Acoustic emission effective signal extraction method, device, equipment and storage medium

Abstract

The application discloses an acoustic emission effective signal extraction method, device and equipment and a storage medium, and comprises the following steps: collecting an initial acoustic emission signal, performing frequency band decomposition on the initial acoustic emission signal through wavelet packet transformation, and obtaining a plurality of frequency band waveforms; inputting each frequency band waveform into a preset autoencoder model for reconstruction feature filtering to obtain an initial filtered signal; inputting the initial filtered signal into a preset feature extraction model for time sequence feature filtering to obtain an acoustic emission effective signal; and the preset feature extraction model is used for time sequence feature analysis on the initial filtered signal. Compared with the prior art, the application firstly performs frequency band decomposition on the initial acoustic emission signal through wavelet packet transformation, then inputs different frequency band waveforms into a preset autoencoder model for reconstruction feature filtering, and finally performs time sequence feature analysis on the filtered signal, effectively eliminates background noise, and realizes extraction of the acoustic emission effective signal in a strong background noise environment.

Description

Technical Field

[0001] This invention relates to the field of civil engineering structure monitoring technology, and in particular to a method, apparatus, equipment and storage medium for extracting effective acoustic emission signals. Background Technology

[0002] Civil engineering structures subjected to dynamic loads are prone to fatigue failure, such as railway and highway steel bridges, offshore oil platforms, wind turbine towers, steel crane beams, and industrial equipment supports. Among these structures, welded steel structures exhibit the most significant fatigue problems. Because fatigue failure often displays highly brittle fracture characteristics, it is extremely sudden and frequently accompanied by catastrophic engineering accidents. To ensure the safe use of building structures and avoid catastrophic accidents, fatigue damage monitoring of welded steel structures prone to fatigue failure is of great importance. Acoustic emission signals are stress waves generated by the evolution of internal defects in materials. Compared with traditional non-destructive testing, acoustic emission methods are dynamic and real-time, and less restricted by the geometry of the specimen or structure. Since acoustic emission signals originate from the damage evolution of structures or materials, the signals themselves carry rich information about damage evolution. Although acoustic emission technology has yielded fruitful research results in laboratory environments since its inception, its practical application in engineering projects is often limited. The main reason is that strong background noise often accompanies the actual engineering environment. This background noise often makes it difficult to identify the effective acoustic emission signal (the signal representing the evolution of structural or material damage), or even completely annihilate it.

[0003] Existing noise reduction methods for acoustic emission signals mostly involve introducing a threshold based on signal decomposition, either through empirical judgment or by using artificial intelligence algorithms. The noise component is then removed, and the effective portion is extracted to reconstruct the signal. These approaches are generally suitable when the frequency difference between noise and the effective signal is significant, or when the noise intensity is relatively low compared to the effective acoustic emission signal in certain frequency bands. However, in the field of civil engineering steel structures, the intensity of background noise is usually greater than the effective signal released by material fatigue damage, and the noise bandwidth is relatively wide, making it difficult to distinguish in the frequency domain.

[0004] Therefore, there is an urgent need for a method to extract effective acoustic emission signals that can solve the technical problem that effective acoustic emission signals are difficult to identify under strong background noise interference in existing technologies. Summary of the Invention

[0005] The main objective of this invention is to provide a method, apparatus, device, and storage medium for extracting effective acoustic emission signals, aiming to solve the technical problem in the prior art that effective acoustic emission signals are difficult to identify under strong background noise interference.

[0006] To achieve the above objectives, the present invention provides a method for extracting effective acoustic emission signals, the method comprising the following steps:

[0007] An initial acoustic emission signal is acquired, and the initial acoustic emission signal is decomposed into multiple frequency band waveforms by wavelet packet transform.

[0008] The waveforms of each frequency band are input into a preset autoencoder model for reconstructed feature filtering to obtain the initial filtered signal;

[0009] The initial filtered signal is input into a preset feature extraction model for time-series feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used to perform time-series feature analysis on the initial filtered signal.

[0010] Optionally, the preset feature extraction model includes a CNN layer, a GRU layer, and an ANN layer. The CNN layer is used to compress data and extract local features of the data, the GRU layer is used to learn temporal features, and the ANN layer is used for signal classification.

[0011] Optionally, before the step of acquiring the initial acoustic emission signal and performing frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms, the method further includes:

[0012] A pure noise signal sample library is obtained, and a target wavelet packet tree transform is performed on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the number of first target waveforms being equal to the number of frequency band waveforms;

[0013] An autoencoder model is established based on each of the first target waveforms. The autoencoder model contains multiple autoencoder modules, and the number of autoencoder modules is equal to the number of the first target waveforms.

[0014] The autoencoder model is trained using the first target waveform to obtain a preset autoencoder model, and a corresponding loss threshold is determined based on the preset autoencoder model.

[0015] Optionally, before the step of acquiring the initial acoustic emission signal and performing frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms, the method further includes:

[0016] A sample library of pure fatigue damage acoustic emission signals is obtained, and the target wavelet packet tree transform is performed on the sample library of pure fatigue damage acoustic emission signals to decompose the pure fatigue damage acoustic emission signals in the sample library of pure fatigue damage acoustic emission signals into multiple second target waveforms, the number of second target waveforms being equal to the number of first target waveforms.

[0017] A feature extraction model is established based on each of the second target waveforms. The feature extraction model contains multiple feature extraction modules. The number of feature extraction modules is equal to the number of second target waveforms. The feature extraction module includes a CNN layer, four GRU layers, and a single ANN layer.

[0018] The feature extraction model is trained using the second target waveform to obtain a preset feature extraction model.

[0019] Optionally, the step of inputting the waveforms of each frequency band into a preset autoencoder model for reconstructed feature filtering to obtain an initial filtered signal includes:

[0020] The waveforms of each frequency band are input into a preset autoencoder model to obtain the reconstructed waveforms;

[0021] Determine the loss value of each reconstructed waveform and the corresponding frequency band waveform;

[0022] Set the reconstructed waveform corresponding to a loss value less than or equal to the loss threshold to zero, and add the reconstructed waveforms corresponding to a loss value greater than the loss threshold to obtain the initial filtered signal.

[0023] Optionally, the step of inputting the initial filtered signal into a preset feature extraction model for temporal feature filtering to obtain an effective acoustic emission signal includes:

[0024] The initial filtered signal is input into a preset feature extraction model to perform time-series feature analysis on the initial filtered signal through the preset feature extraction model, and to obtain the time-series feature analysis results;

[0025] Based on the time-series feature analysis results, the initial filtered signal is classified to obtain noise signals to be removed and fatigue damage signals;

[0026] The noise signal to be removed from the initial filtered signal is removed, and the fatigue damage signal is used as the effective acoustic emission signal.

[0027] Optionally, before the steps of obtaining a pure noise signal sample library and performing a target wavelet packet tree transform on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the method further includes:

[0028] Determine whether the steel structure in the service environment is a steel structure whose construction time has not exceeded the preset time threshold;

[0029] If the steel structure is built within a time limit of a preset time threshold, noise signals in the service environment are collected, and a pure noise signal sample library is generated based on the noise signals in the service environment.

[0030] If the steel structure has been built for a time exceeding a preset time threshold, then the steel structure is checked for fatigue damage.

[0031] If the steel structure does not suffer fatigue damage, noise signals under the service environment are collected, and a pure noise signal sample library is generated based on the noise signals under the service environment.

[0032] Furthermore, to achieve the above objectives, the present invention also proposes an effective acoustic emission signal extraction device, the device comprising:

[0033] The signal acquisition module is used to acquire the initial acoustic emission signal and perform frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms.

[0034] The first filtering module is used to input the waveforms of each frequency band into a preset autoencoder model for reconstructed feature filtering to obtain an initial filtered signal.

[0035] The second filtering module is used to input the initial filtered signal into a preset feature extraction model for time-series feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used to perform time-series feature analysis on the initial filtered signal.

[0036] Furthermore, to achieve the above objectives, the present invention also proposes an acoustic emission effective signal extraction device, the device comprising: a memory, a processor, and an acoustic emission effective signal extraction program stored in the memory and executable on the processor, the acoustic emission effective signal extraction program being configured to implement the steps of the acoustic emission effective signal extraction method described above.

[0037] Furthermore, to achieve the above objectives, the present invention also proposes a storage medium storing an acoustic emission effective signal extraction program, wherein when the acoustic emission effective signal extraction program is executed by a processor, it implements the steps of the acoustic emission effective signal extraction method described above.

[0038] This invention discloses a method for acquiring an initial acoustic emission signal, performing frequency band decomposition on the initial acoustic emission signal using wavelet packet transform to obtain multiple frequency band waveforms, inputting each frequency band waveform into a preset autoencoder model for reconstruction feature filtering to obtain an initial filtered signal, and inputting the initial filtered signal into a preset feature extraction model for temporal feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used to perform temporal feature analysis on the initial filtered signal. Because this invention first uses wavelet packet transform to decompose the initial acoustic emission signal into frequency bands, then inputs different frequency band waveforms into a preset autoencoder model for reconstruction feature filtering, and then performs temporal feature filtering on the filtered signal using a preset feature extraction model, it achieves secondary filtering of the initial acoustic emission signal. Compared with existing technologies, this invention effectively removes background noise and achieves the extraction of effective acoustic emission signals in environments with strong background noise. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating the first embodiment of the acoustic emission effective signal extraction method of the present invention;

[0040] Figure 2 A schematic diagram of typical signals for fatigue crack propagation at a welded joint;

[0041] Figure 3 This is a schematic diagram of a typical noise signal in a real engineering environment;

[0042] Figure 4 This is a schematic diagram of the first filtering effect in the acoustic emission effective signal extraction method of the present invention;

[0043] Figure 5 This is a schematic diagram illustrating the effective acoustic emission signal extraction effect in the effective acoustic emission signal extraction method of the present invention;

[0044] Figure 6 This is a flowchart illustrating the second embodiment of the acoustic emission effective signal extraction method of the present invention;

[0045] Figure 7 This is a schematic diagram of the acoustic emission signal processing flow in the acoustic emission effective signal extraction method of the present invention;

[0046] Figure 8 This is a flowchart illustrating the third embodiment of the acoustic emission effective signal extraction method of the present invention;

[0047] Figure 9 This is a structural block diagram of the first embodiment of the acoustic emission effective signal extraction device of the present invention;

[0048] Figure 10 This is a schematic diagram of the structure of the acoustic emission effective signal extraction device in the hardware operating environment involved in the embodiments of the present invention.

[0049] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0050] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0051] This invention provides a method for extracting effective acoustic emission signals, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the acoustic emission effective signal extraction method of the present invention.

[0052] In this embodiment, the method for extracting effective acoustic emission signals includes the following steps:

[0053] Step S10: Acquire the initial acoustic emission signal and perform frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms.

[0054] It should be noted that the executing entity in this embodiment can be a computer server device with data processing, network communication, and program execution functions, such as a server, tablet computer, or personal computer, or an electronic device or acoustic emission effective signal extraction device capable of performing the above functions. The following uses an acoustic emission effective signal extraction device as an example to illustrate this embodiment and the subsequent embodiments.

[0055] It should be understood that the aforementioned initial acoustic emission signal can be an acoustic emission signal mixed with noise and fatigue damage under actual service conditions, that is, an acoustic emission signal mixed with noise and fatigue damage under actual engineering conditions.

[0056] Understandably, wavelet packet transform is an extension of wavelet transform. Traditional wavelet transform primarily decomposes the low-frequency components of a signal while neglecting the high-frequency components. Wavelet packet transform, however, decomposes not only the low-frequency components but also the high-frequency components, thus providing a more detailed division of all frequency bands. This transform is calculated through a series of basis functions, which include four parameters: frequency, shift, scaling, and inversion, and can describe the local characteristics of the signal at different frequencies.

[0057] It should be noted that, based on the frequency range of the signal to be analyzed, an appropriate wavelet packet basis function can be selected. Then, based on the selected wavelet packet basis function, an optimal wavelet packet tree can be selected. Finally, the initial acoustic emission signal can be decomposed into frequency bands through the wavelet packet transform of the optimal wavelet packet tree.

[0058] Step S20: Input the waveforms of each frequency band into a preset autoencoder model for reconstructed feature filtering to obtain the initial filtered signal.

[0059] It's important to note that a predefined autoencoder model is an unsupervised learning neural network model that works by learning a compressed representation (encoding) of the input data and reconstructing the original input from this representation (decoding). A predefined autoencoder model mainly consists of two parts: an encoder and a decoder. The encoder is responsible for mapping the input data to a low-dimensional latent space representation, aiming to extract effective features from the input data and compress it into a more compact form. The decoder is responsible for mapping the encoder's output (i.e., the latent space representation) back to the original input space, aiming to reconstruct the original input data as accurately as possible.

[0060] In a specific implementation, the waveforms of each frequency band can be input into a preset autoencoder model to obtain reconstructed waveforms; then, the loss value of each reconstructed waveform and the corresponding frequency band waveform can be determined; finally, the reconstructed waveforms whose loss value is less than or equal to the loss threshold of the preset autoencoder model are set to zero, and the reconstructed waveforms whose loss value is greater than the loss threshold of the preset autoencoder model are added together to obtain the initial filtered signal.

[0061] Step S30: Input the initial filtered signal into a preset feature extraction model for time-series feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used to perform time-series feature analysis on the initial filtered signal.

[0062] It's important to explain that in practical engineering, the acoustic emission signals collected by sensors can be divided into two types: one representing ambient background noise, and the other representing the effective acoustic emission signal representing the evolution of fatigue damage at the weld joint. These two signals are usually difficult to distinguish simply by frequency filtering, as both belong to broadband signals. However, the signal representing fatigue damage evolution has its own unique characteristics, which are reflected in both the time and frequency domains, especially its temporal sequence. (Reference) Figure 2 , Figure 2 This diagram illustrates a typical signal of fatigue crack propagation in a welded joint. The horizontal axis represents time (milliseconds, ms), and the vertical axis represents signal intensity (millivolts, mV). During fatigue crack propagation in a steel structure welded joint, the effective acoustic emission signal generated by energy release exhibits typical burst characteristics, displaying a strong triangular time sequence with rapid attenuation. (Reference) Figure 3 , Figure 3 This is a schematic diagram of a typical noise signal in a real engineering environment. The horizontal axis represents time (unit: milliseconds, ms), and the vertical axis represents signal strength (unit: millivolts, mV). According to... Figure 2 and Figure 3 As can be seen, the effective acoustic emission signal and the noise signal have significant differences in timing. Therefore, this embodiment can utilize this difference in timing to perform timing feature analysis on the initial filtered signal using a preset feature extraction model, thereby achieving secondary filtering of the initial acoustic emission signal.

[0063] refer to Figure 4 , Figure 4 This is a schematic diagram of the first filtering effect in the acoustic emission effective signal extraction method of the present invention. In the diagram, the target signal represents the acoustic emission effective signal, the synthesized signal represents the initial acoustic emission signal collected, including background noise and the acoustic emission effective signal, and the signal after the first extraction represents the initial filtered signal obtained after reconstructed feature filtering. Figure 4 It is known that reconstructed feature filtering based on a preset autoencoder model can filter out most noise signals. In order to further eliminate background noise signals, this embodiment performs temporal feature filtering through a preset feature extraction model to achieve secondary filtering of the initial acoustic emission signal, thereby further eliminating background noise signals.

[0064] refer to Figure 5 , Figure 5 This is a schematic diagram of the effective acoustic emission signal extraction effect in the effective acoustic emission signal extraction method of the present invention. In the figure, SNR represents the signal-to-noise ratio, the target signal represents the effective acoustic emission signal, the synthesized signal represents the initial acoustic emission signal collected including the background noise signal and the effective acoustic emission signal, the first extraction represents the reconstruction feature filtering based on the preset autoencoder model, and the second extraction represents the temporal feature filtering based on the preset feature extraction model on the basis of the first extraction. Figure 4 and Figure 5 The horizontal axis represents time (unit: microseconds, μs), and the vertical axis represents signal strength (unit: millivolts, mV).

[0065] It should be explained that the above-mentioned preset feature extraction model includes a CNN (Convolutional Neural Networks) layer, a GRU (Gated Recurrent Unit) layer, and an ANN (Artificial Neural Networks) layer. The CNN layer is used to compress data and extract local features of the data, the GRU layer is used to learn temporal features, and the ANN layer is used for signal classification.

[0066] In a specific implementation, the initial filtered signal can be input into a preset feature extraction model to perform time-series feature analysis on the initial filtered signal and obtain time-series feature analysis results; based on the time-series feature analysis results, the initial filtered signal can be classified to obtain noise signals to be removed and fatigue damage signals; the noise signals to be removed from the initial filtered signal can be removed, and the fatigue damage signals can be used as effective acoustic emission signals.

[0067] This embodiment discloses the acquisition of an initial acoustic emission signal, followed by frequency band decomposition of the initial acoustic emission signal using wavelet packet transform to obtain multiple frequency band waveforms. Each frequency band waveform is then input into a preset autoencoder model for reconstruction feature filtering to obtain an initial filtered signal. Finally, the initial filtered signal is input into a preset feature extraction model for temporal feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used for temporal feature analysis of the initial filtered signal. Because this embodiment first uses wavelet packet transform to decompose the initial acoustic emission signal into frequency bands, then inputs different frequency band waveforms into a preset autoencoder model for reconstruction feature filtering, and then performs temporal feature filtering on the filtered signal using a preset feature extraction model, it achieves secondary filtering of the initial acoustic emission signal. Compared to existing technologies, this embodiment effectively eliminates background noise and achieves the extraction of effective acoustic emission signals even in environments with strong background noise.

[0068] refer to Figure 6 , Figure 6 This is a flowchart illustrating the second embodiment of the acoustic emission effective signal extraction method of the present invention.

[0069] Based on the first embodiment described above, in this embodiment, before step S10, the method further includes:

[0070] Step S101: Obtain a pure noise signal sample library, and perform a target wavelet packet tree transform on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the number of first target waveforms being equal to the number of frequency band waveforms.

[0071] It should be understood that a pure noise signal sample library can be a collection of acoustic emission signal sample data collected under the service environment of steel structures when fatigue damage is ensured. In addition, a pure noise signal sample library can generally cover most noise types under service environments.

[0072] It should be explained that the target wavelet packet tree transform can be the optimal wavelet packet tree transform that takes into account multiple factors such as signal characteristics, analysis target, wavelet packet basis selection and decomposition level, and is adjusted and optimized.

[0073] Understandably, when performing target wavelet packet tree transform on noise signals in a pure noise signal sample library, a suitable wavelet packet basis should be selected. Users can choose according to the actual situation, for example, a wavelet packet basis between Db4 and Db10. Generally, the more layers of the target wavelet tree, the better. However, considering the computational burden, the number of layers of the target wavelet tree in this embodiment can be user-defined, for example, set to 4 layers.

[0074] It should be noted that, through the decomposition of the target wavelet packet tree, the noise signals in the pure noise signal sample library are decomposed into n first target waveforms, where n is equal to the number of terminal nodes of the target wavelet packet tree.

[0075] Step S102: Establish an autoencoder model based on each of the first target waveforms. The autoencoder model contains multiple autoencoder modules, and the number of autoencoder modules is equal to the number of the first target waveforms.

[0076] Step S103: Train the autoencoder model using the first target waveform to obtain a preset autoencoder model, and determine the corresponding loss threshold based on the preset autoencoder model.

[0077] It should be noted that, for n first target waveforms, an autoencoder model is established, which contains n autoencoder modules. Then, the autoencoder model is trained based on the first target waveforms. When the waveform reconstruction effect is the best, the corresponding autoencoder model is used as the preset autoencoder model. Then, the loss threshold distribution of the preset autoencoder model is determined. The quartile with a larger loss can be selected as the loss threshold.

[0078] In a specific implementation, the step of inputting each frequency band waveform into a preset autoencoder model for reconstruction feature filtering to obtain an initial filtered signal includes: inputting each frequency band waveform into a preset autoencoder model to obtain a reconstructed waveform; determining the loss value of each reconstructed waveform and the corresponding frequency band waveform; setting the reconstructed waveform with a loss value less than or equal to the loss threshold to zero, and adding the reconstructed waveform with a loss value greater than the loss threshold to obtain an initial filtered signal.

[0079] It should be understood that the reconstructed waveforms with loss values ​​less than or equal to the loss threshold are background noise signals, while the reconstructed waveforms with loss values ​​greater than the loss threshold are signals that may carry fatigue damage. Therefore, after n waveforms pass through n encoder modules of a preset autoencoder model, those determined to be noise signals are directly zeroed out, while those determined to potentially carry fatigue damage signals are retained. Finally, the retained waveforms are summed to obtain the signal after the first filtering, i.e., the initial filtered signal.

[0080] Furthermore, prior to step S10, the following steps are also included:

[0081] Step S104: Obtain a pure fatigue damage acoustic emission signal sample library, and perform the target wavelet packet tree transform on the pure fatigue damage acoustic emission signal sample library to decompose the pure fatigue damage acoustic emission signals in the pure fatigue damage acoustic emission signal sample library into multiple second target waveforms, the number of second target waveforms being equal to the number of first target waveforms.

[0082] Understandably, a pure fatigue damage acoustic emission signal sample library can be a signal sample library composed of noise-free pure fatigue damage acoustic emission signals collected in a laboratory environment.

[0083] Step S105: Establish a feature extraction model based on each of the second target waveforms. The feature extraction model includes multiple feature extraction modules. The number of feature extraction modules is equal to the number of the second target waveforms. The feature extraction module includes one layer of CNN, four layers of GRU and one layer of ANN.

[0084] It should be noted that the first CNN layer is used to compress data and extract local features of the data, the four GRU layers are used to learn temporal features, and the first ANN layer is used for signal classification.

[0085] Step S106: Train the feature extraction model using the second target waveform to obtain a preset feature extraction model.

[0086] It should be understood that training the feature extraction model using the second target waveform can adjust and optimize the model parameters of the feature extraction model, thereby obtaining a preset feature extraction model.

[0087] refer to Figure 7 , Figure 7 This diagram illustrates the acoustic emission signal processing flow in the effective acoustic emission signal extraction method of this invention. The diagram includes two filtering processes. The first filtering primarily uses optimal wavelet packet tree transform combined with a preset autoencoder model to reconstruct the acoustic emission signal (i.e., the initial acoustic emission signal) under actual service conditions, which is subject to mixed noise and fatigue damage. The second filtering, performed after the first filtering, primarily uses a preset feature extraction model to perform temporal feature filtering to obtain the acoustic emission signal containing the evolution of fatigue damage at the weld joint (i.e., the effective acoustic emission signal). It should be noted that two steps are required before filtering the acoustic emission signal under actual service conditions, which is subject to mixed noise and fatigue damage: Step 1: Learning the background noise characteristics under actual service conditions; Step 2: Learning the temporal characteristics of the fatigue signal at the weld joint.

[0088] In step one, a large number of acoustic emission signals from the service environment of the steel structure need to be collected to obtain a pure noise signal sample library. Then, the pure noise signals in the pure noise signal sample library are decomposed into optimal wavelet packet trees to obtain the waveforms corresponding to the terminal nodes of the optimal wavelet packet trees (i.e., the first target waveforms). The first target waveforms are then input into an autoencoder model for training to obtain a preset autoencoder model. Based on the preset autoencoder model, the corresponding loss threshold is determined. The autoencoder model contains multiple autoencoder modules. The number of autoencoder modules is equal to the number of the first target waveforms. That is, the number of autoencoder modules corresponds to the total number of terminal nodes of the optimal wavelet packet trees. The terminal nodes of n optimal wavelet packet trees correspond to n autoencoder modules (the autoencoder module includes an encoder and a decoder. The encoder and the input layer are on the same side, and the decoder and the output layer are on the same side).

[0089] In step two, signals obtained through fatigue tests in a laboratory environment are collected to obtain a pure fatigue damage acoustic emission signal sample library. Then, the pure fatigue damage acoustic emission signal sample library is decomposed into waveforms corresponding to the terminal nodes of the optimal wavelet packet tree (i.e., the second target waveform) by optimal wavelet packet tree decomposition. The second target waveform is then input into the feature extraction model (the feature extraction model includes CNN+GRU+ANN) for training to obtain the preset feature extraction model, where CNN represents Conv1d in the figure as a convolutional neural network.

[0090] In the specific implementation, after learning the background noise characteristics under actual service conditions in step one and the timing characteristics of fatigue signals at welded joints in step two, acoustic emission signals mixed with noise and fatigue damage under actual service conditions are collected. The initial acoustic emission signals are then decomposed into frequency bands using optimal wavelet packet tree decomposition to obtain multiple frequency band waveforms. Each frequency band waveform is then input into a preset autoencoder model to obtain reconstructed waveforms. The loss value between each reconstructed waveform and its corresponding frequency band waveform is determined. Reconstructed waveforms with loss values ​​less than or equal to a loss threshold are set to zero, while those with loss values ​​greater than the loss threshold are set to zero. The waveforms are summed to obtain an initial filtered signal (i.e., a reconstructed signal); the initial filtered signal is then input into a preset feature extraction model to perform time-series feature analysis on the initial filtered signal, obtaining the time-series feature analysis results; based on the time-series feature analysis results, the initial filtered signal is classified to obtain noise signals to be removed (noise) and fatigue damage signals (damage signals); the noise signals to be removed from the initial filtered signal are removed, and the fatigue damage signals are used as effective acoustic emission signals, finally outputting the effective acoustic emission signal (i.e., the acoustic emission signal containing the evolution of fatigue damage at the weld joint).

[0091] This embodiment discloses a method for acquiring a pure noise signal sample library, performing a target wavelet packet tree transform on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the number of first target waveforms being equal to the number of frequency band waveforms; establishing an autoencoder model based on each first target waveform, the autoencoder model containing multiple autoencoder modules, the number of autoencoder modules being equal to the number of first target waveforms; training the autoencoder model using the first target waveforms to obtain a preset autoencoder model, and determining a corresponding loss threshold based on the preset autoencoder model; acquiring a pure fatigue damage acoustic emission signal sample library, performing the target wavelet packet tree transform on the pure fatigue damage acoustic emission signal sample library to decompose the pure fatigue damage acoustic emission signals in the pure fatigue damage acoustic emission signal sample library into multiple second target waveforms, the number of second target waveforms being equal to the number of first target waveforms; establishing a feature extraction model based on each second target waveform, the feature extraction model containing multiple feature extraction modules, the number of feature extraction modules being equal to the number of second target waveforms, the feature extraction module including a one-layer CNN, a four-layer GRU, and a one-layer ANN; training the feature extraction model using the second target waveforms to obtain a preset feature extraction model. Compared to existing technologies, this embodiment improves the model's ability to remove background noise signals by learning the background noise characteristics in actual service environments and the temporal characteristics of fatigue signals at welded nodes.

[0092] refer to Figure 8 , Figure 8 This is a flowchart illustrating the third embodiment of the acoustic emission effective signal extraction method of the present invention.

[0093] Based on the above embodiments, in this embodiment, before step S101, the method further includes:

[0094] Step S01: Determine whether the steel structure in the service environment is a steel structure whose establishment time has not exceeded the preset time threshold.

[0095] Step S02: If the establishment time is less than the preset time threshold for the steel structure, then collect the noise signal under the service environment and generate a pure noise signal sample library based on the noise signal under the service environment.

[0096] Step S03: If the steel structure has been built for a time exceeding a preset time threshold, then detect whether the steel structure has fatigue damage.

[0097] Step S04: If the steel structure does not show fatigue damage, collect the noise signal under the service environment and generate a pure noise signal sample library based on the noise signal under the service environment.

[0098] Understandably, for newly constructed steel structures (i.e., steel structures in service environments whose construction time has not exceeded a preset time threshold), noise signals from their service environment can be directly collected, and the more noise signal data samples available, the better, if conditions permit. For steel structures that have already been in service for some time (i.e., steel structures in service environments whose construction time has exceeded a preset time threshold), it is necessary to first determine whether the steel structure has already suffered fatigue damage through appropriate inspections. If not, then the process can be followed as for newly constructed steel structures, collecting and building a pure noise signal sample library.

[0099] This embodiment discloses a method for determining whether a steel structure in a service environment has been established within a preset time threshold. If the steel structure has been established within the preset time threshold, noise signals from the service environment are collected, and a pure noise signal sample library is generated based on these noise signals. If the steel structure has been established beyond the preset time threshold, fatigue damage is detected. If the steel structure has not experienced fatigue damage, noise signals from the service environment are collected, and a pure noise signal sample library is generated based on these noise signals. Compared to existing technologies, this embodiment ensures that the pure noise signal sample library contains only pure noise signals from the steel structure in the service environment, thereby improving the accuracy of subsequent model training and ensuring the reliability of effective acoustic emission signal extraction.

[0100] Furthermore, this embodiment of the invention also proposes a storage medium storing an acoustic emission effective signal extraction program, which, when executed by a processor, implements the steps of the acoustic emission effective signal extraction method described above.

[0101] Reference Figure 9 , Figure 9 This is a structural block diagram of the first embodiment of the acoustic emission effective signal extraction device of the present invention.

[0102] like Figure 9 As shown, the acoustic emission effective signal extraction device proposed in this embodiment of the invention includes: a signal acquisition module 901, a first filtering module 902, and a second filtering module 903.

[0103] The signal acquisition module 901 is used to acquire the initial acoustic emission signal and perform frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms.

[0104] The first filtering module 902 is used to input the waveforms of each frequency band to a preset autoencoder model for reconstructed feature filtering to obtain an initial filtered signal.

[0105] The second filtering module 903 is used to input the initial filtered signal into a preset feature extraction model for time-series feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used to perform time-series feature analysis on the initial filtered signal.

[0106] The second filtering module 903 is further configured to input the initial filtered signal into a preset feature extraction model, so as to perform time-series feature analysis on the initial filtered signal through the preset feature extraction model to obtain time-series feature analysis results; classify the initial filtered signal based on the time-series feature analysis results to obtain noise signals to be removed and fatigue damage signals; remove the noise signals to be removed from the initial filtered signal, and use the fatigue damage signals as effective acoustic emission signals.

[0107] This device embodiment discloses the acquisition of an initial acoustic emission signal, followed by frequency band decomposition of the initial acoustic emission signal using wavelet packet transform to obtain multiple frequency band waveforms. Each frequency band waveform is then input into a preset autoencoder model for reconstruction feature filtering to obtain an initial filtered signal. Finally, the initial filtered signal is input into a preset feature extraction model for temporal feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used for temporal feature analysis of the initial filtered signal. Because this device embodiment first uses wavelet packet transform to decompose the initial acoustic emission signal into frequency bands, then inputs different frequency band waveforms into a preset autoencoder model for reconstruction feature filtering, and then performs temporal feature filtering on the filtered signal using a preset feature extraction model, it achieves secondary filtering of the initial acoustic emission signal. Compared to existing technologies, this device embodiment effectively eliminates background noise and achieves the extraction of effective acoustic emission signals even in environments with strong background noise.

[0108] Based on the first embodiment of the acoustic emission effective signal extraction device of the present invention, a second embodiment of the acoustic emission effective signal extraction device of the present invention is proposed.

[0109] In this embodiment, the signal acquisition module 901 is further configured to acquire a pure noise signal sample library, perform a target wavelet packet tree transform on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the number of first target waveforms being equal to the number of frequency band waveforms; establish an autoencoder model based on each first target waveform, the autoencoder model containing multiple autoencoder modules, the number of autoencoder modules being equal to the number of first target waveforms; train the autoencoder model using the first target waveforms to obtain a preset autoencoder model, and determine a corresponding loss threshold based on the preset autoencoder model.

[0110] The signal acquisition module 901 is further configured to acquire a pure fatigue damage acoustic emission signal sample library, perform the target wavelet packet tree transform on the pure fatigue damage acoustic emission signal sample library to decompose the pure fatigue damage acoustic emission signals in the pure fatigue damage acoustic emission signal sample library into multiple second target waveforms, the number of second target waveforms being equal to the number of first target waveforms; establish a feature extraction model based on each second target waveform, the feature extraction model containing multiple feature extraction modules, the number of feature extraction modules being equal to the number of second target waveforms, the feature extraction module including one layer of CNN, four layers of GRU and one layer of ANN; train the feature extraction model using the second target waveforms to obtain a preset feature extraction model.

[0111] Other embodiments or specific implementations of the acoustic emission effective signal extraction device of the present invention can be referred to the above-described method embodiments, and will not be repeated here.

[0112] This application provides an acoustic emission effective signal extraction device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the acoustic emission effective signal extraction method in the above embodiment 1.

[0113] The following is for reference. Figure 10 The diagram illustrates a structural schematic suitable for implementing the acoustic emission effective signal extraction device in the embodiments of this application. The acoustic emission effective signal extraction device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 10 The acoustic emission effective signal extraction device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0114] like Figure 10As shown, the acoustic emission valid signal extraction device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the acoustic emission valid signal extraction device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the acoustic emission effective signal extraction device to communicate wirelessly or wiredly with other devices to exchange data. Although acoustic emission effective signal extraction devices with various systems are shown in the figures, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0115] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0116] The acoustic emission effective signal extraction device provided in this application, employing the acoustic emission effective signal extraction method in the above embodiments, can solve the technical problem in the prior art where the acoustic emission effective signal is difficult to identify under strong background noise interference. Compared with the prior art, the beneficial effects of the acoustic emission effective signal extraction device provided in this application are the same as those of the acoustic emission effective signal extraction method provided in the above embodiments, and other technical features in this acoustic emission effective signal extraction device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0117] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0118] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0119] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0120] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0121] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0122] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for extracting effective acoustic emission signals, characterized in that, The method includes: An initial acoustic emission signal is acquired, and the initial acoustic emission signal is decomposed into multiple frequency band waveforms by wavelet packet transform. The waveforms of each frequency band are input into a preset autoencoder model for reconstructed feature filtering to obtain the initial filtered signal; The initial filtered signal is input into a preset feature extraction model for time-series feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used to perform time-series feature analysis on the initial filtered signal. The preset feature extraction model includes a CNN layer, four GRU layers, and an ANN layer. The CNN layer is used to compress data and extract local features of the data. The four GRU layers are used to learn time-series features. The ANN layer is used for signal classification. Before the step of acquiring the initial acoustic emission signal and performing frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms, the method further includes: A pure noise signal sample library is obtained, and a target wavelet packet tree transform is performed on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the number of first target waveforms being equal to the number of frequency band waveforms; An autoencoder model is established based on each of the first target waveforms. The autoencoder model contains multiple autoencoder modules, and the number of autoencoder modules is equal to the number of the first target waveforms. The autoencoder model is trained using the first target waveform to obtain a preset autoencoder model, and a corresponding loss threshold is determined based on the preset autoencoder model.

2. The method as described in claim 1, characterized in that, Before the step of acquiring the initial acoustic emission signal and performing frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms, the method further includes: A sample library of pure fatigue damage acoustic emission signals is obtained, and the target wavelet packet tree transform is performed on the sample library of pure fatigue damage acoustic emission signals to decompose the pure fatigue damage acoustic emission signals in the sample library of pure fatigue damage acoustic emission signals into multiple second target waveforms, the number of second target waveforms being equal to the number of first target waveforms. A feature extraction model is established based on each of the second target waveforms. The feature extraction model contains multiple feature extraction modules. The number of feature extraction modules is equal to the number of second target waveforms. The feature extraction module includes a CNN layer, four GRU layers, and a single ANN layer. The feature extraction model is trained using the second target waveform to obtain a preset feature extraction model.

3. The method as described in claim 1, characterized in that, The step of inputting the waveforms of each frequency band into a preset autoencoder model for reconstructed feature filtering to obtain an initial filtered signal includes: The waveforms of each frequency band are input into a preset autoencoder model to obtain the reconstructed waveforms; Determine the loss value of each reconstructed waveform and the corresponding frequency band waveform; Set the reconstructed waveform corresponding to a loss value less than or equal to the loss threshold to zero, and add the reconstructed waveforms corresponding to a loss value greater than the loss threshold to obtain the initial filtered signal.

4. The method as described in claim 1, characterized in that, The step of inputting the initial filtered signal into a preset feature extraction model for temporal feature filtering to obtain an effective acoustic emission signal includes: The initial filtered signal is input into a preset feature extraction model to perform time-series feature analysis on the initial filtered signal through the preset feature extraction model, and to obtain the time-series feature analysis results; Based on the time-series feature analysis results, the initial filtered signal is classified to obtain noise signals to be removed and fatigue damage signals; The noise signal to be removed from the initial filtered signal is removed, and the fatigue damage signal is used as the effective acoustic emission signal.

5. The method as described in claim 1, characterized in that, Before the steps of acquiring a pure noise signal sample library and performing a target wavelet packet tree transform on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the method further includes: Determine whether the steel structure in the service environment is a steel structure whose construction time has not exceeded the preset time threshold; If the steel structure is built within a time limit of a preset time threshold, noise signals in the service environment are collected, and a pure noise signal sample library is generated based on the noise signals in the service environment. If the steel structure has been built for a time exceeding a preset time threshold, then the steel structure is checked for fatigue damage. If the steel structure does not suffer fatigue damage, noise signals under the service environment are collected, and a pure noise signal sample library is generated based on the noise signals under the service environment.

6. A device for extracting effective acoustic emission signals, characterized in that, The device includes: The signal acquisition module is used to acquire the initial acoustic emission signal and perform frequency band decomposition on the initial acoustic emission signal through wavelet packet transform to obtain multiple frequency band waveforms. The first filtering module is used to input the waveforms of each frequency band into a preset autoencoder model for reconstructed feature filtering to obtain an initial filtered signal; The second filtering module is used to input the initial filtered signal into a preset feature extraction model for time-series feature filtering to obtain an effective acoustic emission signal. The preset feature extraction model is used to perform time-series feature analysis on the initial filtered signal. The preset feature extraction model includes a CNN layer, four GRU layers, and an ANN layer. The CNN layer is used to compress data and extract local features of the data. The four GRU layers are used to learn time-series features. The ANN layer is used for signal classification. The signal acquisition module is further configured to acquire a pure noise signal sample library, perform a target wavelet packet tree transform on the pure noise signal sample library to decompose the noise signals in the pure noise signal sample library into multiple first target waveforms, the number of first target waveforms being equal to the number of frequency band waveforms; establish an autoencoder model based on each first target waveform, the autoencoder model containing multiple autoencoder modules, the number of autoencoder modules being equal to the number of first target waveforms; train the autoencoder model using the first target waveforms to obtain a preset autoencoder model, and determine a corresponding loss threshold based on the preset autoencoder model.

7. A device for extracting effective acoustic emission signals, characterized in that, The device includes: a memory, a processor, and an acoustic emission effective signal extraction program stored in the memory and executable on the processor, the acoustic emission effective signal extraction program being configured to implement the steps of the acoustic emission effective signal extraction method as described in any one of claims 1 to 5.

8. A storage medium, characterized in that, The storage medium stores a program for extracting effective acoustic emission signals, which, when executed by a processor, implements the steps of the method for extracting effective acoustic emission signals as described in any one of claims 1 to 5.

Images