Mine earthquake time prediction method based on rock deformation localization state

Abstract

The invention belongs to the field of mine dynamic disaster prevention and control, and relates to a mine earthquake time prediction method based on a rock deformation localization state. According to the method, firstly, the relation between mine earthquakes and deformation localization is established through theoretical analysis; and then, obtaining detailed physical information of rock deformation and damage by using a digital image related technology and an acoustic emission monitoring technology. Then, a time prediction model is constructed through an extended Saito theory, and an acoustic emission space aggregation index is introduced to identify a key time point at which deformation localization starts. By analyzing an acoustic emission event, an important precursor feature of a calm period is determined and is used as a key parameter of a prediction model, and the instability failure time of the rock is successfully predicted. Based on the micro-seismic event, the occurrence time of the mine earthquake can be accurately predicted in advance in the field instance application.

Description

Technical Field

[0001] The present invention belongs to the field of mine dynamic disaster prevention and control, and particularly relates to a mine earthquake time prediction method and application, and specifically relates to a mine earthquake time prediction method and application based on the localized state of rock deformation. Background Art

[0002] Mine tremors are dynamic phenomena caused by the instantaneous release of elastic energy accumulated in the overlying rock strata induced by mining activities. Mine tremors often occur in the goaf of adjacent working faces. There is obvious shaking on the ground, but no tunnel damage, equipment damage, or casualties underground. In recent years, with the increase in mining depth and mining intensity, mine tremors have become more frequent. Some mine tremors not only induce secondary disasters such as underground rock pressure and coal and gas outbursts, but also lead to serious consequences such as strong ground shaking and building damage, making the mine tremor problem evolve from a mining safety issue to a public safety issue. Therefore, the management of mine tremor disasters has received more and more attention and importance.

[0003] Deformation localization refers to the deformation behavior of certain areas in the rock mass that is significantly different from the surrounding areas due to stress concentration. This phenomenon is usually regarded as a precursor to the impending occurrence of mine earthquakes. In order to effectively predict coal-rock disasters and mine dynamic disasters, the deformation localization phenomenon is studied in depth, and the key precursor information indicating coal (rock) instability and failure is identified from it. By mining these characteristic information, a reliable scientific basis is provided for the prediction and early warning of coal-rock disasters and mine dynamic disasters. Before rock instability and failure, the deformation localization phenomenon is considered to be a key precursor indicator. Microseismic monitoring data at the engineering site also reveals the deformation localization characteristics, providing a new perspective for the study of coal mine disasters dominated by rock instability and failure. Despite this, an effective prediction model has not yet been developed that can break through the limitations of a single indicator and link the laboratory with the on-site monitoring status based on the localized state of rock deformation to accurately predict the time of mine earthquakes and provide a scientific basis for the early warning and prevention of mine earthquakes. Summary of the invention

[0004] The technical solution adopted by the present invention is: first, the method establishes the connection between mine earthquake and localized deformation through theoretical analysis. Then, the detailed physical information of rock deformation and destruction is obtained by using digital image correlation technology and acoustic emission monitoring technology. Subsequently, a time prediction model is constructed through the extended Saito theory, and the acoustic emission (microseismic) spatial aggregation index is introduced to identify the key time point when deformation localization begins. By analyzing the acoustic emission events, the important precursor feature of "quiet period" is determined, and it is used as the key parameter of the prediction model to successfully predict the instability and destruction time of the rock. Based on the microseismic events, the time of the mine earthquake can be predicted more accurately in advance in the field example application.

[0005] The method for predicting the time of a mine earthquake based on the localized state of rock deformation of the present invention comprises the following steps:

[0006] Step 1: Deformation localization start time calculation

[0007] The spatial distribution types of acoustic emission (microseismic) location points can be divided into three types: clustered, discrete, and random. The spatial distribution characteristics of acoustic emission (microseismic) events over time can be quantitatively analyzed through spatial autocorrelation analysis. Therefore, the spatial aggregation index Moran's I is proposed to track the deformation localization state. When the spatial distribution of acoustic emission (microseismic) events forms a stable aggregation pattern, it indicates that deformation localization has started. The specific calculation method is as follows:

[0008]

[0009] Where: n is the number of regional units; x i is the attribute observation value of regional unit i. ij is the spatial weight coefficient. When regions i and j are adjacent, W ij is equal to 1, when regions i and j are not adjacent, W ij is equal to 0.

[0010] When Moran's I is close to 1, it is positive spatial autocorrelation and presents a clustered distribution; when Moran's I is close to -1, it is negative spatial autocorrelation and presents a discrete distribution; when Moran's I is equal to 0, there is no spatial autocorrelation and presents a random distribution.

[0011] However, due to the large amount of microseismic data collected at the engineering site, according to formula (1), in the traditional algorithm, it is necessary to judge the adjacency relationship of each element in the space, and n elements need to establish an n×n weight matrix, which will lead to low calculation efficiency and may not be realized. Therefore, the improved Moran's I is used for calculation. The improved algorithm formula is as follows:

[0012]

[0013] The improved method adopts the face adjacency model in the spatial model relationship. This method defines the elements with shared boundaries as neighbors, that is, the spatial proximity relationship is determined based on the face adjacency, and the W corresponding to each group (i, j) is calculated. ij When , it is directly substituted into the calculation without storage. In formula (2), N(i) is defined as follows:

[0014] N(i)={(x i +1,y i ), (x i -1,y i ), (x i ,y i+1), (x i ,y i -1)} (3)

[0015] The improved algorithm divides the data into regular grids, and a grid cell may be adjacent to at most four other grid cells. Therefore, when using this spatial model, it is only necessary to perform four adjacency judgments, and the judgment range is shown in formula (3).

[0016] Step 2: Time prediction model

[0017] The pioneering work of Japanese scholar Saito has been influential since the 1960s, and its main academic ideas have been continuously expanded and applied to the prediction of the occurrence time of natural geological disasters such as volcanic eruptions and earthquakes. This paper proposes an expanded mine earthquake time prediction model based on the theory established by Saito. The initial formula is as follows:

[0018]

[0019] In the model of formula (4), t r To predict the time; is the instantaneous event rate. r and The length of rock instability and failure time and the instantaneous event rate during the observation time are used to replace them respectively, and the simplified result is

[0020]

[0021] where t r is the rock instability failure time, t is the real-time observation time, is the instantaneous event rate and a is a constant.

[0022] Assume that at t = t0, Available

[0023]

[0024] here

[0025]

[0026] Then formula (7) can be rewritten as

[0027]

[0028] Among them, Δd is the difference between the relative event values ​​of two time points, and d0 is the event value corresponding to the real-time observation time. Thus, the time prediction model is obtained.

[0029] In actual prediction, the rock instability failure time t is calculated by combining formula (9) within the observation time using a graphical method.r .

[0030] The prediction model formula contains four unknowns a, d0, t r and t0, more reference points need to be taken on the cumulative event curve for calculation. Select three consecutive points D1(d1, t1), D2(d2, t2), D3(d3, t3) on the cumulative event curve. These three points must meet the following conditions:

[0031] d2-d1=d3-d2 (10)

[0032] Substituting the data satisfying the conditions of formula (10) into formula (9), we can obtain three sets of simultaneous equations, which can be simplified to the following relationship:

[0033]

[0034] Simplified

[0035]

[0036] The rock instability failure time can be obtained by moving the term

[0037]

[0038] Formula (13) is the prediction model. In practical application, the rock instability failure time is solved by the graphical method, and the start time of deformation localization is set to a fixed initial time according to the rock deformation evolution process. The solution steps are:

[0039] (1) First, find three consecutive points D1, D2, and D3, whose corresponding times are t1, t2, and t3.

[0040] (2) Select the center point (D2) and draw a perpendicular line to the Y axis. Project points D1 and D3 onto the perpendicular line and record them as D1' and D3' respectively.

[0041] (3) Find the midpoint of D1' and D2 on the vertical line, denoted as M; and the midpoint of D1' and D3 on the vertical line, denoted as N.

[0042] (4) With D2 as the center and MD2 and ND2 as the radius, draw arcs that intersect the perpendicular line of D2 projected onto the X-axis at points M' and N', respectively.

[0043] (5) Connect D1' and N' to make an extension line, which intersects with the line parallel to the X-axis through M' at one point. Draw a perpendicular line to the X-axis to obtain the x-coordinate of the point, which is t. r .

[0044] Step 3: Method reliability verification

[0045] Through rock deformation and failure experiments, the physical information of rock deformation and failure was obtained using digital image correlation technology and acoustic emission monitoring technology. According to the characteristics of deformation field evolution and spatial aggregation parameter curves, the localization start time of rock deformation was analyzed from qualitative and quantitative perspectives. Then, according to the cumulative event frequency parameters of acoustic emission, the time prediction model was used to calculate the prediction trajectory of rock instability and failure time, and the prediction time accuracy was obtained, verifying the reliability of the deformation localization method in predicting rock instability and failure time.

[0046] The beneficial effects created by the present invention are as follows: the model proposed by the present invention can break through the limitations of a single indicator, and based on the localized state of rock deformation, link the laboratory with the on-site monitoring state to accurately predict the time of occurrence of mine earthquakes, providing a scientific basis for the early warning and prevention of mine earthquakes. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 : Flow chart of the method of the present invention;

[0048] Figure 2 : The specimen loading curve of the present invention;

[0049] Figure 3 : The rock deformation field evolution cloud map of the present invention;

[0050] Figure 4 : The start time of the deformation localization zone based on acoustic emission events of the present invention;

[0051] Figure 5 : Schematic diagram of the failure time prediction by graphical method;

[0052] Figure 6 : The rock instability failure time prediction diagram of the present invention;

[0053] Figure 7 : The working face and geological drilling column diagram of the present invention;

[0054] Figure 8 : The microseismic energy evolution curve of the present invention;

[0055] Fig. 9 : The microseismic event evolution curve of the present invention;

[0056] Fig.10 : The start time of the deformation localization zone based on microseismic events of the present invention;

[0057] Fig.11 : The mine earthquake time prediction diagram of the present invention. DETAILED DESCRIPTION

[0058] The embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0059] like Figure 1The mine earthquake time prediction method based on the localized state of rock deformation is shown in the figure:

[0060] Step 1: Deformation localization start time calculation

[0061] The spatial distribution types of acoustic emission (microseismic) location points can be divided into three types: clustered, discrete, and random. The spatial distribution characteristics of acoustic emission (microseismic) events over time can be quantitatively analyzed through spatial autocorrelation analysis. Therefore, the spatial aggregation index Moran's I is proposed to track the deformation localization state. When the spatial distribution of acoustic emission (microseismic) events forms a stable aggregation pattern, it indicates that deformation localization has started. The specific calculation method is as follows:

[0062]

[0063] Where: n is the number of regional units; x i is the attribute observation value of regional unit i. ij is the spatial weight coefficient. When regions i and j are adjacent, W ij is equal to 1, when regions i and j are not adjacent, W ij is equal to 0.

[0064] When Moran's I is close to 1, it is positive spatial autocorrelation and presents a clustered distribution; when Moran's I is close to -1, it is negative spatial autocorrelation and presents a discrete distribution; when Moran's I is equal to 0, there is no spatial autocorrelation and presents a random distribution.

[0065] However, due to the large amount of microseismic data collected at the engineering site, according to formula (1), in the traditional algorithm, it is necessary to judge the adjacency relationship of each element in the space, and n elements need to establish an n×n weight matrix, which will lead to low calculation efficiency and may not be realized. Therefore, the improved Moran's I is used for calculation. The improved algorithm formula is as follows:

[0066]

[0067] The improved method adopts the face adjacency model in the spatial model relationship. This method defines the elements with shared boundaries as neighbors, that is, the spatial proximity relationship is determined based on the face adjacency, and the W corresponding to each group (i, j) is calculated. ij When , it is directly substituted into the calculation without storage. In formula (2), N(i) is defined as follows:

[0068] N(i)={(x i +1,y i ), (x i -1,y i ), (x i ,y i+1), (x i ,y i -1)} (3)

[0069] The improved algorithm divides the data into regular grids, and a grid cell may be adjacent to at most four other grid cells. Therefore, when using this spatial model, it is only necessary to perform four adjacency judgments, and the judgment range is shown in formula (3).

[0070] Step 2: Time prediction model

[0071] The pioneering work of Japanese scholar Saito has been influential since the 1960s, and its main academic ideas have been continuously expanded and applied to the prediction of the occurrence time of natural geological disasters such as volcanic eruptions and earthquakes. This paper proposes an expanded mine earthquake time prediction model based on the theory established by Saito. The initial formula is as follows:

[0072]

[0073] In the model of formula (4), t r To predict the time; is the instantaneous event rate. r and The length of rock instability and failure time and the instantaneous event rate during the observation time are used to replace them respectively, and the simplified result is

[0074]

[0075] where t r is the rock instability failure time, t is the real-time observation time, is the instantaneous event rate and a is a constant.

[0076] Assume that at t = t0, Available

[0077]

[0078] here

[0079]

[0080] Then formula (7) can be rewritten as

[0081]

[0082] Among them, Δd is the difference between the relative event values ​​of two time points, and d0 is the event value corresponding to the real-time observation time. Thus, the time prediction model is obtained.

[0083] In actual prediction, the rock instability failure time t is calculated by combining formula (9) within the observation time using a graphical method.r .

[0084] The prediction model formula contains four unknowns a, d0, t r and t0, more reference points need to be taken on the cumulative event curve for calculation. Select three consecutive points D1(d1, t1), D2(d2, t2), D3(d3, t3) on the cumulative event curve. These three points must meet the following conditions:

[0085] d2-d1=d3-d2 (10)

[0086] Substituting the data satisfying the conditions of formula (10) into formula (9), we can obtain three sets of simultaneous equations, which can be simplified to the following relationship:

[0087]

[0088] Simplified

[0089]

[0090] The rock instability failure time can be obtained by moving the term

[0091]

[0092] Formula (13) is the prediction model. In practical application, the rock instability failure time is solved by the graphical method, and the start time of deformation localization is set to a fixed initial time according to the rock deformation evolution process. The solution steps are:

[0093] (1) First, find three consecutive points D1, D2, and D3, whose corresponding times are t1, t2, and t3.

[0094] (2) Select the center point (D2) and draw a perpendicular line to the Y axis. Project points D1 and D3 onto the perpendicular line and record them as D1' and D3' respectively.

[0095] (3) Find the midpoint of D1' and D2 on the vertical line, denoted as M; and the midpoint of D1' and D3 on the vertical line, denoted as N.

[0096] (4) With D2 as the center and MD2 and ND2 as the radius, draw arcs that intersect the perpendicular line of D2 projected onto the X-axis at points M' and N', respectively.

[0097] (5) Connect D1' and N' to make an extension line, which intersects with the line parallel to the X-axis through M' at one point. Draw a perpendicular line to the X-axis to obtain the x-coordinate of the point, which is t. r .

[0098] Step 3: Method reliability verification

[0099] Through rock deformation and failure experiments, the physical information of rock deformation and failure was obtained using digital image correlation technology and acoustic emission monitoring technology. According to the characteristics of deformation field evolution and spatial aggregation parameter curves, the localization start time of rock deformation was analyzed from qualitative and quantitative perspectives. Then, according to the cumulative event frequency parameters of acoustic emission, the time prediction model was used to calculate the prediction trajectory of rock instability and failure time, and the prediction time accuracy was obtained, verifying the reliability of the deformation localization method in predicting rock instability and failure time.

[0100] In order to verify the present invention, a rock deformation and failure evolution experiment was conducted. The experimental system consists of a loading system and an acquisition system. The acquisition system includes a digital image acquisition system and an acoustic emission monitoring system. The loading and data acquisition parameters in the verification experiment are as follows: uniaxial compression is performed at a displacement control rate of 0.05mm / min, and the digital image acquisition system records the deformation of the specimen surface in real time at a speed of 1 frame / s. The image resolution is as high as 1600 pixels × 1200 pixels, and the object surface resolution reaches 0.13mm / pixel. The acoustic emission monitoring system is set with a preamplifier gain of 40dB and an acquisition threshold value of 10mV, and a sampling rate of 3MHz.

[0101] According to the loading data collected in the experiment, the specimen loading curve is obtained. Figure 2 shown. Figure 3 The deformation field evolution cloud diagram at the corresponding time in is as follows: Figure 3 shown. Figure 3 It can be seen that the start time of deformation localization is approximately the linear rise stage before the peak load of the specimen. The final instability and failure of the rock is caused by the continuous development and penetration of deformation localization.

[0102] First, the time of deformation localization initiation is calculated based on the acoustic emission events, such as Figure 4 As shown. According to the time prediction model, combined with the graphical method, as Figure 5 The predicted value of rock instability failure time is calculated as follows: Figure 6 The prediction accuracy of the 90% confidence interval of the actual failure time was calculated, which confirmed the feasibility of the time prediction model.

[0103] An example of the present invention is given below in conjunction with the accompanying drawings:

[0104] (1) Figure 7 As shown, a mining area in Inner Mongolia -1 105 working surface is located at 3 -1 1. The middle part of the mining area has a strike length of 2671.5m, a working face length of 280 / 291m, and an area of ​​766492m 2. The structural morphology of the working face range is a monocline structure that is generally inclined to the west, with a dip angle of 1 to 7°. The occurrence of the strata varies along the strike and dip, but the changes are not significant. There is a NF40 fault in the excavation range, with a drop of 0 to 3m, as well as faults NF37 and NF39. It is very likely that the extension faults of these two faults will be encountered during the excavation process. The direct roof rock of the mined coal seam is mainly sandy mudstone and fine sandstone, with moderately developed fissures, developed bedding, and blocky structure. The basic roof is mainly medium-coarse-grained sandstone, light gray, mainly containing quartz and feldspar, sub-angular, medium sorting, mud cemented, and semi-hard. The bottom plate of the coal seam is mainly sandy mudstone, gray-black, muddy structure, horizontal bedding, flat or shell-shaped fracture, interbedded with carbonaceous mudstone and thin coal seams, and cemented when encountering.

[0105] (2) The energy of the microseismic monitoring data of the working face from November 1, 2020 to April 2, 2021 was selected for analysis. During the mining period, the working face experienced a strong mine earthquake on April 2, and at the same time, a 2.13×10 6 J energy microseismic events. Figure 8 shown.

[0106] (3) Analysis of the daily and cumulative daily frequencies of microseismic events during the monitoring period, combined with the aforementioned microseismic energy evolution curve, found that the daily frequency of microseismic events around March 22 was relatively large, reaching a maximum of about 3,800 times, and the cumulative frequency curve also showed a sudden increase trend. Around April 1, that is, before the mine earthquake occurred, the microseism showed the characteristics of a "quiet period" and the activity of microseismic events weakened. Therefore, the time before the sudden increase of the microseismic cumulative frequency curve can be used as the prediction calculation cutoff time. Fig. 9 shown.

[0107] (4) Further calculate the deformation localization start time of the microseismic event, such as Fig.10 As shown, the index increased rapidly in the early stage, from negative to positive, until the 38th day during the monitoring period, when the index reached its peak, and then the value began to decline, indicating that the microseismic events had formed a stable spatial aggregation pattern on the 38th day, and deformation localization started at this time.

[0108] (5) The initial time t1 is the start time of deformation localization, i.e., 38 days. The initial time is kept unchanged, and the final deadline is kept consistent with the above, i.e., 123 days. The mine earthquake prediction time is calculated, and the time prediction model trajectory curve based on the cumulative frequency of microseismic events is obtained, as shown in Fig.11As shown in (a) in the figure. The prediction trajectory shows a monotonous upward trend in the initial stage, which is basically consistent with the evolution trend of the cumulative frequency curve. The prediction time increases slowly, and then the prediction time value increases greatly until the 162nd day. Then the prediction time trajectory begins to return continuously to the 97th day, and then continues to return to the 163rd day. The final predicted destruction time is 97 to 163 days. According to the time prediction trajectory, a comparison chart of the predicted time of the mine earthquake and the occurrence time of the mine earthquake is further obtained, as shown in Fig.11 As shown in (b) in the figure. From the above, we can see that the actual time of the mine earthquake is the 150th day, while the predicted time is 98 to 163 days. The error between the earliest predicted time and the actual damage time is 34.6%. If the 90% confidence interval of the actual damage time is defined, 60% of the predicted time points are within the predicted interval, and the last predicted point is exactly within the 90% confidence interval.

Claims

1. A mine earthquake time prediction method based on the localized state of rock deformation, characterized by: Step 1) Calculate the start time of deformation localization: The spatial distribution characteristics of microseismic events over time are quantitatively analyzed through spatial autocorrelation analysis, and the spatial aggregation index Moran's I is proposed to track the localization state of deformation. When Moran's I is close to 1, it is positive spatial autocorrelation and clustered distribution; when Moran's I is close to -1, it is negative spatial autocorrelation and discrete distribution; when Moran's I is equal to 0, there is no spatial autocorrelation and it is randomly distributed. Since the amount of microseismic data collected at the engineering site is large, it is necessary to judge the adjacency relationship of each element in the space, and n elements need to establish an n×n weight matrix. Therefore, the improved Moran's I is used for calculation. The improved algorithm formula is as follows: Where: n is the number of regional units; x i is the attribute observation value of regional unit i, W ij is the spatial weight coefficient. When regions i and j are adjacent, Wi j is equal to 1, when regions i and j are not adjacent, W ij is equal to 0; The improved method adopts the face adjacency model in the spatial model relationship. This method defines the elements with shared boundaries as neighbors, that is, the spatial proximity relationship is determined based on the face adjacency, and the W corresponding to each group (i, j) is calculated. ij When , it is directly substituted into the calculation without storage. In formula (2), N(i) is defined as follows: N(i)={(x i +1,and i ),(x i -1,and i ),(x i ,and i +1),(x i ,and i -1)} (3) The improved algorithm divides the data into regular grids, and a grid unit is adjacent to at most four other grid units. Therefore, when using this spatial model, it is only necessary to perform four adjacency judgments. The judgment range is shown in formula (3); Step 2) Build a time prediction model to obtain the rock instability failure time t r ; Where t1, t2, and t3 are the times corresponding to three consecutive points selected in the cumulative event curve.

2. The method for predicting the time of a mine earthquake based on the localized state of rock deformation according to claim 1 is characterized by: The specific method in step 2 is: The theoretically expanded mine earthquake time prediction model is established, and the initial formula is as follows: In the model of formula (4), t r is the rock instability failure time; is the instantaneous event rate, t r and The length of rock instability and failure time and the instantaneous event rate during the observation time are used to replace them, respectively, and we get where t r is the rock instability failure time, t is the real-time observation time, is the instantaneous event rate, a is a constant, assuming that at t = t0, have to here Then formula (7) can be rewritten as Among them, Δd is the difference between the relative event values ​​of two time points, and d0 is the event value corresponding to the real-time observation time. In actual prediction, the graphical method is used to calculate the rock instability failure time t in the observation time in combination with formula (9): r ; The prediction model formula contains four unknowns a, d0, t r and t0, more reference points need to be taken on the cumulative event curve for calculation. Three consecutive points D1(d1, t1), D2(d2, t2), and D3(d3, t3) are selected on the cumulative event curve. These three points must meet the following conditions: d2-d1=d3-d2 (10) Substituting the data satisfying the conditions of formula (10) into formula (9), we get three sets of simultaneous equations, which can be simplified to the following relationship: Simplified By moving the items, we can get the rock instability failure time.

Images