Rock failure precursor identification method and device, storage medium and computer device

Abstract

The application relates to the technical field of geotechnical engineering, and in particular to a rock failure precursor identification method and device, a storage medium and computer equipment. The method comprises the following steps: acquiring a multimodal acoustic signal of a rock under the superposition of different types of loads; dividing the multimodal acoustic signal according to a preset time window to obtain a multimodal acoustic signal segment; performing time-frequency analysis on the multimodal acoustic signal segment to construct a time-frequency matrix of the multimodal acoustic signal segment; performing singular value decomposition on the time-frequency matrix to obtain a singular value sequence of the time-frequency matrix; calculating the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence; constructing a time-frequency singular entropy sequence of the multimodal acoustic signal according to the time-frequency singular entropy of the multimodal acoustic signal segment; and identifying a rock failure precursor based on the time-frequency singular entropy sequence of the multimodal acoustic signal. The application avoids false positives and false negatives in rock failure precursor identification, and makes early warning more reliable.

Description

Technical Field

[0001] This application relates to the field of geotechnical engineering technology, and in particular to a method, apparatus, storage medium, and computer equipment for identifying precursors of rock failure. Background Technology

[0002] In geotechnical engineering fields such as deep resource extraction, water conservancy and hydropower, slope stabilization, and tunnel excavation, rocks are often subjected to a static-dynamic superposition load environment, where high ground stress (static load) and engineering disturbances (such as dynamic loads such as blasting and mechanical vibration) work together. Under such complex load conditions, rock instability and failure are often sudden and catastrophic.

[0003] In related technologies, most rely on single-mode signals to monitor internal rock damage, making it difficult to comprehensively and accurately depict the entire damage evolution process of rocks under complex loads. Furthermore, the use of single-feature extraction methods in the time or frequency domains is poorly adapted to non-stationary signals, easily losing key time-frequency joint information, resulting in weak anti-interference capabilities, high false alarm rates in noisy environments, and difficulty in achieving early and reliable identification of damage precursors under static-dynamic superimposed load conditions on rocks. Summary of the Invention

[0004] In view of this, this application provides a method, device, storage medium and computer equipment for identifying precursors of rock damage. By combining the complementarity of dual-modal acoustic signals to achieve cross-validation of time-frequency singular entropy sequences, false alarms and missed alarms caused by the failure or interference of a single signal are avoided, making the early warning more advanced and reliable.

[0005] According to one aspect of this application, a method for identifying precursors of rock failure is provided, comprising:

[0006] The multimodal acoustic signals of rocks under superimposed loads of different types are acquired, and the multimodal acoustic signals include acoustic emission signals and active acoustic wave signals;

[0007] The multimodal acoustic signal is divided according to a preset time window to obtain multimodal acoustic signal segments;

[0008] Time-frequency analysis is performed on the multimodal acoustic signal segment to construct the time-frequency matrix of the multimodal acoustic signal segment;

[0009] Singular value decomposition is performed on the time-frequency matrix to obtain a sequence of singular values ​​of the time-frequency matrix;

[0010] Calculate the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence;

[0011] Based on the time-frequency singular entropy of the multimodal acoustic signal segment, construct the time-frequency singular entropy sequence of the multimodal acoustic signal;

[0012] Based on the time-frequency singular entropy sequence of the multimodal acoustic signal, the precursors of rock damage are identified.

[0013] According to another aspect of this application, a corpus management device is provided, comprising:

[0014] The acquisition module is used to acquire multimodal acoustic signals of rocks under superimposed loads of different types, the multimodal acoustic signals including acoustic emission signals and active acoustic wave signals;

[0015] The calculation module is configured to: divide the multimodal acoustic signal into multimodal acoustic signal segments according to a preset time window; perform time-frequency analysis on the multimodal acoustic signal segments to construct a time-frequency matrix of the multimodal acoustic signal segments; perform singular value decomposition on the time-frequency matrix to obtain a sequence of singular values ​​of the time-frequency matrix; calculate the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence; and construct a time-frequency singular entropy sequence of the multimodal acoustic signal based on the time-frequency singular entropy of the multimodal acoustic signal segments.

[0016] The identification module is used to identify the precursors of rock destruction based on the time-frequency singular entropy sequence of the multimodal acoustic signal.

[0017] According to another aspect of this application, a readable storage medium is provided having a program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method for identifying precursors of rock damage.

[0018] According to another aspect of this application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the steps of the above-described rock damage precursor identification method.

[0019] By employing the above technical solutions, this application provides a method, device, storage medium, and computer equipment for identifying precursors to rock failure. By fusing multimodal acoustic signals collected by acoustic emission sensors and acoustic transducers, and combining time-frequency analysis, matrix singular value decomposition, and information entropy theory, a novel nonlinear dynamic characteristic quantity, "time-frequency singular entropy," is constructed. This quantity is used to quantify the evolution of the energy distribution complexity of multimodal acoustic signals in rocks under complex loads within the time-frequency joint domain, thereby accurately capturing precursors to instability. Furthermore, the complementarity of dual-modal acoustic signals enables cross-validation of the time-frequency singular entropy sequence, avoiding false alarms and missed alarms caused by single signal failure or interference. This makes the early warning more advanced and reliable, maintaining high recognition performance even in noisy industrial environments, greatly facilitating engineering applications.

[0020] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0021] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0022] Figure 1 A flowchart illustrating the rock failure precursor identification method provided in an embodiment of this application is shown.

[0023] Figure 2 A flowchart illustrating a method for identifying precursors of rock failure according to another embodiment of this application is shown;

[0024] Figure 3 A schematic diagram of the time spectrum provided in another embodiment of this application is shown;

[0025] Figure 4 A schematic diagram of a time-frequency matrix provided in another embodiment of this application is shown;

[0026] Figure 5 A schematic diagram of a time-frequency singular entropy sequence provided in another embodiment of this application is shown;

[0027] Figure 6 A schematic diagram illustrating real-time monitoring using preset early warning conditions is shown in another embodiment of this application;

[0028] Figure 7 A structural block diagram of the rock damage precursor identification device provided in an embodiment of this application is shown. Detailed Implementation

[0029] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0030] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0031] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “attached” to another element, it can be directly connected or attached to the other element, or there may be intermediate elements. Furthermore, “connected” or “attached” as used herein can include wireless connections or wireless interconnections. The term “and / or” as used herein includes all or any unit and all combinations of one or more associated listed items.

[0032] Exemplary embodiments according to this application will now be described in more detail with reference to the accompanying drawings. However, these exemplary embodiments may be implemented in many different forms and should not be construed as being limited to the embodiments set forth herein. It should be understood that these embodiments are provided so that the disclosure of this application is thorough and complete, and that the concept of these exemplary embodiments is fully conveyed to those skilled in the art.

[0033] This application provides a method for identifying precursors to rock failure, such as... Figure 1 As shown, the method includes:

[0034] Step 101: Obtain multimodal acoustic signals of the rock under superimposed loads of different types.

[0035] Multimodal acoustic signals include acoustic emission signals and active acoustic wave signals.

[0036] In this step, acoustic emission sensors and acoustic transducers are placed on the rock surface to comprehensively collect multimodal acoustic signals emitted by the rock when it is subjected to superimposed loads of different types.

[0037] It is understandable that multimodal acoustic signals include acoustic emission signals obtained from acoustic emission sensors and active acoustic signals obtained from acoustic transducers.

[0038] Step 102: Divide the multimodal acoustic signal according to the preset time window to obtain multimodal acoustic signal segments.

[0039] In this step, the continuously acquired multimodal acoustic signals are divided according to a preset time window to obtain multimodal acoustic signal segments, so that each multimodal acoustic signal segment can be independently subjected to the same time-frequency analysis.

[0040] Step 103: Perform time-frequency analysis on the multimodal acoustic signal segment to construct the time-frequency matrix of the multimodal acoustic signal segment.

[0041] In this step, continuous wavelet transform is used as a time-frequency analysis tool. Taking advantage of its ability to provide optimal time-frequency resolution, time-frequency analysis is performed on non-stationary multimodal acoustic signal segments to construct the time-frequency matrix of the multimodal acoustic signal segments.

[0042] Step 104: Perform singular value decomposition on the time-frequency matrix to obtain the singular value sequence of the time-frequency matrix.

[0043] Step 105: Calculate the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence.

[0044] In this step, time-frequency analysis, matrix singular value decomposition, and information entropy are combined to calculate the time-frequency singular entropy of the multimodal acoustic signal segment. The time-frequency singular entropy, as a nonlinear dynamic characteristic quantity, is used to quantify the evolution of the energy distribution complexity of the multimodal acoustic signal segment of rock in the time-frequency joint domain under complex loads, thereby accurately capturing the precursors of instability.

[0045] Step 106: Construct the time-frequency singular entropy sequence of the multimodal acoustic signal based on the time-frequency singular entropy of the multimodal acoustic signal segment.

[0046] In this step, the time-frequency singular entropies of the acoustic emission signal segments within each preset time window are combined in chronological order to obtain the first time-frequency singular entropy sequence of the acoustic emission signal. Similarly, the time-frequency singular entropies of the active acoustic wave signal within each preset time window are combined in chronological order to obtain the second time-frequency singular entropy sequence of the active acoustic wave signal. Thus, two sequences of entropy values ​​evolving over time are formed.

[0047] Step 107: Identify precursors of rock damage based on time-frequency singular entropy sequences of multimodal acoustic signals.

[0048] This step avoids false alarms caused by single sensor failure or interference through dual-entropy synchronization, and ensures that precursors of different manifestations can be effectively captured through multi-criteria triggering. This greatly reduces the risk of missed alarms and ensures high reliability even when facing complex engineering environment noise and signal fluctuations, truly achieving earlier, more accurate and reliable identification of precursors of damage.

[0049] This embodiment, due to the extreme sensitivity of time-frequency singular entropy to changes in signal complexity, can capture early signs of damage that traditional time-domain or frequency-domain features cannot detect. Furthermore, by combining the complementarity of multimodal acoustic signals collected by acoustic emission sensors and acoustic transducers, cross-validation is achieved, avoiding false alarms or missed alarms caused by single signal failure or interference, making the early warning more advanced and reliable. Moreover, singular value decomposition itself has good denoising and stability, effectively suppressing environmental noise and random interference. Time-frequency analysis focuses signal energy, further highlighting characteristic frequency components related to damage. Therefore, this embodiment maintains high recognition performance even in noisy industrial environments. Simultaneously, this embodiment extracts features from a two-dimensional "time-frequency" space, containing richer information compared to one-dimensional time-domain or frequency-domain features. Through singular value decomposition and entropy calculation, the high-dimensional time-frequency information is condensed into a scalar index with clear physical meaning and easy tracking and analysis—time-frequency singular entropy—greatly facilitating engineering applications.

[0050] Another embodiment of this application provides a method for identifying precursors to rock failure, such as... Figure 2 As shown, the method includes:

[0051] Step 201: Obtain the original multimodal acoustic signals collected by the acoustic emission sensors and acoustic transducers arranged on the rock surface when the rock is subjected to superimposed loads of different types; perform bandpass filtering and wavelet threshold noise reduction processing on the original multimodal acoustic signals in sequence to obtain multimodal acoustic signals.

[0052] It should be noted that in geotechnical engineering fields such as deep resource extraction, water conservancy and hydropower, slope stabilization, and tunnel excavation, rock mass is the core load-bearing structure. Rock mass often endures a static-dynamic superimposed load environment, resulting from the combined effects of high ground stress (static load) and engineering disturbances (such as dynamic loads from blasting and mechanical vibration). Under such complex loading conditions, the instability and failure of the rock composing the rock mass often exhibit suddenness and catastrophic nature.

[0053] In this step, acoustic emission sensors and acoustic transducers are placed on the rock surface to comprehensively collect the original multimodal acoustic signals emitted by the rock during the superposition of different types of loads.

[0054] It is understandable that the original multimodal acoustic signal includes the original acoustic emission signal collected by the acoustic emission sensor and the original active acoustic signal collected by the acoustic transducer.

[0055] Specifically, the rock can be a rock sample in a laboratory environment or an engineering rock in a real-world application scenario, such as a mine slope or tunnel wall; this embodiment does not impose specific limitations. This step applies two different types of loads to the rock: static load and dynamic load, and then superimposes these two loads onto the rock. The static load is a continuous force, such as the constant, enormous pressure from the overlying strata on the engineering rock in a real-world application scenario. The dynamic load is a changing, fluctuating force, such as disturbances to the engineering rock caused by nearby blasting, mechanical vibration, or traffic loads. For example, in a laboratory environment, a servo-controlled testing machine can be used to apply a constant axial pressure to the rock sample to simulate a static load. Based on this constant static load, a dynamic cyclic or disturbance load is then superimposed on the rock sample, thereby simulating a scenario where a static-dynamic superimposed load is applied to the rock.

[0056] During the process of rock being subjected to static and dynamic superimposed loads, acoustic emission sensors can passively acquire the original acoustic emission signals generated inside the rock due to stress, deformation, or the formation of microcracks. The original acoustic emission signals are transient elastic wave signals. The acoustic transducer consists of a transmitting transducer and a receiving transducer in pairs. The transmitting transducer excites ultrasonic pulses at a specific center frequency at fixed intervals, while the receiving transducer is responsible for receiving the original active acoustic wave signals after penetrating the rock.

[0057] Furthermore, after acquiring the original multimodal acoustic signal, this step sequentially performs bandpass filtering and wavelet threshold noise reduction on the original multimodal acoustic signal to improve the signal-to-noise ratio of the original multimodal acoustic signal, thereby obtaining the multimodal acoustic signal. This separates the real and useful acoustic signal from various interference noises, providing a data foundation for subsequent accurate identification.

[0058] It is understandable that multimodal acoustic signals include acoustic emission signals after bandpass filtering and wavelet threshold denoising of the original acoustic emission signals, as well as active acoustic signals after bandpass filtering and wavelet threshold denoising of the original active acoustic signals.

[0059] Specifically, bandpass filtering only allows the original multimodal acoustic signal within a specific frequency range to pass through, while forcibly filtering out signals outside this range. Wavelet thresholding denoising removes noise from the original multimodal acoustic signal after bandpass filtering, eliminating residual noise. In this step, wavelet thresholding uses both hard and soft thresholding to effectively suppress Gaussian white noise.

[0060] For example, bandpass filtering processing represents ,in, The original multimodal acoustic signal, This is the original multimodal acoustic signal after bandpass filtering. This is the Qualcomm cutoff frequency. The low-pass cutoff frequency is (high-pass cutoff frequency, low-pass cutoff frequency), which represents the specific frequency range corresponding to the band-pass filtering process. This is a bandpass filtering process. It should be noted that the low-pass cutoff frequency and high-pass cutoff frequency are specifically set based on the acoustic emission sensor, acoustic transducer, and noise characteristics; this embodiment does not impose specific limitations.

[0061] Furthermore, in the process of wavelet thresholding denoising of the original multimodal acoustic signal after bandpass filtering, firstly, the original multimodal acoustic signal after bandpass filtering is decomposed through wavelet transform to obtain candidate wavelet coefficients. Next, a preset threshold is set based on the noise level of the original multimodal acoustic signal after bandpass filtering. Then, all candidate wavelet coefficients with absolute values ​​less than the preset threshold are directly set to zero, while all candidate wavelet coefficients with values ​​greater than the preset threshold are retained for hard thresholding. Based on the hard thresholding, the retained candidate wavelet coefficients are then shrunk towards zero by a certain amount for soft thresholding. Thus, the soft-thresholded candidate wavelet coefficients are recombinated into a new, smoother signal, which is then used as the multimodal acoustic signal. The soft thresholding shrinkage involves bringing candidate wavelet coefficients greater than the preset threshold closer to zero, and the shrinkage amount is the preset threshold itself. For example, if the candidate wavelet coefficient is 5 and the preset threshold is 2, after soft thresholding, it becomes 3 (5-2). This ensures that the candidate wavelet coefficients change continuously at a preset threshold, avoiding abrupt changes.

[0062] Step 202: Divide the multimodal acoustic signal according to the preset time window to obtain multimodal acoustic signal segments.

[0063] In this step, the continuously acquired multimodal acoustic signals are divided according to a preset time window to obtain multimodal acoustic signal segments, so that each multimodal acoustic signal segment can be independently subjected to the same time-frequency analysis.

[0064] Specifically, a fixed-size time window, known as a preset time window, is pre-set. This preset time window is used to divide the continuous multimodal acoustic signal into a series of shorter multimodal acoustic signal segments in chronological order. In practical applications, the size of the preset time window is set according to the specific segmentation requirements. Preset time windows may or may not overlap; this embodiment does not impose specific restrictions. For example, the preset time window needs to include the key time-frequency structure of the signal, but it cannot be too long to capture dynamic changes. A balance is achieved between time resolution and frequency resolution to meet the required real-time warning requirements. For instance, the preset time window may be set within the range of 1 to 10 seconds.

[0065] It should be noted that, for acoustic emission signals and active acoustic wave signals as multimodal acoustic signals,

[0066] The segmentation process is performed synchronously. That is, starting from the beginning of the time sequence of the acoustic emission signal and the active acoustic wave signal, the acoustic emission signal and the active acoustic wave signal are synchronously segmented at the same position using a preset time window of the same size, thereby obtaining the acoustic emission signal segment and the active acoustic wave signal segment within the same preset time window.

[0067] Step 203: Perform continuous wavelet transform on the multimodal acoustic signal segment to perform time-frequency analysis on the multimodal acoustic signal segment and construct the time-frequency matrix of the multimodal acoustic signal segment.

[0068] In this step, continuous wavelet transform is used as a time-frequency analysis tool. Taking advantage of its ability to provide optimal time-frequency resolution, time-frequency analysis is performed on non-stationary multimodal acoustic signal segments to construct the time-frequency matrix of the multimodal acoustic signal segments.

[0069] For example, the formula for continuous wavelet transform is expressed as:

[0070] ,

[0071] in, The scaling factor controls the mother wavelet. Scaling; when scaling the mother wavelet Scaling alters the mother wavelet. The time width of the mother wavelet is thus altered, thereby changing its oscillation rhythm. The center frequency will also change from its fixed center frequency, thus each scale factor It corresponds to a frequency point. This is the time shift factor, used to shift the mother wavelet. The center moves to the physical time axis of the multimodal acoustic signal segment. In this position, the time shift factor It corresponds to a specific point in time. For multimodal acoustic signal segments, it is understandable that... Including acoustic emission signal segment With active acoustic signal segment . Mother wavelet pass Scaling and The sub-wavelet generated after translation for . for In scale With time translation Continuous wavelet coefficients at the location, It is a complex number, and its modulus is This represents the signal energy density at the corresponding time and frequency point.

[0072] Therefore, based on the continuous wavelet coefficients of the multimodal acoustic signal segment at different scales and time shifts, a continuous wavelet coefficient matrix can be obtained. Taking the modulus of the continuous wavelet coefficient matrix yields the time-frequency energy distribution matrix of the multimodal acoustic signal segment, which is then used as the time-frequency matrix of the multimodal acoustic signal segment. It can be understood that the time-frequency matrix of the multimodal acoustic signal segment includes both the time-frequency matrix of the acoustic emission signal segment and the time-frequency matrix of the active acoustic wave signal segment.

[0073] In this step, the continuous wavelet transform formula uses sub-wavelets to match multimodal acoustic signals through inner product operations, obtaining continuous wavelet coefficients at different scales and locations, thereby reflecting the local time-domain characteristics of the multimodal acoustic signal segment.

[0074] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, step 203, namely, performing continuous wavelet transform on the multimodal acoustic signal segment to perform time-frequency analysis on the multimodal acoustic signal segment and constructing the time-frequency matrix of the multimodal acoustic signal segment, specifically includes: discretizing the multimodal acoustic signal segment according to a preset sampling time point to obtain a discrete signal; scaling a preset mother wavelet based on a preset scaling factor, and matching the scaled preset mother wavelet with the discrete signal through convolution operation to obtain a continuous wavelet coefficient matrix; taking the modulus of the continuous wavelet coefficient matrix to obtain the time-frequency energy distribution matrix of the multimodal acoustic signal segment, and using the time-frequency energy distribution matrix as the time-frequency matrix of the multimodal acoustic signal segment.

[0075] It should be noted that the definition of the continuous wavelet transform formula is to calculate continuous signal waveforms, but computers cannot process infinite continuous data.

[0076] In this step, the continuous process is simulated on a computer through discretization sampling.

[0077] Specifically, the multimodal acoustic signal segments are sampled at equally spaced time points (i.e., preset sampling time points) to obtain discrete signals at each preset sampling time point, thereby converting the continuous multimodal acoustic signal segments into a discrete signal sequence that can be processed by a computer. Similarly, a scale factor is selected within a preset scale range to obtain a preset scale factor. Likewise, in the definition of the continuous wavelet transform formula, the time shift factor is a parameter that can continuously change in the time domain. When entering discrete computation, it is impossible to calculate an infinite number of time shift factors, so it is necessary to discretize the time shift factor in the time domain. In this step, the preset sampling time points are directly used as the discretized time shift factors to discretize the time shift factors and improve computational efficiency.

[0078] Next, the Morlet (Mexhat wavelet) wavelet is selected as the preset mother wavelet. For the Morlet wavelet, the integral in the continuous wavelet transform formula is mathematically equivalent to a convolution operation. Therefore, for any combination of preset scale factor and discretized time shift factor, the multimodal acoustic signal segment is convolved using the aforementioned continuous wavelet transform formula. Taking advantage of its good locality in both the time and frequency domains, a continuous wavelet coefficient matrix is ​​obtained, consisting of continuous wavelet coefficients for the multimodal acoustic signal segment at different combinations of preset scale factor and discretized time shift factor.

[0079] For example, the preset sampling time point is represented as Preset sampling time point The discrete signal below is The preset scaling factor is expressed as: , M The number of preset scale factors.

[0080] Multimodal acoustic signal segment Different preset scale factors With the discretized time shift factor The continuous wavelet coefficient matrix formed by the continuous wavelet coefficients at the combination point is expressed as:

[0081] ,

[0082] in, For multimodal acoustic signal segments In the preset scale factor With the discretized time shift factor Continuous wavelet coefficients at the combination point; It is a continuous wavelet coefficient matrix, which is a A complex matrix.

[0083] For example, the preset scale range is determined based on the frequency range where the main energy of the multimodal acoustic signal is concentrated, so as to avoid missing important frequency bands due to improper scale selection. The preset sampling time point is set according to the requirements of the accuracy of the warning in the actual application scenario. This embodiment does not make specific settings.

[0084] Furthermore, for the continuous wavelet coefficient matrix By performing modulo operation, the time-frequency energy distribution matrix is ​​obtained. Time-frequency energy distribution matrix elements For multimodal acoustic signal segments In scale factor With the discretized time shift factor Signal energy at the junction. Time-frequency energy distribution matrix. It is a size of The real matrix. Therefore, the time-frequency energy distribution matrix... As a time-frequency matrix for multimodal acoustic signal segments.

[0085] Step 204: Perform singular value decomposition on the time-frequency matrix to obtain the singular value sequence of the time-frequency matrix.

[0086] In this step, the time-frequency matrix is... Perform singular value decomposition on the time-frequency matrix. It can be decomposed into the product of three matrices, as follows: .in, For one The unitary matrix, whose column vectors are left singular vectors, forms an orthogonal basis for the frequency space of the time-frequency matrix; For one The unitary matrix, whose column vectors are right singular vectors, forms an orthogonal basis for the time space of the time-frequency matrix; For one A diagonal matrix, whose diagonal elements That is, the time-frequency matrix singular values, .

[0087] It is worth mentioning that singular value decomposition, through the analysis of the time-frequency matrix... Orthogonal decomposition was performed, enabling the redistribution and priority ranking of the contained energy. This decomposition transforms the time-frequency matrix... The matrix contains global, high-energy characteristic structures, namely the dominant modes that persist and concentrate energy in the time-frequency domain, which are mapped to the eigenspace corresponding to the leading singular values. The feature pairs carried by these leading singular values ​​are crucial in reconstructing the time-frequency matrix. The singular values ​​contribute the vast majority of energy and thus characterize the macroscopic dynamics of multimodal acoustic signal segments. Conversely, the later-arriving singular values, with significantly smaller values, often correspond to localized, low-energy fine structures in the time-frequency domain. These structures originate from environmental noise, measurement disturbances, or other non-critical signal components. Therefore, the singular value sequence essentially constitutes a descending energy distribution spectrum, whose distribution directly reflects the energy ratio between the main features and secondary noise in a multimodal acoustic signal segment, providing a scientific data foundation for subsequent steps.

[0088] For example, singular values ​​satisfy nonnegativity and are arranged in descending order: The magnitude of each singular value represents the importance of its corresponding singular vector pair in reconstructing the time-frequency matrix A, i.e., the energy it contributes. Further, a sequence of singular values ​​is extracted. This constitutes a set of stable, orthogonal, and energy-compact eigenvectors describing the characteristics of the original time-frequency matrix A. Large singular values ​​correspond to the main energy patterns in the time-frequency plot corresponding to the time-frequency matrix, while small singular values ​​often correspond to noise or minor components. Thus, the singular value sequence can be used to characterize the rock damage state.

[0089] This embodiment extracts features from a two-dimensional time-frequency matrix, compressing the complex two-dimensional information into a one-dimensional feature sequence to obtain a singular value sequence, which provides a data foundation for calculating the entropy value in subsequent steps.

[0090] Step 205: Calculate the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence.

[0091] In this step, time-frequency analysis, matrix singular value decomposition, and information entropy are combined to calculate the time-frequency singular entropy of the multimodal acoustic signal segment. The time-frequency singular entropy, as a nonlinear dynamic characteristic quantity, is used to quantify the evolution of the energy distribution complexity of the multimodal acoustic signal segment of rock in the time-frequency joint domain under complex loads, thereby accurately capturing the precursors of instability.

[0092] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, and to fully illustrate the specific implementation process of this embodiment, step 205, namely, calculating the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence, specifically includes: taking the singular value sequence as a probability distribution, determining the energy value corresponding to the singular value based on the square of the singular value in the singular value sequence; determining the total energy value corresponding to the singular value sequence based on the energy value corresponding to the singular value; determining the energy proportion of the singular value based on the energy value corresponding to the singular value and the total energy value corresponding to the singular value sequence; determining the energy distribution corresponding to the singular value sequence based on the energy proportion of the singular value; and calculating the information entropy of the energy distribution to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence.

[0093] In this step, the singular value sequence is used as a probability distribution. First, the energy proportion of each singular value in the total energy value corresponding to the singular value sequence is calculated.

[0094] For example, energy density is expressed as: .in, For the first The square of a singular value is proportional to the energy carried by that singular mode, representing the energy value corresponding to the singular value. Let be the sum of squares of all singular values, representing the total energy of the time-frequency matrix A. Here, the square is the square of the Frobenius norm. Understandably, Representing the The percentage of energy carried by each singular value relative to the total energy.

[0095] Furthermore, based on the energy proportion of each singular value, the energy distribution corresponding to the entire singular value sequence is obtained. Therefore, based on information entropy theory, the probability entropy of this energy distribution is calculated, which is the time-frequency singular entropy.

[0096] For example, the time-frequency singular entropy is represented as: .in, For the first k The time-frequency singular entropy of a multimodal acoustic signal segment within a preset time window, expressed in bits. This is understandable. Including the k The first time-frequency singular entropy of the acoustic emission signal segment within a preset time window and the second time-frequency singular entropy of the active acoustic signal segment .

[0097] It is worth noting that in the early stages of rock damage, the multimodal acoustic signals are relatively simple and stable. The acquired multimodal acoustic signal segments mainly consist of a few dominant frequency components, and the energy of the multimodal acoustic signal segments is highly concentrated in the first few large singular values. The energy distribution is concentrated, and the time-frequency singular entropy of the multimodal acoustic signal segments is low. When the rock is close to failure and microfractures are clustered, the multimodal acoustic signals generate rich and variable frequency components. The acquired multimodal acoustic signal segments become complex and chaotic, with the energy distribution dispersed across numerous singular values. The energy distribution tends to be uniform, and the time-frequency singular entropy of the multimodal acoustic signal segments increases significantly. Therefore, a sharp increase in time-frequency singular entropy can serve as an extremely sensitive precursor indicator of rock instability.

[0098] Step 206: Construct the time-frequency singular entropy sequence of the multimodal acoustic signal based on the time-frequency singular entropy of the multimodal acoustic signal segment.

[0099] In this step, the time-frequency singular entropies of the acoustic emission signal segments within each preset time window are combined in chronological order to obtain the first time-frequency singular entropy sequence of the acoustic emission signal. Similarly, the time-frequency singular entropies of the active acoustic wave signal segments within each preset time window are combined in chronological order to obtain the second time-frequency singular entropy sequence of the active acoustic wave signal. Thus, two sequences of entropy values ​​evolving over time are formed.

[0100] For example, the first time-frequency singular entropy sequence is represented as: The second time-frequency singular entropy sequence is represented as: .

[0101] Step 207: Identify precursors of rock damage based on the time-frequency singular entropy sequence of multimodal acoustic signals.

[0102] In this step, for the two constructed time-frequency singular entropy sequences, the synchronicity of their trends is captured, and continuous, dynamic, and multi-condition fusion decision-making is carried out to avoid false alarms and missed alarms, thereby achieving reliable identification and early warning of rock damage precursors.

[0103] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, step 207, namely, identifying rock damage precursors based on the time-frequency singular entropy sequence of multimodal acoustic signals, specifically includes: synchronously determining the current window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence according to the time sequence based on the original window size; determining the first identification index and the second identification index of the current window according to the first time-frequency singular entropy and the second time-frequency singular entropy within the current window respectively; if the first identification index and the second identification index of the current window meet the preset warning conditions, triggering a rock damage precursor warning and stopping the sliding window; if the first identification index and the second identification index of the current window do not meet the preset warning conditions, synchronously determining the next window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence according to the end position of the current window in the time sequence and the original window size, so as to synchronously slide the window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence.

[0104] In this step, a sliding window is synchronously applied to the first time-frequency singular entropy sequence of the acoustic emission signal and the second time-frequency singular entropy sequence of the active acoustic wave signal. Each window covers several first and second time-frequency singular entropies located in the same time sequence within both sequences, used to calculate the first identification index of the first time-frequency singular entropy and the second identification index of the second time-frequency singular entropy within the window. Then, the control window slides forward by a preset step size according to the time sequence, and the calculation is performed again. This allows for the capture of the synchronous dynamic change trend of rock instability precursors from different dimensions within the first and second time-frequency singular entropy sequences.

[0105] The preset step size is set according to the real-time and accuracy requirements of the warning in the actual application scenario, and this embodiment does not impose specific restrictions.

[0106] Specifically, starting from the beginning of the first and second time-frequency singular entropy sequences in time sequence, a current window is synchronously determined on both sequences according to the original window size. This ensures that the current window covers several first and second time-frequency singular entropies that are in the same time sequence within both sequences. Next, a first identification index is calculated based on the first time-frequency singular entropy within the current window, and a second identification index is calculated based on the second time-frequency singular entropy within the current window. Therefore, based on preset warning conditions, a logical judgment is made on the first and second identification indices of the current window. If the first and second identification indices of the current window meet the preset warning conditions, it indicates that the rock is about to be damaged, the sliding window stops, and a rock damage precursor warning is triggered. If the first and second identification indicators of the current window do not meet the preset warning conditions, the window continues to slide synchronously on the first and second time-frequency singular entropy sequences. That is, without changing the window size, the current window continues to slide synchronously for a preset step length from the end position of the current window on the first and second time-frequency singular entropy sequences to obtain the next window on the first and second time-frequency singular entropy sequences. This achieves synchronous sliding of the window on the first and second time-frequency singular entropy sequences for continuous monitoring of the rock.

[0107] This embodiment avoids false alarms caused by single sensor failure or interference through dual-entropy synchronization, and ensures that different manifestations of precursors can be effectively captured through multi-criteria triggering, greatly reducing the risk of missed alarms. It ensures high reliability even when facing complex engineering environment noise and signal fluctuations, and truly achieves earlier, more accurate and reliable identification of precursors of damage.

[0108] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, a first identification index and a second identification index of the current window are determined based on the first time-frequency singular entropy and the second time-frequency singular entropy within the current window, respectively. Specifically, this includes: performing a first-order difference operation on the first time-frequency singular entropy and the second time-frequency singular entropy within the current window to determine the change rate of the first time-frequency singular entropy and the change rate of the second time-frequency singular entropy within the current window; using the change rate of the first time-frequency singular entropy and the change rate of the second time-frequency singular entropy as the first identification index and the second identification index, respectively; and the preset warning conditions include: the first identification index being greater than a preset threshold for the change rate of the first time-frequency singular entropy, and the second identification index being greater than a preset threshold for the change rate of the second time-frequency singular entropy.

[0109] In this step, after obtaining the current window on the first and second time-frequency singular entropy sequences, a first-order difference operation is used to calculate the rate of change of the first time-frequency singular entropy within the current window (i.e., the first time-frequency singular entropy change rate) and the rate of change of the second time-frequency singular entropy within the current window (i.e., the second time-frequency singular entropy change rate). The rate of change of the first time-frequency singular entropy within the current window is used as the first identification index, and the rate of change of the second time-frequency singular entropy within the current window is used as the second identification index.

[0110] At this point, the preset warning conditions are set as follows: the first identification indicator is greater than the preset first time-frequency singular entropy change rate threshold, and the second identification indicator is greater than the preset second time-frequency singular entropy change rate threshold. If the first and second identification indicators meet the preset warning conditions, that is, the first time-frequency singular entropy change rate within the current window is greater than the preset first time-frequency singular entropy change rate threshold, and the second time-frequency singular entropy change rate within the current window is greater than the preset second time-frequency singular entropy change rate threshold, it indicates that the first and second time-frequency singular entropy sequences simultaneously show a rapid increase, indicating that the damage process inside the rock is accelerating, which is a strong instability signal, thus triggering a rock failure precursor warning.

[0111] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, the rock failure precursor identification method further includes: determining a baseline window based on a preset number of windows preceding the current window in chronological order; determining a first safety threshold based on the average value and standard deviation of the first time-frequency singular entropy within the baseline window; determining a second safety threshold based on the average value and standard deviation of the second time-frequency singular entropy within the baseline window; determining a first identification index and a second identification index for the current window based on the first time-frequency singular entropy and the second time-frequency singular entropy within the current window, specifically including: using the average value of the first time-frequency singular entropy and the average value of the second time-frequency singular entropy within the current window as the first identification index and the second identification index, respectively; the preset warning conditions include: the first identification index is greater than the first safety threshold corresponding to the first time-frequency singular entropy within the current window, and the second identification index is greater than the second safety threshold corresponding to the second time-frequency singular entropy within the current window.

[0112] In this step, the baseline levels of the first and second time-frequency singular entropy sequences within the recent sliding window are first calculated. Then, the first and second time-frequency singular entropies within the current window are compared with the recent baseline levels to determine whether they are significantly abnormally high at the same time, thereby determining whether the rock is about to be damaged.

[0113] Specifically, a baseline window is determined based on a preset number of windows preceding the current window in chronological order. Next, the mean and standard deviation of the first time-frequency singular entropy within the baseline window are calculated, and the sum of the mean of the first time-frequency singular entropy and its standard deviation at a preset multiple is set as the first safety threshold for the current window. Similarly, the mean and standard deviation of the second time-frequency singular entropy within the baseline window are calculated, and the sum of the mean of the second time-frequency singular entropy and its standard deviation at a preset multiple is set as the second safety threshold for the current window.

[0114] For example, the preset multiplier can be set to 2 to 3.

[0115] Furthermore, the average value of the first time-frequency singular entropy and the average value of the second time-frequency singular entropy within the current window are calculated, and these average values ​​are used as the first identification index and the second identification index, respectively.

[0116] At this point, the preset warning conditions are set as follows: the first identification indicator is greater than the first safety threshold corresponding to the first time-frequency singular entropy within the current window, and the second identification indicator is greater than the second safety threshold corresponding to the second time-frequency singular entropy within the current window. If the first and second identification indicators within the current window meet the preset warning conditions, that is, the average value of the first time-frequency singular entropy within the current window is greater than the first safety threshold, and the second time-frequency singular entropy within the current window is greater than the second safety threshold, it indicates that the first and second time-frequency singular entropy sequences have significantly deviated from the normal fluctuation range and entered the abnormal interval, which is another strong instability signal, thus triggering a rock failure precursor warning.

[0117] In another embodiment of this application, when a granite specimen is subjected to a uniaxial compression test in a laboratory environment, signs of impending failure are identified.

[0118] First, a standard cylindrical granite specimen (50 mm radius, 100 mm length) was placed on the platform of a rigid testing machine. The machine applied an axial compressive load (static load) at a displacement control rate of 0.5 mm / min to simulate a high ground stress environment. Next, two broadband (50 kHz–1 MHz) acoustic emission sensors and one acoustic transducer pair (including a transmitting transducer T and a receiving transducer R) were uniformly arranged on the side of the granite specimen's center. The acoustic emission sensors were directly coupled to the specimen surface to passively receive the raw acoustic emission signals from the micro-fractures. The acoustic transducers were coupled to the specimen via a couplant. The transmitting transducer T excited an ultrasonic pulse with a center frequency of 250 kHz every 1 second, and the receiving transducer R recorded the raw active acoustic signal after penetrating the specimen. All signals were synchronously acquired by a National Instruments PXIe-1082 data acquisition system with a sampling rate set to 2 MHz and amplified by a preamplifier (40 dB gain). The acquired raw multimodal acoustic signals were then transmitted to an industrial computer equipped with the software specified in this application. The software automatically processes the data according to the following steps: The original multimodal acoustic signal is subjected to a 50kHz high-pass filter to remove low-frequency noise, followed by wavelet denoising to obtain the multimodal acoustic signal. The multimodal acoustic signal is then divided into segments, and time-frequency analysis is performed on these segments. Morlet wavelet transform is applied to the waveforms of each acoustic emission signal segment and each active acoustic wave signal segment to obtain the desired results. Figure 3 and Figure 4 The time-frequency spectrum and time-frequency matrix are shown. As the experiment progresses, the time-frequency singular entropy values ​​of the multimodal acoustic signal segments are continuously calculated and stored, forming a matrix as shown below. Figure 5 The time-series evolution curves are shown, and the two curves are monitored in real time through built-in preset early warning conditions, such as... Figure 6 Monitor the two curves in real time.

[0119] In this embodiment, a first-level warning is triggered when the average of the first time-frequency singular entropy and the second time-frequency singular entropy within the current window (window = 50s) simultaneously exceeds three times the standard deviation of their respective previous baseline average (first 100s); and a highest-level warning is triggered when the rate of change of both simultaneously exceeds a set threshold.

[0120] In this test, the axial load increased steadily. When the load reached approximately 85% of the peak stress (corresponding to...), the load decreased. Figure 5 and Figure 6 Around time step 350, it was recorded that the first and second time-frequency singular entropies began to increase synchronously, rapidly, and continuously (e.g., Figure 5As indicated by the precursor starting point, the fluctuation range far exceeded the previous range. Approximately 10 seconds later (around time step 360), an automatic warning was triggered based on preset warning conditions. Approximately 40 seconds later, the sample underwent macroscopic fracture (at the peak stress).

[0121] This embodiment successfully issued an early warning approximately 40 seconds before macroscopic damage to the rock, and the warning was based on the collaborative verification of dual-modal signals, making the possibility of a false alarm extremely low. This indicates that the method provided in this embodiment can effectively capture the qualitative change process of internal rock damage from stable accumulation to a critical state of instability, achieving true precursor identification.

[0122] It should be noted that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0123] Furthermore, such as Figure 7 As shown, as a specific implementation of the above-mentioned rock damage precursor identification method, this application embodiment provides a rock damage precursor identification device 700, which includes: an acquisition module 701, a calculation module 702 and an identification module 703.

[0124] The acquisition module 701 is used to acquire multimodal acoustic signals of rocks under superimposed loads of different types. The multimodal acoustic signals include acoustic emission signals and active acoustic wave signals.

[0125] The calculation module 702 is used to divide the multimodal acoustic signal according to a preset time window to obtain multimodal acoustic signal segments; and to perform time-frequency analysis on the multimodal acoustic signal segments to construct the time-frequency matrix of the multimodal acoustic signal segments; and to perform singular value decomposition on the time-frequency matrix to obtain the singular value sequence of the time-frequency matrix; and to calculate the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence; and to construct the time-frequency singular entropy sequence of the multimodal acoustic signal based on the time-frequency singular entropy of the multimodal acoustic signal segment.

[0126] The identification module 703 is used to identify precursors of rock damage based on time-frequency singular entropy sequences of multimodal acoustic signals.

[0127] Optionally, the acquisition module 701 is specifically used to acquire the original multimodal acoustic signals collected by the acoustic emission sensor and acoustic transducer arranged on the rock surface when the rock is subjected to different types of superimposed loads; and to perform bandpass filtering and wavelet threshold noise reduction processing on the original multimodal acoustic signals in sequence to obtain multimodal acoustic signals.

[0128] The calculation module 702 is specifically used to discretize the multimodal acoustic signal segment according to the preset sampling time point to obtain a discrete signal; scale the preset mother wavelet based on the preset scaling factor, and match the scaled preset mother wavelet with the discrete signal through convolution operation to obtain a continuous wavelet coefficient matrix; take the modulus of the continuous wavelet coefficient matrix to obtain the time-frequency energy distribution matrix of the multimodal acoustic signal segment, and use the time-frequency energy distribution matrix as the time-frequency matrix of the multimodal acoustic signal segment.

[0129] The calculation module 702 is specifically used to treat the singular value sequence as a probability distribution, determine the energy value corresponding to the singular value based on the square of the singular value in the singular value sequence; determine the total energy value corresponding to the singular value sequence based on the energy value corresponding to the singular value; determine the energy weight of the singular value based on the energy value corresponding to the singular value and the total energy value corresponding to the singular value sequence; determine the energy distribution corresponding to the singular value sequence based on the energy weight of the singular value; calculate the information entropy of the energy distribution, and obtain the time-frequency singular entropy of the multimodal acoustic signal segment corresponding to the singular value sequence.

[0130] The identification module 703 is specifically used to synchronously determine the current window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence according to the original window size and in chronological order; determine the first identification index and the second identification index of the current window according to the first time-frequency singular entropy and the second time-frequency singular entropy within the current window respectively; if the first identification index and the second identification index of the current window meet the preset warning conditions, trigger the rock failure precursor warning and stop the sliding window; if the first identification index and the second identification index of the current window do not meet the preset warning conditions, synchronously determine the next window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence according to the end position of the current window in chronological order and the original window size, so as to synchronously slide the window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence.

[0131] Perform first-order difference operations on the first time-frequency singular entropy and the second time-frequency singular entropy within the current window to determine the change rate of the first time-frequency singular entropy and the change rate of the second time-frequency singular entropy within the current window; use the change rate of the first time-frequency singular entropy and the change rate of the second time-frequency singular entropy as the first identification index and the second identification index, respectively; preset warning conditions include: the first identification index is greater than the preset threshold for the change rate of the first time-frequency singular entropy, and the second identification index is greater than the preset threshold for the change rate of the second time-frequency singular entropy.

[0132] The identification module 703 is specifically used to take the average value of the first time-frequency singular entropy and the average value of the second time-frequency singular entropy within the current window as the first identification index and the second identification index, respectively. The preset warning conditions include: the first identification index is greater than the first safety threshold corresponding to the first time-frequency singular entropy within the current window, and the second identification index is greater than the second safety threshold corresponding to the second time-frequency singular entropy within the current window. The rock failure precursor identification method also includes: determining a baseline window based on a preset number of windows preceding the current window in time sequence; determining a first safety threshold based on the average value and standard deviation of the first time-frequency singular entropy within the baseline window; and determining a second safety threshold based on the average value and standard deviation of the second time-frequency singular entropy within the baseline window.

[0133] Specific limitations regarding the rock failure precursor identification device can be found in the limitations of the rock failure precursor identification method described above, and will not be repeated here. Each module in the aforementioned rock failure precursor identification device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0134] Based on the above, Figures 1 to 2 Accordingly, embodiments of this application also provide a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. Figures 1 to 2 The method for identifying precursors to rock damage is shown.

[0135] Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause a computer device (such as personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0136] Based on the above, Figures 1 to 2 The method shown, and Figure 7 To achieve the above objectives, the present application also provides a computer device, specifically a personal computer, server, network device, etc., as shown in the virtual device embodiment. This computer device includes a storage medium and a processor; the storage medium stores a computer program; the processor executes the computer program to achieve the above-described objectives. Figures 1 to 2 The method for identifying precursors to rock damage is shown.

[0137] Optionally, the computer device may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB ports, card reader ports, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Bluetooth interfaces, Wi-Fi interfaces), etc.

[0138] Those skilled in the art will understand that the computer device structure provided in this embodiment does not constitute a limitation on the computer device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0139] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages and stores the hardware and software resources of a computer device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software within the physical device.

[0140] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platform, or the embodiments of this application can be implemented by hardware.

[0141] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application. Those skilled in the art will understand that the modules in the apparatus of the embodiment can be distributed within the apparatus of the embodiment as described, or can be modified to be located in one or more apparatuses different from this embodiment. The modules of the above-described embodiment can be combined into one module, or further divided into multiple sub-modules.

[0142] The serial numbers in this application are for descriptive purposes only and do not represent the superiority or inferiority of any particular implementation scenario. The above disclosures are merely a few specific implementation scenarios of this application; however, this application is not limited thereto, and any variations conceived by those skilled in the art should fall within the protection scope of this application.

Claims

1. A rock failure precursor identification method characterized by, The method comprises: acquiring multi-modal acoustic signals of rock under superposition of different types of loads, the multi-modal acoustic signals comprising acoustic emission signals and active acoustic wave signals; dividing the multi-modal acoustic signals according to a preset time window to obtain multi-modal acoustic signal segments; performing time-frequency analysis on the multi-modal acoustic signal segments to construct a time-frequency matrix of the multi-modal acoustic signal segments; performing singular value decomposition on the time-frequency matrix to obtain a singular value sequence of the time-frequency matrix; calculating information entropy of the singular value sequence to obtain a time-frequency singular entropy of the multi-modal acoustic signal segment corresponding to the singular value sequence; constructing a time-frequency singular entropy sequence of the multi-modal acoustic signals according to the time-frequency singular entropy of the multi-modal acoustic signal segment; identifying the rock failure precursor based on the time-frequency singular entropy sequence of the multi-modal acoustic signals; the time-frequency analysis performed on the multi-modal acoustic signal segments to construct the time-frequency matrix of the multi-modal acoustic signal segments specifically comprises: discretizing the multi-modal acoustic signal segments according to a preset sampling time point to obtain discrete signals; scaling a preset mother wavelet based on a preset scale factor and matching the scaled preset mother wavelet with the discrete signals through convolution operation to obtain a continuous wavelet coefficient matrix; modulating the continuous wavelet coefficient matrix to obtain a time-frequency energy distribution matrix of the multi-modal acoustic signal segments, and taking the time-frequency energy distribution matrix as the time-frequency matrix of the multi-modal acoustic signal segments; the calculation of the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multi-modal acoustic signal segment corresponding to the singular value sequence specifically comprises: taking the singular value sequence as a probability distribution, and determining an energy value corresponding to the singular value according to the square of the singular value in the singular value sequence; determining a total energy value corresponding to the singular value sequence based on the energy value corresponding to the singular value; determining an energy proportion of the singular value according to the energy value corresponding to the singular value and the total energy value corresponding to the singular value sequence; determining an energy distribution corresponding to the singular value sequence according to the energy proportion of the singular value; calculating the information entropy of the energy distribution to obtain the time-frequency singular entropy of the multi-modal acoustic signal segment corresponding to the singular value sequence; the multi-modal acoustic signal segment comprises an acoustic emission signal segment and an active acoustic wave signal segment, the time-frequency singular entropy of the multi-modal acoustic signal segment comprises a first time-frequency singular entropy of the acoustic emission signal segment and a second time-frequency singular entropy of the active acoustic wave signal segment, the time-frequency singular entropy sequence of the multi-modal acoustic signals comprises a first time-frequency singular entropy sequence of the acoustic emission signal and a second time-frequency singular entropy sequence of the active acoustic wave signal, and the identification of the rock failure precursor based on the time-frequency singular entropy sequence of the multi-modal acoustic signals specifically comprises: synchronously determining a current window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence in time sequence according to an original window size; determining a first identification index and a second identification index of the current window according to the first time-frequency singular entropy and the second time-frequency singular entropy in the current window, respectively; If the first identification index and the second identification index of the current window meet a preset early warning condition, triggering a rock failure precursor early warning, and stopping the sliding window; If the first identification index and the second identification index of the current window do not meet the preset early warning condition, according to the endpoint position of the current window in time sequence and the original window size, synchronously determining a next window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence, to synchronously slide the window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence.

2. The rock failure precursor identification method according to claim 1, characterized by, The method comprises the following steps: obtaining original multi-modal acoustic signals collected by an acoustic emission sensor and a sound wave transducer arranged on the surface of the rock when the rock is subjected to different types of load superposition; sequentially performing band-pass filtering processing and wavelet threshold denoising processing on the original multi-modal acoustic signals to obtain the multi-modal acoustic signals.

3. The rock failure precursor identification method according to claim 1, characterized by, The method comprises the following steps: respectively performing first-order difference operation on the first time-frequency singular entropy and the second time-frequency singular entropy in the current window to determine a first time-frequency singular entropy change rate and a second time-frequency singular entropy change rate in the current window; respectively taking the first time-frequency singular entropy change rate and the second time-frequency singular entropy change rate as the first identification index and the second identification index; The preset early warning condition comprises: the first identification index is greater than a preset first time-frequency singular entropy change rate threshold, and the second identification index is greater than a preset second time-frequency singular entropy change rate threshold.

4. The rock failure precursor identification method according to claim 1, characterized by, The method comprises the following steps: respectively taking the average value of the first time-frequency singular entropy and the average value of the second time-frequency singular entropy in the current window as the first identification index and the second identification index; The preset early warning condition comprises: the first identification index is greater than a first safety threshold corresponding to the first time-frequency singular entropy in the current window, and the second identification index is greater than a second safety threshold corresponding to the second time-frequency singular entropy in the current window; The method further comprises: determining a baseline window according to a preset number of windows in front of the current window in time sequence; determining the first safety threshold according to the average value and the standard deviation of the first time-frequency singular entropy in the baseline window; determining the second safety threshold according to the average value and the standard deviation of the second time-frequency singular entropy in the baseline window.

5. A device for identifying precursors to rock damage, characterized in that, The device comprises: an acquisition module, configured to acquire multi-modal acoustic signals of a rock under superposition of different types of loads, wherein the multi-modal acoustic signals comprise acoustic emission signals and active sound wave signals; The computing module is configured to divide the multi-modal acoustic signal according to a preset time window to obtain a multi-modal acoustic signal segment; perform time-frequency analysis on the multi-modal acoustic signal segment to construct a time-frequency matrix of the multi-modal acoustic signal segment; perform singular value decomposition on the time-frequency matrix to obtain a singular value sequence of the time-frequency matrix; calculate an information entropy of the singular value sequence to obtain a time-frequency singular entropy of the multi-modal acoustic signal segment corresponding to the singular value sequence; and construct a time-frequency singular entropy sequence of the multi-modal acoustic signal according to the time-frequency singular entropies of the multi-modal acoustic signal segments. The time-frequency analysis performed on the multi-modal acoustic signal segment to construct the time-frequency matrix of the multi-modal acoustic signal segment specifically includes: discretizing the multi-modal acoustic signal segment according to a preset sampling time point to obtain a discrete signal; scaling a preset mother wavelet based on a preset scale factor, and matching the scaled preset mother wavelet with the discrete signal through convolution operation to obtain a continuous wavelet coefficient matrix; and taking a modulus of the continuous wavelet coefficient matrix to obtain a time-frequency energy distribution matrix of the multi-modal acoustic signal segment, and taking the time-frequency energy distribution matrix as the time-frequency matrix of the multi-modal acoustic signal segment. The calculation of the information entropy of the singular value sequence to obtain the time-frequency singular entropy of the multi-modal acoustic signal segment corresponding to the singular value sequence specifically includes: taking the singular value sequence as a probability distribution, determining an energy value corresponding to the singular value according to a square of the singular value in the singular value sequence; determining a total energy value corresponding to the singular value sequence based on the energy value corresponding to the singular value; determining an energy proportion of the singular value according to the energy value corresponding to the singular value and the total energy value corresponding to the singular value sequence; determining an energy distribution corresponding to the singular value sequence according to the energy proportion of the singular value; and calculating an information entropy of the energy distribution to obtain the time-frequency singular entropy of the multi-modal acoustic signal segment corresponding to the singular value sequence. The identification module is configured to identify the rock failure precursor based on the time-frequency singular entropy sequence of the multi-modal acoustic signal. The multi-modal acoustic signal segment includes an acoustic emission signal segment and an active acoustic wave signal segment, a time-frequency singular entropy of the multi-modal acoustic signal segment includes a first time-frequency singular entropy of the acoustic emission signal segment and a second time-frequency singular entropy of the active acoustic wave signal segment, a time-frequency singular entropy sequence of the multi-modal acoustic signal includes a first time-frequency singular entropy sequence of the acoustic emission signal and a second time-frequency singular entropy sequence of the active acoustic wave signal, and the rock damage precursor is identified based on the time-frequency singular entropy sequence of the multi-modal acoustic signal, and specifically includes: based on an original window size, a current window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence is synchronously determined in a time sequence; a first identification index and a second identification index of the current window are respectively determined according to the first time-frequency singular entropy and the second time-frequency singular entropy in the current window; if the first identification index and the second identification index of the current window meet a preset early warning condition, a rock damage precursor early warning is triggered, and a sliding window is stopped; and if the first identification index and the second identification index of the current window do not meet the preset early warning condition, a next window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence is synchronously determined according to an end position of the current window in the time sequence and the original window size, so as to synchronously slide the window on the first time-frequency singular entropy sequence and the second time-frequency singular entropy sequence.

6. A readable storage medium, on which a program or instructions are stored, characterized in that, The program or the instruction is executed by the processor to realize the steps of the rock damage precursor identification method in any one of claims 1 to 4.

7. A computer device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, The processor executes the program to realize the rock damage precursor identification method in any one of claims 1 to 4.

Images