0019] The Jacknife method was used to calculate the mean of each window. Deviation estimation specifically includes:

[0020] Remove one sampling point from each window and recalculate the mean to generate a pseudo-value sequence; if the pseudo-value deviates from the original mean by more than 2 standard deviations, remove the window.

[0021] Starting from the second window, loop through and check m=2:M. If the condition is met... If the loop stops, then windows m to M are retained as valid windows;

[0022] Update the sampled points within the selected valid window to the final fitted point set.

[0023] By adopting the above technical solution, the linear dominant region is dynamically screened to avoid exponential decay interference. Since the secondary voltage decays exponentially in the early stages of a power outage, its nonlinear characteristics are significant; while in the later stages (close to the background value), the decay slows down, with linear drift becoming dominant. The rate of change of the window mean is judged iteratively (the numerator is the difference between the current window and the last window, and the denominator is the difference between the first window and the last window). When the rate of change is less than the preset threshold THR, it indicates that the subsequent window has entered the linear dominant region (the change trend is consistent with the last window), avoiding the introduction of early nonlinear region data into the fitting. THR can be dynamically adjusted according to data characteristics (e.g., high-resistivity rock layers decay slowly, requiring an increased THR; low-resistivity mineral layers decay quickly, allowing a decreased THR). Fitting is performed after window mean calculation and window screening to reduce invalid data, lower computational complexity, and balance accuracy and efficiency. Only windows in the later linear region are retained through screening; for example, when M=10, only windows with m=7~10 may be retained. This avoids redundant calculations of a large amount of early nonlinear data, improving algorithm efficiency, and is particularly suitable for massive data scenarios.

[0024] In a preferred embodiment, after extracting a plurality of sampling points based on the number of sampling points in each power-off time base to obtain a set of sampling points, this application further includes:

[0025] Calculate the signal stability index for each sampling point set. :

[0026] in, The voltage value of the j-th sampling point in the i-th sampling point set; Let M be the average voltage of the i-th sampling point set; M is the number of sampling points.

[0027] like Greater than the preset stability threshold Then the extraction range of the corresponding sampling point set is adjusted as follows:

[0028] New starting point = original starting point + k × Δt, new ending point = original ending point − k × Δt, where Δt is the sampling interval, k is the gradually increasing adjustment step size, and k is greater than or equal to 1.

[0029] By adopting the above technical solution and providing a dynamically adjusted sampling mechanism, the stability index of this application is used to measure the fluctuation of the sampling point set within a certain power outage cycle. High variance may originate from residual effective signals, sudden noise, or poor resolution. When the data of a certain cycle is unstable, the sampling window is narrowed to focus on the later and more stable voltage plateau region, effectively avoiding information loss due to excessive clipping.

[0030] In a preferred example, the stability threshold of this application: The calculation formulas include:

[0031] in, For the current survey line The number of sampling points for each effective power outage time base.

[0032] By adopting the above technical solution, adjustments are triggered when the stability threshold deviates significantly from the historical average level.

[0033] In a preferred example, this application includes the following steps before subtracting the linear drift trend from the original time series:

[0034] Calculate the root mean square error of the fitted residuals:

[0035] in, This is the original time series data; The fitted value includes the drift trend; L is the total number of sampling points; t is the discrete time index, representing the sampling point number;

[0036] If the RMSE is greater than the preset error threshold, the preset number of transmission cycles is increased, more sampling point sets are acquired and fitted again, until the RMSE is less than or equal to the preset error threshold.

[0037] By adopting the above technical solution and using the root mean square error analysis model to fit the quality, this application provides an error-driven periodic expansion mechanism. When the existing data is insufficient to accurately fit the drift, the amount of data is actively increased, and more periodic information is used to improve the statistical significance of the trend estimate. Through evaluation, enhancement, and refitting, a high-precision drift removal method is output.

[0038] In a preferred embodiment, this application further includes, after subtracting the linear drift trend from the original time series:

[0039] Mapping the original time series to the interval [0, 1] is expressed by the formula:

[0040] in, To standardize time, and These are the minimum and maximum values ​​of the original time series, respectively;

[0041] A time-dependent drift function containing a linear drift term and a higher-order correction term is constructed, and the time-dependent drift function is applied to correct the drift of the standardized data.

[0042] The corrected data is subjected to inverse normalization transformation based on a preset amplitude compensation factor, expressed by the following formula:

[0043] in, For the final correction data, The corrected data is an intermediate result after correction using the time-related drift function described above; This represents the global peak voltage of the original time series; The peak voltage is the intermediate result of the corrected data.

[0044] By adopting the above technical solutions, a refined processing method and nonlinear compensation are provided after drift correction. Time axis standardization helps eliminate time scale differences and improves data processing consistency. By constructing a time-related drift function containing linear and higher-order terms, the nonlinear components that may exist in the drift are corrected to improve the generalization ability of the time-related drift function. Linear scaling through normalized inverse transform restores the signal amplitude ratio and maintains the physical meaning of the signal. In geological exploration, the secondary field voltage of the time-domain excited polarization method directly reflects the electrical parameters of the underground medium. If the amplitude of the corrected data is distorted (e.g., it is normalized and compressed to the [0, 1] interval), the inverted resistivity and other parameters will deviate from the true value. The linear inverse transform ensures that the physical meaning of the corrected data is consistent with the original data by restoring the original amplitude ratio. Therefore, the linear inverse transform avoids the analysis error caused by the change in amplitude scale.

[0045] Secondly, the objective of this invention is achieved through the following technical solution:

[0046] A data linear drift removal system for time-domain excited polarization (TNP) methods, applied to the data linear drift removal method for time-domain excited polarization methods as described above, the system comprising:

[0047] The data acquisition and filtering module is used to acquire raw time series data through time-domain excited polarization method, and to perform low-pass digital filtering on the raw time series data with a cutoff frequency of 50Hz.

[0048] The sampling point calculation module is used to calculate the number of sampling points for each power-off time base based on the emission period and sampling rate of the time-domain excitation polarization method.

[0049] The sampling point set extraction module is used to extract several sampling points based on the number of sampling points in each power outage time base to obtain a sampling point set;

[0050] The trend fitting module is used to obtain multiple sampling point sets based on a preset number of emission cycles, and to perform least squares fitting on the multiple sampling point sets to obtain a linear drift trend.

[0051] The data repair module is used to subtract the linear drift trend from the original time series to complete the data repair.

[0052] Thirdly, the objective of this invention is achieved through the following technical solution:

[0053] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described data linear drift removal method for time-domain excited polarization.

[0054] Fourthly, the objective of this invention is achieved through the following technical solution:

[0055] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described data linear drift removal method for time-domain excited polarization.

[0056] In summary, this application includes at least one of the following beneficial technical effects:

[0057] 1. This application improves upon the traditional least squares method, effectively solving the problem of the attenuation curve crossing zero. By applying a 50Hz low-pass filter to the original data, power frequency interference is eliminated, enhancing the significance of the drift trend. The number of sampling points for the power outage time base is calculated based on the transmission period and sampling rate, selecting data throughout the power outage period to avoid power supply interference and positive / negative asymmetry. Reasonable setting of sampling start and end points avoids early exponential attenuation and late switching pulse interference. By fitting a linear drift trend using a multi-period sampling point set, the fitting accuracy and stability are significantly improved. After repair, data drift is essentially eliminated, the attenuation curve has a normal shape and remains positive, making the method more accurate and reliable.

[0058] 2. Provides refined processing and nonlinear compensation methods after drift correction. Improves consistency through time axis standardization, constructs a drift function with higher-order terms to enhance the fitting ability of nonlinear components, and uses normalized inverse transform to restore the original amplitude ratio, maintain the physical meaning of the signal, avoid the deviation of inversion parameters due to amplitude distortion, and ensure that the correction results are accurate and reliable. Attached Figure Description

[0059] Figure 1 This is an original time series plot at a certain measurement point obtained by removing the linear drift trend using the traditional time-domain excited polarization method in one embodiment of this application;

[0060] Figure 2 This is the original attenuation curve at a certain measurement point (which crosses the 0 value line) obtained by using the traditional time-domain excited polarization method to remove the linear drift trend in one embodiment of this application.

[0061] Figure 3 This is a flowchart of a data linear drift removal method for time-domain excited polarization in one embodiment of this application;

[0062] Figure 4 This is a repair time series diagram at a certain measurement point in a practical application of the data linear drift removal method for time-domain excited polarization method according to an embodiment of this application;

[0063] Figure 5 This is a repair attenuation curve (not crossing the 0 value line) at a certain measurement point in the actual application of the data linear drift removal method for time-domain excited polarization method in one embodiment of this application.

[0064] Figure 6 This is a schematic diagram of a device according to one embodiment of this application. Detailed Implementation

[0065] The present application will be further described in detail below with reference to the accompanying drawings.

[0066] In one embodiment, such as Figure 3 As shown, this application discloses a method for removing linear data drift using time-domain excited polarization, which specifically includes the following steps:

[0067] S1: Obtain the original time series data by time-domain excited polarization method, and perform low-pass digital filtering on the original time series data with a cutoff frequency of 50Hz.

[0068] In this embodiment, a stainless steel power supply electrode and a non-polarizing copper sulfate measurement electrode are used, with a power supply current intensity of 5-20A and a sampling rate of [missing information]. ≥100Hz. The raw data includes signals during the power-on period (primary field dominance) and the power-off period (secondary field attenuation).

[0069] Specifically, an FIR low-pass filter with a cutoff frequency of 50Hz is used to perform low-pass digital filtering on the acquired raw time series data, resulting in a stopband attenuation of ≥40dB and a passband ripple of ≤0.1dB. The 50Hz low-pass filter can retain the effective frequency band while suppressing ≥98% of power frequency interference, improving the signal-to-noise ratio by more than 20dB.

[0070] S2: Calculate the number of sampling points for each power-off time base based on the emission period and sampling rate of the time-domain excitation polarization method.

[0071] In this embodiment, based on the transmission period T and the sampling rate The number of sampling points N for a single power outage time base is calculated to determine the data segmentation reference. The typical range of the transmission period T is 2-8s.

[0072] Specifically, the formula for calculating the number of sampling points N is: in, The sampling rate is T; the transmission period is T; the extraction range of the sampling point set is set to the position from 3N / 4+1 to 7N / 8 of the power-off time base, and the sampling point set has a total of N / 8 sampling points.

[0073] In the early stages of a power outage, the secondary voltage exhibits an exponential decay trend over time, rather than a completely linear drift trend. For the vast majority of time-domain excitation polarization data, after 3N / 4 sampling points, the secondary voltage has essentially decayed to a very low value, and subsequent sampling points show a linear drift. Therefore, the starting sampling point for each power outage time base is set to 3N / 4+1.

[0074] During the later stages of a power outage, significant pulse interference often occurs due to the imminent resumption of power supply and the activation of the transmitter switch. This pulse interference is very short-lived, lasting no more than N / 8 sampling points. Therefore, the median sampling point for each power outage time base is set to 7N / 8. In practical applications, the sequence numbers of the start and end sampling points in the extracted power outage time base can be set according to specific circumstances.

[0075] In certain situations (such as when the resistivity of the underground is highly anisotropic), the amplitude of the received time series may differ between positive and negative power supply, and is not completely inversely symmetrical. Furthermore, changes in the transmit current also affect the amplitude of the received time series during power supply. Therefore, the algorithm in this scheme does not utilize the sampling points during power supply.

[0076] S3: Based on the number of sampling points, extract several sampling points in each power outage time base to obtain a sampling point set.

[0077] In this embodiment, a fixed number of sampling points are extracted in the stable decay region of the power-off time base (avoiding early transition and late interference) to form a sample set characterizing linear drift.

[0078] In another embodiment, after obtaining the set of sampling points, the method further includes:

[0079] The power outage process is divided into M consecutive windows, and the average value of the secondary field potential in each window is calculated. , where k ;

[0080] The Jacknife method was used to calculate the mean of each window. Deviation estimation specifically includes:

[0081] Remove one sampling point from each window and recalculate the mean to generate a pseudo-value sequence; if the pseudo-value deviates from the original mean by more than 2 standard deviations, remove the window.

[0082] Starting from the second window, loop through and check m=2:M. If the condition is met... If the loop stops, then windows m to M are retained as valid windows;

[0083] Update the sampled points within the selected valid window to the final fitted point set.

[0084] Specifically, the preset threshold THR ranges from 0 to 1. Jacknife uses a "remove-recalculate" mechanism to accurately identify and remove local anomaly windows, such as those where the mean deviates from the normal range due to electrode jitter, thus preventing outliers from contaminating subsequent linear fitting. When the updated sampling point set is obtained, the secondary field potential data within all valid windows can be extracted, and M-estimation iterative calculations can be performed to calculate robust regression coefficients. Based on these robust regression coefficients, a linear drift trend function is obtained. M-estimation iterative calculations are beneficial for solving the problem of outlier interference in field data, such as transient jumps caused by poor electrode contact. In practical applications, the preset threshold THR can be dynamically adjusted according to data characteristics. For example, high-resistivity rock layers decay slowly, so THR needs to be increased; low-resistivity mineral layers decay quickly, so THR can be decreased, solving the problem that traditional fixed-time-window methods may misjudge effective areas under complex geological conditions.

[0085] S4: Obtain multiple sampling point sets based on a preset number of emission cycles; perform least squares fitting on the multiple sampling point sets to obtain a linear drift trend.

[0086] In this embodiment, it is recommended that the preset number of transmission cycles be greater than or equal to 3 cycles. In high-interference regions, the number of cycles can be set even higher.

[0087] Specifically, point sets from multiple transmission cycles are continuously acquired, with each point set corresponding to a stable attenuation region of a power-off time base. The signal stability index for each sampling point set is then calculated. :

[0088] in, The voltage value of the j-th sampling point in the i-th sampling point set; Let M be the average voltage of the i-th sampling point set; M is the number of sampling points; M = N / 8. When M < 3, the power outage time base data is invalid, and the sampling point set needs to be extracted again. Average voltage The calculation needs to exclude outlier sampling points that exceed ±3σ, where σ is the standard deviation.

[0089] like Greater than the preset stability threshold Adjust the extraction range of the corresponding sampling point set as follows:

[0090] New starting point = original starting point + k × Δt, new ending point = original ending point − k × Δt, where Δt is the sampling interval, k is the gradually increasing adjustment step size, and k is greater than or equal to 1.

[0091] For example, when =200Hz, Δt=0.005s, if =0.5mV> =0.4mV, then k=1 is used to adjust the range, and the new range offset is 0.005s.

[0092] Specifically, all sampling points of the p point sets are merged into one dataset, with a total of m = p × (N / 8) points. For example, if p = 5 and N = 200, then m = 125 points.

[0093] Assume the linear drift trend function is ,in Let be the drift slope, b be the intercept, and t be the timestamp. The parameters to be determined are: and .

[0094] Solve and The steps include: first constructing the matrix equation ,in (Design matrix, (where i is the index of the time variable). (Observed voltage vector) (Parameters to be determined), solve the equation .

[0095] Furthermore, to reduce the impact of instability in the early stages of the cycle, weights are introduced. ,in For the periodic number, =1 is the earliest period, and the weighted equation is: ,in Given a diagonal weight matrix, the solution is obtained. and Then, the fitting function is obtained. .

[0096] S5: Subtract the linear drift trend from the original time series to complete the data repair.

[0097] In this embodiment, for each time point t in the original time series, the corresponding repaired data value is equal to the original data value minus the estimated linear drift trend value at that time point t, thus achieving data repair. Figure 4 and Figure 5 The figures shown are the repair time series and the repair attenuation curve of a certain measurement point using the time-domain excited polarization method of this embodiment, respectively. Figure 4 and Figure 1 Compare the processing effects, and Figure 5 and Figure 2 By comparing the processing results, it can be clearly seen that the linear drift removal method of this application almost completely eliminates the linear drift in the time series, the decay curve has a normal shape and all are positive, and the problem of the decay curve crossing the 0 value line is solved.

[0098] Specifically, in practical applications, time-domain excited polarization data acquisition and preprocessing software such as Geo3D and prTDIP can be introduced to remove linear drift in time series data.

[0099] In one embodiment, the stability threshold The calculation formulas include:

[0100] in, For the current survey line The number of sampling points for each effective power outage time base.

[0101] In one embodiment, the data linear drift removal method for time-domain excited polarization further includes:

[0102] Before subtracting the linear drift trend from the original time series, the following is included:

[0103] Calculate the root mean square error of the fitted residuals:

[0104] in, This is the original time series data; The fitted value includes the drift trend; L is the total number of sampling points; t is the discrete time index, representing the sampling point number;

[0105] If the RMSE is greater than the preset error threshold, the preset number of transmission cycles is increased, more sampling points are acquired and fitted again, until the RMSE is less than or equal to the preset error threshold.

[0106] Specifically, when RMSE ≤ 0.5%, the allowable error range is ±0.1%, which is applicable to high-precision scenarios such as fine detection of metal ore bodies; when 0.5% < RMSE ≤ 2%, the allowable error range is ±0.5%, which is applicable to conventional mineral exploration; when RMSE > 2%, a three-level calibration process is triggered, which is applicable to complex geological conditions such as high polarization rate mineralized zones.

[0107] The three-level calibration process includes:

[0108] Step A: Expand the number of emission cycles p (e.g., from 3 to 5), and re-obtain p′ = p + 2 sets of sampling points;

[0109] Step B: Recalculate the stability threshold for the new set of sampling points , and screen the stable point set;

[0110] Step C: Use weighted least squares fitting (at this time, the weight of outliers is reduced) to obtain a new drift trend ;

[0111] Step D: Recalculate RMSE′. If it is still > 2%, start hardware calibration, such as checking the electrode grounding resistance and adjusting the sampling rate.

[0112] In one embodiment, after subtracting the linear drift trend from the original time series, it further includes:

[0113] Map the original time series to the [0, 1] interval, and the formula is expressed as:

[0114] where, is the standardized time, and are the minimum and maximum values of the original time series respectively;

[0115] Construct a time-dependent drift function including a linear drift term and a high-order correction term, and apply the time-dependent drift function to perform drift correction on the standardized data;

[0116] According to the preset amplitude compensation factor, perform a normalized inverse transformation on the corrected data, and the formula is expressed as:

[0117] where, is the finally corrected data, [[ID=4S]] is the corrected data, which is the intermediate result after applying the time-dependent drift function for correction; is the global peak voltage of the original time series; is the peak voltage of the intermediate result of the corrected data; the normalized inverse transformation ensures that the physical dimension of the corrected data is consistent with the original data by maintaining the peak ratio of the original signal.

[0118] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0119] In one embodiment, a data linear drift removal system for time-domain excited polarization is provided, which corresponds to the data linear drift removal method for time-domain excited polarization described in the above embodiments.

[0120] A data linear drift removal system for time-domain excited polarization methods includes a data acquisition and filtering module, a sampling point count calculation module, a sampling point set extraction module, a trend fitting module, and a data repair module. Detailed descriptions of each functional module are as follows:

[0121] The data acquisition and filtering module is used to acquire raw time series data through time-domain excited polarization method and to perform low-pass digital filtering on the raw time series data with a cutoff frequency of 50Hz.

[0122] The sampling point calculation module is used to calculate the number of sampling points for each power-off time base based on the emission period and sampling rate of the time-domain excitation polarization method.

[0123] The sampling point set extraction module is used to extract several sampling points based on the number of sampling points in each power outage time base to obtain a sampling point set;

[0124] The trend fitting module is used to obtain multiple sampling point sets based on a preset number of emission cycles, and to perform least squares fitting on the multiple sampling point sets to obtain a linear drift trend.

[0125] The data repair module is used to subtract the linear drift trend from the original time series to complete the data repair.

[0126] For specific limitations regarding the data linear drift removal system used in time-domain excited polarization (TDP) method, please refer to the limitations of the data linear drift removal method used in TDP method above, which will not be repeated here. Each module in the above-mentioned data linear drift removal system used in time-domain excited polarization method can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in the processor of the computer device in hardware form or independent of the processor, or it can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0127] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores raw time-series data, emission periods, sampling rates, etc. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a data linear drift removal method for time-domain excited polarization.

[0128] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0129] S1: Obtain the original time series data by time-domain excited polarization method, and perform low-pass digital filtering on the original time series data with a cutoff frequency of 50Hz;

[0130] S2: Calculate the number of sampling points for each power-off time base based on the emission period and sampling rate of the time-domain excitation polarization method;

[0131] S3: Based on the number of sampling points, extract several sampling points in each power outage time base to obtain a sampling point set;

[0132] S4: Obtain multiple sampling point sets based on a preset number of emission cycles; perform least squares fitting on the multiple sampling point sets to obtain a linear drift trend;

[0133] S5: Subtract the linear drift trend from the original time series to complete the data repair.

[0134] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0135] S1: Obtain the original time series data by time-domain excited polarization method, and perform low-pass digital filtering on the original time series data with a cutoff frequency of 50Hz;

[0136] S2: Calculate the number of sampling points for each power-off time base based on the emission period and sampling rate of the time-domain excitation polarization method;

[0137] S3: Based on the number of sampling points, extract several sampling points in each power outage time base to obtain a sampling point set;

[0138] S4: Obtain multiple sampling point sets based on a preset number of emission cycles; perform least squares fitting on the multiple sampling point sets to obtain a linear drift trend;

[0139] S5: Subtract the linear drift trend from the original time series to complete the data repair.

[0140] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0141] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0142] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for removing linear data drift in time-domain excited polarization, characterized in that, include: The original time series data was obtained by time-domain excited polarization method, and the original time series data was subjected to low-pass digital filtering with a cutoff frequency of 50Hz. Based on the emission period and sampling rate of the time-domain excitation polarization method, calculate the number of sampling points for each power-off time base; Based on the number of sampling points, several sampling points are extracted in each power outage time base to obtain a sampling point set; Multiple sampling point sets are obtained based on a preset number of transmission cycles; The linear drift trend is obtained by performing least squares fitting on multiple sets of sampling points. Subtract the linear drift trend from the original time series to complete data repair; After subtracting the linear drift trend from the original time series, the method further includes: Mapping the original time series to the interval [0, 1] is expressed by the formula: in, To standardize time, and These are the minimum and maximum values ​​of the original time series, respectively; A time-dependent drift function containing a linear drift term and a higher-order correction term is constructed, and the time-dependent drift function is applied to correct the drift of the standardized data. The corrected data is subjected to inverse normalization transformation based on a preset amplitude compensation factor, expressed by the following formula: in, For the final correction data, The corrected data is an intermediate result after correction using the time-related drift function described above; This represents the global peak voltage of the original time series; The peak voltage is the intermediate result of the corrected data.

2. The data linear drift removal method for time-domain excited polarization method according to claim 1, characterized in that, The starting sampling point of the sampling point set is set in the early stage of power failure based on the acquired secondary voltage decay trend data. At the starting sampling point, the secondary voltage has decayed to close to the background value. The final sampling point of the sampling point set is set in the late stage of power outage, before the power supply pulse interference is generated at the moment of power supply switching, thus avoiding the switching pulse interference range.

3. The data linear drift removal method for time-domain excited polarization method according to claim 2, characterized in that, After obtaining the set of sampling points, it also includes: The power outage process is divided into M consecutive windows, and the average value of the secondary field potential in each window is calculated. , where k ; The Jacknife method was used to calculate the mean of each window. Deviation estimation specifically includes: After removing one sampling point from each window, the mean is recalculated to generate a pseudo-value sequence; windows where the deviation between the pseudo-value and the original mean exceeds 2 standard deviations are removed. Starting from the second window, loop through and check m=2:M. If the condition is met... If the condition is met, the loop stops and windows m to M are retained as valid windows; where Let be the average of the secondary field potentials in the m-th window. Let be the average of the secondary field potentials in the Mth window. This represents the average secondary field potential of the first window; Update the sampled points within the selected valid window to the final fitted point set.

4. The data linear drift removal method for time-domain excited polarization method according to claim 3, characterized in that, After extracting several sampling points based on the stated number of sampling points in each power outage time base to obtain a sampling point set, the method further includes: Calculate the signal stability index for each sampling point set. : in, Let be the voltage value of the j-th sampling point in the i-th sampling point set; Let M be the average voltage of the i-th sampling point set; M is the number of sampling points. like Greater than the preset stability threshold Then the extraction range of the corresponding sampling point set is adjusted as follows: New starting point = original starting point + k × Δt, new ending point = original ending point − k × Δt, where Δt is the sampling interval, k is the gradually increasing adjustment step size, and k is greater than or equal to 1.

5. The data linear drift removal method for time-domain excited polarization method according to claim 4, characterized in that, The stability threshold Calculation formula include: in, For the current survey line The number of sampling points for each effective power outage time base.

6. The data linear drift removal method for time-domain excited polarization method according to claim 4, characterized in that, Before subtracting the linear drift trend from the original time series, the process includes: Calculate the root mean square error of the fitted residuals: in, This is the original time series data; The fitted values ​​include the drift trend; L is the total number of sampling points; t is the discrete-time index, representing the sampling point number; If the RMSE is greater than the preset error threshold, the preset number of transmission cycles is increased, more sampling point sets are acquired and fitted again, until the RMSE is less than or equal to the preset error threshold.

7. A data linear drift removal system for time-domain excited polarization method, characterized in that, The system, applied to the data linear drift removal method for time-domain excited polarization as described in any one of claims 1 to 6, comprises: The data acquisition and filtering module is used to acquire raw time series data through time-domain excited polarization method, and to perform low-pass digital filtering on the raw time series data with a cutoff frequency of 50Hz. The sampling point calculation module is used to calculate the number of sampling points for each power-off time base based on the emission period and sampling rate of the time-domain excitation polarization method. The sampling point set extraction module is used to extract several sampling points based on the number of sampling points in each power outage time base to obtain a sampling point set; The trend fitting module is used to obtain multiple sampling point sets based on a preset number of emission cycles, and to perform least squares fitting on the multiple sampling point sets to obtain a linear drift trend. The data repair module is used to subtract the linear drift trend from the original time series to complete the data repair.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the data linear drift removal method for time-domain excited polarization as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the data linear drift removal method for time-domain excited polarization as described in any one of claims 1 to 6.

Images