Multi-band and multi-scale magnetotelluric noise suppression method
Through the geomagnetic denoising technology of divided frequency bands and multi-scale geomagnetic denoising technology, the medium and high frequency and low frequency band noise are processed by using long reference stations at different distances, effective denoising and signal enhancement of geomagnetic data is achieved, and the noise suppression effect in the existing technology is solved Poor problems improve data quality.
Patent Information
- Application Number
- CN202411116280.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-22
AI Technical Summary
The prior art is difficult to effectively suppress the earth's electromagnetic noise, especially the near-field effects of the related noise and dead frequency band, resulting in the problems of result dispersion and deviation in MT data acquisition and processing.
Using sub-band and multi-scale geomagnetic denoising technology, by setting up far reference stations at different distances, the medium and high frequency and low frequency bands are divided according to the frequency band characteristics of the noise, and the medium and high frequency impedance is obtained by dense measurement points and relatively close far reference stations. Phase curve, the low-frequency impedance phase curve is obtained using large-distance measurement points and relatively far reference stations, and the two are manually spliced to obtain the impedance phase curve of the entire frequency band.
Effectively eliminate noise, enhance effective signals, and improve the quality of earth electromagnetic data, providing a complete set of field data denoising technology processes, suitable for high-interference mineral collection areas, enriching noise suppression method technology.
Smart Images

Figure CN119001883A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of geophysical exploration, and in particular relates to a frequency-band and multi-scale magnetotelluric noise suppression method. Background Art
[0002] With the rapid development of human society, various modern electromagnetic equipment such as high-voltage transmission lines, signal towers, high-speed railways, factories, etc. have brought serious electromagnetic interference, and in actual work, it is impossible to effectively avoid various interference sources. Therefore, electromagnetic noise suppression technology has always been one of the core issues of MT data acquisition and processing. Different electromagnetic signal sources have different characteristics and influence ranges. How to reasonably use and avoid these noises has become a key factor in MT data acquisition and processing.
[0003] Since the early 1950s, when ANTikhonov and L. Cagniard proposed the magnetotelluric sounding method, electromagnetic noise suppression technology has been continuously explored. In order to overcome the influence of uncorrelated noise between channels, Goubau et al. proposed the cross-power spectrum method, which has a certain noise suppression effect, but the defect is that it cannot overcome the correlated noise. The actual electromagnetic noise is often correlated noise on each channel, and its effect is limited. In order to reduce the dispersion and deviation of the results of the least squares impedance tensor estimation of MT data, some scholars such as Egbert proposed the Robust estimation method, which performs weighted processing on the data, can effectively reduce the weight of the "flying point", and minimize the impact on the estimation of the MT impedance function. , which can effectively reduce the dispersion of apparent resistivity and phase, but it cannot eliminate the interference of input and related noise; in order to suppress local and regional electromagnetic noise, Gamble et al. proposed the far reference track technology, which uses the uncorrelated characteristics of noise between far reference points and measuring points to effectively suppress the impact of human interference on MT data. However, there are still great differences among different experts on issues such as the setting distance of far reference stations, and the effects of different reference stations also vary greatly, and this method cannot eliminate the near-field effect of dead bands; some scholars have proposed mathematical morphological filtering, which can well restore the original characteristics of MT signals, but because the selection of structural elements and their thresholds must be obtained through repeated experiments, its promotion still needs further in-depth research.
[0004] The so-called interference noise received during the magnetotelluric measurement refers to electromagnetic signals that do not satisfy the plane electromagnetic wave. However, for a signal source, it may be an interference source for a certain measuring point, but it may be a useful signal for other measuring points. This depends on the distance between the measuring point and the signal source, that is, whether the measuring point is within the far range affected by the signal source. In addition, due to the different distances of the near field and transition field of different frequencies, a signal source may be an interference source for a fixed measuring point at low frequency, but it is a useful signal at high frequency. The total power spectrum of the measured electromagnetic signal is formed by weighted superposition of the autopower spectrum of the segmented time series. Due to the randomness of the noise, the degree of interference pollution of the data in different time series segments is different. There is no processing technology that can achieve perfect denoising effect. Therefore, any denoising process and manual screening are indispensable. However, the human experience factor in the operation process is too large, which is not only time-consuming and labor-intensive, but also relatively difficult to promote. Reasonable selection of different denoising combinations and greatly reducing the workload of manual screening are the key to MT noise suppression.
[0005] Based on the above foundation, it is believed that a frequency-band, multi-scale electromagnetic denoising technology can be introduced in areas with high electromagnetic interference. A relatively close reference station is selected in the medium and high frequency bands, and a relatively distant reference station is selected in the low frequency band. This can not only meet the use of as many effective signals as possible, but also effectively reduce the mutual influence between high-frequency noise and low-frequency noise. Finally, the corrected curves are reasonably spliced to obtain a complete impedance phase curve. Summary of the invention
[0006] The technical problem to be solved by the present invention is to reduce the influence of surrounding noise during magnetotelluric data acquisition, reduce noise signals, and enhance effective information. A new magnetotelluric denoising technical method is provided. By utilizing the difference in the distance of the far zone of different "frequency bands" of noise, remote reference stations of different distances are set up, and the magnetotelluric remote reference method is used to effectively utilize the noise of different distances, so as to achieve the effect of eliminating noise and enhancing effective information.
[0007] The invention specifically adopts a frequency-band, multi-scale magnetotelluric denoising method, which includes the following steps: A. According to the detection target task, combined with the geological background, terrain environment and human electromagnetic noise characteristics of the detection area, classify and count the noise interference in the whole area, design a more reasonable test base station and remote reference station, and conduct experimental tests; B. According to the parameters such as the detection depth, the average resistivity value of the measurement area and the main influencing characteristics of the noise, the main influencing frequencies of the noise are divided into several groups. Generally, a relatively close reference station is set for the medium and high frequency group, and a relatively far reference station is set for the low frequency group. If the acquisition measurement points also adopt multi-scale and segmented measurement, the measurement points with denser point spacing correspond to the medium and high frequency bands, and the measurement points with larger point spacing correspond to the low frequency bands; C. According to the maximum detection depth corresponding to each frequency band, based on the MT field source effect principle, the corresponding skin depth and the distance from the electromagnetic dipole noise source to the "far zone" of the survey area are calculated respectively, and the reference station range is given in the range of 3.5 to 5 times the obtained distance, and a far reference test is carried out in each frequency band to select the best position; D. Carry out field construction according to the designed acquisition time corresponding to each frequency band, and obtain multi-scale and frequency-band magnetotelluric observation data and remote reference station data corresponding to each frequency band; E. Use dense measurement points and relatively close reference stations to obtain the impedance phase curves of medium and high frequencies, use large-distance measurement points and relatively far reference stations to obtain the low-frequency impedance phase curve, and manually splice the two curves to obtain the impedance phase curve of the full frequency band.
[0008] Preferably, the global noise types are investigated and classified, and test base stations are designed according to the characteristics of different types of noise to ensure that there is at least one test point for different types of noise.
[0009] Preferably, the test frequency band is divided into several groups according to the main characteristics of the detection depth, average resistivity of the measurement area and noise. It is recommended to be divided into 2~3 groups, and no more than 3 groups. Usually, 2 groups are appropriate, namely, a medium and high frequency band group, with relatively close reference stations set up, and a low frequency band group, with relatively far reference stations set up. Here, there are two situations. One is that the base station uses full-band data, and sets relatively close reference stations for correcting medium and high frequencies and relatively far reference stations for correcting low frequencies. In this case, multiple stations can be jointly referenced, or they can be referenced separately and then the data can be merged; the second situation is that the base station adopts the "multi-scale, multi-time period" acquisition mode, and the distant reference stations correspond to the scales set by the base station, and the corresponding reference processing is performed respectively. Then, according to the principle of volume effect, the data is manually spliced.
[0010] Preferably, regarding the setting of the far reference distance, according to the frequency band division, the lowest frequency and regional average resistivity corresponding to each frequency band are taken, the corresponding skin depth is calculated, and the distance range of 3.5~5 corresponding to the skin depth of each frequency band is used as the candidate area for selecting the far reference station, and experimental tests are carried out, and at least one suitable reference station is selected in each frequency band.
[0011] Preferably, according to the different data acquisition modes of the base station, for the full-band acquisition mode, not all frequency band references need to be used according to the interference of the data. If the medium and high frequencies are seriously affected, the relatively close reference stations are mainly used to correct the medium and high frequency data, retain the low frequency band data, and then manually splice the data; if the entire frequency band is affected by noise, different frequency band reference stations can be used for correction, and then the data of each frequency band are spliced to obtain the corrected data of the full frequency band. For the "multi-scale, frequency band" acquisition mode, each frequency band is corrected using the corresponding reference station, and then the data is spliced and integrated.
[0012] Preferably, the acquired spliced data usually cannot be effectively corrected within the magnetotelluric "dead band" because the natural effective signal in this frequency band is weak and the data discreteness is large. However, the data acquired by the above technology can provide a good basis for the RhoPlus model fitting method for magnetotelluric data. With this method, the data in the "dead band" can be effectively corrected, thereby obtaining high-quality original data in the full frequency band.
[0013] Compared with the prior art, the technical solution of the present invention has the following beneficial effects: (1) The present invention coordinates magnetotelluric field data collection, remote reference station layout, and noise-affected frequency bands. Based on the noise characteristics of the study area, the present invention uses experimental data to divide the frequency bands for data collection in the entire area, and specifically layouts the reference station locations to eliminate noise and enhance effective signals.
[0014] (2) The present invention constructs a complete set of technical processes for denoising magnetotelluric field data from field data collection, reference station deployment, and indoor data processing. Different from the previous single remote reference processing method or multi-station joint processing technology, it performs "frequency band" processing and makes targeted corrections to noise in different areas.
[0015] (3) The present invention is an effective supplement to the “magnetotelluric multi-scale and multi-period detection method” in suppressing noise and improving the quality of raw data.
[0016] (4) The present invention can be flexibly applied to different magnetotelluric data denoising, enriching the magnetotelluric noise suppression method technology in high-interference mining areas, and verifying its effectiveness through experimental comparison. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0018] Figure 1 is a flow chart of the method of the present invention; Figure 2 It is a schematic diagram of the layout of base stations and remote reference stations for frequency-band multi-scale magnetotelluric denoising technology; Figure 3 It is a simplified distribution diagram of electromagnetic noise sources and experimental base station layout in a certain research area; Figure 4 This is a comparison of the apparent resistivity-phase curves after YC1 processing in the distance test experiment of the far reference station; Figure 5 This is a comparison of the apparent resistivity-phase curves after YC2 processing in the distance test experiment of the far reference station; Figure 6 This is a comparison of the apparent resistivity-phase curves after YC3 processing in the distance test experiment of the far reference station; Figure 7 This is a comparison chart of the effects of different processing methods for the J1 base station in a certain study on frequency band and multi-scale denoising test; Figure 8 This is a comparison chart of the effects of different processing methods for the J2 base station in a certain study on frequency band and multi-scale denoising test; Fig. 9 This is a comparison chart of the effects of different processing methods for the J3 base station in a certain study on frequency band and multi-scale denoising experiment. DETAILED DESCRIPTION
[0019] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.
[0020] In order to solve the problem of noise suppression in magnetotelluric data collection in high-interference mining areas, the idea of this embodiment is to utilize the field source effect of controllable source audio magnetotelluric sounding, assume the electromagnetic interference sources around the magnetotelluric data collection as different electromagnetic dipole sources, and use the different characteristics of the far zone propagation of electromagnetic signals of different frequencies to set corresponding reference stations. Using the far reference station method, part of the noise is converted into effective information, so as to improve the quality of the original magnetotelluric data.
[0021] Specifically, this embodiment is a frequency-band, multi-scale magnetotelluric remote reference channel electromagnetic noise suppression technology, and its specific technical solution is as follows: Figure 1 As shown, it mainly includes the following steps: Step A: Based on the detection target task, combined with the geological background, terrain environment and human electromagnetic noise characteristics of the detection area, classify and count the noise interference in the whole area, design a more reasonable test base station and remote reference station, and conduct experimental tests.
[0022] In the early stage of the project implementation, a statistical analysis was conducted on the possible sources of electromagnetic noise within the study area. The main electromagnetic noise sources include: ①: urban residential areas; ②: trains and other rail transit; ③: airplanes; ④: high-voltage transmission lines; ⑤: communication radars; ⑥: road vehicles; ⑦: signal towers; ⑧: factories; ⑨: power-generating windmills; ⑩: substations; ⑪: underground rail transit; ⑫: underground operation collection cables, etc., and their approximate locations were marked on the regional satellite map, and a simplified version of the regional electromagnetic noise source distribution map was drawn.
[0023] According to the characteristics of electromagnetic noise and the main noise types (square wave noise, triangle wave noise, step noise, pulse noise, charge and discharge noise, periodic noise, etc.), the noise sources are statistically classified, and typical test base stations are selected for noise testing. The typical noise selection criteria include: at least one location with the lowest possible noise interference is selected for comparison; in addition, test points are selected for different noises of the classification, and locations affected by multiple noises can be selected, such as Figure 3 The figure shows the electromagnetic noise source survey and experimental base station layout in a certain research area. The data acquisition device is used to collect full-band data and analyze the original data.
[0024] Step B: Divide the main influencing frequencies of noise into several groups according to the parameters such as the detection depth, the average resistivity value of the measurement area and the main influencing characteristics of noise. Generally, the medium and high frequency groups are set with relatively close reference stations, and the low frequency groups are set with relatively far reference stations. If the acquisition points also adopt multi-scale and segmented measurement, the points with denser point spacing correspond to the medium and high frequency bands, and the points with larger point spacing correspond to the low frequency bands; After completing the analysis of electromagnetic signals and noise data in the study area, the main frequency bands affected by the interference can be roughly obtained, and the frequency bands with greater impact can be grouped. According to past experience, there are no more than three groups. Generally, there are more medium and high frequency groups and low frequency groups, and the frequency range of each frequency band can be determined.
[0025] There are two ways of collecting magnetotelluric data. One is the conventional data collection model, which collects full-band data according to preset measuring points. The other is "a multi-scale, multi-time period magnetotelluric detection method", which is perfectly compatible with the frequency band and multi-scale denoising technology proposed this time.
[0026] C. According to the maximum detection depth corresponding to each frequency band, based on the MT field source effect principle, the corresponding skin depth and the distance from the electromagnetic dipole noise source to the "far zone" of the survey area are calculated respectively, and the reference station range is given in the range of 3.5 to 5 times the obtained distance, and a far reference test is carried out in each frequency band to select the best position; After completing step B, the main electromagnetic noise frequency bands are divided. According to the basic principles of magnetotellurics and the principle of remote reference station method, the lowest frequency skin depth of each frequency band can be obtained by the lowest frequency of each divided frequency band and the average resistivity of the regional formation. The range of different remote reference station layouts can be set within 3.5 to 5 times the skin depth of the frequency band, such as Figure 2 As shown, a remote reference test was carried out and the best reference station location for each frequency band was selected through comparison.
[0027] D. Carry out field construction according to the designed acquisition time corresponding to each frequency band, and obtain multi-scale and frequency-band magnetotelluric observation data and remote reference station data corresponding to each frequency band; This part mainly includes two key points. First, the field base station data acquisition mode adopts full-band data acquisition, that is, under the conventional magnetotelluric acquisition mode, the distant reference station completes data acquisition in a frequency band and multi-scale manner; second, the field base station data acquisition adopts a "multi-scale, multi-time period" mode, that is, a dense dot matrix is used for medium and high frequency data, and a sparse dot matrix is used for low frequency data. According to the magnetotelluric volume effect, the point distribution is reasonably designed to keep the data acquisition frequency band consistent with the distant reference station frequency band. The medium and high frequency bands correspond to relatively close reference stations, and the low frequency bands correspond to relatively far reference stations, and finally complete the data acquisition.
[0028] E. Use dense measurement points and relatively close reference stations to obtain the impedance phase curves of medium and high frequencies, use large-distance measurement points and relatively far reference stations to obtain the low-frequency impedance phase curve, and manually splice the two curves to obtain the impedance phase curve of the full frequency band.
[0029] This part is mainly about data processing and merging, which is mainly divided into two types. One is that the base station data is collected in the whole frequency band. When processing the data, according to the degree of interference of the data, relatively close reference stations can be used to correct the medium and high frequency data, and relatively far reference stations can be used to correct the low frequency band data. Then, the corrected data of different frequency bands are spliced to obtain complete data of the whole frequency band. The original data can also be corrected using a single reference station, or a combination of far and near reference stations. The better method can be selected according to the correction effect.
[0030] The other is that the base station adopts the "multi-scale, frequency band" acquisition mode. At this time, relatively close reference stations are mainly used for correction of medium and high frequency data, and relatively far reference stations are used for correction of low frequency bands. If there are special measuring points where medium and high frequency data cannot be effectively corrected using near references, you can also try to use relatively far reference stations for correction and comparison. However, if the correction effect of low frequency data using relatively far reference stations is not good, relatively close reference stations cannot be used for correction. This is because in this case, relatively close reference stations do not meet the "far zone" characteristics of electromagnetic waves for low frequency data. Finally, after completing the correction of different frequency bands, manual splicing is performed to obtain the completed frequency band data, or splicing is not performed and subsequent processing is performed separately.
[0031] In order to verify the effect of the technical solution of this embodiment, a relevant test study was carried out in a certain area of the Subei Basin using the method of this embodiment, and the construction process and effect thereof were verified as follows: Test 1: A long reference distance test was conducted in a mining area in the Subei Basin of Jiangsu Province, mainly to verify the reference distance theory proposed by this method. The test points were arranged as follows: Figure 3 As shown, two test points, SY1 and SY2, were selected under different noise levels, and three reference stations, YC1, YC2 and YC3, were set up from near to far, and their relative distances from the test points were 3.56 times δ (skin depth), 13 times δ and 27 times δ, respectively.
[0032] By comparing different distance references, such as Figures 4 to 6 As shown in the figure, it can be seen from a single comparison that no matter YC1, YC2 or YC3, after being processed by the remote reference station, the resistivity and phase curves of SY1 and SY2 have been significantly corrected. Among them, the SY1 station is relatively less interfered, and its correction is mainly concentrated in the MT dead band range of about 0.1~10Hz. SY2 is relatively more interfered, and its dead band and medium and high frequency distortion points are significantly improved. When comparing the effects of different reference stations, it can be found that for the SY1 station, as the reference distance becomes farther and farther, the curve jump points in the medium and low frequency parts increase significantly, and the phase also jumps. On the contrary, the reference effect is better at a closer distance. For the SY2 station, the closer the reference station is, the better the correction effect is. It can be seen from the effects of different reference distances of SY1 and SY2 that when there are many and large interference sources around the station, the requirements for the remote reference station are higher. Therefore, when selecting a remote reference station, you should try to choose a location with no interference around it. Because in modern society, there are too many similar interference sources, which are easy to cause situations related to the noise around the station, resulting in the phenomenon that cannot be corrected. The experiments also verified the viewpoints of this article. When the reference station is too far away, some useful signals are often regarded as noise, resulting in poor reference effect.
[0033] Test 2: The test was conducted in the coastal area of Jiangsu Province. The full frequency range of the test was 1000~0.001Hz. Three basic measurement stations were set up at locations with different noise levels. The main noise source around J1 was urban residents’ electricity consumption, J2 was mainly factory and high-voltage transmission line type noise sources, and J3 was a base station with relatively low noise level in the region. Two far reference stations were set up in two frequency bands. The closer reference station YC1 was mainly used for data correction in the medium and high frequency bands (1000-0.1Hz), and the farther reference station YC2 was mainly used for data correction in the medium and low frequency bands (0.1-0.001Hz). The two reference stations also received data in the full frequency band (1000-0.001Hz) for comparative analysis. The data collection time was more than 20h.
[0034] Regarding the calculation of reference distance, the YC1 reference distance is 70km, relative to its main calibration frequency band, the lowest frequency is 0.01Hz, the average resistivity is about 10Ω·m, and the skin depth is about 15km; the YC2 reference distance is 350km, relative to its main calibration frequency band, the lowest frequency is 0.001Hz, and the skin depth is about 50km. Therefore, the far reference station is equivalent to 5~7 times the skin depth relative to its main calibration frequency band.
[0035] First, compare Figure 7 .a. Figure 8 .a and Fig. 9 .a, it can be seen that the original curves of J1 and J2 are chaotic and have no obvious trend, but due to the difference in location and surrounding noise sources, they show different characteristics. J1 is mainly characterized by obvious depression of resistivity curve in the 1-10Hz frequency band, and significant dispersion characteristics in the low frequency band of 0.001-0.1Hz, showing a downward trend; J2 is mainly characterized by segmented fluctuations in resistivity curve, and serious "flying point" phenomenon occurs in the 0.05-0.5Hz segment. The J3 data curve is obviously better than J1 and J2, with good continuity, and only a small amount of dispersion phenomenon exists in the low and high frequency bands, but in the 0.05-0.5Hz frequency band, there is an obvious "near field" effect in which the resistivity drops by 45° and then rises, and the phase tends to 0. This phenomenon also exists in J1 and J2, but due to the poor data quality, the characteristics are not obvious.
[0036] The mid-high frequency band (YC1) and the mid-low frequency band (YC2) were subjected to far reference processing respectively. It can be seen that after far reference processing, the shape of the original curve has been greatly improved, especially in the J1 and J2 curves. However, after YC1 reference station processing, the three base stations all showed obvious "near field" effect characteristics in the 0.05-0.5Hz frequency band, indicating that when the reference station is too close, it cannot effectively eliminate the influence of the near source ( Figure 7 .b. Figure 8 .b and Fig. 9.b); After observing the YC2 far reference processing, it can be found that the "near field" effect characteristics are eliminated in the 0.05-0.5Hz frequency band, but because this frequency band belongs to the MT "dead frequency band" range and the natural signal source is weak, the curve in this section shows obvious fluctuation characteristics after processing, and the data error bar is large. By comparison, it can be found that ( Figure 7 .bc, Figure 8 .bc and Fig. 9 .bc), no matter YC1 or YC2, there is still obvious dispersion in the 400-1000Hz frequency band after remote reference processing. The main reason is that this frequency band belongs to the audio magnetotelluric "dead band" range due to its weak natural field signal.
[0037] By taking the frequency bands of the YC1 reference station (0.1-1000Hz) and the YC2 reference station (0.001-0.1) respectively and physically splicing them, a curve that is greatly improved compared to the original curve can be obtained ( Figure 7 .d. Figure 8 .d and Fig. 9 .d) This curve inherits the improvement of high-frequency data by YC1 and the improvement of low-frequency data by YC2, but the data in the MT and AMT "dead band" frequency bands are still poor and need to be corrected by manual selection and other processing methods.
[0038] Multi-station reference processing using YC1 and YC2 ( Figure 7 .e、 Figure 8 .e and Fig. 9 .e) For J1, the effect of multi-channel reference is significantly worse than that of single-station reference, and the data dispersion in the low-frequency band is significantly increased, which reduces the data quality. For J2 and J3, after multi-station reference processing, the curve shape is improved to a certain extent compared with single-station processing. Therefore, multi-station reference may further improve data quality, or it may reduce data quality. It has different effects for different measurement points and has certain uncertainties.
[0039] Figure 7 .f, Figure 8 .f and Fig. 9 .f is the curve after RhoPlus model fitting correction after the splicing of near and far references, which can better correct the data in the "dead band" frequency band to make the curve shape continuous and smooth. By comparison, it is believed that the curve after splicing of near and far references has a good similarity with the single far reference or multi-station far reference, and there is a certain improvement in some aspects. At the same time, it has the premise of RhoPlus model fitting, which can make the curve with greater interference processed into original data that meets the requirements.
[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A frequency-band, multi-scale magnetotelluric noise suppression method, characterized in that It includes the following steps: A. According to the detection target task, combined with the geological background, terrain environment and human electromagnetic noise characteristics of the detection area, classify and count the noise interference in the whole area, design a more reasonable test base station and remote reference station, and conduct experimental tests; B. According to the parameters such as the detection depth, the average resistivity value of the measurement area and the main influencing characteristics of the noise, the main influencing frequencies of the noise are divided into several groups. Generally, a relatively close reference station is set for the medium and high frequency group, and a relatively far reference station is set for the low frequency group. If the acquisition measurement points also adopt multi-scale and segmented measurement, the measurement points with denser point spacing correspond to the medium and high frequency bands, and the measurement points with larger point spacing correspond to the low frequency bands; C. According to the maximum detection depth corresponding to each frequency band, based on the MT field source effect principle, the corresponding skin depth and the distance from the electromagnetic dipole noise source to the "far zone" of the survey area are calculated respectively, and the reference station range is given in the range of 3.5 to 5 times the obtained distance, and a far reference test is carried out in each frequency band to select the best position; D. Carry out field construction according to the designed acquisition time corresponding to each frequency band, and obtain multi-scale and frequency-band magnetotelluric observation data and remote reference station data corresponding to each frequency band; E. Use dense measurement points and relatively close reference stations to obtain the impedance phase curves of medium and high frequencies, use large-distance measurement points and relatively far reference stations to obtain the low-frequency impedance phase curve, and manually splice the two curves to obtain the impedance phase curve of the full frequency band.
2. A frequency-band, multi-scale magnetotelluric noise suppression method according to claim 1, characterized in that Investigate and classify the global noise types in step A, and design test base stations according to the characteristics of different types of noise to ensure that there is at least one test point for different types of noise.
3. The method for suppressing magnetotelluric noise in a frequency-band and multi-scale manner according to claim 1, characterized in that In step B, according to the detection depth of each reference station, the average resistivity of the survey area and the main characteristics of noise, the test frequency band is divided into several groups. It is recommended to be divided into 2~3 groups, and no more than 3 groups. Usually, 2 groups are appropriate, namely, the medium and high frequency band group, which sets relatively close reference stations, and the low frequency band group, which sets relatively far reference stations. Here, it is divided into two cases. One is that the base station uses full-band data, and sets relatively close reference stations for correcting medium and high frequencies and relatively far reference stations for correcting low frequencies. In this case, multiple stations can be jointly referenced, or they can be referenced separately and then the data can be merged; the second case is that the base station adopts the "multi-scale, multi-time period" acquisition mode, and the distant reference stations set the scales corresponding to the base station, and perform corresponding reference processing respectively. Then, according to the principle of volume effect, the data are manually spliced.
4. The method for suppressing magnetotelluric noise in a frequency-band and multi-scale manner according to claim 1, characterized in that Regarding the distance setting of the far reference in step c, according to the frequency band division, take the corresponding lowest frequency and regional average resistivity of each frequency band, calculate the corresponding skin depth, and use the distance range of 3.5~5 corresponding to the skin depth of each frequency band as the candidate area for selecting the far reference station, conduct experimental tests, and finally select at least one suitable reference station in each frequency band.
5. The frequency-band, multi-scale magnetotelluric noise suppression method according to claim 1, characterized in that In step d, according to the difference between the multi-scale, frequency-band magnetotelluric observation data and the distant reference station data corresponding to each frequency band, for the full-band acquisition mode, not all frequency band references need to be used according to the interference of the data. If the medium and high frequencies are seriously affected, the relatively close reference stations are mainly used to correct the medium and high frequency data, retain the low-frequency data, and then manually splice the data; if the entire frequency band is affected by noise, different frequency band reference stations can be used for correction respectively, and then the data of each frequency band are spliced to obtain the corrected data of the entire frequency band. For the "multi-scale, frequency-band" acquisition mode, each frequency band is corrected using the corresponding reference station, and then the data are spliced and integrated.
6. The frequency-band, multi-scale magnetotelluric noise suppression method according to claim 1, characterized in that The impedance phase curve data of the full frequency band obtained after splicing in step E cannot usually be effectively corrected within the magnetotelluric "dead band" range. Due to the weak natural effective signal in this frequency band, the data discreteness is large. However, the data obtained by the above technology can provide a good basis for the RhoPlus model fitting method for magnetotelluric data. With this method, the data in the "dead band" can be effectively corrected, thereby obtaining high-quality original data of the full frequency band.
Citation Information
Cited By
Method for suppressing magnetotelluric mixed noise in shallow water area
CN120447068A
Magnetotelluric sounding data splicing processing method and system
CN120630314A
Multi-station remote reference magnetotelluric denoising method based on heavy coherence degree
CN121596409A
