Rock mass instability acoustic emission precursor signal extraction method and instability precursor discrimination method
By real-time detection and SVD decomposition of rock mass acoustic emission signals, and extraction and identification of precursory characteristic signals, the problem of difficulty in identifying precursory signals of rock mass instability has been solved, realizing real-time and accurate early warning of rock mass stability monitoring, which is applicable to multiple engineering fields.
Patent Information
- Application Number
- CN202211355662.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-01
- Publication Date
- 2026-02-03
- Estimated Expiration
- 2042-11-01
AI Technical Summary
In existing technologies, it is difficult to accurately identify precursor signals of rock mass instability, and there are many noise signals, which makes it difficult to monitor and warn of rock mass stability. Moreover, existing precursor indicators fluctuate greatly and are difficult to apply in actual engineering.
By real-time detection of acoustic emission signals from rock masses, a trajectory matrix is constructed and SVD decomposition is performed. The contribution rate and kurtosis of eigenvalues are calculated, precursor eigenvalues are extracted, precursor eigensignatures are reconstructed, and precursors are identified using average frequency and acoustic emission rise time parameters.
It enables real-time and accurate extraction and identification of precursor signals of rock mass instability, improves the accuracy of early warning, simplifies the operation process, reduces computational complexity, is suitable for conventional processors, and is applicable to rock mass early warning in tunnels, slopes, mining, and water conservancy and hydropower projects.
Smart Images
Figure CN115950959B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of geotechnical engineering monitoring and early warning, and particularly relates to a rock mass instability acoustic emission precursor signal extraction method and an instability precursor discrimination method. BACKGROUND
[0002] With the continuous development of engineering construction towards Tibet and other high-stress and tectonically active areas, a series of engineering problems related to rock mass stability will be encountered, such as large deformation, rock burst, collapse, landslide, etc. Therefore, real-time and accurate monitoring of rock mass stability, advanced prediction and early warning of its instability and failure are of great significance to protect people's lives and property.
[0003] Rock mass instability and failure often have the characteristics of suddenness and small deformation due to its brittleness, which leads to poor reliability of deformation as a precursor indicator and unobvious precursors. Acoustic emission is a transient elastic wave generated by the rapid release of strain energy in the interior or surface of a material, which reflects the real-time changes in the state of the rock mass. At present, a large number of studies use acoustic emission frequency, b value, fractal dimension, RA / AF, etc. as precursor indicators. However, the precursor indicators used by these existing technologies have large fluctuations, and the precursor signals often contain a lot of noise signals, making it difficult to apply in actual engineering. SUMMARY
[0004] The purpose of the present application is to provide a rock mass instability acoustic emission precursor signal extraction method and an instability precursor discrimination method, which can track and warn in real time the state of the rock mass, and the extracted precursor signals are easy to identify and the judgment method is simple.
[0005] The technical solution of the present application to solve the above technical problems is as follows:
[0006] The present application provides a rock mass instability acoustic emission precursor signal extraction method, which comprises:
[0007] S1: Real-time detection and recording of relevant parameters of the current monitoring rock mass acoustic emission signal;
[0008] S2: Embedding the time series of the relevant parameters according to a preset window length to obtain a trajectory matrix;
[0009] S3: SVD decomposition of the trajectory matrix to obtain a decomposed trajectory matrix;
[0010] S4: Calculation of the contribution rate of the eigenvalues in the decomposed trajectory matrix, and arrangement in descending order to obtain the arranged eigenvalue contribution rate;
[0011] S5: Extraction of precursor eigenvalues according to the kurtosis of the arranged eigenvalue contribution rate and the mean value of the kurtosis to obtain a precursor eigenvalue set.
[0012] S6: Based on the set of precursor feature values, reconstruct the matrix corresponding to the precursor feature values to obtain the precursor feature signal.
[0013] Alternatively, in step S1, the relevant parameters include acoustic emission rise time / amplitude parameters and average frequency parameters.
[0014] Alternatively, in step S2, the preset window length L satisfies: 2≤L≤N-1, where N represents the length of the time series data.
[0015] Alternatively, in step S2, the trajectory matrix X is:
[0016]
[0017] Where K is the number of columns in the trajectory matrix, and K = N - L + 1, N represents the length of the time series data, and L is the window length parameter.
[0018] Alternatively, step S3 may include:
[0019] S31: Calculate XX T The initial eigenvalues are obtained and arranged in descending order to obtain the eigenvalues;
[0020] S32: Obtain the left and right eigenvectors of the trajectory matrix X;
[0021] S33: Based on the left and right eigenvectors of the trajectory matrix X, and the eigenvalues, the decomposed trajectory matrix is obtained.
[0022] Alternatively, the decomposed trajectory matrix is:
[0023]
[0024] Where d is the number of non-zero eigenvalues, and d = max{i,λ} i >0}=rank X, For singular values, λ i (i = 1, 2, ..., L) is a matrix S = XX T Eigenvalues sorted in descending order, U i and V i These are the left and right eigenvectors of the trajectory matrix X, respectively.
[0025] Alternatively, in step S4, the contribution rate η of the feature value i for:
[0026]
[0027] Where, λ i (i = 1, 2, ..., L) is a matrix S = XX T The eigenvalues are sorted in descending order, where d is the number of non-zero eigenvalues, and d = max{i,λ} i >0} = rank X, where X is the trajectory matrix.
[0028] Alternatively, in step S4, the kurtosis Ku of the eigenvalue contribution rate is:
[0029]
[0030] Where, λ i (i = 1, 2, ..., L) is a matrix S = XX T The eigenvalues are sorted in descending order, where d is the number of non-zero eigenvalues, and d = max{i,λ} i >0} = rank X, where X is the trajectory matrix. This represents the mean of the contribution rates of the eigenvalues.
[0031] Alternatively, step S6 is to reconstruct each precursor feature value into a new time series of length N.
[0032] This invention also provides a method for identifying precursors of rock mass instability, the method comprising the above-mentioned method for extracting acoustic emission precursor signals of rock mass instability, and further comprising:
[0033] The precursor characteristic signal is judged using one of the following instability precursor conditions to obtain the judgment result:
[0034] The precursor characteristic signal is valid when the average frequency parameter drops suddenly and / or the acoustic emission rise time / amplitude parameter increases suddenly; otherwise, the precursor characteristic signal is invalid.
[0035] The present invention has the following beneficial effects:
[0036] Acoustic emission waveforms can be directly measured by acoustic emission monitoring equipment, and acoustic emission rise time parameters can be directly obtained from acoustic emission waveform diagrams. The acquisition is very convenient, quick, and easy. The calculation process and subsequent judgment of acoustic emission rise time coefficient of variation, skewness, and kurtosis are very simple. They can be achieved by embedding calculation programs in conventional processors without building complex algorithm models. The precursor points of rock mass instability and failure are easy to identify, and the calculation error is small and the stability is high. Most engineering construction units have the conditions to apply this technology.
[0037] Furthermore, the precursor indicators of this invention are diverse and closely related to the type of rock and soil fracture, which can improve the accuracy of early warning.
[0038] In summary, this invention can characterize the development trend of rock mass failure types (crack types) in real time, and can be widely used in rock mass early warning and forecasting in fields such as tunnel engineering, slope engineering, mining engineering, and water conservancy and hydropower engineering. Attached Figure Description
[0039] Figure 1 This is a flowchart of the method for extracting precursor signals of acoustic emission due to rock mass instability according to the present invention;
[0040] Figure 2 The image shows the acoustic emission (AF) curves collected during the instability and failure of a rock mass with pre-fabricated cracks under uniaxial loading.
[0041] Figure 3 The image shows acoustic emission (RA) curves collected during the instability and failure of a rock mass with pre-fabricated cracks under uniaxial loading.
[0042] Figure 4 This is a schematic diagram of the extraction of precursor features of acoustic emission;
[0043] Figure 5 The extracted acoustic emission (AF) and RA precursor signals and precursor point discrimination diagram are shown. Detailed Implementation
[0044] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.
[0045] This invention provides a method for extracting precursor signals of acoustic emission due to rock mass instability, with reference to... Figure 1 As shown, the method for extracting precursor signals of acoustic emission due to rock mass instability includes:
[0046] S1: Real-time detection and recording of relevant parameters of acoustic emission signals from the monitored rock mass;
[0047] This invention does not impose specific limitations on the real-time detection equipment used to monitor relevant parameters of acoustic emission signals from rock masses. As one implementation method, this invention employs an acoustic emission monitoring device to acquire relevant parameters, including the acoustic emission rise time / amplitude parameter RF and the average frequency parameter AF. After installing and fixing the acoustic emission probe, the acoustic emission signal of the rock mass can be monitored, and the RA and AF parameters can be acquired; the operation is relatively convenient.
[0048] S2: Embed the time series of the relevant parameters according to a preset window length to obtain a trajectory matrix;
[0049] In this invention, a preset window length L is set to satisfy: 2≤L≤N-1, where N represents the length of the time series data.
[0050] In addition, embedding is the process of embedding a one-dimensional time series X = (x1, x2, x3, ..., x...).N ), mapped to a multidimensional time series X i =(x I ,…,x i+L-1 ) T ∈R L The method is as follows. Where K = N - L + 1, and L is the window length parameter. The trajectory matrix X can be represented as follows:
[0051]
[0052] S3: Perform SVD decomposition on the trajectory matrix to obtain the decomposed trajectory matrix;
[0053] Alternatively, step S3 may include:
[0054] S31: Calculate XX T The initial eigenvalues are obtained by arranging them in descending order, and then representing the eigenvalues as: (λ1≥λ2…λ). L ≥0);
[0055] S32: Obtain the left and right eigenvectors of the trajectory matrix X;
[0056] S33: Based on the left and right eigenvectors of the trajectory matrix X, and the eigenvalues, the decomposed trajectory matrix is obtained.
[0057] The decomposed trajectory matrix is:
[0058]
[0059] Where d is the number of non-zero eigenvalues, and d = max{i,λ} i >0}=rank X, For singular values, λ i (i = 1, 2, ..., L) is a matrix S = XX T Eigenvalues sorted in descending order, U i and V i These are the left and right eigenvectors of the trajectory matrix X, respectively.
[0060] S4: Calculate the contribution rate of the eigenvalues in the decomposed trajectory matrix and arrange them in descending order to obtain the arranged eigenvalue contribution rates;
[0061] The contribution rate η of the feature value i for:
[0062]
[0063] Where, λ i (i = 1, 2, ..., L) is a matrix S = XXT The eigenvalues are sorted in descending order, where d is the number of non-zero eigenvalues, and d = max{i,λ} i >0} = rank X, where X is the trajectory matrix.
[0064] S5: Extract precursor eigenvalues based on the kurtosis and mean kurtosis of the eigenvalue contribution rates after sorting, and obtain a set of precursor eigenvalues;
[0065] The kurtosis Ku of the eigenvalue contribution rate is:
[0066]
[0067] Where, λ i (i = 1, 2, ..., L) is a matrix S = XX T The eigenvalues are sorted in descending order, where d is the number of non-zero eigenvalues, and d = max{i,λ} i >0} = rank X, where X is the trajectory matrix. This represents the mean of the contribution rates of the eigenvalues.
[0068] The mean of the contribution rate and kurtosis of eigenvalues The precursor feature set I obtained by extracting precursor features is represented as:
[0069]
[0070] The matrix corresponding to the set of precursor eigenvalues of I can also be represented as:
[0071]
[0072] For the matrix X of the i-th group I The corresponding eigenvalue contribution rate can be regarded as an important component in the signal.
[0073] S6: Based on the set of precursor feature values, reconstruct the matrix corresponding to the precursor feature values to obtain the precursor feature signal.
[0074] That is, to assign each precursor feature value Reconstruct it into a new time series of length N.
[0075] Let Y be an L×K matrix, where L is the number of rows and K is the number of columns. The element in the i-th row and j-th column is represented as y. ij Let L = 1 ≤ i ≤ L, 1 ≤ j ≤ K. * =min(L,K), K * =max(L,K), N = L + K - 1. By diagonally bisecting the matrix Y, it can be transformed into a time series y1, y2, ... y3 of length N. N The expression is:
[0076]
[0077] The above equation is equivalent to inverse diagonalizing the elements of the diagonal matrix Y, satisfying the condition i+j=k+1. Here, k is the order of the diagonal matrix, and when k=1, y1=y 1,1 When k = 2, y1 = (y 1,2 +y 2,1 ) / 2, and so on. Applying the above formula to the matrix Time series can be obtained Therefore, the original time series X = (x1, x2, x3, ..., x...) can be represented as X = (x1, x2, x3, ..., x...) N The decomposition is the sum of m reconstructed sequences, expressed as:
[0078]
[0079] In the formula: (n=1,2,…N), N is the length of the time series data.
[0080] Based on the above technical solution, the present invention also provides a method for identifying precursors of rock mass instability, the method comprising the above-mentioned method for extracting acoustic emission precursor signals of rock mass instability, and further comprising:
[0081] The precursor characteristic signal is judged using one of the following instability precursor conditions to obtain the judgment result:
[0082] The precursor characteristic signal is valid when the average frequency parameter drops suddenly and / or the acoustic emission rise time / amplitude parameter increases suddenly; otherwise, the precursor characteristic signal is invalid.
[0083] This method of judgment can be performed manually or automatically by equipment.
[0084] Precursor signals and their effects test
[0085] Joints and fissures are widely distributed in rock masses in nature. Therefore, a pre-fractured rock mass was used as the current monitoring rock mass. A uniaxial loading test was conducted on this rock mass, and an acoustic emission (AE) instrument was used for simultaneous detection to record and acquire the acoustic emission (RA) and acoustic emission (AF) parameters during the rock mass's instability and failure process. The acquired parameters are as follows: Figure 2 , Figure 3 As shown.
[0086] The window length was set to 400, and time series embedding was performed. Precursor features were extracted using the skewness of the feature contribution rate and the mean of the skewness. The results are as follows: Figure 4 As shown; the feature vectors corresponding to the extracted feature values are reconstructed, and the result is as follows. Figure 5 As shown.
[0087] If RA suddenly increases and AF suddenly decreases, it is considered a precursor signal. In this experiment, the rock mass containing pre-fabricated cracks became unstable and failed at 336.80s. After using the above-mentioned precursor signal extraction method, precursor 1 appeared at 276.39s and precursor 2 appeared at 293.63s. The precursors occurred before the rock mass became unstable and failed, so the precursor information is valid.
[0088] In summary, the present invention has the following advantages:
[0089] 1. Diverse precursor signals can improve the accuracy of early warnings;
[0090] 2. The precursor points of rock mass instability and failure are easy to identify. This algorithm only requires the input of the embedding window length, making it easy to operate and low in implementation cost.
[0091] 3. It can characterize the development trend of shear cracks and tension cracks in rock mass in real time, and can be widely used in rock mass early warning and forecasting in fields such as tunnel engineering, slope engineering, mining engineering, and water conservancy and hydropower engineering.
[0092] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for extracting precursor signals of acoustic emission due to rock mass instability, characterized in that, The method for extracting precursor signals of acoustic emission due to rock mass instability includes: S1: Real-time detection and recording of relevant parameters of acoustic emission signals from the monitored rock mass; The relevant parameters include acoustic emission rise time / amplitude parameters and average frequency parameters; S2: Embed the time series of the relevant parameters according to a preset window length to obtain a trajectory matrix; S3: Perform SVD decomposition on the trajectory matrix to obtain the decomposed trajectory matrix; S4: Calculate the contribution rate of the eigenvalues in the decomposed trajectory matrix and arrange them in descending order to obtain the arranged eigenvalue contribution rates; The contribution rate of the eigenvalue for: in, For matrix Eigenvalues sorted in descending order d The number of non-zero eigenvalues, and , For the trajectory matrix, For window length parameter, For the first One non-zero eigenvalue, Represents the transpose of a matrix; kurtosis of the eigenvalue contribution rate for: in, The mean of the contribution rates of the eigenvalues; S5: Extract precursor eigenvalues based on the kurtosis and mean kurtosis of the eigenvalue contribution rates after sorting, and obtain a set of precursor eigenvalues; S6: Based on the set of precursor feature values, reconstruct the matrix corresponding to the precursor feature values to obtain the precursor feature signal.
2. The method for extracting precursor signals of acoustic emission due to rock mass instability according to claim 1, characterized in that, In step S2, the preset window length parameter L satisfy: ,in, N This represents the length of the time series data.
3. The method for extracting precursor signals of acoustic emission due to rock mass instability according to claim 1, characterized in that, In step S2, the trajectory matrix for: in, K Let be the number of columns in the trajectory matrix, and , N This is expressed as the length of the time series data. This is the window length parameter.
4. The method for extracting precursor signals of acoustic emission due to rock mass instability according to claim 3, characterized in that, Step S3 includes: S31: Calculation The initial eigenvalues are obtained and arranged in descending order to obtain the eigenvalues; S32: Obtain the trajectory matrix The left and right eigenvectors; S33: According to the trajectory matrix The left and right eigenvectors, along with the eigenvalues, are used to obtain the decomposed trajectory matrix.
5. The method for extracting precursor signals of acoustic emission due to rock mass instability according to claim 1 or 4, characterized in that, The decomposed trajectory matrix is: in, d The number of non-zero eigenvalues, and , It is a singular value. For matrix Eigenvalues sorted in descending order and Trajectory matrices The left and right eigenvectors.
6. The method for extracting precursor signals of acoustic emission due to rock mass instability according to claim 4, characterized in that, Step S6 involves reconstructing each precursor feature value into a form of length [length missing]. N The new time series.
7. A method for identifying precursors of rock mass instability, characterized in that, The method for identifying precursors of rock mass instability includes the method for extracting acoustic emission precursor signals of rock mass instability according to any one of claims 1-6, and further includes: The precursor characteristic signal is judged using one of the following instability precursor conditions to obtain the judgment result: The precursor characteristic signal is valid when the average frequency parameter drops suddenly and / or the acoustic emission rise time / amplitude parameter increases suddenly; otherwise, the precursor characteristic signal is invalid.