Real-time prediction method and device for coal rock failure based on acoustic emission signal analysis

Abstract

The application provides a coal rock damage real-time prediction method and device based on acoustic emission signal analysis, which comprises the following steps: in the case of performing coal rock damage test, obtaining statistical information of multiple acoustic emission signals; the statistical information comprises waveform data and the number of acoustic emission events; determining the fracture type of each acoustic emission signal according to the waveform data, and determining the b value and b absolute change rate of the value of each acoustic emission signal according to the waveform data and the number of acoustic emission events; determining the shear fracture proportion corresponding to the multiple acoustic emission signals according to the fracture type, and analyzing the statistical feature evolution information of the waveform data to obtain a target statistical feature; determining the warning threshold according to the b value, b absolute change rate of the value, the shear fracture proportion and the target statistical feature, and performing coal rock damage real-time prediction through the warning threshold. The method and device provided by the application can realize warning only by using the signals before the current time, and improve the accuracy and real-time performance of coal rock damage real-time prediction.

Description

Technical Field

[0001] This invention relates to the field of coal and rock sample testing technology, and in particular to a method and device for real-time prediction of coal and rock damage based on acoustic emission signal analysis. Background Technology

[0002] Prediction and early warning of coal and rock failure are of great significance for the prediction and forecasting of natural and engineering disasters. The realization of early warning of rock failure at the laboratory scale is the basis for prediction and early warning of rock instability at the engineering scale. Among them, acoustic emission signals originate from the internal failure of the loaded coal and rock mass and are an important source of early warning information for coal and rock failure.

[0003] In related technologies, when using acoustic emission characteristics for early warning, predictions are often made by counting acoustic emission rings or by using energy, based on whether acoustic emission has entered a quiet period or if acoustic emission has entered a quiet period. b Using a stable decrease in the value as an early warning signal, this type of method has a single judgment standard. In actual operation, the judgment results vary from person to person and are prone to human error. Moreover, the early warning methods of related technologies are usually based on data monitored throughout the entire process, resulting in poor real-time performance. Summary of the Invention

[0004] This invention provides a method and device for real-time prediction of coal and rock damage based on acoustic emission signal analysis, which solves the shortcomings of existing early warning methods, such as single discrimination criteria, large error, and poor real-time performance due to relying on data monitored throughout the loading process.

[0005] This invention provides a real-time prediction method for coal and rock failure based on acoustic emission signal analysis, comprising: In the case of performing a coal and rock failure test, statistical information of multiple acoustic emission signals is obtained; the statistical information includes waveform data and the number of acoustic emission events; The breakage type of each acoustic emission signal is determined based on the waveform data, and the number of acoustic emission events is determined based on the waveform data and the number of acoustic emission events. b Value and b The absolute rate of change of the value; the fracture type includes shear failure or tensile failure; The proportion of shear fracture corresponding to the plurality of acoustic emission signals is determined according to the fracture type, and the statistical characteristics of the waveform data are analyzed over time to obtain target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity. According to the above b value, b The absolute rate of change of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

[0006] According to the present invention, a real-time prediction method for coal and rock failure based on acoustic emission signal analysis is provided, wherein the method is based on the... b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics includes: Calculate the following based on the target feature function and the first window parameters. b value, b The absolute rate of change of value, the proportion of shear fracture, and the large window information corresponding to the target statistical features are used to obtain multiple large window features. The target feature function and the second window parameters are then used to calculate... b value, b Multiple small-window features are obtained by combining the absolute rate of change of the value, the proportion of shear fracture, and the small-window information corresponding to the target statistical features; the target feature function is shown in the following formula: ; in, F For the target feature function, X ( i ) is the first i Relevant data of acoustic emission signals X ( i +1) is the first i +1 related data of acoustic emission signals, the related data including b value, b The absolute rate of change of value, the proportion of shear fracture, and one of the target statistical characteristics; K is the weighting coefficient; Calculate the ratio between the mean values ​​corresponding to the features of the multiple small windows and the mean values ​​corresponding to the features of the multiple large windows to obtain the window-to-size ratio; The sum and warning threshold corresponding to the window size ratio are determined by the following formula: ; in, The ratio of the size window, the For the first i -1 All before The sum of values For the first i -1 Value before all Standard deviation of the value x The data values ​​are the relevant data.

[0007] According to the real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided by the present invention, after obtaining the window size ratio, the method further includes: Multiple window size ratios are normalized using a dimensionless method to obtain a sequence of window size ratios. The sequence of window size ratios includes multiple normalized window size ratios, and different normalized window size ratios correspond to different weighting coefficients. Calculate the Euclidean distance of the processed window size ratio sequence from the safety reference center to obtain the distance calculation result; The warning levels for different areas of coal and rock are determined based on the distance calculation results and the preset safe zone division threshold.

[0008] According to the present invention, a real-time prediction method for coal and rock failure based on acoustic emission signal analysis is provided, wherein the statistical information is the linear elastic stage data of the coal and rock failure test; According to the b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics also includes: Determine the upper and lower threshold values; the upper threshold value is the warning threshold corresponding to the relevant parameter with a sudden increase in value, and the lower threshold value is the warning threshold corresponding to the relevant parameter with a sudden decrease in value. The warning threshold is used to generate a warning signal when the numerical variation of the relevant parameters corresponding to the continuous time-series acoustic emission signals exceeds three times the standard deviation.

[0009] According to the present invention, a real-time prediction method for coal and rock failure based on acoustic emission signal analysis is provided, wherein the method is based on the... b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics also includes: Regarding the b value, b The parameter sequence corresponding to any one of the absolute change rate of value, shear fracture ratio, and target statistical features is averaged and smoothed to obtain a smoothed sequence; the parameter sequence includes the values ​​of the same parameter at different time series; The abruptness is calculated based on the rate of change of the slope of the curve corresponding to the smooth sequence, and the warning threshold is determined based on the abruptness.

[0010] According to the present invention, a real-time prediction method for coal and rock failure based on acoustic emission signal analysis is provided, wherein the method is based on the... b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics also includes: As stated b value, bTraining samples are constructed using the absolute change rate of values, the proportion of shear fracture, and the values ​​of the target statistical features at different time points. The number of clusters is set to n, and the K-Means clustering model is iteratively calculated to determine the safe cluster center and the dangerous cluster center, where n≥2 and n is a positive integer. The warning threshold is determined based on the midpoint of the Euclidean distance between the safe cluster center and the dangerous cluster center.

[0011] According to the present invention, a real-time prediction method for coal and rock damage based on acoustic emission signal analysis is provided, wherein the statistical information further includes the location information of acoustic emission events of different energies, and the acoustic emission events include high-energy acoustic emission events and low-energy acoustic emission events; After performing real-time prediction of coal and rock damage through the aforementioned early warning threshold, the method further includes: In the event of a high-energy acoustic emission event during a coal and rock failure test, a region where low-energy acoustic emission events accumulate is determined in the coal and rock based on the location information of the low-energy acoustic emission events. A line is drawn connecting the overlapping region of the high-energy acoustic emission events and the region where low-energy events accumulate to form the potential failure surface of the coal and rock.

[0012] The present invention also provides a real-time prediction device for coal and rock failure based on acoustic emission signal analysis, comprising: The information acquisition module is used to acquire statistical information of multiple acoustic emission signals during coal and rock destructive tests; the statistical information includes waveform data and the number of acoustic emission events. The first calculation module is used to determine the breakage type of each acoustic emission signal based on the waveform data, and to determine the number of acoustic emission events for each acoustic emission signal based on the waveform data and the number of acoustic emission events. b Value and b The absolute rate of change of the value; the fracture type includes shear failure or tensile failure; The second calculation module is used to determine the proportion of shear fracture corresponding to the plurality of acoustic emission signals according to the fracture type, and to analyze the statistical characteristics of the waveform data over time to obtain target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity. The early warning module is used to... b value, b The absolute rate of change of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the real-time prediction method for coal and rock damage based on acoustic emission signal analysis as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the real-time prediction method for coal and rock damage based on acoustic emission signal analysis as described above.

[0015] The present invention provides a real-time prediction and device for coal and rock failure based on acoustic emission signal analysis. This device acquires statistical information from multiple acoustic emission signals of coal and rock, and determines the magnitude of each acoustic emission signal based on waveform data and the number of acoustic emission events. b value, b The absolute rate of change and fracture type are determined, and then the proportion of shear fracture corresponding to multiple acoustic emission signals is determined based on the fracture type. The statistical characteristics of the waveform data over time are analyzed to obtain the target statistical characteristics. Finally, based on... b value, b The absolute change rate of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, thereby realizing real-time prediction of coal and rock failure. By analyzing the acoustic emission signals during the loading process in real time, abnormal signals before failure are found and the abnormal signals are used as early warning indicators. Moreover, the judgment of abnormal signals does not depend on the acoustic emission data of the entire loading process, but only on the signals before the current moment, which improves the accuracy and real-time performance of real-time prediction of coal and rock failure. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the real-time prediction method for coal and rock damage based on acoustic emission signal analysis provided by the present invention.

[0018] Figure 2 This is a schematic diagram illustrating the relevant parameters of the acoustic emission signal provided by the present invention.

[0019] Figure 3 This is a schematic diagram of the number of window samples and the sliding step size provided by the present invention.

[0020] Figure 4This is a schematic diagram of the structure of the real-time prediction device for coal and rock damage based on acoustic emission signal analysis provided by the present invention.

[0021] Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0023] The following is combined Figures 1-4 This invention describes a method and apparatus for real-time prediction of coal and rock damage based on acoustic emission signal analysis.

[0024] Figure 1 This is a flowchart illustrating the real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided by the present invention, as shown below. Figure 1 As shown, the method includes the following steps: Step 110: Under the condition of performing coal and rock destruction test, obtain statistical information of multiple acoustic emission signals; the statistical information includes waveform data and the number of acoustic emission events.

[0025] In this step, waveform data refers to the raw physical signals acquired in real time by an acoustic emission probe deployed on the surface of the coal and rock sample. Specifically, it includes: the amplitude of the acoustic emission waveform (in decibels dB), rise time (the time from the first time the waveform exceeds the threshold value to the peak value, in milliseconds ms), duration (the time interval between the first and last times the waveform exceeds the threshold value, in milliseconds ms), and ring count (the number of times the waveform exceeds the threshold value, dimensionless).

[0026] In this embodiment, the acoustic emission signal acquisition scenario is arranged through the following steps: A rectangular specimen is selected for mounting the acoustic emission probe and accurately positioning it. The specimen size can be determined based on the size of the acoustic emission probe and the loading capacity of the testing machine, and should not be less than 50×50×100mm. At least eight acoustic emission probes should be used (the number is not limited if the failure location is not predicted), evenly distributed on the specimen surface, with a coupling agent such as petroleum jelly applied between the probes and the specimen. Alternatively, the sample wave velocity can be measured using the built-in wave velocity testing function of the acoustic emission instrument, or other instruments can be used. The obtained wave velocity is used as the wave velocity for acoustic emission positioning. After the acoustic emission probes are accurately positioned, a lead is broken on the specimen surface to determine if the probe contact is good and if the probe placement is accurate. If inaccurate, adjustments should be made promptly.

[0027] In this embodiment, after the acoustic emission signal acquisition scenario is set, the loading can be achieved by displacement control or load control, and the sample is continuously loaded at a constant rate until it is destroyed. During the loading process, the received statistical information, such as the time, waveform data and positioning information of the acoustic emission signal, is recorded in real time. Subsequently, parameters such as acoustic emission amplitude, rise time, duration and ring count can be obtained based on the waveform data.

[0028] In this embodiment, the number of acoustic emission events refers to the total frequency of valid acoustic emission signals detected within a unit time window. Waveform data is used to extract the mechanical mechanism features of microfractures (such as fracture type and energy release intensity), and the number of acoustic emission events is used to characterize the intensity of microfracture activity. Together, they constitute the basic data source for subsequent multi-parameter analysis.

[0029] Figure 2 This is a schematic diagram illustrating the relevant parameters of the acoustic emission signal provided by the present invention. Figure 2 In the illustrated embodiment, the following relevant parameters of the current acoustic emission signal can be extracted from the waveform data: amplitude, rise time, duration, energy (calculated based on amplitude), ring count, and threshold voltage, etc.

[0030] Step 120: Determine the breakup type of each acoustic emission signal based on the waveform data, and determine the breakup characteristics of each acoustic emission signal based on the waveform data and the number of acoustic emission events. b Value and b The absolute rate of change of value; the type of fracture includes shear failure or tensile failure.

[0031] In this step, the amplitude, rise time, duration, and ring count of the acoustic emission signal can be calculated based on the waveform data; then, the amplitude and the number of acoustic emission events can be calculated... b Value, and according to b Information on how values ​​change over time is obtained. b The absolute rate of change of the value.

[0032] Specifically, b The value can be calculated using the following formula: ; in, A dB The amplitude of acoustic emission is expressed in decibels, i.e., the amplitude mentioned above. N acoustic emission amplitude at A dB The number of acoustic emission events is as follows; a and b It is a constant. b That is b The value and the goodness of fit are R 2 .

[0033] In this embodiment, when determining the calculation b Value window size L When selecting, you can choose based on the time step or a fixed number of acoustic emission events; if selecting based on the time step, the window size... L The choice is related to the loading rate; the slower the loading rate, the smaller the window... L The larger the interval, the better; regardless of whether it can be selected based on the time step or a fixed number of acoustic emission events, the interval should not be less than 100, but should not exceed 1000; the selection of the sliding step size l should be set to... b The value is calculated as 1 / 5 of the window L; there are advantages to choosing based on the time step and based on a fixed number of acoustic emission events, but the prediction effect of the calculation result is better when the window size is determined by a fixed number of acoustic emission events.

[0034] Figure 3 This is a schematic diagram of the number of window samples and the sliding step size provided by the present invention. Figure 3 In the embodiment shown, b The evolution of the value over loading time is achieved through the following steps: ① Divide the collected data into several groups, each group containing L A single acoustic emission event. L That is, the number of samples in the window; ② Determine L an event b Value, Result b The value corresponds to the time of the last acoustic emission event in that group; ③ Slide to the right l From the acoustic emission data, a new set of data is obtained, and the data volume remains the same. L Repeat step ② to obtain a new group. L Data b The value corresponds to the time of the last sound emission event in the new set; ④ And so on, until the last group. L The data obtained b The sequence of values ​​over time is b The value is determined as the loading process evolves; it should be noted that the last set may not be sufficient. L Use as many data points as you have.

[0035] In this embodiment, b The value exhibits strong volatility, but near the point of destruction, b The frequency and amplitude of the value fluctuations are significantly enhanced. This embodiment achieves this through the following formula per unit time. b rate of change of value : ; in, For the first unit of time i +1 b value, For the first unit of time i indivual b value, for and The difference between corresponding times; the difference within a unit of time. b absolute rate of change of value That is, all within a unit of time. b The sum of the absolute values ​​of the slopes of the changes in the values.

[0036] In this embodiment, the unit time can be 1 second to achieve real-time early warning at the second level, or the unit time can be set to a longer time period.

[0037] In this embodiment, source mechanism analysis is performed on the acoustic emission signals collected from the pre-experimental sample to determine the fracture type of each signal, that is, to determine whether each signal is shear failure or tensile failure.

[0038] In this embodiment, the break type can be determined by the ratio of the RA value (rise time to amplitude, RT / A, in ms / dB) and the AF value (ring count to duration, RC / DT, in dimensionless / ms) of the acoustic emission waveform.

[0039] Specifically, the shear fracture ratio is determined through the following steps: (1) The RA value is calculated based on the rise time and amplitude of the acoustic emission signal, and is expressed by the following formula: ; Where RT is the rise time in milliseconds (ms), and A is the amplitude of the acoustic emission waveform in dB. (2) The AF value is calculated based on the ring count data (the number of times the waveform exceeds the threshold value) and the duration, and is expressed by the following formula: ; Where RC is the ring count data; DT is the duration, which is the time interval between the first and last time the waveform exceeds the threshold value, in milliseconds; (3) Calculate the ratio k of RA to FA. When the ratio k is less than the critical value, it means that the rupture that releases the acoustic emission signal is a tension rupture; when the ratio k is greater than the critical value, it means that the rupture that releases the acoustic emission signal is a shear rupture.

[0040] Specifically, first calculate the k value of each acoustic emission signal, find the minimum k value corresponding to tensile failure and the maximum k value corresponding to shear failure, and denote them as k. 拉 and k 剪 Then, calculate the k value for each acoustic emission signal in chronological order. When k is less than k... 拉 When k is greater than k, it is recorded as tensile failure; 剪 When the time is reached, it is recorded as shear failure, thereby determining the fracture type of each acoustic emission signal.

[0041] Step 130: Determine the proportion of shear fracture corresponding to multiple acoustic emission signals according to the fracture type, and analyze the statistical characteristics of the waveform data over time to obtain the target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity.

[0042] In this step, the shear fracture ratio information refers to the percentage of shear failure events out of the total number of events within the sliding time window. This parameter reflects the process of the macroscopic failure of coal and rock shifting from tension-dominated to shear-dominated. A sudden increase in the ratio indicates that a through-shear zone is about to form.

[0043] In this embodiment, the variance of the waveform data is calculated by the squared average of the deviations of the ringing count data of the first N acoustic emission events from their mean. This variance is used to measure the dispersion of the intensity of micro-fracture activity. A sharp increase in variance indicates that the fracture activity has lost its uniformity and entered a localized concentration stage.

[0044] In this embodiment, the autocorrelation coefficient is a statistic used to evaluate the autocorrelation of the time series of acoustic emission parameters. A decrease in the autocorrelation coefficient after a lag of k steps indicates that the system's memory is weakened and its randomness is enhanced, which is an important indicator of instability precursors.

[0045] In this embodiment, the activity level is represented by the ratio of the average amplitude of the first N events to the maximum amplitude, which is used to normalize and characterize the overall activity level of micro-fractures in the current period. The continuous increase in activity level reflects the accelerated release of energy.

[0046] In this embodiment, after determining the fracture type of each acoustic emission signal, the number of acoustic emission signals caused by shear failure during the coal and rock failure test is recorded, and the ratio of shear fracture to the total number of fractures is calculated to obtain the proportion of shear fracture.

[0047] In this embodiment, the variances of parameters such as ring count data, acoustic emission energy (determined based on amplitude), RA, and AF at any given time are calculated using the following formula: ; in, D For the front N The variance of the data; S For the front N The standard deviation of each data point; For the first i One data value; For the front N The mean of the data; In this embodiment, the time corresponding to the calculated variance is the [number]th [time]. N By analyzing the time of each data point, a dataset is obtained showing the evolution of the variance of parameters such as ring count, acoustic emission energy (determined based on amplitude), RA, and AF at any given moment over time. This dataset includes the values ​​of the variance of each parameter at different time points.

[0048] In this embodiment, the autocorrelation coefficients of the variances of parameters such as ring count data, acoustic emission energy (determined based on amplitude), RA, and AF at any given time are calculated using the following formula: ; in, S For the front N The standard deviation of each data point; For the first i One data value; For the front N The mean of the data points, k is the number of lag steps, and the value of k is positively correlated with the amount of acoustic emission data.

[0049] In this embodiment, the time corresponding to the calculated autocorrelation coefficient is the [number]th [time]. N The time of each data point is used to obtain a dataset that calculates the time evolution of the autocorrelation coefficients of parameters such as ring count data, acoustic emission energy (determined based on amplitude), RA, and AF at any given time, including the values ​​of the autocorrelation coefficients of each parameter at different time points.

[0050] In this embodiment, activity S The activity level comprehensively considers the temporal, spatial, and intensity factors of internal rock fracture. It is a physical quantity reflecting the concentration and energy scale of acoustic emission sources within the rock mass, composed of the frequency, average energy level, and maximum energy level of acoustic emission events. SIt is calculated by the following formula: ; in, m i The amplitude of each acoustic emission waveform; m N For the front N The largest value among the data.

[0051] In this embodiment, the calculated activity level S The corresponding time is the N The time of each data point is obtained. S The dataset evolves over time, including the activity values ​​of each parameter at different points in time.

[0052] Step 140, according to b value, b The absolute change rate of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

[0053] In this step, due to the acoustic emission of coal and rock before destruction... b The value dropped suddenly. b The absolute rate of change of values ​​and the proportion of shear cracks surge; in addition, since the variance, autocorrelation coefficient and activity of parameters such as ring count data, acoustic emission energy (determined based on amplitude), RA and AF will also change abruptly before failure, this embodiment can set corresponding early warning thresholds for different parameters, or determine a unique early warning threshold for each parameter, so as to realize real-time prediction of coal and rock failure based on acoustic emission signal analysis. That is, the key to this embodiment is to find the mutation point (corresponding to the early warning threshold), and the determination of the mutation point must be unique.

[0054] Based on this, this embodiment can determine the warning threshold through various methods, such as the window size ratio method, the statistical deviation method based on the 3σ criterion of the stable period, the geometric inflection point method based on the sudden change of tangent slope, and the adaptive boundary method based on K-Means clustering.

[0055] In this embodiment, a multi-indicator fusion early warning strategy can also be used to predict the early warning threshold, thereby determining the early warning level corresponding to different coal and rock regions.

[0056] The real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided in this invention obtains statistical information of multiple acoustic emission signals of coal and rock, and determines the value of each acoustic emission signal based on waveform data and the number of acoustic emission events. b value, bThe system analyzes the absolute rate of change of b-values ​​and fracture types, then determines the proportion of shear fractures corresponding to multiple acoustic emission signals based on the fracture types, and analyzes the statistical characteristics of waveform data over time to obtain target statistical characteristics. Finally, it determines the warning threshold based on the b-value, the absolute rate of change of b-values, the proportion of shear fractures, and the target statistical characteristics, thereby achieving real-time prediction of coal and rock failure. By analyzing the acoustic emission signals during the loading process in real time, abnormal signals before failure are found, and these abnormal signals are used as warning indicators. Moreover, the identification of abnormal signals does not depend on the acoustic emission data throughout the loading process; only signals before the current moment are needed to achieve warning, thus improving the accuracy and real-time performance of real-time prediction of coal and rock failure.

[0057] In some embodiments, according to the real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided by the present invention, based on... b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the target statistical characteristics includes: calculating based on the target characteristic function and the first window parameters. b value, b The absolute change rate of the value, the proportion of shear fracture, and the target statistical characteristics are each represented by a large window of information, resulting in multiple large window features. These features are then calculated based on the target feature function and the parameters of the second window. b value, b The absolute rate of change of value, the proportion of shear fracture, and the target statistical features are each represented by a small window, resulting in multiple small window features; the target feature function is shown in the following formula: ; in, F For the target feature function, X ( i ) is the first i Relevant data of acoustic emission signals X ( i +1) is the first i +1 related data for acoustic emission signals, including related data b value, b One of the following: absolute rate of change of value, proportion of shear fracture, and target statistical characteristics; K The weighting coefficient is used; the ratio between the mean values ​​corresponding to multiple small window features and the mean values ​​corresponding to multiple large window features is calculated to obtain the small window ratio. The sum and warning threshold corresponding to the ratio of large and small windows are determined by the following formula: ; in, The ratio of the size window, For the first i -1 All before The sum of values For the firsti -1 Value before all Standard deviation of the value x The data values ​​are the relevant data.

[0058] In this embodiment, the abrupt change points of various acoustic emission signals are determined according to the window size ratio method, which is specifically achieved through the following steps: (1) Determine the data volume of the large and small windows as Nbig and Nsmall, respectively. Nbig should not be less than 100, but should not exceed 1000. Set Nsmall to 1 / 5 of Nbig. (2) Calculate the large window information corresponding to each parameter through the target feature function; where, K Calculated using the following formula: ; In the target feature function, It can reflect the amplitude of signal fluctuations. It can reflect the frequency of signal vibration, and can capture both amplitude signal changes and frequency signal changes; (3) Calculate the small window information corresponding to each parameter through the target feature function, and calculate the mean value corresponding to multiple small window features and the mean value corresponding to multiple large window features; wherein, the mean value corresponding to multiple large window features is expressed by the following formula: ; The mean of multiple small window features is represented by the following formula: ; (4) The ratio of the large window to the small window is obtained by the following formula. : ; (5) The conditions for triggering the early warning are set as follows: ; in, For the first i indivual value.

[0059] The real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided in this invention embodiment uses the large and small window ratio method according to... b value, b The absolute change rate of values, the proportion of shear fracture, and the statistical characteristics of the target are used to determine the early warning threshold, which can accurately quantify the abrupt change points of various acoustic emission signals, thereby improving the accuracy of early warning of coal and rock damage.

[0060] In some embodiments, after obtaining the size window ratio, the real-time prediction method for coal and rock damage based on acoustic emission signal analysis further includes: performing dimensionless normalization on multiple size window ratios to obtain a processed size window ratio sequence, wherein the size window ratio sequence includes multiple normalized size window ratios, and different normalized size window ratios correspond to different weighting coefficients; calculating the Euclidean distance of the processed size window ratio sequence from the safety reference center to obtain the distance calculation result; and determining the early warning level of different coal and rock regions based on the distance calculation result and a preset safety region division threshold.

[0061] In this embodiment, a vector modulus analysis method based on multidimensional feature space is used to expand the "point-like" early warning of a single indicator into a comprehensive "area-like" early warning in multidimensional space. The specific implementation steps are as follows: (1) Construct a multidimensional early warning feature vector, that is, define various acoustic emission parameters as the dimensions of the early warning feature space.

[0062] For example, select b value, b′ Value, shear crack ratio, ringing count data, variance, autocorrelation coefficient r, and activity. S A total of 6 parameters are used as feature components. At any time t, according to the method described in Section 6, the window size ratios (feature function ratios) corresponding to the above 6 parameters have been calculated and denoted as α1, α2, ..., α6. The multidimensional feature state vector V(t) at time t is constructed as follows. .

[0063] (2) Perform dimensionless normalization on the above multidimensional feature state vectors.

[0064] Due to different parameters α Values ​​may fluctuate within different ranges (e.g., activity level). α The value could be as high as 10, and b Value α The value may be only 2. To eliminate the influence of dimensions, a weighting coefficient needs to be introduced. Make corrections.

[0065] The weight coefficient of the i-th parameter The coefficient of variation method is used to determine this, specifically by calculating the following formula: ; in, α is the coefficient of variation (standard deviation divided by mean) of the sequence of values ​​of the i-th parameter α in the preliminary experimental stage. α is the coefficient of variation of the j-th parameter α value sequence in the preliminary experimental phase, used to reflect the sensitivity to mutations. The higher the sensitivity, the greater the weight.

[0066] (3) Calculate the comprehensive early warning modulus M(t); In the multidimensional feature space, when the coal and rock mass is in the stable loading stage, the ratio of the size windows α of each parameter theoretically approaches 1; Therefore, in this embodiment, the point (1,1,1,1,1,1) is first defined as the safety reference center, and then the Euclidean distance of the current state vector V(t) from the safety reference center is calculated, which is the comprehensive early warning modulus M(t), specifically calculated by the following formula: ; in, Let αi be the ratio of the size window of the i-th parameter at time t. If a parameter αi < 1 (indicating that the signal is below the background value and there is no abnormality), then its term is set to 0 in the calculation, and only positive abnormal deviations are accumulated.

[0067] (4) Determine the fusion warning level and threshold; In this embodiment, the larger the value of M(t), the further the system deviates from the stable state and the higher the risk of damage. Through backtracking analysis of the pre-experimental data, the critical threshold M(t) is determined. cr .

[0068] In this embodiment, the following graded early warning criteria may be adopted: Blue safe zone: M(t)≤M cr This indicates that the micro-fractures inside the coal and rock mass are within a range of random fluctuations. Yellow Alert Zone: M cr <M(t)≤1.5M cr This indicates that multiple indicators have deviated and micro-fractures have begun to cluster, requiring a reduction in the loading rate or enhanced monitoring. Red destruction zone: M(t) > 1.5M cr This indicates a sudden change in the eigenvector modulus, signifying that the coal and rock mass is about to undergo macroscopic instability and failure, immediately triggering a shutdown or evacuation signal.

[0069] In this embodiment, M cr The value is taken as 3 times the mean of M(t) throughout the entire preliminary experiment.

[0070] The real-time prediction method for coal and rock damage based on acoustic emission signal analysis provided in this invention successfully integrates six heterogeneous parameters into a single risk metric M(t) by constructing an Euclidean distance model based on the coefficient of variation weighting. Furthermore, it provides clear operational guidance through multi-level early warning classification, enabling on-site personnel to make accurate responses based on color levels without needing to understand complex parameters, thereby improving the practicality and reliability of real-time prediction of coal and rock damage.

[0071] In some embodiments, the statistical information is linear elastic stage data from coal and rock failure tests; according to b value, bThe determination of the warning threshold based on the absolute rate of change of value, the proportion of shear fracture, and the statistical characteristics of the target also includes: determining the upper limit threshold and the lower limit threshold; the upper limit threshold is the warning threshold corresponding to the relevant parameter with a sudden increase in value, and the lower limit threshold is the warning threshold corresponding to the relevant parameter with a sudden decrease in value; the warning threshold is used to generate a warning signal when the numerical change of the relevant parameter corresponding to the continuous time series acoustic emission signal exceeds 3 times the standard deviation.

[0072] In this embodiment, the linear elastic stage data refers to the statistical information of acoustic emission signals collected when the coal and rock sample is in the elastic deformation stage at the initial stage of loading. It corresponds to the 20%~50% time interval or the 20%~40% peak load interval of the entire loading process. During this stage, the micro-fracture activity is in a random and stable state and can be used as a background noise benchmark.

[0073] In this embodiment, the upper limit threshold applies to parameters with sudden numerical increases, including b The absolute rate of change of values, the proportion of shear cracks, variance, and activity are all considered. An early warning is triggered when the monitored values ​​of these parameters at multiple consecutive points exceed the upper threshold. The lower threshold is applicable to parameters with sudden numerical decreases, primarily... b The values ​​and autocorrelation coefficients are used to trigger an alert when these parameters continuously fall below the lower limit threshold.

[0074] In this embodiment, the 3-times standard deviation is determined based on the mean μ of the linear elastic stage. ref Based on μ ref ±3σ ref As a statistical boundary for anomaly detection, this method is based on the principle of normal distribution to ensure that the early warning trigger has statistical significance and avoid false alarms caused by random fluctuations.

[0075] In this embodiment, the initial elastic phase of loading is defined as the background noise region, and a baseline statistic is calculated based on this. Subsequent signals deviating from the baseline by more than three standard deviations are considered abnormal.

[0076] Specifically, linear elastic stage data of the coal and rock loading process (e.g., the 20%~50% range of loading time) is extracted as a reference dataset R; the acoustic emission parameters in the reference dataset R are calculated. b Value, variance D, activity S The mean μ of (etc.) ref and standard deviation σ ref For parameters with sudden increases in value (such as activity level) S (Shear crack ratio, variance D), with an upper limit threshold T. upper =μ ref +3σ ref For parameters with a sudden drop in value (such as...) b (Value), set the lower limit threshold T lower =μ ref -3σref Then monitor subsequent data X(t) in real time. If m consecutive points (m≥5 recommended) satisfy X(t)>T upper , or X(t) <T lower If the threshold is triggered, a corresponding warning signal will be generated.

[0077] The real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided in this invention generates an early warning signal when the numerical variation of relevant parameters corresponding to continuous time-series acoustic emission signals exceeds three times the standard deviation by determining upper and lower thresholds. This method relies only on historical data before triggering, meets the requirements for real-time early warning, and sets a discrimination boundary based on statistical significance level, so that the time relationship between the early warning signal and the failure time can be deterministically estimated through statistical analysis of pre-experimental data, thereby improving the reliability and engineering applicability of the early warning signal.

[0078] In some embodiments, according to b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the target statistical characteristics also includes: b value, b The parameter sequence corresponding to any one of the absolute change rate of value, shear fracture ratio, and target statistical characteristics is averaged and smoothed to obtain a smoothed sequence; the parameter sequence includes the values ​​of the same parameter in different time series; the abruptness is calculated based on the rate of change of the slope of the curve corresponding to the smoothed sequence, and the warning threshold is determined based on the abruptness.

[0079] In this embodiment, the real-time acquired acoustic emission parameter sequence is subjected to moving average filtering through average smoothing to eliminate random fluctuations and measurement noise in the original data, preserve its macroscopic evolution trend, and ensure the stability of subsequent slope calculation.

[0080] In this embodiment, the parameter sequence includes the same acoustic emission characteristic parameter (such as...). b Value, variance D, activity S Time series data composed of data at different loading times (etc.).

[0081] In this embodiment, the rate of change of the curve slope can be determined by the difference of the first derivatives of the parametric curves, which is essentially the second derivative or curvature of the curve, and is used to quantify the degree of deflection of the curve tangent angle.

[0082] In this embodiment, the mutation degree is the absolute value of the slope difference between adjacent time steps, which is used to characterize the geometric feature of the parameter evolution curve showing a "sharp turn". The corresponding warning threshold can be taken as 60% to 80% of the maximum mutation degree in the entire loading process. When the real-time mutation degree exceeds the threshold, it is determined that the parameter has entered the nonlinear instability stage and a warning is triggered.

[0083] Specifically, when rocks are nearing failure, their parameter evolution curves often exhibit a "sharp inflection point." This embodiment uses the rate of change of the curve's tangent slope (i.e., curvature or second derivative) to capture the curve's "inflection point" as a warning sign. The specific implementation process is as follows: (1) Due to the large fluctuation of acoustic emission data, the parameter sequence X(t) acquired in real time is first smoothed by moving average to obtain the smoothed sequence Xˉ(t).

[0084] (2) Calculate the slope k(t) of the change between the current time t and the previous time t-Δt using the following formula: ; (3) The absolute value of the difference between the slopes of adjacent time steps, i.e., the degree of abrupt change J(t), is calculated by the following formula: ; (4) Determine the critical angle and set the slope change threshold Jcr. This can be understood as the curve tangent angle has changed significantly. The value of Jcr is 60% to 80% of the maximum slope change value during the entire loading process.

[0085] (5) Trigger judgment: When J(t)>Jcr, it means that the parameter curve has a sharp "rise" or "plunge", which is confirmed as a sudden change point and triggers an early warning.

[0086] The real-time prediction method for coal and rock damage based on acoustic emission signal analysis provided in this invention directly identifies the geometric inflection point of coal and rock transitioning from stable damage to accelerated failure by capturing the abrupt change characteristics of the slope of the tangent line of the parameter evolution curve. Compared with statistical methods, this method is more sensitive to abrupt changes in the nonlinearity of parameters and does not require the assumption that the data follows a normal distribution. It is applicable to non-stationary acoustic emission sequences, enhances the robustness and adaptability of the early warning mechanism in complex working conditions, and ensures that early warning can be triggered without lag when the shape of the parameter curve undergoes a fundamental change.

[0087] In some embodiments, according to b value, b The determination of early warning thresholds based on absolute change rate of value, proportion of shear fracture, and target statistical characteristics also includes: b value, b Training samples were constructed using the values ​​of absolute change rate, shear fracture ratio, and target statistical characteristics at different time points. The number of clusters was set to n, and the K-Means clustering model was iteratively calculated to determine the safe cluster center and the dangerous cluster center, where n≥2 and n is a positive integer. The warning threshold was determined based on the midpoint of the Euclidean distance between the safe cluster center and the dangerous cluster center.

[0088] In this embodiment, the training samples may include six acoustic emission feature parameters ( b value, b'Value, shear fracture ratio, variance D, autocorrelation coefficient r, activity S The values ​​at the same time are combined into a 6-dimensional feature vector, which is used as the input sample points for the K-Means clustering algorithm.

[0089] In this embodiment, all samples are automatically divided into two categories: "safe cluster" and "dangerous cluster," which correspond to the linear elastic stability stage and the nonlinear instability precursor stage in the coal and rock loading process, respectively, without the need for manual pre-setting of classification labels.

[0090] In this embodiment, the safe cluster center Csafe is a data center representing a stable state determined after the algorithm iterative convergence, located in a region of low value and dense samples in the feature space; the dangerous cluster center Cdanger represents an unstable state, located in a region of high value and discrete samples.

[0091] In this embodiment, the midpoint of the line connecting the two center points can be taken as the classification boundary between the two clusters. This position is the adaptively determined warning threshold surface. When a new data point is located on the boundary near the dangerous cluster side, a warning is triggered.

[0092] Specifically, unsupervised machine learning algorithms can be used to automatically divide data points into "safe clusters" and "dangerous clusters" without the need for manual setting of specific values ​​(such as greater than 100 or less than 0.5). The algorithm will automatically find the boundary between the two types of data. The steps for determining the warning threshold using the K-Means clustering method are as follows: (1) Constructing the feature space: Select N acoustic emission parameters (e.g., b value, S The real-time data vector consists of values, variance, etc.

[0093] (2) Real-time clustering analysis: All data points from the start of loading to the current time t are used as the sample set, and the number of clusters is set to k=n (for example, n=2, which represent "stable state" and "instability precursors" respectively). The K-Means algorithm is used for iterative calculation.

[0094] (3) Calculate the cluster center distance: The algorithm will output two cluster centers, namely Csafe (safe center, usually corresponding to the area with large data volume and low value) and Cdanger (danger center, usually corresponding to the area with high value and discrete).

[0095] (4) Set the discrimination boundary: take the midpoint of the Euclidean distance between the two cluster centers as the dividing hyperplane (threshold surface).

[0096] (5) Warning trigger judgment: When the latest collected data point is classified into the Cdanger cluster by the algorithm, and the distance of the data point from Csafe exceeds twice the average distance within the Csafe cluster (warning threshold), a warning is triggered.

[0097] In this embodiment, n can also take the values ​​3, 4, 5, etc. That is, by setting the number of clusters in the K-Means clustering model to n (n≥2), multiple early warning levels (such as safety, alert, damage, etc.) can be provided for the real-time prediction of current coal and rock damage.

[0098] The real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided in this invention embodiment, through... b value, b Training samples were constructed using the values ​​of absolute change rate, shear fracture ratio, and target statistical features at different time points. The number of clusters was set to 2, and the K-Means clustering model was iteratively calculated to determine the safe cluster center and the dangerous cluster center. The warning threshold was determined based on the midpoint of the Euclidean distance between the safe cluster center and the dangerous cluster center. This method automatically learns the inherent distribution pattern of safe and dangerous states from historical data through unsupervised machine learning algorithms, without the need for manual setting of specific numerical thresholds. It can intelligently identify complex patterns of instability precursors from high-dimensional feature space, and is particularly sensitive to nonlinear interactions between parameters, significantly improving the adaptability and universality of the warning mechanism.

[0099] In some embodiments, the statistical information also includes the location information of acoustic emission events of different energies, including high-energy acoustic emission events and low-energy acoustic emission events; after performing real-time prediction of coal and rock failure through an early warning threshold, the real-time prediction method of coal and rock failure based on acoustic emission signal analysis further includes: in the case of high-energy acoustic emission events occurring in the coal and rock failure test, determining the low-energy event clustering area in the coal and rock according to the location information of the low-energy acoustic emission events, and connecting the overlapping part of the high-energy acoustic emission event corresponding area and the low-energy event clustering area as the potential failure surface of the coal and rock.

[0100] In this embodiment, the location information of the acoustic emission event can be determined by the time difference of signals received by multiple (at least 8) probes arranged on the sample surface, combined with the pre-determined wave velocity, to calculate the three-dimensional spatial coordinates (x, y, z) of the acoustic emission source inside the sample.

[0101] In this embodiment, high-energy and low-energy acoustic emission events are classified based on relative thresholds from pre-experimental or historical data. High energy typically corresponds to the penetration or propagation of macroscopic cracks, while low energy typically corresponds to the initiation or friction of microcracks. The potential failure surface is the specific spatial geometric location at which macroscopic fracturing or shear slip is predicted to occur in the coal and rock mass.

[0102] Specifically, after releasing the warning signal, a dataset of acoustic emission event locations is collected, and high-energy events are selected (e.g., events greater than 1000 aJ are selected as the threshold for high-energy events; the threshold value is selected with reference to the maximum energy collected during the pre-experiment, and it is recommended to be 2-4 orders of magnitude lower than the maximum energy). The high-energy events are then selected by time period, the length of which is determined based on the loading time during the pre-experiment, and it is recommended to take 1 / 10 or 5 seconds of the loading time. If high-energy acoustic emission events cluster in a certain time period after the warning signal is released, the cluster area of ​​high-energy events is the high-risk area for damage. The cluster area of ​​low-energy events in this time period is marked, and the line connecting each cluster area of ​​high-energy events through the cluster area of ​​low-energy events is the potential damage surface.

[0103] The real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided in this invention determines the clustering area of ​​low-energy acoustic emission events in coal and rock by using the location information of low-energy acoustic emission events. The overlapping part of the high-energy acoustic emission event region and the low-energy event clustering region is connected to form the potential failure surface of coal and rock. This method can accurately delineate the shear zone or tensile fracture surface that is forming inside the rock, which helps engineers to take more targeted reinforcement or risk avoidance measures.

[0104] The real-time prediction device for coal and rock damage based on acoustic emission signal analysis provided by the present invention will be described below. The real-time prediction device for coal and rock damage based on acoustic emission signal analysis described below can be referred to in correspondence with the real-time prediction method for coal and rock damage based on acoustic emission signal analysis described above.

[0105] Figure 4 This is a schematic diagram of the real-time prediction device for coal and rock failure based on acoustic emission signal analysis provided by the present invention, as shown below. Figure 4 As shown, the real-time prediction device for coal and rock damage based on acoustic emission signal analysis includes: an information acquisition module 410, a first calculation module 420, a second calculation module 430, and an early warning module 440.

[0106] The information acquisition module 410 is used to acquire statistical information of multiple acoustic emission signals during the coal and rock failure test; the statistical information includes waveform data and the number of acoustic emission events. The first calculation module 420 is used to determine the breakage type of each acoustic emission signal based on the waveform data, and to determine the breakage rate of each acoustic emission signal based on the waveform data and the number of acoustic emission events. b Value and b The absolute rate of change of value; the fracture type includes shear failure or tensile failure; The second calculation module 430 is used to determine the proportion of shear fracture corresponding to multiple acoustic emission signals according to the fracture type, and to analyze the statistical characteristics of the waveform data over time to obtain the target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity. Early warning module 440, used to... b value, b The absolute change rate of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

[0107] The real-time prediction device for coal and rock failure based on acoustic emission signal analysis provided in this invention acquires statistical information of multiple acoustic emission signals from coal and rock, and determines the magnitude of each acoustic emission signal based on waveform data and the number of acoustic emission events. b value, b The system analyzes the absolute rate of change of b-values ​​and fracture types, then determines the proportion of shear fractures corresponding to multiple acoustic emission signals based on the fracture types, and analyzes the statistical characteristics of waveform data over time to obtain target statistical characteristics. Finally, it determines the warning threshold based on the b-value, the absolute rate of change of b-values, the proportion of shear fractures, and the target statistical characteristics, thereby achieving real-time prediction of coal and rock failure. By analyzing the acoustic emission signals during the loading process in real time, abnormal signals before failure are found, and these abnormal signals are used as warning indicators. Moreover, the identification of abnormal signals does not depend on the acoustic emission data throughout the loading process; only signals before the current moment are needed to achieve warning, thus improving the accuracy and real-time performance of real-time prediction of coal and rock failure.

[0108] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call logical instructions in the memory 530 to execute a real-time prediction method for coal and rock failure based on acoustic emission signal analysis. This method includes: acquiring statistical information of multiple acoustic emission signals during a coal and rock failure test; the statistical information includes waveform data and the number of acoustic emission events; determining the failure type of each acoustic emission signal based on the waveform data; and determining the failure rate of each acoustic emission signal based on the waveform data and the number of acoustic emission events. b Value and bThe absolute rate of change of the value; the fracture type includes shear failure or tensile failure; the proportion of shear failure corresponding to multiple acoustic emission signals is determined according to the fracture type, and the statistical characteristics of the waveform data are analyzed over time to obtain the target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity; according to b value, b The absolute change rate of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

[0109] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0110] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program performs the real-time prediction method for coal and rock failure based on acoustic emission signal analysis provided by the methods described above. This method includes: acquiring statistical information of multiple acoustic emission signals during a coal and rock failure test; the statistical information includes waveform data and the number of acoustic emission events; determining the fracture type of each acoustic emission signal based on the waveform data; and determining the fracture type of each acoustic emission signal based on the waveform data and the number of acoustic emission events. b Value and b The absolute rate of change of the value; the fracture type includes shear failure or tensile failure; the proportion of shear failure corresponding to multiple acoustic emission signals is determined according to the fracture type, and the statistical characteristics of the waveform data are analyzed over time to obtain the target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity; according to b value, b The absolute change rate of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

[0111] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0112] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0113] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A real-time prediction method for coal and rock failure based on acoustic emission signal analysis, characterized in that, include: Statistical information on multiple acoustic emission signals was obtained during coal and rock failure tests. The statistical information includes waveform data and the number of acoustic emission events; The breakage type of each acoustic emission signal is determined based on the waveform data, and the number of acoustic emission events is determined based on the waveform data and the number of acoustic emission events. b Value and b The absolute rate of change of the value; the fracture type includes shear failure or tensile failure; The proportion of shear fracture corresponding to the plurality of acoustic emission signals is determined according to the fracture type, and the statistical characteristics of the waveform data are analyzed over time to obtain target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity. According to the above b value, b The absolute rate of change of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

2. The real-time prediction method for coal and rock failure based on acoustic emission signal analysis according to claim 1, characterized in that, According to the b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics includes: Calculate the following based on the target feature function and the first window parameters. b value, b The absolute rate of change of value, the proportion of shear fracture, and the large window information corresponding to the target statistical features are used to obtain multiple large window features. The target feature function and the second window parameters are then used to calculate... b value, b Multiple small-window features are obtained by combining the absolute rate of change of the value, the proportion of shear fracture, and the small-window information corresponding to the target statistical features; the target feature function is shown in the following formula: ； in, F For the target feature function, X ( i ) is the first i Relevant data of acoustic emission signals X ( i +1) is the first i +1 related data of acoustic emission signals, the related data including b value, b The absolute rate of change of value, the proportion of shear fracture, and one of the target statistical characteristics; K is the weighting coefficient; Calculate the ratio between the mean values ​​corresponding to the features of the multiple small windows and the mean values ​​corresponding to the features of the multiple large windows to obtain the window-to-size ratio; The sum and warning threshold corresponding to the window size ratio are determined by the following formula: ； in, The ratio of the size window, the For the first i -1 All before The sum of values For the first i -1 Value before all Standard deviation of the value x The data values ​​are the relevant data.

3. The real-time prediction method for coal and rock failure based on acoustic emission signal analysis according to claim 2, characterized in that, After obtaining the window size ratio, the method further includes: Multiple window size ratios are normalized using a dimensionless method to obtain a sequence of window size ratios. The sequence of window size ratios includes multiple normalized window size ratios, and different normalized window size ratios correspond to different weighting coefficients. Calculate the Euclidean distance of the processed window size ratio sequence from the safety reference center to obtain the distance calculation result; The warning levels for different areas of coal and rock are determined based on the distance calculation results and the preset safe zone division threshold.

4. The real-time prediction method for coal and rock failure based on acoustic emission signal analysis according to claim 1, characterized in that, The statistical information refers to the linear elastic stage data of the coal and rock failure test; According to the b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics also includes: Determine the upper and lower threshold values; the upper threshold value is the warning threshold corresponding to the relevant parameter with a sudden increase in value, and the lower threshold value is the warning threshold corresponding to the relevant parameter with a sudden decrease in value. The warning threshold is used to generate a warning signal when the numerical variation of the relevant parameters corresponding to the continuous time-series acoustic emission signals exceeds three times the standard deviation.

5. The real-time prediction method for coal and rock failure based on acoustic emission signal analysis according to claim 1, characterized in that, According to the b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics also includes: Regarding the b value, b The parameter sequence corresponding to any one of the absolute change rate of value, shear fracture ratio, and target statistical features is averaged and smoothed to obtain a smoothed sequence; the parameter sequence includes the values ​​of the same parameter at different time series; The abruptness is calculated based on the rate of change of the slope of the curve corresponding to the smooth sequence, and the warning threshold is determined based on the abruptness.

6. The real-time prediction method for coal and rock failure based on acoustic emission signal analysis according to claim 1, characterized in that, According to the b value, b The determination of the early warning threshold based on the absolute rate of change of the value, the proportion of shear fracture, and the aforementioned target statistical characteristics also includes: As stated b value, b Training samples are constructed using the absolute change rate of values, the proportion of shear fracture, and the values ​​of the target statistical features at different time points. The number of clusters is set to n, and the K-Means clustering model is iteratively calculated to determine the safe cluster center and the dangerous cluster center, where n≥2 and n is a positive integer. The warning threshold is determined based on the midpoint of the Euclidean distance between the safe cluster center and the dangerous cluster center.

7. The real-time prediction method for coal and rock failure based on acoustic emission signal analysis according to claim 1, characterized in that, The statistical information also includes location information for acoustic emission events of different energies, including high-energy acoustic emission events and low-energy acoustic emission events; After performing real-time prediction of coal and rock damage through the aforementioned early warning threshold, the method further includes: In the event of a high-energy acoustic emission event during a coal and rock failure test, a region where low-energy acoustic emission events accumulate is determined in the coal and rock based on the location information of the low-energy acoustic emission events. A line is drawn connecting the overlapping region of the high-energy acoustic emission events and the region where low-energy events accumulate to form the potential failure surface of the coal and rock.

8. A real-time prediction device for coal and rock failure based on acoustic emission signal analysis, characterized in that, include: The information acquisition module is used to acquire statistical information of multiple acoustic emission signals during coal and rock destructive tests. The statistical information includes waveform data and the number of acoustic emission events; The first calculation module is used to determine the fracture type of each acoustic emission signal based on the waveform data, and to determine the b-value and absolute rate of change of each acoustic emission signal based on the waveform data and the number of acoustic emission events; the fracture type includes shear failure or tensile failure; The second calculation module is used to determine the proportion of shear fracture corresponding to the plurality of acoustic emission signals according to the fracture type, and to analyze the statistical characteristics of the waveform data over time to obtain target statistical characteristics; the target statistical characteristics include variance, autocorrelation coefficient and activity. The early warning module is used to... b value, b The absolute rate of change of the value, the proportion of shear fracture, and the target statistical characteristics are used to determine the early warning threshold, and the early warning threshold is used to perform real-time prediction of coal and rock damage.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the real-time prediction method for coal and rock damage based on acoustic emission signal analysis as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the real-time prediction method for coal and rock damage based on acoustic emission signal analysis as described in any one of claims 1 to 7.

Images