Multi-Attribute Analysis of Transmission Channel Waves: Applications in Mine Water Damage Prevention

Abstract

In-seam seismics is one of the main methods for detecting small geostructures in a coal seam working face, using the velocity and energy attenuation characteristics. When the receiver coupling is poor or geological anomaly is strong, the stability and accuracy of the inversion results are affected greatly. In order to improve the stability and enrich the parameters of the inversion, this paper proposes a channel wave multi-attribute tomography method. Based on the theoretical analysis, the formulas for calculating the spectral ratio, arc length, and bandwidth parameters of the channel wave were obtained. A two-dimensional numerical simulation was used to analyze the attenuation characteristic and exploration potential of three attribute parameters of the channel wave. Through the field measured experiment, the exploration effect of the three attributes was realized and compared. In conclusion, the effectiveness of the channel wave multi-attribute tomography method in characterizing geological structures within coal seams was successfully demonstrated and verified. This approach offers a novel and robust solution for channel wave data processing and exploration, significantly enhancing the prevention of mine water damage.

1. Introduction

The primary challenge in achieving green and intelligent coal mining lies in constructing detailed and transparent geological models that account for factors such as faults, goaf areas, and in situ stress and mine water [1]. The occurrence and migration of mine water in coal seams and the surrounding rock masses are complex. It will interact with other concealed disaster-causing bodies, further intensifying safety risks [2,3]. Accidents caused by these hidden disaster-inducing factors are the primary barriers to the intelligent mining and safe production of coal mines [4]. Minor faults within coal seams are a significant factor in coal mine safety incidents, making an accurate prediction of these faults an urgent necessity [5,6,7]. While 3D seismic exploration is effective for detecting faults with offsets greater than 3 m, new methods are needed to detect faults with offsets of less than 3 m [8,9,10]. Channel wave exploration, with its high resolution and extended detection range, has proven effective for identifying smaller faults [11,12,13,14].

Channel wave signals, however, often suffer from noise interference due to factors like geophone coupling during data acquisition, resulting in poor amplitude preservation [15]. Simultaneously, the presence of mine water can alter the electrical and acoustic properties of coal seams and surrounding rocks, further disrupting the propagation and reception of channel wave signals. Large geological anomalies can cause severe channel wave energy attenuation, creating excessive energy discrepancies between adjacent signal channels. This, in turn, leads to imaging results that disproportionately highlight certain anomalies while neglecting smaller ones. Geological anomalies in a coal mining face will cause differences in channel wave velocity and energy attenuation. At present, channel wave exploration faults are mainly velocity and attenuation coefficient tomography. This method is widely used in the exploration of geological anomalies in working faces such as faults, collapse columns, and stress concentration areas. By analyzing and comparing the exploration of small faults by transmission and reflection channel wave methods, the response characteristics of in-seam wave to small and medium faults in thick coal seams and extremely thick coal seams are obtained [16,17]. Through the joint exploration of seismic in-seam wave transmission and reflection, the internal structure of coal seam can be better obtained [18]. But they lack the precision needed for accurate time synchronization in channel wave transmission. These issues reduce the accuracy of both energy attenuation and velocity imaging. Channel wave single-attribute imaging suffers from ambiguity, failing to provide definitive judgments on coal seam structures. Moreover, the interference and influences from factors related to mine water further complicate the process of making such judgments [19].

To address these challenges, this paper proposes a multi-attribute analysis and imaging approach for channel waves. Building upon previous work, we selected three attributes of channel waves for research: namely, the spectral ratio, arc length, and bandwidth [20,21]. The multi-attribute characteristics of channel waves aid in identifying geological structures within coal seams, thereby enhancing the potential of channel wave exploration and offering new technologies for preventing mine water damage.

2. Materials and Methods

Figure 1 illustrates the technical route and successive stages of the study.

2.1. Theoretical Approach

2.1.1. Spectral Ratio

The spectral ratio is one of the commonly used attributes in reflected-wave seismic exploration and is often applied to reservoir interpretation. The spectral ratio refers to the ratio relationship between different frequency components in the spectrum analysis of seismic signals. Anomalous bodies are geological formations that exhibit significant differences in physical properties, such as density and elastic modulus, when compared to the surrounding medium. Anomalous bodies produce frequency-selective absorption and scattering of seismic waves. High-frequency seismic waves are typically more susceptible to absorption or scattering, whereas low-frequency seismic waves tend to pass through more readily. This results in a change in the seismic wave spectrum following the abnormal body, with a relative reduction in the high-frequency component and a relative increase in the low-frequency component, thereby altering the spectral ratio [22]. The channel wave signal undergoes FFT (Fast Fourier Transform) to obtain the frequency-amplitude spectrum $R(f_j)$, $f_j=(j-1)/T$, Here, j represents the number of sampling points, where j = 1, …, N, T is the total sampling time [23]. The frequency corresponding to the maximum energy in the amplitude spectrum is the dominant frequency of the seismic wave $f_p$. The spectral ratio represents the ratio of low-frequency energy to high-frequency energy. The calculation formula for a single-trace signal is as follows [24]:

$$H={\displaystyle \sum _{f=1}^{f_p}{R}^2(f_j)}/{\displaystyle \sum _{f=f_p}^{(N-1)/T}{R}^2(f_j)}$$

(1)

2.1.2. Arc Length

Arc length is one of the commonly used seismic waveform attributes and is often used in the interpretation of sandstone and shale interfaces in reflected seismic waves [25]. In geological structures, such as faults and collapse columns, the propagation path of seismic waves becomes complexly affected. The arc length attribute can highlight these structural areas because seismic waves near the structure will exhibit phenomena such as diffraction and scattering, causing an abnormal change in arc length [26]. The arc length is defined as the waveform length of a seismic trace, which is a proportional measurement of the variation range of the seismic trace within a time window. Imagine using a rope to draw the curve of the seismic trace, and the arc length of the seismic trace is the total length when the rope is stretched out. Practice has shown that this attribute parameter is highly applicable to high-frequency signals. Channel wave signals exhibit dispersion, with abundant amplitude information in different frequency bands, and their characteristics fall within the applicable range of the arc length attribute.

The calculation formula for arc length is as follows [27]:

$$S={\displaystyle \frac{1}{N\Delta T}}{\displaystyle \sum _{i=1}^N\sqrt{{\left[A\left(i+1\right)-A\left(i\right)\right]}^2+\Delta T^2}}$$

(2)

ΔT is the sampling interval time, N is the total number of sampling points, and A(i) is the amplitude value of the i-th sampling point.

2.1.3. Bandwidth

Bandwidth is a commonly used seismic attribute in reflection seismic exploration, particularly for interpreting oil and gas reservoirs [28]. Due to the development of internal fissures and loose fillings within the fault or fracture zone, there is strong scattering and absorption of seismic waves during their propagation. This results in significant attenuation of the high-frequency components, leading to a narrower bandwidth (with a reduced main frequency and a downward shift in the high-frequency cut-off frequency) [29]. The bandwidth parameter reflects the frequency range in which the vast majority of the seismic wave energy is concentrated on the frequency spectrum curve. The channel wave energy disperses as the propagation distance increases, and the bandwidth parameter can well describe this process.

Bandwidth is defined as the absolute value of the change in the logarithm of the amplitude. Its mathematical calculation formula is as follows [30]:

$$\sigma ={\displaystyle \frac{1}{2\pi }}{\displaystyle \sum _{i=1}^N\left|\frac{d}{di}\ln A\left(i\right)\right|}$$

(3)

A(i) represents the amplitude value of the i-th sampling point.

2.2. Experimental Approach

2.2.1. Numerical Simulation Experiment

To investigate the theoretical characteristics of the transmission channel wave attributes, this study employs a high-order staggered-grid finite-difference method to simulate the 3D elastic wave equation. We designed two 3D models, each measuring 1000 m × 200 m × 200 m (Figure 2), and used numerical simulation data to calculate and analyze the three attributes of channel wave described in the previous section. This analysis aims to explore the relationship between channel wave attribute parameters and propagation distance. To minimize calculation errors in the numerical simulation, the coal seam thickness in each model is set to 10 m.

Figure 2a shows a model with a uniform horizontal coal seam, while Figure 2b includes a fault structure located at X = 500 m, with a fault throw of 5 m. The specific physical parameters of the models are detailed in

Table 1. Physical parameters of the numerical model.

Medium	P-Wave Velocity	S-Wave Velocity	Densities	Quality Factor
Coal seam	1700 m/s	1000 m/s	1300 Kg/m3	50
Rock	3500 m/s	2000 m/s	2600 Kg/m3	150

.

2.2.2. Field Experiment

To verify the effectiveness of the attenuation feature imaging based on the transmission channel wave attribute parameters, an in-field experiment was conducted at a coal mine working face where channel waves are developed.

The experiment was conducted at the No. 4108 working face, located in the southwest part of the mining area. The working face is in the first mining zone, with a mining level at the first horizon. The strike length is 1444 m, the dip length is 274 m, and the available mining length is 1315 m. The primary sources of direct water inrush at the 4108 working face are attributed to fissure water in the sandstone roof and floor of the No. 4 coal seam. The indirect sources of water inrush include the floor limestone water and the deep Ordovician limestone water. Therefore, identifying minor faults in the working face that may lead to water inrush is crucial for preventing mine water damage in that area.

The experimental setup involved placing seismic sources and geophones along the tunnel in the working face. Both the geophone and shot point spacing were set to 10 m. Considering the tunnel equipment and production conditions, a total of 133 receivers and 76 effective sources were arranged (Figure 3). Data collection was carried out using the “one shot, multiple recordings” method, and a total of 76 Common Shot Point gathers were collected.

3. Results

3.1. Numerical Simulation Results and Analysis

The channel wave records of three-dimensional models of two types of coal-rock media were obtained through simulation, as shown in Figure 4. By comparing the channel wave records of the two models, it was found that in the channel wave signal of the model with a fault, there was a sudden decrease in energy starting from the 40th trace.

To validate the reliability of the numerical simulation data, we analyzed the total energy characteristics of the simulated channel wave signals. Figure 5 shows the total energy calculation results for the channel wave. Since the source energy in the numerical simulation was not assigned units, the total energy data is dimensionless. The total energy of the channel wave exhibits a significant linear decay with propagation distance, confirming the reliability of the numerical simulation data. Further analysis can be conducted on other attribute characteristics.

Based on Formula (1), the spectral ratio of the numerical simulation signals is depicted in Figure 6. Figure 6a presents the calculation results of the numerical simulation signals from the faultless model depicted in Figure 2a, whereas Figure 6b showcases the results from the fault-bearing model as illustrated in Figure 2b. The black dots represent the spectral ratio of the channel wave signal, and the red dashed line indicates its trend. In Figure 6a, the spectral ratio exhibits a linear growth pattern, increasing with propagation distance. Figure 6b shows a significant increase in the spectral ratio at X = 400. These phenomena indicate that the growth characteristics of the channel wave spectral ratio exhibit a linear relationship with propagation distance and are sensitive to faults, making them useful parameters for channel wave tomography.

According on Formula (2), The arc length calculation results are shown in Figure 7. Since the arc length is a measure of the channel wave waveform length without a defined unit, the arc length data is dimensionless. Figure 7a presents the calculation results for the model without a fault (as in Figure 2a), while Figure 7b shows the results for the model with a fault (as in Figure 2b). In both figures, black dots represent the arc lengths of the channel wave signals, and the red dashed line indicates the trend line. As seen in Figure 7a, the arc length of the channel wave signals shows a linear decay characteristic, decreasing with increasing propagation distance. In Figure 7b, a clear arc length attenuation anomaly appears at X = 400 m. These phenomena suggest that the arc length decay characteristics of channel wave signals are also sensitive to geological anomalies and maintain a linear relationship with propagation distance, making it another valuable parameter for channel wave tomography.

The bandwidth calculation results are shown in Figure 8. Figure 8a presents the calculation results for the model without a fault (as in Figure 2a), while Figure 8b shows the results for the model with a fault (as in Figure 2b). In both figures, black dots represent the bandwidth of the channel wave signals, and the red dashed line indicates its trend. As seen in Figure 8a, the bandwidth of the channel wave signals exhibits a linear increase with increasing propagation distance. In Figure 8b, a clear bandwidth increase anomaly appears at X = 400 m. This indicates that the bandwidth is also a useful parameter for channel wave tomography.

3.2. Experimental Data and Tomographic Imaging Results

A group of common source point data was selected for preliminary analysis. The measured common shot point gathers and attribute calculation results of the 37th shot are shown in Figure 9.

Figure 9a shows a typical shot gathered from the coal seam transmission seismic data, normalized for display. The signal has a high overall signal-to-noise ratio, and the transmission channel wave is well-developed, meeting the data quality requirements for transmission channel wave imaging. Figure 9b, Figure 9c, and Figure 9d show the spectral ratio, arc length, and bandwidth of the measured channel wave data from Shot 37, respectively. Their variations with propagation distance exhibit a clear linear relationship, making them suitable for direct use in linear tomography imaging. Since seismic source signals are typically not collected in transmission channel wave exploration, the spectral ratio, arc length, and bandwidth of the source cannot be accurately estimated and are usually assigned empirical values. From Figure 9b, Figure 9c, and Figure 9d, The empirical values for the seismic source’s spectral ratio, arc length, and bandwidth are found to be 40, 1000, and 10, respectively. We then performed imaging processing on the three parameters and total energy of the measured transmission channel wave data, resulting in tomography images for each parameter, as shown in Figure 10.

Figure 10a–c illustrate the tomography results for the three-channel wave attribute parameters, whereas Figure 10d depicts the tomography result for the conventional total energy attenuation coefficient. In Figure 10d, three anomalies are predicted. However, TF-1 and TF-3 were not revealed during the actual mining, and the results in Figure 10d indicate false positives. When compared with the Total energy attenuation coefficient tomography, the spectral ratio tomography result (Figure 10a) accurately predicts the faults TF-1, TF-2, TF-4, TF-5, but fails to predict TF-3 and TF-6. The tomography results for the arc length (Figure 10b) successfully predict faults TF-3, TF-4, and TF-5. However, the bandwidth attenuation tomography results (Figure 10c) also successfully predict faults TF-3, TF-4, and TF-5 but fail to predict other faults.

4. Discussion

Through theoretical analysis and numerical simulation experiments, it was discovered that the spectral ratio, arc length, and bandwidth of channel waves exhibit significant linear correlations with propagation distance and are quite sensitive to the fault of coal seams. When the channel wave passed through a fault, all three attributes exhibited a positive response to it. In field experiments, the spectral ratio and arc length have proven effective in predicting large-scale faults. However, bandwidth and total energy are more sensitive to minor faults in coal seam working faces. It is important to emphasize that this study has only verified the response characteristics of three attributes of channel waves to faults, while the sensitivity and differences to other geological anomalous bodies, such as collapse columns and coal seam thinning areas, have not yet been supported by experimental data. Utilizing the multi-attributes of channel waves for the qualitative identification of geological anomalies in coal seams still requires further effort.

This study has, for the first time, achieved the calculation of various attributes of channel waves, obtained the response characteristics of three attribute parameters to coal seam faults, and conducted a preliminary evaluation of their effectiveness on faults. No further research has been conducted on the various attribute combinations, multi-attribute fusion, and intelligent detection models for fault identification. However, the authors acknowledge that with the deep integration of artificial intelligence, the multi-attribute technology of channel waves holds significant potential for intelligent qualitative identification and quantitative prediction of coal seam geological anomalies. It will play a more important role in coal seam geological structure and floor water prevention and control.

5. Conclusions

(1)

Based on the analysis of attribute calculations from numerical simulation and measured data, the spectral ratio, arc length, and bandwidth of channel waves exhibited an obvious linear variation pattern with the propagation distance. The variation characteristics of the spectral ratio, arc length, and bandwidth of channel waves were highly sensitive to geological anomalies, making them effective parameters for tomographic imaging in channel wave exploration.

(2)

This paper proposes a method for multi-attribute calculation and tomography of transmitted channel waves. The feasibility of this method has been verified through numerical simulation experiments and field measurements. This method addresses issues related to traditional single-attribute channel wave exploration, enhancing the stability and reliability of geological interpretation, and provides new technologies for mine water damage prevention.