Rock damage early warning method and device based on acoustic emission b value

Abstract

The invention discloses a rock damage early warning method and device based on an acoustic emission b value, and the method comprises the steps: responding to the arrangement of at least one acoustic emission probe, carrying out the control loading of a rock sample, and monitoring at least one acoustic emission probe at the same time, recording the number and time of acoustic emission events collected by each acoustic emission probe from the beginning of loading to the occurrence of damage of the rock sample; determining a target amplitude interval and a target acoustic emission probe based on the number of acoustic emission events collected by at least one acoustic emission probe; based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe, determining corresponding b value sequences under different parameter conditions; and based on the b value sequences corresponding to different parameter conditions, determining a target b value sequence by using the target analysis model, and carrying out rock damage early warning based on the target b value sequence. According to the method, the scientificity of determining the b value and the evolution trend of the b value along with the loading process is improved, and the accuracy of rock damage early warning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of coal and rock testing, and in particular to a rock damage early warning method and device based on acoustic emission b-value. Background Art

[0002] Rock acoustic emission characteristics are a key tool in the field of rock mechanics for studying internal rock deformation and failure. The b-value patterns associated with these characteristics are crucial for studying rock failure mechanisms and preventing disasters caused by rock failure and instability. Therefore, determining the b-value of rock acoustic emission is crucial for providing early warning of rock failure based on this b-value.

[0003] The acquisition of the acoustic emission b-value is affected by multiple factors (such as the number of window samples and the sliding step size). However, in the existing technology, the impact of each factor on the b-value is still uncertain, and how to unify the b-value results obtained by different factors has not been solved. This reduces the scientific nature of the determination of the b-value and the evolution trend of the b-value with the loading process, and reduces the accuracy of early warning of rock failure based on the acoustic emission b-value. Summary of the Invention

[0004] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0005] To this end, the present invention proposes a rock damage early warning method based on acoustic emission b-values. The method can determine the b-value sequences corresponding to different parameter conditions based on the number of acoustic emission events collected by at least one acoustic emission probe and the determined target amplitude interval and target acoustic emission probe, and use a target analysis model to determine the target b-value sequences corresponding to the b-value sequences under different parameter conditions. Rock damage early warning is performed based on the target b-value sequence, thereby improving the scientific nature of the determination of the b-value and the evolution trend of the b-value with the loading process, and improving the accuracy of rock damage early warning.

[0006] Another object of the present invention is to provide a rock damage early warning device based on the acoustic emission b value.

[0007] To achieve the above objectives, the present invention provides a rock damage early warning method based on acoustic emission b-value, the method comprising:

[0008] In response to the deployment of at least one acoustic emission probe, the rock sample is controlled loaded while the at least one acoustic emission probe is monitored, and the number and time of acoustic emission events collected by each acoustic emission probe from the start of loading to the occurrence of failure of the rock sample are recorded;

[0009] determining a target amplitude interval and a target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe;

[0010] Determining a b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe;

[0011] Based on the b-value sequences corresponding to the different parameter conditions, a target b-value sequence is determined using a target analysis model, and rock damage early warning is performed based on the target b-value sequence.

[0012] The rock damage early warning method based on acoustic emission b-value according to the embodiment of the present invention may also have the following additional technical features:

[0013] In one embodiment of the present invention, determining the target amplitude interval and the target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe includes:

[0014] Performing distribution statistics on the number of acoustic emission events collected by the at least one acoustic emission probe based on amplitude statistics and initial amplitude intervals to determine an amplitude distribution graph corresponding to the acoustic emission events;

[0015] Performing a linear fit on the data in the amplitude distribution graph to obtain a goodness of fit of a corresponding first fitting curve, and determining the absolute value of the slope of the first fitting curve as a b value;

[0016] Changing the initial amplitude interval based on the threshold range of the amplitude interval, repeating the above steps to obtain corresponding multiple different b values;

[0017] determining a target amplitude interval based on the plurality of different b-values;

[0018] A target acoustic emission probe is determined based on the number of acoustic emission events collected by the at least one acoustic emission probe and the target amplitude interval.

[0019] In one embodiment of the present invention, determining a target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe and the target amplitude interval includes:

[0020] determining a target threshold based on the number of acoustic emission events collected by the at least one acoustic emission probe;

[0021] Eliminating, from the at least one acoustic emission probe, acoustic emission probes whose number of collected acoustic emission events is less than the target threshold, to obtain a set of candidate acoustic emission probes;

[0022] Based on the target amplitude interval, the number of acoustic emission events collected by the probes or probe combinations in the candidate acoustic emission probe set is used to determine the goodness of fit and b-value corresponding to different probes or probe combinations;

[0023] The probe or probe combination corresponding to the goodness of fit and b value that meets the preset conditions is determined as the target acoustic emission probe.

[0024] In one embodiment of the present invention, determining the b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe includes:

[0025] Determining a first b-value sequence under different window sample numbers based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe;

[0026] Determining a second b-value sequence under different sliding step sizes based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe;

[0027] Based on the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe, a third b-value sequence in different amplitude intervals is determined.

[0028] In one embodiment of the present invention, determining the first b-value sequence under different window sample numbers based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe includes:

[0029] Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the first b value corresponding to the first preset window sample number in the arrangement result by the target amplitude interval, wherein the time corresponding to the first b value is the time of the last acoustic emission event in the first preset window sample number;

[0030] Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine a second b value corresponding to the number of samples in the first preset window;

[0031] Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and determine the obtained sequence of the corresponding first b-value, second b-value and remaining b-values ​​as the first b-value subsequence;

[0032] Changing the first preset window sample number based on the window sample number value range, repeating the above steps, to obtain at least one first b-value subsequence under different window sample numbers;

[0033] The at least one first b-value subsequence is fused, and the obtained fused sequence is determined as the first b-value sequence under different window sample numbers.

[0034] In one embodiment of the present invention, determining the second b-value sequence at different sliding step sizes based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe includes:

[0035] Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the third b-value corresponding to the second preset window sample number in the arrangement result by the target amplitude interval, wherein the time corresponding to the third b-value is the time of the last acoustic emission event in the second preset window sample number;

[0036] Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine a fourth b value corresponding to the preset number of window samples;

[0037] Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and determine the obtained sequence corresponding to the third b-value, the fourth b-value and the remaining b-values ​​as the second b-value subsequence;

[0038] Changing the preset number of slides based on the sliding step value range, repeating the above steps to obtain at least one second b-value subsequence under different sliding step sizes;

[0039] The at least one second b-value subsequence is fused, and the obtained fused sequence is determined as the second b-value sequence under different sliding step sizes.

[0040] In one embodiment of the present invention, determining the third b-value sequence in different amplitude intervals based on the number and time of acoustic emission events collected by the target acoustic emission probe includes:

[0041] Arranging the acoustic emission events collected by the target acoustic emission probe, and performing distribution statistics based on the initial amplitude interval and the target amplitude interval to obtain a second fitting curve corresponding to the number of samples in the third preset window in the arrangement result;

[0042] determining at least one target amplitude interval based on the second fitting curve, and determining a fifth b value corresponding to each target amplitude interval through a third fitting curve graph corresponding to each target amplitude interval;

[0043] Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine the sixth b value corresponding to each target amplitude interval corresponding to the number of samples in the third preset window;

[0044] Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and determine the obtained sequence of the fifth b-value, the sixth b-value, and the remaining b-values ​​corresponding to each target amplitude interval as the third b-value subsequence corresponding to each target amplitude interval;

[0045] The third b-value subsequences corresponding to the target amplitude intervals are fused, and the obtained fused sequence is determined as the third b-value sequence under different amplitude intervals.

[0046] In one embodiment of the present invention, the target analysis model includes a first LLM-Mixer module, a fusion module, a feedforward layer module, multiple encoder modules, a decoder module and a second LLM-Mixer module.

[0047] In one embodiment of the present invention, determining a target b-value sequence using a target analysis model based on the b-value sequences corresponding to the different parameter conditions includes:

[0048] Inputting the b-value sequences corresponding to the different parameter conditions into the first LLM-Mixer module to obtain at least one long-term trend time series and at least one short-term trend time series;

[0049] Inputting the at least one long-term trend time series into the fusion module to obtain at least one corresponding fused long-term trend time series;

[0050] Inputting the at least one fused long-term trend time series into the feedforward layer module to obtain a corresponding target long-term trend time series;

[0051] Inputting the at least one short-term trend time series and the at least one fused long-term trend time series into the multiple encoder modules to obtain a deep representation sequence;

[0052] Inputting the deep representation sequence and the target long-term trend time series into the Decoder module to obtain the corresponding target short-term trend time series;

[0053] The target long-term trend time series and the target short-term trend time series are input into the second LLM-Mixer module to obtain a target b-value sequence.

[0054] Another aspect of the present invention provides a rock damage early warning device based on acoustic emission b-value, the device comprising:

[0055] a collection module for controlling the loading of the rock sample and simultaneously monitoring the at least one acoustic emission probe in response to the deployment of the at least one acoustic emission probe, and recording the number and time of acoustic emission events collected by each acoustic emission probe from the start of loading to the occurrence of failure of the rock sample;

[0056] A first determination module is configured to determine a target amplitude interval and a target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe;

[0057] A second determination module is configured to determine a b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe;

[0058] The early warning module is used to determine a target b-value sequence based on the corresponding b-value sequence under the different parameter conditions using a target analysis model, and to perform rock damage early warning based on the target b-value sequence.

[0059] The rock damage early warning method and device based on acoustic emission b-values ​​of the embodiments of the present invention can determine the b-value sequence corresponding to different parameter conditions based on the number of acoustic emission events collected by at least one acoustic emission probe and the determined target amplitude interval and target acoustic emission probe, and use the target analysis model to determine the target b-value sequence corresponding to the b-value sequence under different parameter conditions. Rock damage early warning is performed based on the target b-value sequence, thereby improving the scientific nature of the determination of the b-value and the evolution trend of the b-value with the loading process, and improving the accuracy of the rock damage early warning.

[0060] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0062] Figure 1 is a flow chart of a rock damage early warning method based on acoustic emission b-value according to one embodiment of the present invention;

[0063] Figure 2 is a schematic diagram of an amplitude distribution diagram according to an embodiment of the present invention;

[0064] Figure 3 is a schematic diagram of a first fitting curve according to one embodiment of the present invention;

[0065] Figure 4 Schematic diagram of the number of window samples and sliding step size according to one embodiment of the present invention;

[0066] Figure 5 is a schematic diagram of a linear fitting curve corresponding to a target b-value time series according to one embodiment of the present invention;

[0067] Figure 6is a schematic diagram of early warning based on a target b-value sequence according to one embodiment of the present invention;

[0068] Figure 7 1 is a structural diagram of a rock damage early warning device based on acoustic emission b-value according to an embodiment of the present invention. DETAILED DESCRIPTION

[0069] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0070] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0071] The following describes a rock damage early warning method and device based on acoustic emission b-value according to an embodiment of the present invention with reference to the accompanying drawings.

[0072] Figure 1 This is a flow chart of a rock damage early warning method based on acoustic emission b-value according to an embodiment of the present invention.

[0073] like Figure 1 As shown, the method may include the following steps:

[0074] Step 101 , in response to at least one acoustic emission probe being deployed, the rock sample is controlled to be loaded and at least one acoustic emission probe is monitored simultaneously, and the number and time of acoustic emission events collected by each acoustic emission probe from the start of loading to the occurrence of failure of the rock sample are recorded.

[0075] In one embodiment of the present invention, the rock sample can be a rectangular parallelepiped specimen, thereby facilitating the installation and accurate positioning of the acoustic emission probe. Furthermore, in one embodiment of the present invention, the size of the rock sample can be determined based on the size of the acoustic emission probe and the loading capacity of the testing machine, and is generally no less than 50 mm x 50 mm x 100 mm.

[0076] In one embodiment of the present invention, the number and position of the acoustic emission probes can be determined based on test requirements, and a coupling agent such as vaseline can be applied between the acoustic emission probes and the sample. Furthermore, in one embodiment of the present invention, after the acoustic emission probes are accurately deployed, a probe can be cut on the sample surface to determine whether the probes are in good contact and whether the deployment position can cover the rock sample. If the probes are inaccurate or cannot cover the rock sample, adjustments can be made in a timely manner until at least one acoustic emission probe is deployed.

[0077] Furthermore, in one embodiment of the present invention, in response to the deployment of the at least one acoustic emission probe, the rock sample can be subjected to controlled loading while simultaneously monitoring the at least one acoustic emission probe, recording the number and time of acoustic emission events collected by each acoustic emission probe from the start of loading to the occurrence of failure of the rock sample. In one embodiment of the present invention, the controlled loading of the rock sample can include displacement control, load control, or loading rate control, with continuous loading until the rock sample fails, and the number and time of acoustic emission events collected by each acoustic emission probe can be recorded in real time during the loading process.

[0078] Step 102: Determine a target amplitude interval and a target acoustic emission probe based on the number of acoustic emission events collected by at least one acoustic emission probe.

[0079] In one embodiment of the present invention, after obtaining the number of acoustic emission events collected by at least one acoustic emission probe through the above steps, the target amplitude interval and the target acoustic emission probe can be determined based on the number of acoustic emission events collected by the at least one acoustic emission probe.

[0080] Specifically, in one embodiment of the present invention, the method for determining the target amplitude interval and the target acoustic emission probe based on the number of acoustic emission events collected by at least one acoustic emission probe may include the following steps:

[0081] Step 1021: performing distribution statistics on the number of acoustic emission events collected by at least one acoustic emission probe based on amplitude statistics and initial amplitude intervals to determine an amplitude distribution graph corresponding to the acoustic emission events;

[0082] Step 1022: Perform a linear fit on the data in the amplitude distribution graph to obtain the goodness of fit of the corresponding fitting curve, and determine the absolute value of the slope of the fitting curve as the b value;

[0083] Step 1023, changing the initial amplitude interval based on the threshold range of the amplitude interval, repeating the above steps to obtain corresponding multiple different b values;

[0084] Step 1024, determining a target amplitude interval based on a plurality of different b values;

[0085] Step 1025 : Determine a target acoustic emission probe based on the number of acoustic emission events collected by at least one acoustic emission probe and the target amplitude interval.

[0086] In one embodiment of the present invention, the number of all acoustic emission events collected by at least one acoustic emission probe is distributed and counted based on the amplitude (db) to obtain the number of acoustic emission events in the amplitude interval, where the amplitude interval is ΔA. Figure 2 A schematic diagram of an amplitude distribution diagram proposed for the implementation of the present invention is shown in FIG. Figure 2 As shown, when ΔA=2, we can get Figure 2 Amplitude distribution diagram of the statistical histogram.

[0087] Furthermore, in one embodiment of the present invention, it is possible to count the number of cells greater than a certain amplitude A. db The number of acoustic emission events N, and A db / 20 is the horizontal axis and lgN is the vertical axis for linear fitting to obtain the first fitting curve (such as Figure 3 As shown). In one embodiment of the present invention, after the first fitting curve is obtained through the above steps, the goodness of fit R of the first fitting curve can be obtained. 2 , and the absolute value of the slope of the first fitting curve is determined as the b value.

[0088] Furthermore, in one embodiment of the present invention, the initial amplitude interval ΔA can be changed based on the amplitude interval threshold range, and steps 1021 to 1022 can be repeated to obtain corresponding multiple different b values. In one embodiment of the present invention, the amplitude interval threshold range can be set as needed, for example, the amplitude interval threshold range is [1, 10).

[0089] Furthermore, in one embodiment of the present invention, after obtaining multiple different b values ​​through the above steps, the ΔA corresponding to the b value closest to the first preset value can be determined as the candidate target amplitude interval ΔA. At the same time, the goodness of fit R of the fitting curve corresponding to the candidate target amplitude interval is determined. 2 Is it greater than the second preset value? If the goodness of fit R 2 is greater than the second preset value, the candidate target amplitude interval is determined to be the target amplitude interval ΔA; if the goodness of fit R 2 If the value of b is less than or equal to the second preset value, the next ΔA corresponding to the value of b close to the first preset value is found, and the above steps are repeated until the target amplitude interval ΔA is obtained. In one embodiment of the present invention, the second preset value can be set as needed, such as 0.8.

[0090] Furthermore, in one embodiment of the present invention, after determining the target amplitude interval through the above steps, the method for determining the target acoustic emission probe may include the following steps based on the number of acoustic emission events collected by at least one acoustic emission probe and the target amplitude interval:

[0091] Step 10251: determining a target threshold based on the number of acoustic emission events collected by at least one acoustic emission probe;

[0092] Step 10252: Remove from at least one acoustic emission probe those acoustic emission probes whose number of collected acoustic emission events is less than a preset target threshold, to obtain a set of candidate acoustic emission probes.

[0093] Step 10253: Based on the target amplitude interval and the number of acoustic emission events collected by the probes or probe combinations in the candidate acoustic emission probe set, determine the goodness of fit and b-value corresponding to different probes or probe combinations;

[0094] In step 10254, the probe or probe combination corresponding to the goodness of fit and b value that meet the preset conditions is determined as the target acoustic emission probe.

[0095] In one embodiment of the present invention, the method for determining the target threshold based on the number of acoustic emission events collected by at least one acoustic emission probe may include: dividing the number of all acoustic emission events collected by at least one acoustic emission probe by the number of acoustic emission probes to obtain the mean m of the number of collected acoustic emission events.

[0096] In one embodiment of the present invention, acoustic emission probes whose number of collected acoustic emission events is less than a preset target threshold are eliminated from at least one acoustic emission probe to obtain a set of candidate acoustic emission probes. In one embodiment of the present invention, the preset threshold can be set as needed, such as 1 / 2.

[0097] And, in one embodiment of the present invention, based on the target amplitude interval, the number of acoustic emission events collected by the probes or probe combinations in the candidate acoustic emission probe set is used to determine the goodness of fit R corresponding to different probes or probe combinations through the above steps 1021 to 1022. 2 and b value.

[0098] Furthermore, in one embodiment of the present invention, the goodness of fit R that meets the preset conditions is 2 The probe or probe combination corresponding to the b value is determined as the target acoustic emission probe. The above preset conditions can be set as needed, such as the goodness of fit R 2 >0.85 and b value ∈ (0.85, 1.2).

[0099] And, in one embodiment of the present invention, the above probe combination can be from 2 probes to n probes, traversing all combinations, a total of A combination of methods.

[0100] Step 103 : determining a b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe.

[0101] In one embodiment of the present invention, after determining the target amplitude interval and the target acoustic emission probe through the above steps, the corresponding b-value sequence under different parameter conditions is determined based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe.

[0102] In one embodiment of the present invention, the method for determining the corresponding b-value sequence under different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe may include the following steps:

[0103] Step 1031 , determining a first b-value sequence under different window sample numbers based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe;

[0104] Step 1032 , determining a second b-value sequence under different sliding step sizes based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe;

[0105] Step 1033 : determining a third b-value sequence in different amplitude intervals based on the number and time of acoustic emission events collected by the target acoustic emission probe.

[0106] In one embodiment of the present invention, the method for determining the first b-value sequence under different window sample numbers based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe may include the following steps:

[0107] Step 10311: Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the first b value corresponding to the number of samples in the first preset window in the arrangement result through the target amplitude interval;

[0108] In one embodiment of the present invention, the time corresponding to the first b value is the time of the last acoustic emission event in the first preset window sample quantity.

[0109] Step 10312: Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above step 10311, and determine the second b value corresponding to the number of samples in the first preset window;

[0110] Step 10313: Repeat steps 10311 to 10312 until all acoustic emission events collected by the target acoustic emission probe are traversed, and the obtained sequence of the corresponding first b-value, second b-value, and remaining b-values ​​is determined as a first b-value subsequence;

[0111] Step 10314: Change the first preset window sample number based on the window sample number value range, and repeat steps 10311 to 10313 to obtain at least one first b-value subsequence under different window sample numbers;

[0112] Step 10315: fuse at least one first b-value subsequence, and determine the obtained fused sequence as the first b-value sequence under different window sample numbers.

[0113] In one embodiment of the present invention, the acoustic emission events collected by the target acoustic emission probe are arranged, the number of acoustic emission events corresponding to the first preset window sample number L is taken, and the first b value corresponding to the first preset window sample number is obtained through the target amplitude interval and the above steps 1021 to 1022.

[0114] And, in one embodiment of the present invention, the last starting event in the arrangement result is slid to the right by a preset number of l acoustic emission events (such as Figure 4 ), and the second b-value corresponding to the number of samples in the first preset window is re-determined. Furthermore, in one embodiment of the present invention, steps 10311 to 10312 are repeated until all acoustic emission events collected by the target acoustic emission probe are traversed, and the resulting sequence of the corresponding first b-value, second b-value, and remaining b-values ​​is determined as a first b-value subsequence. The number of acoustic emission events in the last group may be less than L data in the number of samples in the first preset window. In this case, the corresponding b-value can be determined based on the last group of acoustic emission events.

[0115] Furthermore, in one embodiment of the present invention, the value range of the number of window samples can be set as needed, such as (200, 10% of the total number of acoustic emission events).

[0116] Furthermore, in one embodiment of the present invention, the method of fusing at least one first b-value subsequence and determining the obtained fused sequence as the first b-value sequence under different window sample numbers may include the following steps:

[0117] Step 1: linearly interpolate each first b-value subsequence at a specified time interval (e.g., 0.1s);

[0118] Step 2: Calculate the average of the b-value data of the first preset time to obtain the target b-value at each moment in the first preset time;

[0119] In one embodiment of the present invention, the first preset time can be set as needed, such as 20 seconds.

[0120] Step 3: For the b values ​​of the remaining times except the first preset time, the mean square error σ is calculated using the first b value subsequences of the second preset time before the target time, and the median x is calculated using the values ​​of the first b value subsequences of the target time. i ;

[0121] Step 4: Based on the mean square error and the median, determine the preset range of the b value at the target time (such as [x i -3σ,x i +3σ]);

[0122] Step 5: Calculate the average of the b-values ​​within the preset range in each first b-value subsequence at the target moment, and use the obtained average result as the target b-value at the target moment;

[0123] Step 6: Generate a first b-value sequence with different window sample numbers based on the target b-value at each moment.

[0124] Furthermore, in one embodiment of the present invention, the method for determining the second b-value sequence under different sliding step sizes based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe may include the following steps:

[0125] Step 10321: Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the third b value corresponding to the number of samples in the second preset window in the arrangement result through the target amplitude interval;

[0126] In one embodiment of the present invention, the time corresponding to the third b value is the time of the last acoustic emission event in the second preset window sample quantity.

[0127] Step 10322: Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above step 10321, and determine the fourth b value corresponding to the second preset window sample number;

[0128] Step 10323: Repeat steps 10321 to 10322 until all acoustic emission events collected by the target acoustic emission probe are traversed, and the obtained sequence corresponding to the third b-value, the fourth b-value, and the remaining b-values ​​is determined as a second b-value subsequence;

[0129] Step 10324: Change the preset number of slides based on the sliding step value range, and repeat steps 10321 to 10323 to obtain at least one second b-value subsequence under different sliding step sizes;

[0130] Step 10325: fuse at least one second b-value subsequence, and determine the obtained fused sequence as the second b-value sequence under different sliding step sizes.

[0131] In one embodiment of the present invention, the acoustic emission events collected by the target acoustic emission probe are arranged, the number of acoustic emission events corresponding to the number of samples in the second preset window is taken, and the third b value corresponding to the number of samples in the second preset window is obtained through the target amplitude interval and the above steps 1021 to 1022.

[0132] Furthermore, in one embodiment of the present invention, the previous starting event in the arrangement result is slid to the right by a preset number of l acoustic emission events, and the fourth b-value corresponding to the second preset window sample number is re-determined. Furthermore, in one embodiment of the present invention, steps 10321 to 10322 are repeated until all acoustic emission events collected by the target acoustic emission probe are traversed, and the resulting sequence of corresponding second b-values, third b-values, and remaining b-values ​​is determined as a second b-value subsequence, wherein the number of acoustic emission events in the last group may be insufficient for the number of samples in the second preset window. In this case, the corresponding b-value can be determined based on the number of acoustic emission events in the last group.

[0133] Furthermore, in one embodiment of the present invention, the sliding step value range can be set as needed, such as (20, 50% of the number of samples in the second preset window).

[0134] Furthermore, in one embodiment of the present invention, the method of fusing at least one second b-value subsequence and determining the obtained fused sequence as a second b-value sequence under different sliding step sizes may include the following steps:

[0135] Step 1: The maximum value l based on the sliding step value range max and the minimum value l min , calculate the maximum change multiple of the sliding step size

[0136] Step 2: interpolate the calculated b-value results of each sliding step length l to obtain at least one interpolated second b-value subsequence;

[0137] In one embodiment of the present invention, the above-mentioned mean interpolation method can be adopted, that is, the inserted b value is equal to the mean of the b values ​​before and after the insertion point, so that the number of the interpolated second b value subsequences under each sliding step result is equal.

[0138] Furthermore, in one embodiment of the present invention, other interpolation methods may also be selected.

[0139] Step 3: calculate the weighted average of the interpolated second b-value subsequence at the same time, and determine the obtained weighted average at each time as the target b-value at each time;

[0140] In one embodiment of the present invention, if the b-values ​​at the target moment are all actually calculated, the weights of the b-values ​​are consistent; if the b-values ​​at the target moment include b-values ​​inserted by interpolation, the sum of the weights of all b-values ​​inserted by the interpolation method is 0.4, the weights of the b-values ​​inserted by each interpolation method are consistent, the sum of the weights of the actually calculated b-values ​​is 0.6, and the weights of the actually calculated b-values ​​are consistent.

[0141] Step 4: Generate a second b-value sequence with different sliding step sizes based on the target b-value at each moment.

[0142] Furthermore, in one embodiment of the present invention, the method for determining the third b-value sequence in different amplitude intervals based on the number and time of acoustic emission events collected by the target acoustic emission probe may include the following steps:

[0143] Step 10331: Arrange the acoustic emission events collected by the target acoustic emission probe, and perform distribution statistics based on the initial amplitude interval and the target amplitude interval to obtain a second fitting curve corresponding to the number of samples in the third preset window in the arrangement result;

[0144] Step 10332: Determine at least one target amplitude interval based on the second fitting curve, and determine a fifth b value corresponding to each target amplitude interval using the third fitting curve corresponding to each target amplitude interval;

[0145] Step 10333: Slide the previous starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat steps 10331 to 10332, and determine the sixth b value corresponding to each target amplitude interval corresponding to the number of samples in the third preset window;

[0146] Step 10334: Repeat steps 10331 to 10333 until all acoustic emission events collected by the target acoustic emission probe are traversed, and the obtained sequence of the fifth b-value, the sixth b-value, and the remaining b-values ​​corresponding to each target amplitude interval is determined as the third b-value subsequence corresponding to each target amplitude interval;

[0147] Step 10335: fuse the third b-value subsequences corresponding to the target amplitude intervals, and determine the obtained fused sequence as the third b-value sequence under different amplitude intervals.

[0148] In one embodiment of the present invention, the method for determining at least one target amplitude range based on the second fitting curve may include: dividing the horizontal axis A by dBThe starting point of / 20 is gradually increased from the starting point (such as 2) to perform linear fitting until the goodness of fit R 2 >0.85, discard the subsequent horizontal coordinates to obtain the corresponding first target amplitude interval, where A dB / 20 can be increased by 0.05 each time; based on the second fitting curve, the horizontal axis A dB The starting point of / 20 is gradually reduced from the end point to perform linear fitting until the goodness of fit R 2 >0.85, discard the previous horizontal coordinate to obtain the corresponding second target amplitude range, where A dB / 20 is reduced by 0.05 each time; based on the second fitting curve, the horizontal axis A dB The starting point of / 20 gradually increases from the starting point and the horizontal axis A dB The starting point of / 20 is gradually reduced from the end point to perform linear fitting until the goodness of fit R 2 >0.85, and the corresponding third target amplitude interval is obtained, A dB The amount of increase and decrease at both ends of / 20 is 0.05 each time.

[0149] Furthermore, in one embodiment of the present invention, a method for fusing the third b-value subsequences corresponding to each target amplitude interval and determining the obtained fused sequence as a third b-value sequence under different amplitude intervals may include: determining the average of the b-values ​​at each target moment as the target b-value at each target moment based on the third b-value subsequences corresponding to each target amplitude interval, and generating the third b-value sequence under different amplitude intervals based on the target b-value at each moment.

[0150] In one embodiment of the present invention, the number of samples in the second preset window and the number of samples in the third preset window may be the same.

[0151] Step 104 : Based on the corresponding b-value sequences under different parameter conditions, a target b-value sequence is determined using a target analysis model, and rock damage warning is performed based on the target b-value sequence.

[0152] Among them, in one embodiment of the present invention, after obtaining the corresponding b-value sequence under different parameter conditions through the above steps, the target b-value sequence can be determined based on the corresponding b-value sequence under different parameter conditions using the target analysis model, and rock damage warning can be performed based on the target b-value sequence.

[0153] And, in one embodiment of the present invention, the above-mentioned target analysis model may include a first LLM-Mixer module, a fusion module, a feedforward layer module, multiple encoder modules, a decoder module and a second LLM-Mixer module.

[0154] Specifically, in one embodiment of the present invention, the method for determining a target b-value sequence using a target analysis model based on corresponding b-value sequences under different parameter conditions may include the following steps:

[0155] Step 1041: Input the b-value sequences corresponding to different parameter conditions into the first LLM-Mixer module to obtain at least one long-term trend time series and at least one short-term trend time series;

[0156] In one embodiment of the present invention, the multivariate time b-value sequence is projected onto the depth feature X by the first LLM-Mixer module. 0 In, X 0 =Embed(X), thus decomposing the multivariate time series into the long-term trend time series x l and short-term trend time series x m , so that the subsequent model can better process and learn. In one embodiment of the present invention, after the b-value sequence is input into the first LLM-Mixer module, the long-term trend time series x corresponding to the b-value sequence can be obtained. l and short-term trend time series x m , and the long-term trend time series x l and short-term trend time series x m Same dimension as the input b-value sequence.

[0157] Step 1042: inputting at least one long-term trend time series into a fusion module to obtain at least one corresponding fused long-term trend time series;

[0158] In one embodiment of the present invention, the lengths corresponding to at least one long-term trend time series obtained through the above steps may be different. Based on this, it is necessary to fill the position of the shorter long-term trend time series so that the lengths of the long-term trend time series after filling are consistent.

[0159] In one embodiment of the present invention, a randomly selected position in the long-term trend time series to be filled can be filled, and the filling range can be traversed to determine the filling value, and the filled long-term trend time series that minimizes the objective function L is determined as the target filled long-term trend time series. The objective function is:

[0160]

[0161] Among them, T is the number of long-term trend time series that need to be filled, is the long-term trend time series after the k-th filling, is a long-term trend time series that is similar to the k-th long-term trend time series that needs to be filled. is a long-term trend time series that is far from the k-th long-term trend time series that needs to be filled. It should be noted that the long-term trend time series with similar conditions is the long-term trend time series that needs to be filled, and the corresponding number of different parameter conditions is 1; the long-term trend time series with distant conditions is the long-term trend time series that needs to be filled, and the corresponding number of different parameter conditions is 2.

[0162] Step 1043: Input at least one fused long-term trend time series into the feedforward layer module to obtain the corresponding target long-term trend time series.

[0163] Step 1044: Input the at least one short-term trend time series and the at least one fused long-term trend time series into multiple encoder modules to obtain a deep representation sequence;

[0164] In one embodiment of the present invention, at least one fused long-term trend time series is input as Q, thereby obtaining a deep representation of the long-term trend and short-term fluctuations.

[0165] And, in one embodiment of the present invention, the above-mentioned rules of long-term trends and short-term trends under different conditions are mixed by stacking multiple encoder modules. For the nth layer, the input is X n-1 , the processing process is:

[0166] X n =Encoder(X n-1 ), l∈{0,…,N}

[0167] Among them, N is the total number of layers, M is the number of b-value sequences corresponding to different parameter conditions, Represents the past representation of the mth short-term trend time series after mixing n layers, with d model channels,

[0168] Step 1045: Input the deep representation sequence and the target long-term trend time series into the Decoder module to obtain the corresponding target short-term trend time series;

[0169] In one embodiment of the present invention, the target long-term trend time series Input Q input and deep representation sequence into the Decoder module to obtain the corresponding target short-term trend time series

[0170] Step 1046: Time series of target long-term trend and target short-term trend timing Input the second LLM-Mixer module to obtain the target b value sequence.

[0171] In one embodiment of the present invention, the target long-term trend time series and target short-term trend timing Input the second LLM-Mixer module for reverse fusion to obtain the target b-value sequence. It should be noted that the function of the first LLM-Mixer module is time series decomposition, and the function of the second LLM-Mixer module is time series fusion.

[0172] Furthermore, in one embodiment of the present invention, the above steps can be used to determine a target b-value sequence based on the b-value sequences under different parameter conditions, thereby improving the scientific nature of the determination of the b-value and the evolution trend of the b-value during the loading process.

[0173] Furthermore, in one embodiment of the present invention, after obtaining the target b-value sequence through the above steps, an early warning of rock damage can be issued based on the temporal trend of the b-values ​​in the target b-value sequence.

[0174] In one embodiment of the present invention, the method for providing early warning of rock damage based on the temporal trend of b-values ​​in a target b-value sequence may include performing a linear fit on the b-value time series within a preset time period in a sliding manner, and when the linear fit slope is less than a third preset value, setting the initial time within the preset time period as the warning time, and providing an early warning based on the warning time. In one embodiment of the present invention, the preset time period may be set based on experience, such as 5 seconds; and the third preset value may be set based on experience, such as -0.03.

[0175] For example, Figure 5 This is a schematic diagram of a linear fitting curve corresponding to a target b value time series proposed in an embodiment of the present disclosure. Figure 5 As shown, when the linear fitting slope is less than -0.03, the initial time within 5 seconds is determined as the warning time.

[0176] Furthermore, in another embodiment of the present invention, the method for early warning of rock damage based on the temporal trend of the b-value in the target b-value sequence may include the following steps:

[0177] Step a, calculating the b-value fluctuation characteristic parameter WIb within the unit time corresponding to the target moment, and determining the b-value fluctuation characteristic parameter WIb within the unit time as the WIb value at the target moment;

[0178] Step b: determining the warning threshold corresponding to the target time, and issuing a warning when the WIb value at the target time exceeds the corresponding warning threshold.

[0179] In one embodiment of the present invention, the b-value fluctuation characteristic parameter WIb within a unit time corresponding to the target moment can be calculated using a first formula, wherein the first formula is:

[0180]

[0181] Among them, b i is the b value corresponding to time i, t i is the time corresponding to moment i. The difference between two adjacent b values ​​is calculated by the above formula, and the quotient of the time difference between the two points is obtained, that is, the slope k between the two adjacent b values i , for k in unit time i The absolute value sum of is the fluctuation characteristic parameter WIb of the unit time corresponding to the target moment. And, the unit time corresponding to the target moment can be the unit time before the target moment, and the unit time can be set according to experience, such as 1s.

[0182] In one embodiment of the present invention, the method for determining the warning threshold corresponding to the target time may include: determining the warning threshold corresponding to the target time by a second formula, wherein the second formula is:

[0183] Th i =AVG(WIb)+2×SD(WIb)

[0184] Among them, Th i is the warning threshold at the target moment, AVG(WIb) is the mean of the fluctuation characteristic parameter WIb corresponding to all moments before the target moment, and SD(WIb) is the standard deviation of the fluctuation characteristic parameter WIb corresponding to all moments before the target moment.

[0185] For example, Figure 6 This is a schematic diagram of an early warning based on a target b value sequence proposed in an embodiment of the present disclosure. Figure 6 As shown, each moment has a corresponding WIB value and warning threshold. When the WIB value corresponding to the moment is greater than the warning threshold, a warning is issued.

[0186] The rock damage early warning method based on acoustic emission b-values ​​in an embodiment of the present invention can determine the b-value sequences corresponding to different parameter conditions based on the number of acoustic emission events collected by at least one acoustic emission probe, the determined target amplitude interval, and the target acoustic emission probe, and use a target analysis model to determine the target b-value sequences corresponding to the b-value sequences under different parameter conditions. Rock damage early warning is performed based on the target b-value sequence, thereby improving the scientific nature of the determination of b-values ​​and the evolution trend of b-values ​​with the loading process, and improving the accuracy of rock damage early warning.

[0187] Figure 7 1 is a schematic structural diagram of a rock damage early warning device 700 based on acoustic emission b-value according to an embodiment of the present invention.

[0188] like Figure 7 As shown, the device may include:

[0189] The acquisition module 701 is configured to control loading of the rock sample and simultaneously monitor the at least one acoustic emission probe in response to the deployment of the at least one acoustic emission probe, recording the number and time of acoustic emission events collected by each acoustic emission probe from the start of loading to the occurrence of failure of the rock sample;

[0190] A first determining module 702 is configured to determine a target amplitude interval and a target acoustic emission probe based on the number of acoustic emission events collected by at least one acoustic emission probe;

[0191] The second determination module 703 is used to determine the b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe;

[0192] The early warning module 704 is used to determine a target b-value sequence based on the corresponding b-value sequences under different parameter conditions using a target analysis model, and to perform rock damage early warning based on the target b-value sequence.

[0193] In one embodiment of the present invention, the first determining module 702 is specifically configured to:

[0194] Performing distribution statistics on the number of acoustic emission events collected by at least one acoustic emission probe based on amplitude statistics and initial amplitude intervals to determine an amplitude distribution graph corresponding to the acoustic emission events;

[0195] Performing a linear fit on the data in the amplitude distribution graph to obtain a goodness of fit of a corresponding first fitting curve, and determining the absolute value of the slope of the first fitting curve as a b value;

[0196] Changing the initial amplitude interval based on the threshold range of the amplitude interval, repeating the above steps to obtain corresponding multiple different b values;

[0197] determining a target amplitude interval based on the plurality of different b-values;

[0198] A target acoustic emission probe is determined based on the number of acoustic emission events collected by the at least one acoustic emission probe and the target amplitude interval.

[0199] In one embodiment of the present invention, the first determining module 702 is further configured to:

[0200] determining a target threshold based on the number of acoustic emission events collected by at least one acoustic emission probe;

[0201] Eliminating, from at least one acoustic emission probe, acoustic emission probes whose number of collected acoustic emission events is less than a preset target threshold, to obtain a set of candidate acoustic emission probes;

[0202] Based on the target amplitude interval, the number of acoustic emission events collected by the probes or probe combinations in the candidate acoustic emission probe set is used to determine the goodness of fit and b-value corresponding to different probes or probe combinations;

[0203] The probe or probe combination corresponding to the goodness of fit and b value that meets the preset conditions is determined as the target acoustic emission probe.

[0204] In one embodiment of the present invention, the second determining module 703 is specifically configured to:

[0205] Determine the first b-value sequence under different window sample numbers based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe;

[0206] Determine the second b-value sequence under different sliding step sizes based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe;

[0207] Based on the number and time of acoustic emission events collected by the target acoustic emission probe, a third b-value sequence under different amplitude intervals is determined.

[0208] In one embodiment of the present invention, the second determining module 703 is further configured to:

[0209] Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the first b value corresponding to the first preset window sample number in the arrangement result through the target amplitude interval, wherein the time corresponding to the first b value is the time of the last acoustic emission event in the first preset window sample number;

[0210] Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine the second b value corresponding to the number of samples in the first preset window;

[0211] Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and the obtained sequence of the corresponding first b-value, second b-value, and remaining b-values ​​is determined as the first b-value subsequence;

[0212] Changing the first preset window sample number based on the window sample number value range, repeating the above steps, to obtain at least one first b-value subsequence under different window sample numbers;

[0213] At least one first b-value subsequence is fused, and the obtained fused sequence is determined as the first b-value sequence under different window sample numbers.

[0214] In one embodiment of the present invention, the second determining module 703 is further configured to:

[0215] Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the third b-value corresponding to the second preset window sample number in the arrangement result by the target amplitude interval, wherein the time corresponding to the third b-value is the time of the last acoustic emission event in the second preset window sample number;

[0216] Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine the fourth b value corresponding to the number of samples in the second preset window;

[0217] Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and the obtained sequence corresponding to the third b-value, the fourth b-value, and the remaining b-values ​​is determined as the second b-value subsequence;

[0218] Changing the preset number of slides based on the sliding step value range, repeating the above steps, and obtaining at least one second b-value subsequence under different sliding step sizes;

[0219] At least one second b-value subsequence is fused, and the obtained fused sequence is determined as the second b-value sequence under different sliding step sizes.

[0220] In one embodiment of the present invention, the second determining module 703 is further configured to:

[0221] Arrange the acoustic emission events collected by the target acoustic emission probe, and obtain a second fitting curve corresponding to the number of samples in the third preset window in the arrangement result by performing distribution statistics based on the initial amplitude interval and the target amplitude interval;

[0222] determining at least one target amplitude interval based on the second fitting curve, and determining a fifth b value corresponding to each target amplitude interval using a third fitting curve corresponding to each target amplitude interval;

[0223] Slide the previous starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine the sixth b value corresponding to each target amplitude interval corresponding to the number of samples in the third preset window;

[0224] Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and determine the sequences of the fifth b-value, the sixth b-value, and the remaining b-values ​​corresponding to each target amplitude interval as the third b-value subsequence corresponding to each target amplitude interval;

[0225] The third b-value subsequences corresponding to the target amplitude intervals are fused, and the obtained fused sequence is determined as the third b-value sequence under different amplitude intervals.

[0226] In one embodiment of the present invention, the target analysis model includes a first LLM-Mixer module, a fusion module, a feedforward layer module, multiple encoder modules, a decoder module and a second LLM-Mixer module.

[0227] In one embodiment of the present invention, the warning module 704 is specifically configured to:

[0228] Input the b-value sequences corresponding to different parameter conditions into the first LLM-Mixer module to obtain at least one long-term trend time series and at least one short-term trend time series;

[0229] Inputting at least one long-term trend time series into a fusion module to obtain at least one corresponding fused long-term trend time series;

[0230] Input at least one fused long-term trend time series into the feedforward layer module to obtain the corresponding target long-term trend time series;

[0231] Input at least one short-term trend time series and at least one fused long-term trend time series into multiple encoder modules to obtain a deep representation sequence;

[0232] Input the deep representation sequence and the target long-term trend time series into the Decoder module to obtain the corresponding target short-term trend time series;

[0233] The target long-term trend time series and the target short-term trend time series are input into the second LLM-Mixer module to obtain the target b-value sequence.

[0234] The rock damage early warning device based on acoustic emission b-values ​​in an embodiment of the present invention can determine the b-value sequences corresponding to different parameter conditions based on the number of acoustic emission events collected by at least one acoustic emission probe, the determined target amplitude interval, and the target acoustic emission probe, and use a target analysis model to determine the target b-value sequences corresponding to the b-value sequences under different parameter conditions. Rock damage early warning is performed based on the target b-value sequence, thereby improving the scientific nature of the determination of the b-value and the evolution trend of the b-value with the loading process, and improving the accuracy of the rock damage early warning.

[0235] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0236] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

Claims

1. A rock damage early warning method based on acoustic emission b value, characterized in that: The method comprises: In response to the deployment of at least one acoustic emission probe, the rock sample is controlled loaded while the at least one acoustic emission probe is monitored, and the number and time of acoustic emission events collected by each acoustic emission probe from the start of loading to the occurrence of failure of the rock sample are recorded; determining a target amplitude interval and a target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe; Determining a b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe; Based on the b-value sequences corresponding to the different parameter conditions, a target b-value sequence is determined using a target analysis model, and rock damage early warning is performed based on the target b-value sequence.

2. The method according to claim 1, characterized in that The determining of the target amplitude interval and the target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe comprises: Performing distribution statistics on the number of acoustic emission events collected by the at least one acoustic emission probe based on amplitude statistics and initial amplitude intervals to determine an amplitude distribution graph corresponding to the acoustic emission events; Performing a linear fit on the data in the amplitude distribution graph to obtain a goodness of fit of a corresponding first fitting curve, and determining the absolute value of the slope of the first fitting curve as a b value; Changing the initial amplitude interval based on the threshold range of the amplitude interval, repeating the above steps to obtain corresponding multiple different b values; determining a target amplitude interval based on the plurality of different b-values; A target acoustic emission probe is determined based on the number of acoustic emission events collected by the at least one acoustic emission probe and the target amplitude interval.

3. The method according to claim 2, characterized in that The step of determining a target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe and the target amplitude interval includes: determining a target threshold based on the number of acoustic emission events collected by the at least one acoustic emission probe; Eliminating, from the at least one acoustic emission probe, acoustic emission probes whose number of collected acoustic emission events is less than a preset target threshold value, to obtain a set of candidate acoustic emission probes; Based on the target amplitude interval, the number of acoustic emission events collected by the probes or probe combinations in the candidate acoustic emission probe set is used to determine the goodness of fit and b-value corresponding to different probes or probe combinations; The probe or probe combination corresponding to the goodness of fit and b value that meets the preset conditions is determined as the target acoustic emission probe.

4. The method according to claim 1, wherein The determining of the b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe includes: Determining a first b-value sequence under different window sample numbers based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe; Determining a second b-value sequence under different sliding step sizes based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe; Based on the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe, a third b-value sequence in different amplitude intervals is determined.

5. The method according to claim 4, characterized in that The determining of the first b-value sequence under different numbers of window samples based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe includes: Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the first b value corresponding to the first preset window sample number in the arrangement result by the target amplitude interval, wherein the time corresponding to the first b value is the time of the last acoustic emission event in the first preset window sample number; Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine a second b value corresponding to the number of samples in the first preset window; Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and determine the obtained sequence of the corresponding first b-value, second b-value and remaining b-values ​​as the first b-value subsequence; Changing the first preset window sample number based on the window sample number value range, repeating the above steps, to obtain at least one first b-value subsequence under different window sample numbers; The at least one first b-value subsequence is fused, and the obtained fused sequence is determined as the first b-value sequence under different window sample numbers.

6. The method according to claim 4, characterized in that The determining of a second b-value sequence under different sliding step sizes based on the target amplitude interval and the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe includes: Arrange the acoustic emission events collected by the target acoustic emission probe, and calculate the third b-value corresponding to the second preset window sample number in the arrangement result by the target amplitude interval, wherein the time corresponding to the third b-value is the time of the last acoustic emission event in the second preset window sample number; Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine a fourth b value corresponding to the number of samples in the second preset window; Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and determine the obtained sequence corresponding to the third b-value, the fourth b-value and the remaining b-values ​​as the second b-value subsequence; Changing the preset number of slides based on the sliding step value range, repeating the above steps to obtain at least one second b-value subsequence under different sliding step sizes; The at least one second b-value subsequence is fused, and the obtained fused sequence is determined as the second b-value sequence under different sliding step sizes.

7. The method according to claim 4, characterized in that The determining of a third b-value sequence in different amplitude intervals based on the number and time of acoustic emission events correspondingly collected by the target acoustic emission probe includes: Arranging the acoustic emission events collected by the target acoustic emission probe, and performing distribution statistics based on the initial amplitude interval and the target amplitude interval to obtain a second fitting curve corresponding to the number of samples in the third preset window in the arrangement result; determining at least one target amplitude interval based on the second fitting curve, and determining a fifth b value corresponding to each target amplitude interval using a third fitting curve corresponding to each target amplitude interval; Slide the last starting event in the arrangement result to the right by a preset number of acoustic emission events, repeat the above steps, and determine the sixth b value corresponding to each target amplitude interval corresponding to the number of samples in the third preset window; Repeat the above steps until all acoustic emission events collected by the target acoustic emission probe are traversed, and determine the obtained sequence of the fifth b-value, the sixth b-value, and the remaining b-values ​​corresponding to each target amplitude interval as the third b-value subsequence corresponding to each target amplitude interval; The third b-value subsequences corresponding to the target amplitude intervals are fused, and the obtained fused sequence is determined as the third b-value sequence under different amplitude intervals.

8. The method according to claim 1, characterized in that The target analysis model includes a first LLM-Mixer module, a fusion module, a feedforward layer module, multiple encoder modules, a decoder module and a second LLM-Mixer module.

9. The method according to claim 8, characterized in that Determining a target b-value sequence using a target analysis model based on the b-value sequences corresponding to the different parameter conditions includes: Inputting the b-value sequences corresponding to the different parameter conditions into the first LLM-Mixer module to obtain at least one long-term trend time series and at least one short-term trend time series; Inputting the at least one long-term trend time series into the fusion module to obtain at least one corresponding fused long-term trend time series; Inputting the at least one fused long-term trend time series into the feedforward layer module to obtain a corresponding target long-term trend time series; Inputting the at least one short-term trend time series and the at least one fused long-term trend time series into the multiple encoder modules to obtain a deep representation sequence; Inputting the deep representation sequence and the target long-term trend time series into the Decoder module to obtain the corresponding target short-term trend time series; The target long-term trend time series and the target short-term trend time series are input into the second LLM-Mixer module to obtain a target b-value sequence.

10. A rock damage early warning device based on acoustic emission b value, characterized in that: The device comprises: a collection module for controlling the loading of the rock sample and simultaneously monitoring the at least one acoustic emission probe in response to the deployment of the at least one acoustic emission probe, and recording the number and time of acoustic emission events collected by each acoustic emission probe from the start of loading to the occurrence of failure of the rock sample; A first determination module is configured to determine a target amplitude interval and a target acoustic emission probe based on the number of acoustic emission events collected by the at least one acoustic emission probe; A second determination module is configured to determine a b-value sequence corresponding to different parameter conditions based on the target amplitude interval and the number and time of acoustic emission events collected by the target acoustic emission probe; The early warning module is used to determine a target b-value sequence based on the corresponding b-value sequence under the different parameter conditions using a target analysis model, and to perform rock damage early warning based on the target b-value sequence.

Images