A method and system for dynamic monitoring and early warning of underground media based on electromagnetic signals.

Abstract

This invention relates to the field of geophysical exploration and engineering monitoring and early warning technology. Addressing the problems of background drift, high computational load, and susceptibility to false triggering in existing methods, it discloses a method and system for dynamic monitoring and early warning of underground media based on electromagnetic signals. The method includes: continuously acquiring and processing time-domain electromagnetic data in segments; calculating the apparent resistivity at each target frequency at the monitoring location; constructing and updating a background benchmark; calculating the relative rate of change and statistical characteristics of apparent resistivity, constructing multi-dimensional anomaly judgment conditions, generating an inversion trigger identifier for the monitoring location when the condition is met, otherwise proceeding to the next monitoring period; after triggering, limiting the inversion data range for resistivity inversion imaging, identifying anomalies and issuing an early warning by comparing the differences between the inverted resistivity model and the background resistivity model. This invention avoids drift and missed detections through an anomaly exclusion mechanism, reduces computational consumption through conditional triggering inversion, and reduces false triggering through multi-dimensional joint judgment based on temporal persistence, frequency correlation, and spatial continuity.

Description

Technical Field

[0001] This invention belongs to the field of geophysical exploration and engineering monitoring and early warning technology, specifically relating to a method and system for dynamic monitoring and early warning of underground media based on electromagnetic signals. Background Technology

[0002] With the continuous increase in engineering activities such as urban underground space development, tunnel and foundation pit construction, dam and slope treatment, and risk prevention and control in mining subsidence and karst areas, the underground medium often undergoes dynamic changes such as changes in water content, fissure expansion, local loosening, cavity development, and weak zone evolution under the influence of factors such as excavation disturbance, groundwater changes, stress redistribution, and structural deterioration. These changes are often hidden, gradual, and sudden, and can easily induce engineering disasters such as collapse, water inrush, and instability. Therefore, there is an urgent need for monitoring and early warning technologies that can continuously reflect the time-varying characteristics of the underground medium and promptly identify abnormal development trends.

[0003] While current methods such as drilling, in-situ testing, manual inspection, and phased geophysical exploration can identify underground anomalies to some extent, they mostly employ point-based, cross-sectional, or low-frequency observation methods, resulting in limited spatial coverage and insufficient temporal continuity. This makes it difficult to capture the incubation and evolution of anomalies, and in particular, it is difficult to meet the needs of continuous, real-time, and regional monitoring during the construction and operation phases of underground engineering projects.

[0004] Electromagnetic methods are highly sensitive to changes in the electrical properties of underground media and offer advantages such as wide coverage, repeatable observation, and convenient continuous data acquisition. After time-frequency conversion, the time-domain data can yield frequency domain responses and apparent resistivity at multiple frequencies. Different frequencies exhibit varying sensitivities to different depths underground. Continuous comparison of frequency domain responses and apparent resistivity across multiple time periods provides an effective data foundation and technical support for identifying dynamic changes in underground media, determining anomalies, inverting and locating locations, and issuing risk warnings.

[0005] However, the existing electromagnetic monitoring technology still has the following problems in engineering applications: (1) Long-term monitoring background drift problem. If the background benchmark is fixed, normal drift is easily misjudged as anomaly. If all data is absorbed indiscriminately, the real abnormal signal will be gradually absorbed by the background, resulting in missed reports. There is a lack of effective abnormal data exclusion mechanism. (2) Computational resource consumption problem. If the existing method performs indiscriminate inversion of each monitoring period, the computational cost is high and it is difficult to meet the real-time requirements of continuous monitoring. (3) Noise false triggering and insufficient utilization of multidimensional information problem. Occasional noise at a single frequency point or in a single time period may trigger false anomalies. The existing method does not make sufficient use of multidimensional information such as time continuity, frequency correlation and spatial continuity. Summary of the Invention

[0006] To address the issues of background drift during long-term monitoring, high resource consumption in inversion calculations, noise-induced false triggering, and insufficient utilization of multi-dimensional information, this invention provides a method and system for dynamic monitoring and early warning of underground media based on electromagnetic signals. To achieve continuous monitoring, anomaly identification, spatial location, and dynamic early warning of the electrical state of underground media, this invention uses controllable source electromagnetic continuous observation data as its primary implementation basis. Continuous time-domain data is divided into multiple monitoring periods according to a preset time window. Frequency domain amplitudes corresponding to multiple target frequencies are extracted and used for initial screening of candidate anomalies. Simultaneously, the apparent resistivity of multiple frequencies is calculated for qualified frequency domain data and compared with a background benchmark across multiple time periods. Based on this, a joint inversion trigger identifier is formed by combining temporal continuity, frequency correlation, and spatial continuity. According to the triggering results, the inversion period, target frequency, and anomaly region are selected for inversion imaging. The difference between the current inversion model and the background inversion model is used to locate underground anomalies, generating and updating early warning results. This achieves dynamic monitoring and early warning capabilities with long-term stable monitoring, efficient calculation, and accurate early warning.

[0007] This invention provides a method for dynamic monitoring and early warning of underground media based on electromagnetic signals, comprising the following steps: Step 110: Continuously collect time-domain electromagnetic data of the underground medium under electromagnetic field, and divide it into different monitoring periods according to a preset time window; based on the time-domain electromagnetic data of the current monitoring period, calculate the apparent resistivity of the underground medium at each monitoring location and each target frequency in the current monitoring period; Step 120, constructing and maintaining a background reference for apparent resistivity, including: comparing the apparent resistivity of the current monitoring period with the background reference to obtain the relative rate of change of apparent resistivity; and updating the background reference using the apparent resistivity that meets the preset background update conditions in the current monitoring period. Step 130: Based on the comparison results between the relative rate of change of apparent resistivity and the threshold of change of apparent resistivity, calculate the statistical characteristics, including the number of consecutive time periods exceeding the limit, the number of frequencies of anomalies in the same direction, and the number of monitoring locations of anomalies in the same direction. Step 140: Construct multi-dimensional anomaly determination conditions using the relative rate of change of apparent resistivity and the statistical characteristics; when the multi-dimensional anomaly determination conditions are met at least at one target frequency, generate an inversion trigger identifier for the corresponding monitoring location; otherwise, return to step 110 to obtain the time-domain electromagnetic data for the next monitoring period and calculate the apparent resistivity. Step 150: Based on the inversion trigger identifier, limit the data range for inversion; extract the effective electromagnetic data within the data range for resistivity inversion imaging to obtain the inversion resistivity model for the current monitoring period; Step 160: Compare the inverted resistivity model of the current monitoring period with the preset background resistivity model to obtain the difference area; determine the spatial location and range of the anomaly in the underground medium based on the difference area, and generate or update the early warning information.

[0008] Preferably, in step 120, the preset background update condition includes: the absolute value of the relative change rate of apparent resistivity is less than the apparent resistivity change threshold; when the time-domain electromagnetic data is collected using a controllable source electromagnetic observation method, the preset background update condition further includes: the current monitoring period, monitoring location, and target frequency are not marked as candidate anomalies by the initial screening of frequency domain amplitude. The step of updating the background reference using the apparent resistivity that meets the preset background update conditions during the current monitoring period includes: When the preset background update condition is met, the background update coefficient is applied. The background benchmark is then updated using a weighted method: ; in, For the current number after the update Background benchmark for each monitoring period For the first time before the update Background benchmark for each monitoring period For the current number Apparent resistivity for each monitoring period, background update coefficient satisfy ; Otherwise, the background reference remains unchanged: .

[0009] Preferably, in step 130: The number of consecutive out-of-limit periods is calculated by backtracking from the current monitoring period to count the number of monitoring periods in which the absolute value of the relative change rate of apparent resistivity is continuously greater than or equal to the apparent resistivity change threshold at the same monitoring location and target frequency. When the first monitoring period that does not meet the out-of-limit condition is encountered during the backtracking process, the count is reset to zero and the count restarts from the next monitoring period. The out-of-limit condition refers to the absolute value of the relative change rate of apparent resistivity being greater than or equal to the apparent resistivity change threshold. The number of unidirectional abnormal frequencies is taken as the maximum value between the number of positively exceeding target frequencies and the number of negatively exceeding target frequencies. ; in, It is the monitoring location and monitoring period The number of target frequencies exceeding the internal positive limit. It is the monitoring location and monitoring period The number of target frequencies exceeding the negative limit is given by the following formulas: ; ; In the formula, For indicator functions, It is the total number of target frequencies participating in anomaly detection. It is the first Monitoring time period at monitoring location and target frequency The relative rate of change of apparent resistivity under these conditions It is the threshold for apparent resistivity change; The number of unidirectional anomaly monitoring locations is given by the following formula: ; in, It includes monitoring locations and the preset range set of adjacent monitoring positions, It is the first Monitoring time period at monitoring location and target frequency The relative rate of change of apparent resistivity under these conditions .

[0010] Preferably, in step 140, the multi-dimensional anomaly determination conditions include the following necessary conditions: Necessary condition 1: The current monitoring period, monitoring location, and target frequency are determined to pass the initial screening of frequency domain amplitude; Second necessary condition: The absolute value of the relative rate of change of apparent resistivity is greater than or equal to the threshold value of apparent resistivity change; Necessary condition three: The number of consecutive over-limit periods is greater than or equal to a preset threshold for the number of consecutive over-limit periods; Necessary condition four: For scenarios with multiple monitoring locations, the number of unidirectional abnormal frequencies is greater than or equal to a preset threshold for the number of unidirectional abnormal frequencies, or the number of unidirectional abnormal monitoring locations is greater than or equal to a preset threshold for the number of unidirectional abnormal monitoring locations. For scenarios where only a single monitoring location is set, the number of unidirectional abnormal frequencies is greater than or equal to a preset threshold for the number of unidirectional abnormal frequencies.

[0011] On the other hand, the present invention also provides a dynamic monitoring and early warning system for underground media based on electromagnetic signals, comprising: The first module is used to continuously collect time-domain electromagnetic data of underground media under electromagnetic fields and divide it into different monitoring periods according to a preset time window; based on the time-domain electromagnetic data of the current monitoring period, it calculates the apparent resistivity of the underground media at each monitoring location and each target frequency in the current monitoring period. The second module is used to construct and maintain a background reference for apparent resistivity, including: comparing the apparent resistivity of the current monitoring period with the background reference to obtain the relative rate of change of apparent resistivity; and updating the background reference using the apparent resistivity that meets the preset background update conditions in the current monitoring period. The third module is used to calculate statistical characteristics based on the comparison results between the relative rate of change of apparent resistivity and the threshold of change of apparent resistivity, including the number of consecutive time periods exceeding the limit, the number of frequencies of anomalies in the same direction, and the number of monitoring locations of anomalies in the same direction. The fourth module is used to construct multi-dimensional anomaly judgment conditions using the relative change rate of apparent resistivity and the statistical characteristics; when the multi-dimensional anomaly judgment conditions are met at least at one target frequency, an inversion trigger identifier for the corresponding monitoring location is generated; otherwise, the module returns to the first module to obtain the time-domain electromagnetic data for the next monitoring period and calculate the apparent resistivity. The fifth module is used to limit the data range for inversion based on the inversion trigger identifier; extract the effective electromagnetic data within the data range for resistivity inversion imaging, and obtain the inversion resistivity model for the current monitoring period; The sixth module is used to compare the inverted resistivity model of the current monitoring period with the preset background resistivity model to obtain the difference area; based on the difference area, the spatial location and range of the anomaly in the underground medium are determined, and early warning information is generated or updated.

[0012] Compared with the prior art, the specific beneficial effects achieved by the solution of the present invention include: (1) This invention constructs a background reference apparent resistivity and calculates the difference between the multi-frequency apparent resistivity of each monitoring period and the background reference, highlighting the change information of the underground medium relative to the steady state and weakening the influence of static geological differences and environmental background on anomaly identification. Only data that has not triggered candidate anomaly markers and whose rate of change is within the normal fluctuation range are used to participate in the background reference update. Candidate anomalies, data with excessive rate of change or data that have entered the inversion process are excluded, effectively preventing abnormal signals from being slowly absorbed into the background reference, avoiding the problem of missed reports caused by "anomaly drift" in long-term monitoring, and tracking confirmed instrument drift, environmental changes or other non-target background changes.

[0013] (2) By using multi-dimensional joint judgment, an inversion trigger flag for the corresponding monitoring location is generated and inversion imaging is initiated only when a significant anomaly is confirmed. If no trigger is triggered, only low-cost data acquisition and statistical judgment are performed, and the next monitoring period continues. This invention repositions resistivity inversion from a conventional processing step to a precise positioning method after anomaly confirmation, avoiding the entry of a large amount of data from non-anomaly periods into the high-cost inversion step, significantly reducing computational resource consumption, and meeting the real-time requirements of continuous monitoring.

[0014] (3) This invention combines three dimensions—temporal persistence, frequency correlation, and spatial continuity—for joint judgment, which significantly reduces the false trigger rate. Among them, temporal persistence is used to reduce the impact of occasional interference in a single monitoring period; frequency correlation is verified by the consistency of the direction of change of multiple adjacent or related target frequencies to reduce false triggers caused by occasional anomalies in a single frequency; spatial continuity is used to reduce the impact of anomalies in isolated monitoring locations. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a method for dynamic monitoring and early warning of underground media based on electromagnetic signals in one embodiment of the present invention. Figure 2 The following is a schematic diagram of the inversion results for different monitoring periods in one embodiment of the present invention: (a) is a schematic diagram of the inversion results for the monitoring period numbered Line01_0711_01h, (b) is a schematic diagram of the inversion results for the monitoring period numbered Line01_0713_03h, (c) is a schematic diagram of the inversion results for the monitoring period numbered Line01_0713_20h, (d) is a schematic diagram of the inversion results for the monitoring period numbered Line01_0714_21h, and (e) is a schematic diagram of the inversion results for the monitoring period numbered Line01_0715_14h. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] In one embodiment, such as Figure 1 As shown, this invention provides a method for dynamic monitoring and early warning of underground media based on electromagnetic signals, comprising the following steps: Step 110: Continuously collect time-domain electromagnetic data of the underground medium under an electromagnetic field, and divide it into different monitoring periods according to a preset time window; based on the time-domain electromagnetic data of the current monitoring period, calculate the apparent resistivity of the underground medium at each monitoring location and each target frequency during the current monitoring period. Specifically, this includes: Step 111: Deploy electromagnetic observation devices in the area to be monitored to conduct electromagnetic observations, continuously collect time-domain electromagnetic data generated by the underground medium under the action of electromagnetic fields, and divide the data into different monitoring periods according to a preset time window.

[0019] Depending on the monitoring range, target depth, and on-site conditions, the electromagnetic observation method can be either a controlled-source electromagnetic observation method or a natural-source electromagnetic observation method.

[0020] When using controlled-source electromagnetic observation, an artificial electromagnetic field source applies electromagnetic excitation to the underground medium according to a preset transmission waveform and working cycle. Electric field, magnetic field, or received voltage time-domain data are continuously acquired at one or more fixed monitoring locations. Simultaneously, observation parameters such as transmission current, field source location, receiving location, transmission-receiver distance, and acquisition time are recorded. It is preferable to maintain consistency in the field source deployment, receiving location, and observation parameters across different monitoring periods. When the transmission current changes, the frequency domain amplitude is normalized and corrected based on the measured transmission current.

[0021] When using natural source electromagnetic observation, mutually orthogonal time-domain electric and magnetic field data are simultaneously collected in the monitoring area, and the coordinates of the monitoring points, channel direction, sampling rate, and acquisition time are recorded. The same or spatially corresponding observation setup is used for each monitoring period to ensure comparability of data from different monitoring periods.

[0022] In a preferred embodiment, a controlled-source electromagnetic observation method is employed, in which a grounded long-wire transmitting source (transmitting electric dipole AB) and multiple receiving stations (including receiving electric dipole MN) are deployed in the monitoring area. The length of the transmitting electric dipole... and receiving electric dipole length Configure according to site conditions. For example, configure... Within the range of 100m to 1000m, The distance is set from 1m to 20m. Each receiving station is set at equal intervals along the survey line, and the distance between adjacent receiving stations can be set from 1m to 55m depending on the monitoring accuracy requirements.

[0023] In one embodiment, after the electromagnetic observation device is activated, the transmitting source applies electromagnetic excitation to the underground medium according to a preset transmission waveform and working cycle. This embodiment uses a bipolar square wave as the transmission waveform, and the transmission frequency is selected within the range of 0.1Hz to 65kHz based on the target detection depth.

[0024] The acquired time-domain electromagnetic data is segmented into different monitoring periods, using a preset time window as the unit. The current [number]th [period / section]... Each monitoring period can be represented as: ; in, This is the start time of the monitoring period. The time window length is specified. The time window can be set to 2s, 4s, 8s, or other time lengths, and can be a continuous non-overlapping window or a sliding window with a preset overlap rate.

[0025] The time-domain electromagnetic data of each monitoring period are correlated with the monitoring location, target frequency, transmitted current, field source parameters, received parameters and acquisition time to form a multi-time period monitoring data sequence with a unified time and space reference.

[0026] Step 112: Based on the time-domain electromagnetic data for the current monitoring period, calculate the apparent resistivity of the subsurface medium at each monitoring location and each target frequency during the current monitoring period. This includes: (1) Preprocess the time-domain electromagnetic data, including quality screening of the time-domain electromagnetic data at each monitoring location to obtain effective time-domain data.

[0027] For the current monitoring period The time-domain electromagnetic data of each monitoring location are time-synchronized, de-DC and de-trend processed to suppress power frequency and its harmonics, pulse interference and local high-amplitude noise.

[0028] The time-domain electromagnetic data at each monitoring location undergoes quality screening to identify and remove instrument saturation, missing data, abrupt jumps, and abnormal waveforms. Valid data segments are selected based on the effective sampling ratio, signal stability, and transmitted current records to obtain valid time-domain data. The same or corresponding preprocessing parameters are used for different monitoring periods to avoid introducing additional variations due to differences in processing conditions.

[0029] (2) Extracting frequency domain amplitude, including performing time-frequency conversion or target frequency extraction operations on effective time domain data, and extracting frequency domain amplitude corresponding to multiple target frequencies.

[0030] Specifically, for the first Monitoring locations within each monitoring period Effective time-domain data Perform time-frequency conversion or target frequency extraction to obtain the frequency domain complex response: ; In the formula, This indicates Fourier transform, short-time Fourier transform, digital correlation detection, synchronization demodulation, or other target frequency extraction operations, with the time variable... ; For the first One target frequency, ; This represents the total number of target frequencies extracted during time-frequency conversion at each monitoring location.

[0031] By taking the modulo value (absolute value) ), to obtain the frequency domain amplitude corresponding to the target frequency: ; The initial screening phase focuses on using the frequency domain amplitude corresponding to the target frequency.

[0032] (3) If the data is collected by the controllable source electromagnetic observation method, after extracting the frequency domain amplitude, the frequency domain amplitude is normalized and corrected according to the measured transmitted current amplitude to obtain the normalized frequency domain amplitude.

[0033] For controlled-source electromagnetic observation, to eliminate the influence of transmitted current fluctuations on apparent resistivity calculation, after extracting the frequency domain amplitude, the frequency domain amplitude is also normalized and corrected based on the measured transmitted current amplitude to obtain the normalized frequency domain amplitude: ; In the formula, For the first Each monitoring period at frequency The amplitude of the transmitted current, To prevent positive adjustment quantities with a denominator of zero, a normalized frequency domain amplitude dataset is thus formed, with each monitoring location, monitoring period, and target frequency corresponding to a specific value.

[0034] (4) If the data is collected by the controllable source electromagnetic observation method, after extracting the frequency domain amplitude, the frequency domain amplitude is initially screened and candidate anomalies are marked.

[0035] First, select an initial monitoring period where the subsurface medium condition is relatively stable and the data quality is acceptable (e.g., the first 3 to 20 monitoring periods after monitoring starts). Alternatively, multiple monitoring periods with relatively stable subsurface medium condition and data quality can be selected, and monitoring locations can be established separately for each period. and each target frequency Corresponding background normalized frequency domain amplitude .

[0036] Then calculate the rate of change of the normalized frequency domain amplitude of the current monitoring period relative to the normalized frequency domain amplitude of the background: ; in, This is the positive adjustment factor used in the calculation of the rate of change of amplitude.

[0037] When the absolute value of the frequency domain amplitude change rate ( When the amplitude initial screening threshold is met and the corresponding signal-to-noise ratio and integrity satisfy the preset quality requirements, the current (number) threshold will be selected. (number) monitoring time periods, monitoring locations ( ) and target frequency ( The current monitoring period, monitoring location, and target frequency are marked as candidate anomalies and included in the candidate anomaly set. At this time, the current monitoring period, monitoring location, and target frequency are determined to pass the initial screening of frequency domain amplitude.

[0038] Otherwise, if If the current monitoring period, monitoring location, and target frequency are not selected, the frequency domain amplitude will be deemed not to have passed the initial screening.

[0039] In an optional embodiment, the amplitude initial screening threshold in the frequency domain amplitude initial screening. The parameters can be set based on the statistical results of normal fluctuations during the stable monitoring phase, for example, 2 to 3 times the standard deviation of normal fluctuations. The quality requirements for signal-to-noise ratio and data integrity can be preset according to the instrument performance and the on-site noise level. The signal-to-noise ratio threshold can be set to 30 dB to 60 dB, and the data validity threshold can be set to 80% to 90%.

[0040] The initial screening of frequency domain amplitude is used to mark candidate anomalies for the current monitoring period, monitoring location and target frequency, and to assist in constructing subsequent multi-dimensional anomaly judgment conditions. It is not used as the final confirmation basis for underground medium anomalies. For frequency domain data with qualified quality, the corresponding apparent resistivity is still calculated.

[0041] (5) Calculate the apparent resistivity, including calculating the apparent resistivity for each monitoring location based on the extracted frequency domain amplitude and the parameters of the electromagnetic observation device. At each target frequency apparent resistivity The apparent resistivity is calculated using the wide-area apparent resistivity formula or the Caniah apparent resistivity formula.

[0042] In one embodiment, the apparent resistivity is calculated using a wide-area apparent resistivity formula. For a horizontal electric dipole source and a ground-based receiving electric dipole MN, the wide-area apparent resistivity formula can be obtained based on the ratio of the received potential difference amplitude to the transmitted current amplitude: ; in, It is the first Monitoring time period at monitoring location receiving electric dipole The amplitude of the frequency domain potential difference between them It is the first One target frequency; For the first Monitoring period and target frequency The amplitude of the transmitting current; The electromagnetic observation device is at the monitoring location The device coefficient at the location, The electromagnetic response function for the entire region is given by the following formulas: ; ; In the formula: The imaginary unit; The wavenumber, determined by the target frequency and the apparent resistivity to be determined over the wide area, is solved using an iterative method. It is angular frequency. The magnetic permeability of the underground medium. The dielectric constant of the underground medium; From the center of the transmitting electric dipole to the monitoring position The distance from the center of the receiving electric dipole; For the direction of the transmitting electric dipole and the monitoring position The included angle between the corresponding transmitting and receiving radial directions; This is merely a function marker used to identify whether the defined device coefficients or response functions are for a preset observation direction (e.g., along a Cartesian coordinate system). In terms of electromagnetic field components (direction) observation.

[0043] Since the electromagnetic response function of the entire region includes the apparent resistivity to be determined over the wide area, the apparent resistivity calculated by the formula for wide area apparent resistivity can be obtained by using iterative methods, inverse interpolation methods, or a pre-established response function lookup table. .

[0044] In another embodiment, when the monitoring device simultaneously acquires mutually orthogonal electric and magnetic field components, and the observation conditions satisfy the far-field or wave region approximation, the apparent resistivity can also be calculated using the following Carnia apparent resistivity formula: ; in, The vacuum permeability; To observe along a predetermined direction (e.g., in a rectangular coordinate system) The electric field amplitude (in the axial direction); The direction orthogonal to the preset observation direction (e.g., in a Cartesian coordinate system) The magnitude of the magnetic field strength along the axial direction. When the magnetic sensor outputs a magnetic induction intensity of... At that time, it can be based on After conversion, we can obtain .

[0045] For continuous monitoring data from the same monitoring location, it is preferable to uniformly adopt the wide-area apparent resistivity formula or the Cagnia apparent resistivity formula, and maintain consistency in device parameters, numerical solution methods, and frequency sets to ensure the comparability of apparent resistivity results at different time periods. When using the wide-area apparent resistivity formula, it can cover the near-field, transitional, and far-field regions; when using the Cagnia apparent resistivity formula, the corresponding applicable conditions for the far-field region should be met.

[0046] The above steps achieve a complete conversion from time-domain electromagnetic data to apparent resistivity, providing accurate physical property parameters for subsequent background benchmark construction, rate of change calculation, and multi-dimensional statistics. Each monitoring period, each monitoring location, and each target frequency is calculated independently, ensuring data integrity and independence.

[0047] Step 120, constructing and maintaining a background reference for apparent resistivity, includes: comparing the apparent resistivity of the current monitoring period with the background reference to obtain the relative rate of change of apparent resistivity; and updating the background reference using the apparent resistivity that meets the preset background update conditions during the current monitoring period. Specifically, this includes: Step 121: Construct a background reference for apparent resistivity.

[0048] An initial background baseline was constructed from apparent resistivity data during the initial stable monitoring period. .

[0049] The background benchmark can be composed of apparent resistivity data from a single stable period, or it can be composed of the average, median, or robust statistical values ​​of apparent resistivity from multiple stable periods. The corresponding standard deviation, absolute median deviation, or historical fluctuation range can be statistically analyzed to provide a basis for setting the apparent resistivity anomaly threshold.

[0050] In one embodiment, the initial stable monitoring period refers to the initial monitoring period when the underground medium is relatively stable (e.g., the first 3 to 5 monitoring periods after monitoring starts) or the historical stable monitoring period (e.g., multiple monitoring periods when the underground medium is relatively stable and the data quality is relatively stable).

[0051] The above process determines the initial value of the background baseline and provides a benchmark reference for subsequent maintenance processes by constructing the background baseline.

[0052] Step 122: Calculate the relative rate of change of apparent resistivity.

[0053] As monitoring continues, the current number will be... Apparent resistivity for each monitoring period Compared with the background benchmark before the judgment of this period, such as the first time before the update. Background benchmark for each monitoring period By comparing the corresponding monitoring locations and target frequencies, the relative rate of change of apparent resistivity is calculated. ; in, This is the positive adjustment factor used in the calculation of the apparent resistivity change rate.

[0054] Step 123: Update the background reference by using the apparent resistivity that meets the preset background update conditions during the current monitoring period, thereby maintaining the background reference.

[0055] Typically, the preset background update conditions include: the absolute value of the relative rate of change of apparent resistivity. Less than the apparent resistivity change threshold .

[0056] When the time-domain electromagnetic data is acquired using a controllable source electromagnetic observation method, the preset background update condition further includes: the current monitoring period, monitoring location, and target frequency are not marked as candidate anomalies by the initial screening of frequency domain amplitude.

[0057] It should be noted that the candidate anomaly markers are generated by the initial screening of frequency domain amplitude in step 110. When the absolute value of the frequency domain amplitude change rate exceeds the initial screening threshold and the quality meets the requirements, the monitoring period, monitoring location and target frequency are marked as candidate anomalies and included in the candidate anomaly set.

[0058] "Not triggered candidate anomaly labeling" refers to a state where the data for the current monitoring period (corresponding to the monitoring period, monitoring location, and target frequency) has not been labeled as a candidate anomaly. In this state, the data can be further evaluated based on the magnitude of the relative rate of change of apparent resistivity to determine whether it participates in the background benchmark update, specifically including: When the preset background update condition is met, the background update coefficient is applied. The background benchmark is then updated using a weighted method: ; in, For the current number after the update Background benchmark for each monitoring period For the first time before the update Background benchmark for each monitoring period The apparent resistivity for the current monitoring period, and the background update coefficient. satisfy ; Otherwise, when the preset background update condition is not met, the background reference remains unchanged: .

[0059] Background update coefficient This is used to control the degree of influence of current monitoring data on the background benchmark update results. A larger background update coefficient results in a faster tracking speed of the background benchmark to recent monitoring data; a smaller background update coefficient results in a smoother change in the background benchmark. The background update coefficient can be preset based on the normal fluctuation level of apparent resistivity during the stable monitoring phase and the required background tracking speed, and remains unchanged within the same monitoring phase, or it can be adjusted according to the fluctuation characteristics of the stable monitoring data. In a preferred embodiment, It takes values ​​between 0.02 and 0.8.

[0060] Furthermore, if a candidate anomaly flag has been triggered, the background baseline remains unchanged.

[0061] The background baseline update operation adopts a recursive approach based on each monitoring period. The background baseline for each monitoring period is established solely based on data from previous historical monitoring periods that did not trigger candidate anomaly markers and whose relative rate of change of apparent resistivity was normal. No additional data is introduced for the first monitoring period. Data from the monitoring period and subsequent monitoring periods were used. Therefore, the background baseline before the update was used when calculating the relative change rate of apparent resistivity, ensuring the independence of anomaly detection, avoiding circular dependencies where the current period's data participates in the detection, and ensuring the purity of the background baseline.

[0062] By constructing and maintaining a background benchmark, the differences between the multi-frequency apparent resistivity of each monitoring period and the background benchmark are calculated. This highlights the changes in the underground medium relative to its steady state, weakens the influence of static geological differences and environmental background on anomaly identification, and makes the monitoring results focus more on the analysis of changes. Only data that have not triggered candidate anomaly markers and whose rate of change is within the normal fluctuation range are used to participate in the background benchmark update. This effectively prevents anomaly signals from being slowly absorbed into the background benchmark, avoiding the problem of missed reports caused by "anomaly drift" in long-term monitoring. At the same time, it tracks normal slow drifts such as seasonal water level changes, reducing the false alarm rate.

[0063] Step 130: Calculate statistical characteristics based on the comparison results between the relative rate of change of apparent resistivity and the threshold of change of apparent resistivity, including the number of consecutive time periods exceeding the limit, the number of frequencies of anomalies in the same direction, and the number of monitoring locations of anomalies in the same direction.

[0064] (1) Calculate the number of consecutive periods of exceeding limits (statistical characteristics in the time dimension) In one embodiment, the same monitoring location can be statistically analyzed over multiple recent monitoring periods. and the same target frequency Under these conditions, the absolute value of the relative rate of change of apparent resistivity is continuously greater than or equal to the threshold value of apparent resistivity change. The number of monitoring periods (i.e., exceeding limits) is used to calculate the number of consecutive periods exceeding limits: ; in, This is an indicator function; it takes the value 1 if the condition inside the parentheses is true, and 0 otherwise. For variables summed over historical monitoring periods, The number of retrospective monitoring periods involved in ongoing assessments; For the first Monitoring time period at monitoring location and target frequency The relative rate of change of apparent resistivity.

[0065] The above-mentioned window accumulation method for calculating the number of consecutive over-limit periods is simple to implement and can reflect the overall frequency of apparent resistivity anomalies within a certain time range.

[0066] Preferably, in order to further reduce false triggering caused by sporadic, time-dispersed interference, the number of consecutive over-limit periods is defined as continuous, i.e., based on the current period... Starting from a given monitoring period and moving backward (through historical monitoring periods), the number of monitoring periods in which the absolute value of the relative rate of change of apparent resistivity is continuously greater than or equal to the apparent resistivity change threshold (continuous exceedance) is counted; when a monitoring period does not meet the exceedance condition ( When the count reaches zero, the count is reset and restarted from the next monitoring period.

[0067] The above definition of continuity can be expressed in the following recursive form: ; in, For monitoring location and target frequency The number of consecutive periods exceeding the limit in the previous monitoring period is counted below.

[0068] The apparent resistivity changes caused by anomalies in the underground medium are usually continuous and persistent over time, remaining above the limit for multiple consecutive monitoring periods. In contrast, random disturbances, instrument transients, and occasional environmental changes often manifest as isolated or scattered short-term anomalies. Window accumulation methods may aggregate multiple discontinuous occasional disturbances within a window, leading to a higher number of consecutive periods exceeding the limit. Reaching a preset threshold triggers inversion, increasing the risk of false triggering. The continuity definition requires that exceeding the limit must occur continuously in time, which can more accurately reflect the physical meaning of time persistence and significantly reduce the impact of single-period or scattered interference. This is consistent with the technical purpose of using time persistence in multi-dimensional joint judgment to reduce false triggering by occasional interference.

[0069] In one embodiment, the number of backtracking monitoring periods Set the threshold to 5-10 for the number of consecutive periods exceeding the limit. The value is 3 to 5. For example, if the exceedance status of the most recent 5 monitoring periods is: Yes, Yes, Yes, No, Yes, then the number of consecutive exceedance periods in the current period is 1 (counting backwards from the current monitoring period to the monitoring period without exceedance), not 4 (the total number of exceedances). Only the number of consecutive exceedance periods is considered. Only then will the "time continuity" condition be met.

[0070] Using strict continuous counting rather than total count ensures the persistence of anomalies over time. While occasional noise (such as transient electromagnetic interference) in a single monitoring period may cause limits to be exceeded in that period, it is not continuous and will not meet the requirement of a number of consecutive periods of exceeding limits, thus being effectively excluded.

[0071] (2) Calculate the number of anomalies in the same direction (statistical characteristics of the frequency dimension) At the same monitoring location and the same monitoring period Within this context, let the total number of target frequencies participating in anomaly detection be... Count the number of target frequencies that exceed the positive and negative limits of apparent resistivity respectively: ; ; Take the larger of the two values ​​as the number of anomalies in the same direction: ; This statistical characteristic is used to determine the consistency of apparent resistivity change direction across multiple target frequencies. The actual electrical changes in the underground medium typically exhibit a relatively consistent response across multiple adjacent or related target frequencies, while the stability of single-frequency sporadic anomalies is relatively low. Therefore, by requiring multiple target frequencies to exceed limits in the same direction for frequency-dimensional consistency verification, false triggering caused by single-frequency sporadic anomalies can be reduced.

[0072] (3) Calculate the number of anomaly monitoring locations in the same direction (statistical characteristics of spatial dimension) For survey lines or areas with multiple continuous monitoring locations, in the area containing the monitoring locations and the preset range set of adjacent monitoring positions. All monitoring locations within Statistical analysis of the frequency of the same target The number of monitoring locations where the relative rate of change of the apparent resistivity is in the same direction and exceeds the limit is considered as the number of anomaly monitoring locations in the same direction: ; In the above formula, through Characterizing the consistency of the direction of change of the relative rate of change of apparent resistivity, that is, the relative rate of change of apparent resistivity at the current monitoring location. , and the preset range set Internal monitoring location The relative rate of change of apparent resistivity The signs are the same (both positive or both negative). The direction of change of the relative rate of change of apparent resistivity includes: a positive value indicates an increase in apparent resistivity relative to the background reference; a negative value indicates a decrease in apparent resistivity relative to the background reference. This is achieved through the indicator function. As a product term, the statistics satisfy the excess condition. The number of monitoring locations.

[0073] Genuine water-bearing anomalies typically have a certain spatial distribution range (such as fracture zones or aquifers), and will produce similar resistivity change responses at adjacent monitoring locations; while isolated point noise (such as spurious anomalies caused by poor contact of a certain electric dipole) usually does not have spatial continuity. By requiring the number of anomaly monitoring locations in the same direction to reach a preset threshold, false triggering caused by isolated point noise is further eliminated.

[0074] For scenarios where only a single monitoring location is set, no calculation is performed. The system uses only the apparent resistivity change amplitude, the number of consecutive over-limit periods, and the number of unidirectional abnormal frequencies for judgment, ensuring the adaptability of the solution in single-point monitoring scenarios.

[0075] This implementation method uses a joint judgment basis based on the apparent resistivity change amplitude, time duration, multi-frequency correlation, and spatial continuity of the monitoring location to reduce false triggering caused by occasional interference from a single time period, single frequency, or single measuring point. Cross-validation across these three dimensions significantly improves the reliability of early warning.

[0076] Step 140: Construct multi-dimensional anomaly determination conditions using the relative change rate of apparent resistivity and the statistical characteristics; when the multi-dimensional anomaly determination conditions are met at least at one target frequency, generate an inversion trigger identifier for the corresponding monitoring location; otherwise, return to step 110 to obtain the time-domain electromagnetic data for the next monitoring period and calculate the apparent resistivity.

[0077] When using a controllable source electromagnetic observation method, the multi-dimensional anomaly determination condition is a logical AND of the following necessary conditions: Prerequisite 1: Initial screening based on frequency domain amplitude, current monitoring period Monitoring location and target frequency It has been determined to have passed the initial screening of frequency domain amplitude (i.e., it has been marked as a candidate anomaly). This condition utilizes the results of the initial screening of frequency domain amplitude as the first entry threshold for the inversion trigger determination.

[0078] Necessary condition two: The apparent resistivity change exceeds the limit, that is, the absolute value of the relative rate of change of apparent resistivity is greater than or equal to the apparent resistivity change threshold. .

[0079] Necessary condition three: The duration of the violation must be met, meaning the number of consecutive periods exceeding the limit must be greater than or equal to the preset threshold for the number of consecutive periods exceeding the limit. .

[0080] Fourth necessary condition: Frequency unidirectionality or spatial continuity meets the requirements, that is, the following conditions are satisfied: For scenarios with multiple monitoring locations, the number of anomalies in the same direction is greater than or equal to a preset threshold for the number of anomalies in the same direction, or the number of anomaly monitoring locations in the same direction is greater than or equal to a preset threshold for the number of anomaly monitoring locations in the same direction: ,or ; For scenarios where only a single monitoring location is set, the number of anomalies in the same direction is greater than or equal to the preset threshold for the number of anomalies in the same direction: ; in, It is a preset threshold for the number of abnormal frequencies in the same direction.

[0081] When all four necessary conditions for multi-dimensional anomaly determination are met simultaneously at at least one target frequency, an inversion trigger identifier for the corresponding monitoring location is generated, denoted as . Otherwise, record as .

[0082] when If no inversion trigger flag is generated for the corresponding monitoring location, return to step 110 and proceed to the next monitoring period (the next monitoring period). Data collection and processing (for each monitoring period).

[0083] when At that time, it will directly proceed to the selective inversion process in the subsequent step 150.

[0084] For the implementation method using natural source electromagnetic observation, the multi-dimensional anomaly determination conditions do not include necessary condition one (frequency domain amplitude initial screening), but only necessary condition two, necessary condition three, and necessary condition four.

[0085] The above implementation introduces a threshold triggering mechanism based on anomaly identification. Through multi-dimensional joint judgment, an inversion trigger flag for the corresponding monitoring location is generated and inversion imaging is initiated only when a significant anomaly is confirmed. If no trigger is triggered, only low-cost data acquisition and statistical judgment are performed, and the next monitoring period continues. Resistivity inversion is repositioned from a conventional processing step to a precise location method after anomaly confirmation, avoiding the ingestion of a large amount of data from anomaly-free periods into the high-cost inversion process, significantly reducing computational resource consumption, and meeting the real-time requirements of continuous monitoring.

[0086] Step 150: Based on the inversion trigger identifier, limit the data range for inversion; extract the effective electromagnetic data within the data range for resistivity inversion imaging to obtain the inversion resistivity model for the current monitoring period.

[0087] (1) Responding to the inversion trigger flag Limit the range of data for inversion: Time range: from the start of the continuous over-limit period to the current period. The monitoring period is defined as the range of monitoring periods to be inverted; among them, the continuous exceedance start period refers to the first monitoring period in which the absolute value of the relative change rate of apparent resistivity is continuously greater than or equal to the apparent resistivity change threshold. Frequency range: The target frequency that meets the abnormal conditions and its adjacent target frequencies that have qualified data quality are taken as the target frequency range to be inverted; among them, the target frequency that meets the abnormal conditions is the target frequency corresponding to the multi-dimensional abnormal judgment conditions. Spatial range: The area between the first and last monitoring positions among multiple consecutive anomaly monitoring positions is taken as the range of the position to be inverted, and the range is extended to both sides by a preset distance as needed; wherein, the anomaly monitoring position is the monitoring position that generates the inversion trigger identifier.

[0088] Only valid electromagnetic data within the aforementioned time, frequency, and location ranges are extracted for inversion processing, avoiding indiscriminate inversion of all continuous monitoring data.

[0089] (2) Perform resistivity inversion imaging: Valid electromagnetic data within the data range are extracted for resistivity inversion imaging to obtain the inversion resistivity model for the current monitoring period.

[0090] In the inversion process, the background resistivity model (corresponding to the model parameter vector denoted as) is used. ) and the resistivity model retrieved from the previous monitoring period (the corresponding model parameter vector is denoted as ) () is used as a reference model for constraint inversion.

[0091] make This is the model parameter vector for the resistivity model retrieved during the current monitoring period. The inversion dimension is determined based on the layout of the monitoring lines: when monitoring is conducted using a single line, a two-dimensional inversion is used, and the retrieved resistivity model corresponds to the two-dimensional resistivity distribution of the profile where the line is located (i.e., along the line direction and the vertical depth direction); when multiple parallel lines, intersecting lines, or grid-like lines are laid out, a three-dimensional inversion can be used, and the retrieved resistivity model corresponds to the three-dimensional resistivity distribution of the monitoring area. Accordingly, the model parameter vector... and spatial smoothing constraint matrix The dimensions are adjusted based on a two-dimensional or three-dimensional discrete grid.

[0092] Specifically, the inversion objective function of the inversion resistivity model is: ; in: Data fitting term Ensuring that the inversion results can fit the effective electromagnetic data collected during the current monitoring period is the basic constraint for the inversion. Spatial smoothing constraint This ensures that the inversion results are spatially continuous and smooth, avoids excessive resistivity abrupt changes between adjacent grid cells, and conforms to the spatial continuity law of geological bodies; Background Difference Constraints This constrains the current inversion model from deviating excessively from the background resistivity model, in order to prevent non-physical solutions caused by multiple solutions in the inversion, while ensuring the physical rationality of the inversion results. Time constraints To constrain the current inversion model from deviating excessively from the resistivity model of the previous monitoring period, and to suppress the influence of random noise on the inversion results by utilizing the physical law that geological changes are usually gradual. This represents the square value of the L2 norm. For the currently triggered data, Forward model response; For the data weight matrix, The space smoothing constraint matrix is, The background difference constraint matrix, The time constraint matrix, , and These are the corresponding weighting coefficients.

[0093] In one embodiment, a logarithmic resistivity model is used to provide the model parameter vector. , and ,in: , , ;here It is the resistivity distribution under background conditions. and They are the first The and the first Resistivity distribution during each monitoring period.

[0094] In one embodiment, the objective function is iteratively solved using the Gauss-Newton method or a quasi-Gauss-Newton method until the data fitting residual drops below a preset threshold or the maximum number of iterations is reached, thus obtaining the inversion resistivity model for the current monitoring period (providing the corresponding model parameter vector). And resistivity distribution).

[0095] Figure 2 This is a schematic diagram illustrating the inversion results for different monitoring periods in one embodiment of the present invention. Figure 2 As shown, this paper presents the resistivity inversion results of monitoring line L01 during different monitoring periods in a certain monitoring area during tunnel shield construction. The L01 monitoring line is 40 m long, with a monitoring point spacing of 1 m, and a total of 41 monitoring points were set up. Thirty-six target frequencies were selected for inversion analysis, ranging from 5.12 kHz to 65 kHz. Figure 2 (a) through (e) correspond to the monitoring results at 01:00 on July 11 (monitoring period number: Line01_0711_01h), 03:00 on July 13 (monitoring period number: Line01_0713_03h), 20:00 on July 13 (monitoring period number: Line01_0713_20h), 21:00 on July 14 (monitoring period number: Line01_0714_21h), and 14:00 on July 15 (monitoring period number: Line01_0715_14h), respectively. The figure uses... Characterizing the resistivity distribution of underground media, where, Regions with larger values ​​correspond to regions with relatively high resistivity. The smaller the value, the lower the resistivity of the region.

[0096] Depend on Figure 2 As shown in (a), in the initial monitoring stage, the area approximately 30m to 40m to the right of the survey line, with an elevation of approximately -28m to -10m, exhibits relatively high resistivity. Combined with existing engineering geological data, this can be interpreted as an unfilled or low-water-content karst cave, a fractured area, or a relatively dry rock mass. As the tunnel boring machine construction continues... Figure 2The underground resistivity distribution in (b) to (e) shows significant changes, and the spatial morphology and influence range of the low resistivity anomaly area gradually adjust. During this monitoring stage, the tunnel boring machine is located near 0 m to the left of the survey line. Its construction disturbance may change the connectivity of the surrounding rock fissures and the groundwater migration conditions, causing groundwater to migrate to adjacent areas along existing fissures, karst channels, or seepage channels formed by the disturbance, and resulting in a decrease in the resistivity of the relevant areas.

[0097] Figure 2 The inversion results from different monitoring periods exhibit temporally continuous changes in electrical structure, indicating that the underground medium does not remain static under construction disturbance conditions, but rather undergoes a dynamic electrical response that can be identified by continuous electromagnetic observation. Combining the apparent resistivity variation characteristics of each monitoring period, the persistence of the abnormal changes can be further determined. When the anomaly reaches the preset triggering conditions, the location and impact range of the anomaly area can be spatially determined using the inversion results, thus providing a basis for monitoring and early warning of abnormal evolution of the underground medium.

[0098] This implementation introduces a threshold triggering mechanism based on anomaly identification. When the apparent resistivity change meets preset conditions, resistivity inversion imaging is initiated, improving the spatial resolution capability of anomalies while ensuring continuous monitoring efficiency. Responding to the trigger flag, resistivity inversion is performed only on data within the trigger time, trigger location, and adjacent frequency bands, achieving "precise positioning on demand."

[0099] In other embodiments, the background resistivity model can be used as the initial model, and the current data can be inverted before model difference analysis is performed.

[0100] Step 160: Compare the inverted resistivity model of the current monitoring period with the preset background resistivity model to obtain the difference area; determine the spatial location and range of the anomaly in the underground medium based on the difference area, and generate or update the early warning information.

[0101] (1) Calculate the relative difference between the current inverted resistivity model and the background resistivity model: ; in, It is the position of the current inverted resistivity model relative to the background resistivity model in three-dimensional space. The relative change in resistivity at a given point is used to identify and locate anomalies in the underground medium; The resistivity values ​​are retrieved from the resistivity model for the current monitoring period. The resistivity values ​​are for the background resistivity model. This is the positive adjustment factor used in calculating the rate of change in the inversion model.

[0102] This section uses As a unified representation of spatial location: when using two-dimensional inversion, this expression actually degenerates into two-dimensional coordinates within the survey profile. ,in The horizontal distance along the survey line direction. For burial depth or elevation, it is used to characterize the relative change in resistivity within the profile and the location of anomalies; when using three-dimensional inversion, This indicates that the monitoring area is on the horizontal plane ( Direction) and vertical depth ( The three-dimensional coordinates (direction) are used to characterize the relative change in resistivity and the spatial distribution of anomalies in three-dimensional space. In practical applications, the appropriate inversion dimension is selected according to the survey line layout method, and areas of difference are marked to meet the requirements. The grid cells are combined and adjacent connected regions are merged to determine the spatial location and extent of the anomaly, in order to accurately identify the anomaly. This is the threshold for changes in the inversion model.

[0103] Relative change in resistivity A negative value indicates a low resistivity change relative to the background resistivity, while a positive value indicates a high resistivity change relative to the background resistivity. The absolute value reflects the magnitude of the change. The specific geological causes of low or high resistivity changes need to be determined comprehensively by considering the monitoring object, geological conditions, and other monitoring information. Under specific conditions, low resistivity changes may be related to increased water content or the development of water-filled fissures, while high resistivity changes may be related to decreased water content, drying shrinkage, or air-filled cavities.

[0104] Will satisfy The grid cells are marked as candidate anomalous grid cells; then all adjacent candidate anomalous grid cells are merged into a continuous spatially connected region to form a difference region, and a continuous anomalous body is determined based on the difference region.

[0105] The horizontal position, top interface burial depth, bottom interface burial depth, center depth, vertical thickness, and lateral extension range of the anomaly are determined based on the continuous anomaly. The symbol can identify the current resistivity change or the trend of low resistivity change relative to the background resistivity; the specific geological causes of low resistivity change or high resistivity change need to be determined by combining the monitoring object, geological conditions and other monitoring information. Under specific conditions, low resistivity change may be related to increased water content or the development of water-filled fissures, while high resistivity change may be related to decreased water content, drying shrinkage or air-filled cavities.

[0106] (2) Tracking dynamic evolution trends: based on the current inversion model Inversion model with the previous monitoring period The differences determine the dynamic evolution trend of the anomalous body: whether it is in an expanding, contracting, or migrating state.

[0107] Specifically, the spatial extent of a continuous anomaly is compared across different time periods. If the extent of the anomaly expands, it is classified as "expansion"; if the extent of the anomaly shrinks, it is classified as "contraction"; and if the center of the anomaly shifts significantly, it is classified as "migration".

[0108] (3) Determine the warning level and generate or update warning information: Comprehensive abnormal amplitude ( The anomaly status and warning level of the underground medium are determined by the following factors: size, duration (number of consecutive periods exceeding the limit), burial depth (depth location of the anomaly), scope of influence (spatial range of the anomaly, such as the lateral / vertical extension size or cross-sectional area of ​​the anomaly in the profile during two-dimensional inversion, and the spatial volume of the anomaly during three-dimensional inversion), and development trend (expansion, contraction, or migration).

[0109] The warning level is divided into three levels based on the scale and degree of danger of the anomaly: Level 1 warning: The anomaly is small in scale and spreads slowly, indicating that attention should be paid to the abnormal area and it is recommended to increase the frequency of monitoring; Level 2 Warning: The abnormality is of moderate scale and shows a continuous expansion trend, prompting the initiation of a special investigation and recommending the development of an emergency plan; Level 3 warning: The anomaly is widespread, spreading rapidly, and poses a safety risk, prompting the immediate activation of emergency response procedures.

[0110] The system ultimately generates monitoring and early warning information that includes the location of the anomaly (horizontal coordinates, burial depth), the spatial range of the anomaly, the magnitude of the anomaly, the dynamic evolution trend, and the warning level, and presents it to monitoring personnel in a visual manner.

[0111] In step 160 above, the current inverted resistivity model is compared with the background resistivity model. Through the difference analysis between the current inverted model and the background inverted model, the spatial location, burial depth, and influence range of the underground anomaly are located. Simultaneously, the current inverted model is compared with both the background model and the model from the previous time period. This comparison with the background model pinpoints the spatial location and range of the anomaly, while the comparison with the model from the previous time period tracks the expansion, contraction, or migration trend of the anomaly. This upgrades from static location to dynamic tracking, providing more complete temporal evolution information for disaster early warning. The background baseline, anomaly location, influence range, and warning level are dynamically updated with newly added monitoring data, forming a closed-loop processing procedure for dynamic monitoring and early warning of underground media.

[0112] In one embodiment, the present invention also provides a dynamic monitoring and early warning system for underground media based on electromagnetic signals, the system comprising the following modules: The first module is used to continuously collect time-domain electromagnetic data of underground media under electromagnetic fields and divide it into different monitoring periods according to a preset time window; based on the time-domain electromagnetic data of the current monitoring period, it calculates the apparent resistivity of the underground media at each monitoring location and each target frequency in the current monitoring period. The second module is used to construct and maintain a background reference for apparent resistivity, including: comparing the apparent resistivity of the current monitoring period with the background reference to obtain the relative rate of change of apparent resistivity; and updating the background reference using the apparent resistivity that meets the preset background update conditions in the current monitoring period. The third module is used to calculate statistical characteristics based on the comparison results between the relative rate of change of apparent resistivity and the threshold of change of apparent resistivity, including the number of consecutive time periods exceeding the limit, the number of frequencies of anomalies in the same direction, and the number of monitoring locations of anomalies in the same direction. The fourth module is used to construct multi-dimensional anomaly judgment conditions using the relative change rate of apparent resistivity and the statistical characteristics; when the multi-dimensional anomaly judgment conditions are met at least at one target frequency, an inversion trigger identifier for the corresponding monitoring location is generated; otherwise, the module returns to the first module to obtain the time-domain electromagnetic data for the next monitoring period and calculate the apparent resistivity. The fifth module is used to limit the data range for inversion based on the inversion trigger identifier; extract the effective electromagnetic data within the data range for resistivity inversion imaging, and obtain the inversion resistivity model for the current monitoring period; The sixth module is used to compare the inverted resistivity model of the current monitoring period with the preset background resistivity model to obtain the difference area; based on the difference area, the spatial location and range of the anomaly in the underground medium are determined, and early warning information is generated or updated.

[0113] The specific workflow of each module corresponds to the steps in the aforementioned method embodiments, and will not be repeated here.

[0114] On the other hand, in one embodiment, the present invention provides a computer device including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the underground medium dynamic monitoring and early warning method based on electromagnetic signals provided in any of the above embodiments. The computer device may be a server. The computer device includes a processor, a memory, a network interface, and a database connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device stores sample data. The network interface of the computer device is used for communication with an external PC via a network connection.

[0115] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of the method for dynamic monitoring and early warning of underground media based on electromagnetic signals provided in any of the above embodiments.

[0116] 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 hardware related to computer program instructions. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it 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 can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various 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.

[0117] Matters not covered in this invention are common knowledge.

[0118] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0119] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application.

[0120] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for dynamic monitoring and early warning of underground media based on electromagnetic signals, characterized in that, include: Step 110: Continuously collect time-domain electromagnetic data of the underground medium under electromagnetic field, and divide it into different monitoring periods according to a preset time window; Based on the time-domain electromagnetic data of the current monitoring period, calculate the apparent resistivity of the underground medium at each monitoring location and each target frequency during the current monitoring period; Step 120, constructing and maintaining a background reference for apparent resistivity, including: comparing the apparent resistivity of the current monitoring period with the background reference to obtain the relative rate of change of apparent resistivity; and updating the background reference using the apparent resistivity that meets the preset background update conditions in the current monitoring period. Step 130: Based on the comparison results between the relative rate of change of apparent resistivity and the threshold of change of apparent resistivity, calculate the statistical characteristics, including the number of consecutive time periods exceeding the limit, the number of frequencies of anomalies in the same direction, and the number of monitoring locations of anomalies in the same direction. Step 140: Construct multi-dimensional anomaly determination conditions using the relative rate of change of apparent resistivity and the statistical characteristics; when the multi-dimensional anomaly determination conditions are met at least at one target frequency, generate an inversion trigger identifier for the corresponding monitoring location; otherwise, return to step 110 to obtain the time-domain electromagnetic data for the next monitoring period and calculate the apparent resistivity. Step 150: Based on the inversion trigger identifier, limit the data range for inversion; extract the effective electromagnetic data within the data range for resistivity inversion imaging to obtain the inversion resistivity model for the current monitoring period; Step 160: Compare the inverted resistivity model of the current monitoring period with the preset background resistivity model to obtain the difference area; determine the spatial location and range of the anomaly in the underground medium based on the difference area, and generate or update the early warning information.

2. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 1, characterized in that, In step 110, calculating the apparent resistivity of the underground medium at different monitoring locations and different target frequencies based on the time-domain electromagnetic data of the current monitoring period includes: The temporal electromagnetic data at each monitoring location are screened for quality to obtain valid temporal data; Perform time-frequency conversion or target frequency extraction operations on the effective time-domain data to extract the frequency domain amplitude corresponding to multiple target frequencies; Based on the frequency domain amplitude and combined with the parameters of the electromagnetic observation device, the apparent resistivity of each monitoring location at each target frequency is calculated; the parameters of the observation device include at least one of the following: transmitting current, field source length, transmitting and receiving distance, receiving electric dipole length, and observation direction.

3. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 2, characterized in that, In step 110, the apparent resistivity is calculated using the wide-area apparent resistivity formula or the Kania apparent resistivity formula. The formula for the wide-area apparent resistivity is: ； in, For the current number Apparent resistivity for each monitoring period It is the first Monitoring time period at monitoring location receiving electric dipole The amplitude of the frequency domain potential difference between them It is the first One target frequency; For the first Monitoring period and target frequency The amplitude of the transmitting current; The electromagnetic observation device is at the monitoring location The device coefficient at the location, The electromagnetic response function for the entire region is given by the following formulas: ； ； In the formula: The imaginary unit; The wavenumber, determined by the target frequency and the apparent resistivity to be determined over the wide area, is solved using an iterative method. It is angular frequency. The magnetic permeability of the underground medium. The dielectric constant of the underground medium; From the center of the transmitting electric dipole to the monitoring position The distance from the center of the receiving electric dipole; For the direction of the transmitting electric dipole and the monitoring position The included angle between the corresponding transmitting and receiving radial directions; Function marker only; The length of the transmitting electric dipole; The length of the receiving electric dipole; The apparent resistivity formula of Cania is as follows: ； in, The vacuum permeability; The electric field amplitude along the preset observation direction; The magnitude of the magnetic field strength in the direction orthogonal to the preset observation direction.

4. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 3, characterized in that, In step 110: The time-domain electromagnetic data were acquired using a controlled-source electromagnetic observation method. After extracting the frequency domain amplitude, the following steps are also included: The frequency domain amplitude is normalized and corrected based on the measured transmitted current amplitude to obtain the normalized frequency domain amplitude. Perform initial screening of frequency domain amplitude: Establish background normalized frequency domain amplitudes corresponding to monitoring locations and target frequencies; Calculate the rate of change of the normalized frequency domain amplitude of the current monitoring period relative to the normalized frequency domain amplitude of the background; When the absolute value of the frequency domain amplitude change rate is greater than or equal to the initial amplitude screening threshold and the corresponding signal-to-noise ratio and integrity meet the preset quality requirements, the current monitoring period, monitoring location and target frequency are marked as candidate anomalies, included in the candidate anomaly set, and determined to have passed the initial frequency domain amplitude screening; otherwise, they are determined to have failed the initial frequency domain amplitude screening. The initial screening of frequency domain amplitude is also used to assist in constructing the multi-dimensional anomaly determination conditions.

5. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 4, characterized in that, In step 120, constructing the background benchmark includes: Select an initial stable monitoring period or multiple historical stable monitoring periods as the initial stable monitoring period; An initial background baseline is constructed from the apparent resistivity data during the initial stable monitoring period; The background baseline is formed by using the apparent resistivity of a single stable monitoring period, or by using the average, median, or robust statistical value of the apparent resistivity of multiple stable monitoring periods.

6. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 4, characterized in that, In step 120, the preset background update conditions include: the absolute value of the relative change rate of apparent resistivity is less than the apparent resistivity change threshold; when the time-domain electromagnetic data is collected using a controllable source electromagnetic observation method, the preset background update conditions also include: the current monitoring period, monitoring location and target frequency are not marked as candidate anomalies by the initial screening of frequency domain amplitude. The step of updating the background reference using the apparent resistivity that meets the preset background update conditions during the current monitoring period includes: When the preset background update condition is met, the background update coefficient is applied. The background benchmark is then updated using a weighted method: ； in, For the current number after the update Background benchmark for each monitoring period For the first time before the update Background benchmark for each monitoring period For the current number Apparent resistivity for each monitoring period, background update coefficient satisfy ; Otherwise, the background reference remains unchanged: 。 7. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 6, characterized in that, In step 130: The number of consecutive out-of-limit periods is calculated by backtracking from the current monitoring period to count the number of monitoring periods in which the absolute value of the relative change rate of apparent resistivity is continuously greater than or equal to the apparent resistivity change threshold at the same monitoring location and target frequency. When the first monitoring period that does not meet the out-of-limit condition is encountered during the backtracking process, the count is reset to zero and the count restarts from the next monitoring period. The out-of-limit condition refers to the absolute value of the relative change rate of apparent resistivity being greater than or equal to the apparent resistivity change threshold. The number of unidirectional abnormal frequencies is taken as the maximum value between the number of positively exceeding target frequencies and the number of negatively exceeding target frequencies. ； in, It is the monitoring location and monitoring period The number of target frequencies exceeding the internal positive limit. It is the monitoring location and monitoring period The number of target frequencies exceeding the negative limit is given by the following formulas: ； ； In the formula, For indicator functions, It is the total number of target frequencies participating in anomaly detection. It is the first Monitoring time period at monitoring location and target frequency The relative rate of change of apparent resistivity under these conditions It is the threshold for apparent resistivity change; The number of unidirectional anomaly monitoring locations is given by the following formula: ； in, It includes monitoring locations and the preset range set of adjacent monitoring positions, It is the first Monitoring time period at monitoring location and target frequency The relative rate of change of apparent resistivity under these conditions .

8. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 7, characterized in that, In step 140, the multi-dimensional anomaly determination conditions include the following necessary conditions: Necessary condition 1: The current monitoring period, monitoring location, and target frequency are determined to pass the initial screening of frequency domain amplitude; Second necessary condition: The absolute value of the relative rate of change of apparent resistivity is greater than or equal to the threshold value of apparent resistivity change; Necessary condition three: The number of consecutive over-limit periods is greater than or equal to a preset threshold for the number of consecutive over-limit periods; Necessary condition four: For scenarios with multiple monitoring locations, the number of unidirectional abnormal frequencies is greater than or equal to a preset threshold for the number of unidirectional abnormal frequencies, or the number of unidirectional abnormal monitoring locations is greater than or equal to a preset threshold for the number of unidirectional abnormal monitoring locations. For scenarios where only a single monitoring location is set, the number of unidirectional abnormal frequencies is greater than or equal to a preset threshold for the number of unidirectional abnormal frequencies.

9. The method for dynamic monitoring and early warning of underground media based on electromagnetic signals according to claim 1, characterized in that, In step 150, the defined data range for the inversion includes: The monitoring period range to be inverted is defined as the period from the start of the continuous excess of the relative rate of change of apparent resistivity to the current monitoring period; the start of the continuous excess of the relative rate of change of apparent resistivity refers to the first monitoring period in which the absolute value of the relative rate of change of apparent resistivity is continuously greater than or equal to the threshold value of the apparent resistivity change. The target frequency range to be inverted is defined as the target frequency that triggers an anomaly and the target frequency adjacent to the target frequency that triggers an anomaly and has qualified data quality; the target frequency that triggers an anomaly refers to the target frequency corresponding to the condition that the multi-dimensional anomaly determination is met. The abnormal monitoring location and the area extending to both sides by a preset distance from the abnormal monitoring location are used as the spatial location range to be inverted; the abnormal monitoring location is the monitoring location that generates the inversion trigger identifier.

10. A dynamic monitoring and early warning system for underground media based on electromagnetic signals, characterized in that, include: The first module is used to continuously collect time-domain electromagnetic data of underground media under electromagnetic fields and divide them into different monitoring periods according to preset time windows; Based on the time-domain electromagnetic data of the current monitoring period, calculate the apparent resistivity of the underground medium at each monitoring location and each target frequency during the current monitoring period; The second module is used to construct and maintain a background reference for apparent resistivity, including: comparing the apparent resistivity of the current monitoring period with the background reference to obtain the relative rate of change of apparent resistivity; and updating the background reference using the apparent resistivity that meets the preset background update conditions in the current monitoring period. The third module is used to calculate statistical characteristics based on the comparison results between the relative rate of change of apparent resistivity and the threshold of change of apparent resistivity, including the number of consecutive time periods exceeding the limit, the number of frequencies of anomalies in the same direction, and the number of monitoring locations of anomalies in the same direction. The fourth module is used to construct multi-dimensional anomaly judgment conditions using the relative change rate of apparent resistivity and the statistical characteristics; when the multi-dimensional anomaly judgment conditions are met at least at one target frequency, an inversion trigger identifier for the corresponding monitoring location is generated; otherwise, the module returns to the first module to obtain the time-domain electromagnetic data for the next monitoring period and calculate the apparent resistivity. The fifth module is used to limit the data range for inversion based on the inversion trigger identifier; extract the effective electromagnetic data within the data range for resistivity inversion imaging, and obtain the inversion resistivity model for the current monitoring period; The sixth module is used to compare the inverted resistivity model of the current monitoring period with the preset background resistivity model to obtain the difference area; based on the difference area, the spatial location and range of the anomaly in the underground medium are determined, and early warning information is generated or updated.

Images