A concrete damage detection method, system, device, medium and product

Abstract

The application discloses a concrete damage detection method, system, device, medium and product, and relates to the field of concrete structure health monitoring. The method comprises the following steps: acquiring one-dimensional piezoelectric sensing signals of a concrete test piece at different damage stages and real labels of the damage stages; preprocessing the one-dimensional piezoelectric sensing signals; determining a delay time and an embedding dimension of a phase space based on a phase space reconstruction theory, mapping the preprocessed one-dimensional piezoelectric sensing signals to a multi-dimensional phase space, forming an attractor reflecting chaotic characteristics, constructing a multi-dimensional phase space matrix, and extracting chaotic characteristics of the attractor in the multi-dimensional phase space; the chaotic characteristics comprise a center distance, a vector radius offset and a correlation dimension; and inputting the chaotic characteristics and the real labels of the damage stages into a feedforward neural network model to output a concrete damage category. The application can improve damage recognition rate and effectively distinguish different damage stages.

Description

Technical Field

[0001] This application relates to the field of concrete structure health monitoring, and in particular to a method, system, equipment, medium, and product for detecting concrete damage. Background Technology

[0002] During long-term service, concrete structures are prone to developing microcracks due to impact loads, environmental erosion, and other factors. These microcracks gradually expand into macrocracks, leading to a decrease in structural stiffness and load-bearing capacity, ultimately threatening overall safety and durability. Therefore, developing efficient and accurate concrete structure damage monitoring technologies is of great significance for ensuring engineering safety and extending structural lifespan.

[0003] Currently, non-destructive testing methods for concrete structures mainly include acoustic emission, electromagnetic induction, ultrasonic waves, and infrared thermography. However, these methods still have certain limitations in practical applications. Most rely on a single characteristic index for evaluation, and the physical correlation between the characteristic and the damage mechanism is weak, affecting the reliability of the diagnosis.

[0004] In recent years, active sensing technology based on piezoelectric ceramics has received widespread attention in the field of structural health monitoring, especially wave propagation methods and electromechanical impedance methods. These methods use piezoelectric sensors to excite and receive stress waves, and utilize the modulation effect of damage on the wave propagation characteristics to achieve damage identification. However, the damage evolution of concrete under impact load is a strongly nonlinear and non-stationary dynamic process. Traditional time-domain, frequency-domain, and time-frequency analysis methods (such as wavelet packet analysis and HHT transform) are often based on the assumption of signal stationarity or system linearity, making it difficult to fully extract and characterize the complex dynamic characteristics contained in the piezoelectric response signal during the damage evolution process, resulting in limited accuracy in identifying damage stages. Summary of the Invention

[0005] The purpose of this application is to provide a concrete damage detection method, system, equipment, medium, and product to solve the problem of low accuracy in identifying damage stages.

[0006] To achieve the above objectives, this application provides the following solution.

[0007] In a first aspect, this application provides a method for detecting concrete damage, comprising the following steps.

[0008] One-dimensional piezoelectric sensing signals and true labels of different damage stages of concrete specimens were obtained; the different damage stages included the undamaged stage, the budding stage, the stable propagation stage, and the unstable propagation stage.

[0009] The one-dimensional piezoelectric sensing signal is preprocessed.

[0010] Calculate the maximum Lyapunov exponent of the preprocessed one-dimensional piezoelectric sensing signal and determine whether the preprocessed one-dimensional piezoelectric sensing signal has chaotic characteristics.

[0011] Based on the phase space reconstruction theory, the pre-processed one-dimensional piezoelectric sensing signal is mapped to a multi-dimensional phase space, the delay time and the embedding dimension are determined, an attractor reflecting chaotic characteristics is formed, and a multi-dimensional phase space matrix is ​​constructed; the multi-dimensional phase space matrix includes a two-dimensional phase space matrix and a three-dimensional phase space matrix.

[0012] Based on the multidimensional phase space matrix, chaotic features of attractors in the multidimensional phase space are extracted; the chaotic features include center distance, vector offset, and correlation dimension.

[0013] The chaotic features and the true labels of the damage stages are input into the feedforward neural network model to output the concrete damage category; the concrete damage category includes no damage, slight damage, moderate damage and complete damage.

[0014] Secondly, this application provides a concrete damage detection system, including the following modules.

[0015] The signal acquisition module is used to acquire one-dimensional piezoelectric sensing signals and real labels of different damage stages of concrete specimens; the different damage stages include the undamaged stage, the budding stage, the stable propagation stage, and the unstable propagation stage.

[0016] A one-dimensional piezoelectric sensing signal determination module is used to preprocess the one-dimensional piezoelectric sensing signal.

[0017] The chaos analysis module is used to calculate the maximum Lyapunov exponent of the preprocessed one-dimensional piezoelectric sensing signal and determine whether the preprocessed one-dimensional piezoelectric sensing signal has chaotic characteristics.

[0018] The multidimensional phase space matrix construction module is used to map the pre-processed one-dimensional piezoelectric sensing signal to a multidimensional phase space based on the phase space reconstruction theory, determine the delay time and embedding dimension, form an attractor reflecting chaotic characteristics, and construct a multidimensional phase space matrix; the multidimensional phase space matrix includes a two-dimensional phase space matrix and a three-dimensional phase space matrix.

[0019] The chaotic feature extraction module for attractors is used to extract chaotic features of attractors in the multidimensional phase space based on the multidimensional phase space matrix; the chaotic features include center distance, vector offset and correlation dimension.

[0020] The concrete damage detection module is used to input the chaotic features and the true labels of the damage stages into the feedforward neural network model and output the concrete damage category; the concrete damage category includes no damage, slight damage, moderate damage and complete damage.

[0021] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described concrete damage detection method.

[0022] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described concrete damage detection method.

[0023] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described concrete damage detection method.

[0024] According to the specific embodiments provided in this application, this application has the following technical effects.

[0025] This application acquires one-dimensional piezoelectric sensing signals and true labels of damage stages from concrete specimens at different damage stages. The one-dimensional piezoelectric sensing signals are preprocessed, and based on phase space reconstruction theory, they are mapped to a multi-dimensional phase space to form attractors reflecting chaotic characteristics. A multi-dimensional phase space matrix is ​​constructed to extract three types of features: center distance, radius offset, and correlation dimension. These three features comprehensively characterize the chaotic features of the attractors in the multi-dimensional phase space from "macroscopic complexity to microscopic dynamic behavior," enabling the differentiation and prediction of the damage degree in concrete at four stages: "undamaged - budding - stable expansion - unstable expansion." This allows for accurate differentiation of similar stages. Furthermore, the feature differentiation between the undamaged stage and the unstable expansion stage reaches over 85%. This application can accurately capture changes in chaotic characteristics through center distance, radius offset, and correlation dimension, improving the damage recognition rate. In addition, by inputting the chaotic features in the multi-dimensional phase space into a feedforward neural network to output the concrete damage category, the accuracy of damage stage identification is improved. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram of a concrete damage detection method provided in an embodiment of this application.

[0028] Figure 2 An attractor distribution diagram in the undamaged phase space provided in an embodiment of this application.

[0029] Figure 3 An attractor distribution diagram in two-dimensional phase space during the germination stage provided in an embodiment of this application.

[0030] Figure 4 An attractor distribution diagram in two-dimensional phase space during the stable expansion phase provided in an embodiment of this application.

[0031] Figure 5 A diagram showing the attractor distribution in a two-dimensional phase space during the instability propagation stage, provided in an embodiment of this application.

[0032] Figure 6 An attractor distribution diagram in the three-dimensional phase space during the undamaged phase, provided as an embodiment of this application.

[0033] Figure 7 An attractor distribution diagram in three-dimensional phase space during the germination stage provided in an embodiment of this application.

[0034] Figure 8 An attractor distribution diagram in three-dimensional phase space during the stable expansion phase provided in an embodiment of this application.

[0035] Figure 9 An attractor distribution diagram in three-dimensional phase space during the instability propagation stage provided in an embodiment of this application.

[0036] Figure 10 This is a schematic diagram of the center distance in two-dimensional phase space provided in an embodiment of this application.

[0037] Figure 11 This is a schematic diagram of the three-dimensional phase space center distance provided in an embodiment of this application.

[0038] Figure 12 This is a schematic diagram of two-dimensional phase space offset provided in an embodiment of this application.

[0039] Figure 13 This is a schematic diagram of three-dimensional phase space offset provided in an embodiment of this application.

[0040] Figure 14 This is a schematic diagram of the correlation dimension of a two-dimensional phase space provided in an embodiment of this application.

[0041] Figure 15 This is a schematic diagram of the correlation dimension of a three-dimensional phase space provided in an embodiment of this application.

[0042] Figure 16 This is a schematic diagram of the sample loss curve provided in an embodiment of this application.

[0043] Figure 17 This is a schematic diagram of the accuracy curve provided in one embodiment of this application.

[0044] Figure 18 This is a schematic diagram of one-dimensional piezoelectric sensing signal processing provided in an embodiment of this application.

[0045] Figure 19 This is a schematic diagram of the L-index of a one-dimensional piezoelectric sensing signal in the undamaged stage provided in an embodiment of this application.

[0046] Figure 20 This is a schematic diagram of the L-index of a one-dimensional piezoelectric sensing signal in the budding stage provided in an embodiment of this application.

[0047] Figure 21 This is a schematic diagram of the L-index of a one-dimensional piezoelectric sensing signal in the stable expansion phase provided in an embodiment of this application.

[0048] Figure 22 This is a schematic diagram of the L-index of a one-dimensional piezoelectric sensing signal during the instability propagation stage provided in an embodiment of this application. Detailed Implementation

[0049] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] To make the objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0051] like Figure 1 As shown in the figure, this application provides a concrete damage detection method, which includes the following steps.

[0052] S1: Obtain one-dimensional piezoelectric sensing signals and true labels of different damage stages of concrete specimens; among which, different damage stages include the undamaged stage, the budding stage, the stable propagation stage, and the unstable propagation stage.

[0053] S2: Preprocess the one-dimensional piezoelectric sensing signal.

[0054] S3: Calculate the maximum Lyapunov exponent of the preprocessed one-dimensional piezoelectric sensing signal and determine whether the preprocessed one-dimensional piezoelectric sensing signal has chaotic characteristics.

[0055] S4: Based on the phase space reconstruction theory, the pre-processed one-dimensional piezoelectric sensing signal is mapped to a multi-dimensional phase space, the delay time and the embedding dimension are determined, an attractor reflecting chaotic characteristics is formed, and a multi-dimensional phase space matrix is ​​constructed; the multi-dimensional phase space matrix includes a two-dimensional phase space matrix and a three-dimensional phase space matrix.

[0056] S5: Based on the multidimensional phase space matrix, extract the chaotic features of attractors in the multidimensional phase space; the chaotic features include center distance, vector offset and correlation dimension.

[0057] S6: Input the chaotic features and the true labels of the damage stages into the feedforward neural network model and output the concrete damage category; the concrete damage category includes no damage, slight damage, moderate damage and complete failure.

[0058] This application maps one-dimensional piezoelectric sensing signals to a multi-dimensional phase space based on phase space reconstruction theory, thereby eliminating the assumptions of signal linearity and stationarity and directly revealing the nonlinear nature of damage evolution. Building upon this, multi-scale chaotic features such as correlation dimension, center distance, and vector offset are extracted from the reconstructed attractor trajectories, achieving feature quantification based on the mechanism. Finally, these features are input into a feedforward neural network for automatic classification, achieving high-precision and objective identification of the four damage stages: "no damage, crack initiation, stable propagation, and unstable propagation." This method as a whole achieves a seamless integration from signal perception, dynamic analysis, feature construction to intelligent decision-making, providing a practical new approach for real-time monitoring and early warning of concrete structure damage.

[0059] In an exemplary embodiment, S1 is performed using a piezoelectric sensor and a data acquisition device to obtain one-dimensional piezoelectric sensing signals and true labels of different damage stages, providing a data foundation for the subsequent construction of a concrete structure detection model.

[0060] The actual operation involved placing piezoelectric ceramic (specifically a lead zirconate titanium, PZT) sensors near the impact zone of the concrete structure. A test system was constructed using a National Instruments (NI) data acquisition card and accompanying LabVIEW software. The time series of one-dimensional piezoelectric sensing signals was acquired based on the Wave Propagation Method (WPM). i|i=1,2,…,N}: One PZT sensor acts as an actuator, applying a sweeping alternating voltage with an initial frequency of 1000Hz, a final frequency of 400kHz, an amplitude of 10V, and a duration of 1s via LabVIEW software. This utilizes the inverse piezoelectric effect of the piezoelectric material to convert electrical energy into mechanical energy, generating a solid stress wave that propagates along the concrete medium. Another PZT sensor acts as a receiver, positioned along the stress wave propagation path. It converts the mechanical energy of the received stress wave into electrical energy through the direct piezoelectric effect. The NI data acquisition card then synchronously records the one-dimensional piezoelectric sensing signal at a sampling frequency of 1MHz, completing the signal acquisition.

[0061] In an exemplary embodiment, S2 is executed by a computer processing system to map the one-dimensional piezoelectric sensing signal to a uniform scale, thereby accelerating neural network convergence, focusing on the relative relationships and distribution patterns of features, and improving the model training stability and generalization ability.

[0062] The actual operation is as follows: 1) Min-Max normalization: x'=(x-min) / (max-min), which linearly maps the data to the interval [0,1]; where x' is the normalized one-dimensional piezoelectric sensing signal, x is the original data of the one-dimensional piezoelectric sensing signal, min is the minimum value of the one-dimensional piezoelectric sensing signal, and max is the maximum value of the one-dimensional piezoelectric sensing signal.

[0063] 2) Z-Score standardization: z=(x-μ) / σ; where z is the standardized one-dimensional piezoelectric sensing signal, μ is the mean, and σ is the standard deviation. Both μ and σ are calculated from the training set.

[0064] 3) Test set normalization: use min, max, μ and σ calculated from the training set to ensure data distribution consistency.

[0065] In practical applications, the L-exponent of the four-stage one-dimensional piezoelectric sensing signal is as follows: Figures 19-22 As shown, all values ​​are basically greater than 0, indicating that the signal has obvious chaotic characteristics.

[0066] In an exemplary embodiment, S4 is executed by a computer processing system to map the pre-processed one-dimensional piezoelectric sensing signal to a multi-dimensional phase space to form an attractor that reflects the chaotic characteristics of the system. S4 specifically includes the following steps.

[0067] S41: Map the pre-processed one-dimensional piezoelectric sensing signal to a multi-dimensional phase space, and use the average mutual information function method to determine the delay time.

[0068] S42: The embedding dimension is determined using the pseudo-nearest neighbor method.

[0069] S43: Construct a multidimensional phase space matrix based on the delay time and the embedding dimension.

[0070] The actual operation is as follows:

[0071] (1) Determine the delay time τ: The average mutual information function (AMI) method is adopted, and the first local minimum of mutual information is taken as τ.

[0072]

[0073] Where I(T) is the average mutual information value, and I is the average autocorrelation function of T; p(x(n)) and p(x(n+T)) are the univariate probability densities of x(n) and x(n+T) respectively; p(x(n), x(n+T)) is the joint probability density; n is any possible position in the time series, x(n) is the value of the time series at time n, x(n+T) is the value of the time series at time n+T, T is a candidate delay time, and τ∈T. The optimal delay time τ is calculated to be 7.

[0074] (2) Determine the embedding dimension m: The pseudo-nearest neighbor method is adopted, and the phase vector is defined according to Takens' theorem (m≥2d+1, where d is the dimension of the original dynamic system).

[0075] In the d-dimensional phase space, the phase space reconstruction vector As shown below.

[0076]

[0077] And there exists a nearest neighbor, whose vector expression is... As shown below.

[0078]

[0079] in, For the time series at time 10:00 The value, , For x(n), x(n+T), The nearest neighbor.

[0080] Using the Euclidean norm, calculate the squared distance between a point in the phase space trajectory and its nearest neighbor. As shown below.

[0081]

[0082] in, For the time series at time 10:00 The value, for The nearest neighbor, k is the summation index variable.

[0083] To find the optimal embedding dimension, we define As shown below.

[0084]

[0085]

[0086]

[0087] Among them, y i (m+1)=(x i ,x i+T ,…,x i+mT ), where n(i,m) is a positive integer satisfying the condition 1≤n(i,m)≤N-mT, and E(m) is independent of the embedding dimension m and the delay time τ. This is a stability index corresponding to the embedding dimension m, used to determine the trend of E as the embedding dimension increases. The average distance change rate when the embedding dimension is m+1. Let y be the average distance change rate when the embedding dimension is m, N be the total length of the original time series, a(i,m) be the distance ratio at the i-th phase point, and y be the average distance change rate when the embedding dimension is m. i (m) represents the i-th state vector in m-dimensional space. Let y be in m+1 dimensional space i The nearest neighbor of (m), where M is the number of state points in phase space. For an embedding dimension of m+1, the squared distance between the phase space vector and its nearest neighbor vector.

[0088] Define the proportion of pseudo-nearest neighbors: ,when When stable, m+1 is the optimal value.

[0089] (3) Time series The phase point Y embedded in m-dimensional space is shown below.

[0090] (4)

[0091] Where Y is the reconstructed m-dimensional phase space matrix; Let x be the i-th phase point, i = 1, 2, ..., N; i x i+τ x i+2τ ... The coordinates of the state point are given, and the state point is a one-dimensional piezoelectric sensing signal; m ≥ 2d + 1; d is the dimension of the prime mover system.

[0092] like Figures 2-5As shown, by setting τ=7 and m=2, two-dimensional phase space reconstruction is performed to obtain the attractor distribution map in two-dimensional phase space of the one-dimensional piezoelectric sensing signal collected at different stages of damage development of concrete specimens, where t is the time series of the one-dimensional piezoelectric sensing signal.

[0093] Constructing a three-dimensional phase space matrix: such as Figures 6-9 As shown, the measured signal is reconstructed in three-dimensional phase space. The embedding dimension in Y is set to 3, and the delay time τ is set to 7. With τ=7 and m=3 set in Y, the three-dimensional phase space is reconstructed to obtain the attractor distribution map in three-dimensional phase space of the one-dimensional piezoelectric sensing signal collected at different stages of damage development in the concrete specimen.

[0094] In an exemplary embodiment, S4 is executed based on a computer processing system to extract chaotic features of attractors in a multidimensional phase space, so as to establish a mapping relationship between chaotic features and different damage stages of the damaged specimen.

[0095] The actual operation is as follows:

[0096] (1) The center distance is the sum of the distances from each vector point in the phase plane to the origin. The center distance H in two-dimensional phase space is:

[0097]

[0098] Center distance in three-dimensional phase space for:

[0099]

[0100] The attractor center distances of all samples in different dimensional phase spaces are as follows: Figures 10-11 As shown.

[0101] (2) Radius shift (geometric feature, reflecting the degree of phase point aggregation).

[0102] Two-dimensional phase space radius shift for:

[0103]

[0104] 3D phase space radius shift for:

[0105]

[0106] in, The position vector of the state point; The angle between the state vector and the main diagonal.

[0107] The attractor vector offsets of all samples in different dimensional phase spaces are as follows: Figures 12-13 As shown.

[0108] (3) Association dimension (attribute features, reflecting attractor complexity).

[0109] In the reconstructed phase space, a pair of phase points (i.e., attractors) , The mathematical expression of ) is shown below.

[0110]

[0111]

[0112] in, For phase point and phase point The Euclidean distance between them Let be the coordinates of the state point. Given a critical distance r, the proportion of vector pairs with a distance less than r among all vector pairs is defined as the correlation integral. As shown below.

[0113]

[0114] Where M = N - (m - 1)τ, and θ is the step function: x is the independent variable. .

[0115] The power law is followed as shown below.

[0116]

[0117] in, This represents the quantitative relationship between r and D(m).

[0118] The correlation dimension D(m) is shown below.

[0119]

[0120] The attractor correlation dimension of all samples in different dimensional phase spaces is as follows: Figures 14-15 As shown.

[0121] In one exemplary embodiment, S5 is performed based on a computer processing system to perform neural network classification and prediction.

[0122] The actual operation is as follows:

[0123] 1) Sample construction (data preparation), which involves calculating the center distance, vector offset, and correlation dimension feature values ​​of all samples to obtain a CSV file.

[0124] Data source: CSV file, each row has 7 values, the first 3 are the chaotic features of the attractor, and the last 4 are one-hot encoded labels, i.e. concrete damage categories.

[0125] Data cleaning: Outliers are removed using the Z-score (standard deviation method), meaning that samples whose values ​​for each chaotic feature differ from the mean by more than 3 times the standard deviation are removed.

[0126] Data partitioning: Using 5-fold cross-validation, the data is randomly divided into 5 parts, with 4 parts used for training and 1 part for testing each time. This ensures that each sample appears in the test set exactly once.

[0127] 2) Model building (neural network structure):

[0128] Input layer: There are 3 chaotic features with 3 attractors, so the input dimension is 3. The first layer has 10 neurons.

[0129] Hidden layers: Five hidden layers were added consecutively in the code, each with 10 neurons. The ReLU activation function was used. Each hidden layer was followed by a Dropout layer with a dropout rate of 0.3 (that is, 30% of the neurons were randomly dropped during training to reduce overfitting).

[0130] Output layer: The softmax activation function in the softmax classifier is used to output four concrete damage categories.

[0131] Training and Validation: Stochastic gradient descent optimization was performed, with 5 repeated random training iterations, achieving an average accuracy of ≥85%. The accuracy of determining the concrete's stage based on the feature values ​​at this point is then assessed. The loss and accuracy of the feedforward neural network model output samples are shown below. Figures 16-17 As shown.

[0132] The processing and output flowchart for one-dimensional piezoelectric sensing signals used in this application is as follows: Figure 18 As shown.

[0133] Therefore, this application can achieve the following effects.

[0134] 1. High accuracy in feature extraction.

[0135] Advantages: The AMI method (instead of the autocorrelation method) is used to determine τ, which can capture nonlinear signal correlations; the pseudo-nearest neighbor method is used to determine the dimension to ensure that the attractor is fully expanded; three types of features comprehensively characterize the attractor from the perspectives of "distribution-aggregation-complexity", and the feature distinction between the intact and unstable stages is over 85%, solving the problem of single feature extraction in traditional methods.

[0136] Principle: This scheme ensures that τ and m are optimal and can achieve quantitative calculation of features without subjective error.

[0137] 2. Sensitive damage detection.

[0138] Advantages: Traditional methods have low recognition rates for micro-damage. This application can capture chaotic characteristic changes more accurately by extracting multi-scale feature sets, with a recognition rate of ≥85%.

[0139] Principle: The extracted correlation dimension, center distance, and vector offset, as well as other multi-scale features, can effectively characterize the evolution of the attractor with damage. The complexity of the attractor has already increased in the early stages of damage, and this change can be quantified.

[0140] 3. High classification accuracy.

[0141] Advantages: Multiple training sessions with 105 sets of samples, average accuracy ≥85%, avoiding randomness in sample partitioning.

[0142] Principle: A 5-layer feedforward neural network adapts to the nonlinear mapping of 3D features, and a Softmax classifier achieves accurate multi-stage output, conforming to... Figure 18 Robust design of the algorithm framework.

[0143] 4. The project is highly feasible.

[0144] Advantages: The sensor model (PZT-5H), sampling frequency (1000Hz), τ=7, m=3 and other parameters are clearly defined and have been verified by actual measurement (when τ=7, the phase trajectory has no overlap and no divergence), which can directly guide the on-site deployment.

[0145] Principle: The parameters are determined by formula derivation, not empirical values, which lowers the threshold for engineering applications.

[0146] This application provides a concrete damage detection system, which includes the following modules.

[0147] The signal acquisition module is used to acquire one-dimensional piezoelectric sensing signals and real labels of different damage stages of concrete specimens; the different damage stages include the undamaged stage, the budding stage, the stable propagation stage, and the unstable propagation stage.

[0148] A one-dimensional piezoelectric sensing signal determination module is used to preprocess the one-dimensional piezoelectric sensing signal.

[0149] The chaos analysis module is used to calculate the maximum Lyapunov exponent of the preprocessed one-dimensional piezoelectric sensing signal and determine whether the preprocessed one-dimensional piezoelectric sensing signal has chaotic characteristics.

[0150] The multidimensional phase space matrix construction module is used to map the pre-processed one-dimensional piezoelectric sensing signal to a multidimensional phase space based on the phase space reconstruction theory, determine the delay time and embedding dimension, form an attractor reflecting chaotic characteristics, and construct a multidimensional phase space matrix; the multidimensional phase space matrix includes a two-dimensional phase space matrix and a three-dimensional phase space matrix.

[0151] The chaotic feature extraction module for attractors is used to extract chaotic features of attractors in the multidimensional phase space based on the multidimensional phase space matrix; the chaotic features include center distance, vector offset and correlation dimension.

[0152] The concrete damage detection module is used to input the chaotic features and the true labels of the damage stages into the feedforward neural network model and output the concrete damage category; the concrete damage category includes no damage, slight damage, moderate damage and complete damage.

[0153] This application also includes the hardware components of a concrete damage monitoring system, comprising: 1) a piezoelectric sensor module (PZT-5H): for acquiring vibration signals; 2) a data acquisition module: connecting the sensor to achieve analog-to-digital conversion and transmission; 3) an image acquisition module: a high-resolution camera to acquire damage labels; and 4) a computer processing module: containing built-in signal preprocessing, phase space reconstruction, and neural network calculation programs. All modules are connected via USB / wireless connection. The sensors are placed in key areas of the concrete, and the computer outputs the damage stage and prediction results.

[0154] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments. The computer device can be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (I / O), and a communication interface. The processor, memory, and I / O interface are connected via a system bus, and the communication interface is connected to the system bus via the I / O interface. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device stores data to be processed. The I / O interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communicating with an external terminal via a network connection. When the computer program is executed by the processor, it implements the above-described methods.

[0155] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0156] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0157] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0158] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0159] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by hardware related to computer program instructions. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0160] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0161] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0162] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting concrete damage, characterized in that, include: One-dimensional piezoelectric sensing signals and true labels of damage stages were obtained from concrete specimens at different damage stages; the different damage stages included the undamaged stage, the budding stage, the stable propagation stage, and the unstable propagation stage. The one-dimensional piezoelectric sensing signal is preprocessed; Calculate the maximum Lyapunov exponent of the preprocessed one-dimensional piezoelectric sensing signal and determine whether the preprocessed one-dimensional piezoelectric sensing signal has chaotic characteristics. If so, based on phase space reconstruction theory, the pre-processed one-dimensional piezoelectric sensing signal is mapped to a multi-dimensional phase space, the delay time and embedding dimension are determined, an attractor reflecting chaotic characteristics is formed, and a multi-dimensional phase space matrix is ​​constructed; the multi-dimensional phase space matrix includes a two-dimensional phase space matrix and a three-dimensional phase space matrix; the multi-dimensional phase space matrix Y is: in, Let m be the i-th phase point, i = 1, 2, ..., N, where N is the total length of the time series of the one-dimensional piezoelectric sensing signal; m is the embedding dimension; x i x i+τ x i+2τ ... Here are the coordinates of the attractor, and the attractor is a phase point in phase space; For delay time; When m=2, the multidimensional phase space matrix is ​​a two-dimensional phase space matrix; when m=3, the multidimensional phase space matrix is ​​a three-dimensional phase space matrix. Based on the multidimensional phase space matrix, chaotic features of attractors in the multidimensional phase space are extracted; the chaotic features include center distance, vector offset, and correlation dimension. The center distance includes the two-dimensional phase space center distance and the three-dimensional phase space center distance; The two-dimensional phase space center distance for: The three-dimensional phase space center distance for: Where M is the number of state points in phase space; The radius vector offset includes two-dimensional phase space radius vector offset and three-dimensional phase space radius vector offset; The two-dimensional phase space vector offset for: The three-dimensional phase space radius offset for: in, The position vector of the state point; The angle between the state vector and the main diagonal; The correlation dimension for: in, represents the proportion of vector pairs whose distance is less than r among all vector pairs; r is the critical distance. The chaotic features and the true labels of the damage stages are input into a feedforward neural network model to output concrete damage categories; the concrete damage categories include no damage, slight damage, moderate damage, and complete damage; If not, reacquire one-dimensional piezoelectric sensing signals and true labels of damage stages for different damage stages of the concrete specimen.

2. The concrete damage detection method according to claim 1, characterized in that, Based on phase space reconstruction theory, the pre-processed one-dimensional piezoelectric sensing signal is mapped to a multi-dimensional phase space. The delay time and embedding dimension are determined, an attractor reflecting chaotic characteristics is formed, and a multi-dimensional phase space matrix is ​​constructed, specifically including: The pre-processed one-dimensional piezoelectric sensing signal is mapped to a multi-dimensional phase space, and the delay time is determined by the average mutual information function method. The embedding dimension is determined using the pseudo-nearest neighbor method. A multidimensional phase space matrix is ​​constructed based on the delay time and the embedding dimension.

3. A concrete damage detection system, characterized in that, The concrete damage detection system performs the concrete damage detection method according to any one of claims 1-2, and the concrete damage detection system comprises: The signal acquisition module is used to acquire one-dimensional piezoelectric sensing signals and real labels of different damage stages of concrete specimens; the different damage stages include the undamaged stage, the budding stage, the stable propagation stage, and the unstable propagation stage. A one-dimensional piezoelectric sensing signal determination module is used to preprocess the one-dimensional piezoelectric sensing signal; The chaos analysis module is used to calculate the maximum Lyapunov exponent of the preprocessed one-dimensional piezoelectric sensing signal and determine whether the preprocessed one-dimensional piezoelectric sensing signal has chaotic characteristics. The multidimensional phase space matrix construction module is used to map the pre-processed one-dimensional piezoelectric sensing signal to a multidimensional phase space based on the phase space reconstruction theory, determine the delay time and embedding dimension, form an attractor reflecting chaotic characteristics, and construct a multidimensional phase space matrix; the multidimensional phase space matrix includes a two-dimensional phase space matrix and a three-dimensional phase space matrix. The chaotic feature extraction module for attractors is used to extract chaotic features of attractors in the multidimensional phase space based on the multidimensional phase space matrix; the chaotic features include center distance, vector offset, and correlation dimension; The concrete damage detection module is used to input the chaotic features and the true labels of the damage stages into the feedforward neural network model and output the concrete damage category; the concrete damage category includes no damage, slight damage, moderate damage and complete damage.

4. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the concrete damage detection method according to any one of claims 1-2.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the concrete damage detection method according to any one of claims 1-2.

6. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the concrete damage detection method according to any one of claims 1-2.

Images