Memory, hydrogen-induced cracking process identification method, device and equipment

Abstract

The application discloses a kind of identification method and device of hydrogen induced cracking process, the method includes the following steps: A, when monitoring hydrogen induced cracking process to metal tensile test piece under the condition of constant load and temperature control hydrogen charging, acoustic emission signal of whole process is collected;B, the time domain parameter extracted in acoustic emission signal and the wavelet energy spectrum coefficient obtained by calculation are used as input variable to carry out cluster analysis, and the cluster class corresponding to bubble generation and rupture, material surface cracking and separation, hydrogen-induced crack growth three stages is obtained;C, the time domain parameter and wavelet energy spectrum coefficient of different regions in cluster class are used as input, and the acoustic emission source of three stages is used as output, and hydrogen-induced cracking process identification model is trained;D, the three stages of hydrogen-induced cracking process corresponding to different acoustic emission signals are identified by hydrogen-induced cracking process identification model after training.The method and device of the application can quickly and effectively process and analyze acoustic emission signals under different modes in hydrogen induced cracking process, and provide a theoretical basis for hydrogen induced cracking on-site monitoring of hydrogen pressure vessel.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydrogen pressure vessel monitoring, in particular to a hydrogen-induced cracking process identification and evaluation method and device. BACKGROUND

[0002] Hydrogen-induced cracking is a high-frequency and serious damage phenomenon of hydrogen pressure vessels. Hydrogen pressure vessels are widely used in petrochemical enterprises and often operate under harsh conditions such as high temperature, high pressure, and strong corrosion. The medium is often flammable, explosive, or corrosive and toxic. Once a leak or explosion occurs, it will directly affect people's life and property safety, national economic security, and social stability. Therefore, effective online real-time monitoring of hydrogen pressure vessels is an important research topic.

[0003] Acoustic emission detection technology is a dynamic non-destructive testing method that can effectively represent the internal damage evolution information of materials by monitoring the elastic waves generated by energy release inside the material. It has been widely applied to the damage detection of various types of materials. The damage mode and process of hydrogen-induced cracking are very complex, and a large number of acoustic emission signals are generated during the damage process. Therefore, fast and effective processing and analysis of hydrogen-induced cracking signals is the key to online monitoring.

[0004] Non-destructive testing using acoustic emission methods is also used in existing technologies. For example, Chinese patent application CN110376289A discloses a composite material fiber weaving layer damage identification method based on acoustic emission methods, which belongs to the field of non-destructive testing. It uses acoustic emission non-destructive testing methods to monitor the damage status of different fiber weaving layers of fiber-reinforced composite materials. By combining acoustic emission technology with wavelet analysis methods for multi-scale analysis signals and effective pattern recognition technology, it can achieve effective, accurate, and real-time online non-destructive testing of different fiber weaving layers of fiber-reinforced composite structures. The scheme includes the following steps: establishing a damage characteristic analysis table for the monitored material; and establishing an acoustic emission online monitoring system based on the obtained damage characteristic analysis table. This method can improve the accuracy of detection, and is more detailed and reliable for damage evaluation of composite materials.

[0005] How to effectively identify acoustic emission signals and obtain qualitative and quantitative information of acoustic emission sources from these acoustic signals, and extract the characteristic parameters of the principal components, is of great significance to the online real-time monitoring of hydrogen pressure vessels. At present, the accurate identification of the hydrogen-induced cracking process still needs further research. Therefore, by testing under laboratory conditions, combining visual observation, electrochemistry, mechanics, and acoustics, and other multi-source information, the different modal conditions of the hydrogen-induced cracking process are studied, and the corresponding characteristic parameters are extracted, providing a theoretical basis and technical support for the online monitoring of hydrogen pressure vessels.

[0006] The information disclosed in this Background section is only for the purpose of increasing an understanding of the general context of the present application and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art that is already known in any country in the world. SUMMARY

[0007] The purpose of the present application is to provide a method and device for identifying hydrogen-induced cracking process, which can quickly and effectively process and analyze acoustic emission signals in different modes during hydrogen-induced cracking process under laboratory conditions, and provide a theoretical basis for on-site monitoring of hydrogen-induced cracking of hydrogen pressure vessels in use.

[0008] To achieve the above-mentioned purpose, according to a first aspect of the present application, a method for identifying hydrogen-induced cracking process is provided, comprising the following steps: A, collecting acoustic emission signals in the whole process when monitoring hydrogen-induced cracking process of a metal tensile specimen under constant load and temperature-controlled hydrogen charging experimental conditions; B, performing cluster analysis on time domain parameters extracted from the acoustic emission signals and wavelet energy spectrum coefficients obtained by calculation as input variables, and obtaining cluster categories corresponding to three stages of bubble generation and rupture, material surface cracking and separation, and hydrogen-induced crack growth, respectively; C, taking time domain parameters and wavelet energy spectrum coefficients in different regions of the cluster categories as input and acoustic emission sources in the three stages as output, training a hydrogen-induced cracking process identification model; D, identifying the three stages of hydrogen-induced cracking process corresponding to different acoustic emission signals through the trained hydrogen-induced cracking process identification model.

[0009] Further, in the above technical solution, the time domain parameters can include: rise time, amplitude, energy, count and duration of the acoustic emission signal.

[0010] Further, in the above technical solution, the calculation of the wavelet energy spectrum coefficient can specifically include: calculating a wavelet decomposition scale for determining the number of wavelet decomposition; calculating energy values in each level of the wavelet function and a total energy value according to the number of wavelet decomposition; and calculating the wavelet energy spectrum coefficient through the energy values in each level and the total energy value.

[0011] Further, in the above technical solution, the calculation formula of the wavelet decomposition scale is:

[0012]

[0013] Wherein, f s is the sampling frequency, L f is the filter length, and N is the sampling length.

[0014] The wavelet function is specifically expressed as:

[0015] f(t)=f 0 (t)+f1 (t)+…+f j (t) Formula (2);

[0016] wherein, f 0 (t), f 1 (t), …, f j (t) is each level of the decomposition signal;

[0017] The energy in each level is:

[0018] The total energy is:

[0019] The calculation formula of the wavelet energy spectrum coefficient of each level under the wavelet scale is:

[0020] R (j) (t) = E (j) (t) / E (T) (t) Formula (5).

[0021] Further, in the technical scheme, the clustering analysis in step B is based on the k-means algorithm for iterative calculation, and specifically includes: taking the normalized time domain parameters and the wavelet energy spectrum coefficients as inputs of the k-means, using a weighted Euclidean distance to classify the three-stage acoustic emission signals, and obtaining clustering centers corresponding to the three stages respectively; and performing reinforcement processing on the edge data of each clustering center through a compensation algorithm to obtain a clustering category with clear boundaries.

[0022] Further, in the process of obtaining the clustering centers corresponding to the three stages respectively, the condition for determining the clustering effectiveness is that the Euclidean distance as a similarity measure between acoustic emission signals is greater than a preset threshold; and the similarity measure formula is:

[0023]

[0024] wherein, a(l) represents the average distance between the lth point and other points in the same category; and b(l,k) is a vector representing the average distance between the 1st point and points in different categories.

[0025] Further, in the technical scheme, the reinforcement processing of the edge data through the compensation algorithm can specifically include: selecting a cluster center generated by clustering as an anchor point, taking the cluster center as the center and the maximum value of the Euclidean distance as the radius, segmenting the Euclidean distance, dividing the clustering signals under a classification mode into multiple regions, performing different degrees of compensation on different regions, obtaining corresponding compensation coefficients based on the dichotomy, and making the edge data of each region equivalent to the clustering center.

[0026] Further, in the technical scheme, the hydrogen induced cracking process identification model in step C can be a BP neural network based algorithm model.

[0027] Further, in the technical scheme, during the establishment of the hydrogen induced cracking process identification model, a compensation coefficient can be added to the algorithm of the connection weight of the hydrogen induced cracking process identification model.

[0028] To achieve the above-mentioned purposes, according to the second aspect of the present application, the present application provides a hydrogen induced cracking process identification device, comprising: an acoustic emission signal acquisition module, which is used for acquiring acoustic emission signals in the whole process when monitoring the hydrogen induced cracking process of a metal tensile specimen under constant load and temperature-controlled hydrogen charging experimental conditions; a clustering analysis module, which is used for performing clustering analysis on time domain parameters extracted from the acoustic emission signals and wavelet energy spectrum coefficients obtained by calculation, and obtaining clustering categories corresponding to three stages of bubble generation and rupture, material surface cracking and separation, and hydrogen induced crack growth, respectively; a model training module, which takes time domain parameters and wavelet energy spectrum coefficients in different regions of the clustering categories as inputs and acoustic emission sources in the three stages as outputs, and trains a hydrogen induced cracking process identification model; and a process identification module, which is used for identifying the three stages of the hydrogen induced cracking process corresponding to different acoustic emission signals through the trained hydrogen induced cracking process identification model.

[0029] To achieve the above-mentioned purposes, according to the third aspect of the present application, the present application provides a memory comprising a set of instructions adapted to be executed by a processor to perform the steps in the hydrogen induced cracking process identification method.

[0030] To achieve the above-mentioned purposes, according to the fourth aspect of the present application, the present application provides a hydrogen induced cracking process identification device, comprising a bus, an input device, an output device, a processor and a memory as described above; the bus is used to connect the memory, the input device, the output device and the processor; the input device and the output device are used to realize the interaction with the user; and the processor is used to execute the set of instructions in the memory.

[0031] Compared with the prior art, the present application has the following beneficial effects:

[0032] 1) The present application can more effectively represent the characteristic parameters of the hydrogen induced cracking process identification model through the time domain parameters of the five acoustic emission signals and the calculated wavelet energy spectrum coefficients, and the identification accuracy is higher;

[0033] 2) The clustering analysis process of the present application adds a compensation algorithm, which can strengthen the data at the edge of each clustering center, so that the boundary of the clustering category is clearer, and the classification identification error is minimized;

[0034] 3) The present application can provide direct basis for the damage evolution and failure mechanism research of the material in the hydrogen induced cracking process.

[0035] 4) Through the hydrogen induced cracking process recognition model established based on the BP neural network, the hydrogen induced cracking acoustic emission signal can be effectively recognized, so that the nature of the acoustic emission signal source is determined, that is, the failure mechanism of the material is judged;

[0036] 5) The present application can distinguish different stages of hydrogen induced cracking of the material, and has great significance for real-time monitoring of in-service hydrogen pressure vessels.

[0037] Other features and aspects of the present application will become apparent from the following detailed description of exemplary embodiments, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0038] One or more embodiments are illustrated by way of example in the drawings in which like reference numerals indicate like elements, and in which:

[0039] Figure 1 It is a flowchart of the hydrogen induced cracking process recognition method of the present application embodiment 1;

[0040] Figure 2 It is a structural schematic diagram of the hydrogen induced cracking process recognition device of the present application embodiment 2;

[0041] Figure 3 It is a structural schematic diagram of the hydrogen induced cracking process recognition device of the present application embodiment 4. DETAILED DESCRIPTION

[0042] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings, but it should be understood that the scope of protection of the present application is not limited by the specific embodiments.

[0043] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme of the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application. Unless otherwise explicitly stated, in the entire specification and claims, the term "comprises" or its variants such as "includes" or "comprises" will be understood to include the stated elements or components, and will not exclude other elements or components.

[0044] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0045] In addition, numerous specific details are provided in the following detailed description to better illustrate the present invention. Those skilled in the art will appreciate that the present invention can be practiced without certain specific details. In some instances, methods, means, and components well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present invention.

[0046] Example 1

[0047] like Figure 1 As shown, Example 1 of the method for identifying the hydrogen-induced cracking process of the present invention specifically includes the following steps:

[0048] Step S101 involves monitoring the hydrogen-induced cracking process of a metal tensile specimen under constant load and temperature-controlled hydrogen charging conditions, collecting acoustic emission signals throughout the entire process. Specifically, a metal tensile specimen is prepared, and the experimental conditions meet the requirements of a constant load and temperature-controlled hydrogen charging environment. The process of hydrogen-induced cracking of the target material specimen is measured under different constant loads and high-temperature environments. A PCI-2 fully digital acoustic emission instrument is used to collect acoustic emission signals throughout the entire experimental process. The hydrogen-induced cracking process primarily includes three distinct stages: bubble generation and collapse during electrolytic hydrogen charging, surface cracking and detachment of the material, and hydrogen-induced crack growth.

[0049] Step S102: Record experimental data. Throughout the experiment, observe the acoustic emission correlation diagram obtained from the acoustic emission instrument. Visually observe the rise and collapse of bubbles in the environmental chamber and changes in the corrosion layer on the surface of the target material specimen. These changes are recorded. This experiment provides acoustic sample data for three different stages of the entire process of hydrogen-induced cracking of metal materials under different loading conditions, temperatures, and hydrogen concentrations.

[0050] Step S103 extracts time-domain parameters from the collected acoustic emission signal and calculates wavelet energy spectrum coefficients. Specifically, the time-domain parameters of the acoustic emission signal are extracted from the acoustic emission correlation diagram in step S102, primarily including five characteristic parameters: rise time, amplitude, energy, count, and duration. Simultaneously, the experimentally acquired acoustic emission signal undergoes db wavelet decomposition and reconstruction to obtain a reconstructed acoustic emission waveform signal and spectrum. Wavelet energy spectrum coefficients are calculated for typical signals of three types of acoustic emission sources (i.e., the three stages of the entire hydrogen-induced cracking process). These six characteristic parameters are normalized using mean-variance normalization.

[0051] The calculation of the wavelet energy spectrum coefficient can specifically include the following steps: first, calculating a wavelet decomposition scale for determining the number of wavelet decomposition levels; second, calculating the energy value in each level of the wavelet function and the total energy value according to the number of wavelet decomposition levels; and calculating the wavelet energy spectrum coefficient by the energy value in each level and the total energy value. The calculation formula of the wavelet decomposition scale can be specifically expressed as:

[0052]

[0053] wherein f s is the sampling frequency, L f is the filter length, and N is the sampling length.

[0054] The wavelet function can be specifically expressed as:

[0055] f(t) = f 0 (t) + f 1 (t) + … + f j (t) Formula (2).

[0056] wherein f 0 (t), f 1 (t), …, f j (t) are each level of the decomposed signal.

[0057] The energy in each level of the wavelet function is:

[0058] The total energy is expressed as:

[0059] The calculation formula of the wavelet energy spectrum coefficient in each level under the wavelet scale can be expressed as:

[0060] R (j) (t) = E (j) (t) / E (T) (t) Formula (5).

[0061] As can be seen from Formula (5), the wavelet energy spectrum coefficient can be calculated by the ratio of the energy value in each level of the wavelet function to the total energy value. The wavelet energy spectrum coefficient obtained by calculation is used as the acoustic emission signal characteristic parameter for identifying the hydrogen-induced cracking damage mode of the material, and together with the other five time domain parameters, it can be used as the input parameter for subsequent cluster analysis. The above six characteristic parameters need to be normalized before cluster analysis. In this embodiment, the mean-variance standardization method is used for the normalization of the acoustic emission signal characteristic parameters, and the data is specifically transformed into a standard normal distribution with a mean of 0 and a standard deviation of 1. That is, for a sample with n acoustic emission signals, the calculation of the mean-variance standardization can be performed by the following formula:

[0062]

[0063] where x i is the original signal parameter; is the mean of the original signal parameter; x′ i is the normalized signal parameter; σ is the standard deviation.

[0064] In step S104, the six characteristic parameters after normalization are taken as input variables for cluster analysis, so as to obtain the cluster categories corresponding to the three stages of bubble generation and rupture, material surface cracking and separation, and growth of hydrogen-induced cracks. Specifically, the cluster analysis can be based on k-means algorithm for iterative calculation, and the calculation process can specifically include:

[0065] First, the time domain parameters and wavelet energy spectrum coefficients after normalization are taken as inputs of k-means, and the k-means algorithm iterative process and sample division can use weighted Euclidean distance to classify the acoustic emission signals of the three stages, so as to obtain the cluster centers corresponding to the three stages. The acoustic emission signal cluster analysis method based on unsupervised pattern recognition adopted in this embodiment is to select characteristic parameters, and according to a certain similarity measurement method and clustering algorithm between parameters, the acoustic emission signals with similar categories are divided into a category, under the condition that the damage mode and each signal category are unknown. Specifically, the number of categories k is defined in advance, and k samples are selected as initial cluster centers; the distance between each input vector and the cluster center is calculated, the input vector is assigned to the category with the minimum distance to the cluster center, an initial classification scheme is obtained, and the cluster center is recalculated according to the mean of each category; the samples are reclassified according to the new cluster center; the calculation and reclassification process is repeated, and when the cluster center converges, the clustering ends.

[0066] In the process of obtaining the cluster centers corresponding to the three stages, the condition for determining the validity of clustering is that the Euclidean distance as the similarity measure between acoustic emission signals is greater than a preset threshold; the similarity measure formula is:

[0067]

[0068] where a(l) represents the average distance between the lth point and other points in the same category; b(l,k) is a vector representing the average distance between the 1st point and points in different categories. After iterative calculation, when the value of s satisfies the condition of being greater than the preset threshold, the clustering is determined to be valid.

[0069] Secondly, the edge data of each cluster center is enhanced by a compensation algorithm to obtain the cluster category with clear boundaries. Because the signal classification of the edges between different stages of cluster centers and cluster centers is not clear and the features are not obvious, it is easy to be blurred in the classification process, thereby leading to misidentification. Therefore, the edge data of each cluster center is enhanced, which can make the classification boundary more clear. In this embodiment, a compensation algorithm for edge data is used, that is, the cluster center generated by the k-means algorithm clustering is selected as an anchor point, the cluster center is taken as the center of a circle, the maximum value of the Euclidean distance is taken as the radius, the distance is segmented, the cluster signal under a classification mode is divided into several regions, different degrees of compensation are performed on each different region, the corresponding compensation coefficient is given based on the dichotomy, so that the distinctness of the source feature (i.e. edge data) of each region and the cluster center is the same, thereby weakening the unclear status of the edge data and improving the accuracy of identification. Specifically, the maximum distance x can be segmented based on the dichotomy, and the dichotomy is divided into n segments, the iteration number of the dichotomy is n-1, n regions can be divided in turn, and the inverse of the distance length of each region is compared, and the weight of each proportion is calculated according to the comparison value. The weight is the final compensation coefficient.

[0070] In step S105, the time domain parameters and wavelet energy spectrum coefficients of different regions in the cluster category are taken as inputs, and the three stages of acoustic emission sources are taken as outputs to train the hydrogen-induced cracking process identification model. The hydrogen-induced cracking process identification model of this embodiment preferably adopts an algorithm model based on BP neural network. Specifically, the six characteristic parameters of the rise time, amplitude, energy, count, duration of the acoustic emission signal extracted in the foregoing and the calculated wavelet energy spectrum coefficients are taken as the input layer of the hydrogen-induced cracking process identification model, and the three different types of acoustic emission sources are taken as the output layer to establish a three-layer BP neural network composed of an input layer, a hidden layer and an output layer. The compensation coefficient calculated in step S104 is added to the algorithm of the connection weight of the model, the training sample set and the test sample set of the model are established by using the typical data obtained in the experiment, and the model is trained and tested.

[0071] In step S106, the three stages of the hydrogen-induced cracking process corresponding to different acoustic emission signals are identified by the trained hydrogen-induced cracking process identification model. An intelligent identification method is provided for analyzing the hydrogen-induced cracking damage process, and a theoretical basis is provided for the on-site monitoring of hydrogen-induced cracking of in-service hydrogen pressure vessels.

[0072] The application provides direct basis for damage evolution and failure mechanism research in the hydrogen-induced cracking process of materials, the hydrogen-induced cracking process recognition model established based on the BP neural network can effectively recognize the acoustic emission signals in the hydrogen-induced cracking process, so that the nature of the acoustic emission signal source is determined, that is, the failure mechanism of the material is judged, different stages of the hydrogen-induced cracking of the material can be distinguished, and real-time monitoring of the in-service hydrogen pressure container has great significance. The five time domain parameters of the acoustic emission signals and the wavelet energy spectrum coefficients calculated therefrom can more effectively represent the feature parameters of the hydrogen-induced cracking process recognition model, and the recognition accuracy is higher. The clustering analysis process adds a compensation algorithm, which can strengthen the data at the edges of each cluster center, so that the boundaries of the cluster categories are clearer, and errors in classification and recognition are avoided to the maximum extent.

[0073] Example 2

[0074] As shown in Figure 2 The hydrogen-induced cracking process recognition device of the embodiment is a device corresponding to the recognition method in Embodiment 1, that is, the method in Embodiment 1 is realized in the form of a virtual device, and each virtual module constituting the hydrogen-induced cracking process recognition device can be executed by an electronic device, such as a network device, a terminal device, or a server.

[0075] The hydrogen-induced cracking process recognition device provided in the embodiment mainly includes an acoustic emission signal acquisition module 201, a clustering analysis module 202, a model training module 203, and a process recognition module 204. The acoustic emission signal acquisition module 201 is used to acquire acoustic emission signals in the whole process when monitoring the hydrogen-induced cracking process of a metal tensile specimen under constant load and temperature-controlled hydrogen charging experimental conditions. The clustering analysis module 202 is used to perform clustering analysis on the time domain parameters extracted from the acoustic emission signals and the wavelet energy spectrum coefficients obtained by calculation as input variables, and obtain clustering categories corresponding to three stages of bubble generation and rupture, material surface cracking and separation, and hydrogen-induced crack growth, respectively. The model training module 203 is used to train the hydrogen-induced cracking process recognition model by taking the time domain parameters and the wavelet energy spectrum coefficients in different regions of the clustering categories as inputs and taking the acoustic emission sources in the three stages as outputs. The process recognition module 204 is used to recognize the three stages of the hydrogen-induced cracking process corresponding to different acoustic emission signals through the trained hydrogen-induced cracking process recognition model.

[0076] Example 3

[0077] The embodiment provides a memory, which can be a non-transient (non-volatile) computer storage medium, and the computer storage medium stores computer executable instructions. The computer executable instructions can execute each step of the hydrogen-induced cracking process recognition method in any method embodiment described above, and achieve the same technical effects.

[0078] Example 4

[0079] This embodiment provides a device for identifying a hydrogen-induced cracking process. The device includes a memory containing a corresponding computer program product. When the program instructions included in the computer program product are executed by a computer, the computer can execute the hydrogen-induced cracking process identification method described in the above aspects and achieve the same technical effects.

[0080] Figure 3 This is a schematic diagram of the hardware structure of the electronic device in this embodiment. Figure 3 As shown, the device includes one or more processors 610 and a memory 620. Taking one processor 610 as an example, the device may further include an input device 630 and an output device 640.

[0081] The processor 610, the memory 620, the input device 630 and the output device 640 may be connected via a bus or other means. Figure 3 The bus connection is taken as an example.

[0082] Memory 620, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs, and modules. Processor 610 executes the non-transitory software programs, instructions, and modules stored in memory 620 to execute various functional applications and data processing of the electronic device, thereby implementing the processing method of the above-mentioned method embodiment.

[0083] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data, etc. In addition, the memory 620 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely located relative to the processor 610, and these remote memories may be connected to the processing device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0084] The input device 630 can receive input digital or character information and generate signal input. The output device 640 can include a display device such as a display screen.

[0085] The one or more modules are stored in the memory 620 and, when executed by the one or more processors 610, perform a method of identifying a hydrogen induced cracking process of the present application. The product described above can perform the method provided by the embodiments of the present application, has the corresponding function modules and beneficial effects of performing the method. Technical details not described in detail in the embodiments can be seen from the method provided by the embodiments of the present application.

[0086] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for identifying hydrogen-induced cracking processes, characterized in that: The steps include: A. When monitoring the hydrogen-induced cracking process of metal tensile specimens under constant load and temperature-controlled hydrogen charging experimental conditions, collect the acoustic emission signals of the entire process; B. performing cluster analysis using the time domain parameters extracted from the acoustic emission signal and the calculated wavelet energy spectrum coefficients as input variables to obtain cluster categories corresponding to the three stages of bubble generation and bursting, material surface cracking and detachment, and hydrogen-induced crack growth; The cluster analysis is based on an iterative calculation of the k-means algorithm, specifically including: taking the normalized time domain parameters and wavelet energy spectrum coefficients as the input of the k-means, using the weighted Euclidean distance to classify the acoustic emission signals of the three stages, and obtaining the cluster centers corresponding to the three stages respectively; strengthening the marginalized data of each cluster center by a compensation algorithm to obtain the cluster category with clear boundaries; the strengthening of the marginalized data by the compensation algorithm is specifically as follows: selecting the cluster center generated by clustering as the anchor point, taking the cluster center as the center of the circle, and taking the maximum value of the Euclidean distance as the radius, segmenting the Euclidean distance, dividing the cluster signal under a classification mode into multiple regions, performing different degrees of compensation on each different region, and obtaining the corresponding compensation coefficient based on the dichotomy method, so that the marginalized data of each region is equal to the cluster center; C. Training a hydrogen-induced cracking process identification model using the time domain parameters and wavelet energy spectrum coefficients of different regions in the cluster category as input and the acoustic emission sources of the three stages as output; D. Identifying the three stages of the hydrogen-induced cracking process corresponding to different acoustic emission signals through the trained hydrogen-induced cracking process identification model.

2. The method for identifying hydrogen-induced cracking according to claim 1, wherein: The time domain parameters include: rise time, amplitude, energy, count and duration of the acoustic emission signal.

3. The method for identifying hydrogen-induced cracking according to claim 1, wherein: The calculation of the wavelet energy spectrum coefficient specifically includes: Calculate the wavelet decomposition scale to determine the order of wavelet decomposition; Calculating the energy value at each level of the waveform function and the total energy value according to the level of the wavelet decomposition; The wavelet energy spectrum coefficients are calculated according to the energy values ​​in each level and the total energy value.

4. The method for identifying hydrogen-induced cracking according to claim 3, wherein: The calculation formula of the wavelet decomposition scale is: Formula (1); Among them, f s is the sampling frequency, L f is the filter length, N is the sampling length; The waveform function is specifically expressed as: Formula (2); in, To decompose the signal at all levels; The energy in each stage is: Formula (3); The total energy is: Formula (4); The calculation formula of the wavelet energy spectrum coefficient at each level under the wavelet scale is: Formula (5).

5. The method for identifying hydrogen-induced cracking according to claim 1, wherein: In the process of obtaining the cluster centers corresponding to the three stages, the condition for determining whether the cluster is valid is that the value of the Euclidean distance as the similarity measure between acoustic emission signals is greater than a preset threshold; the similarity measure formula is: Formula (6): in, Indicates the l The average distance between a point and other points of the same type; is a vector representing the 1 The average distance between a point and points in different categories.

6. The method for identifying hydrogen-induced cracking according to claim 1, wherein: The hydrogen-induced cracking process identification model in step C is an algorithm model based on BP neural network.

7. The method for identifying hydrogen-induced cracking according to claim 1, wherein: During the process of establishing the hydrogen-induced cracking process identification model, the compensation coefficient is added to the algorithm of the connection weight of the hydrogen-induced cracking process identification model.

8. A device for identifying hydrogen-induced cracking, characterized in that: include: Acoustic emission signal acquisition module, which is used to collect acoustic emission signals during the entire process of monitoring hydrogen-induced cracking of metal tensile specimens under constant load and temperature-controlled hydrogen charging experimental conditions; A cluster analysis module is used to perform cluster analysis using the time domain parameters extracted from the acoustic emission signal and the calculated wavelet energy spectrum coefficients as input variables to obtain cluster categories corresponding to the three stages of bubble generation and rupture, material surface cracking and detachment, and hydrogen-induced crack growth; The cluster analysis is based on an iterative calculation of the k-means algorithm, specifically including: taking the normalized time domain parameters and wavelet energy spectrum coefficients as the input of the k-means, using the weighted Euclidean distance to classify the acoustic emission signals of the three stages, and obtaining the cluster centers corresponding to the three stages respectively; strengthening the marginalized data of each cluster center by a compensation algorithm to obtain the cluster category with clear boundaries; the strengthening of the marginalized data by the compensation algorithm is specifically as follows: selecting the cluster center generated by clustering as the anchor point, taking the cluster center as the center of the circle, and taking the maximum value of the Euclidean distance as the radius, segmenting the Euclidean distance, dividing the cluster signal under a classification mode into multiple regions, performing different degrees of compensation on each different region, and obtaining the corresponding compensation coefficient based on the dichotomy method, so that the marginalized data of each region is equal to the cluster center; a model training module, which trains a hydrogen-induced cracking process identification model by taking the time domain parameters and wavelet energy spectrum coefficients of different regions in the cluster category as input and the acoustic emission sources of the three stages as output; A process identification module is used to identify the three stages of the hydrogen-induced cracking process corresponding to different acoustic emission signals through the trained hydrogen-induced cracking process identification model.

9. A memory, characterized in that: The method comprises an instruction set, wherein the instruction set is suitable for a processor to execute the steps of the method for identifying a hydrogen-induced cracking process according to any one of claims 1 to 7.

10. A device for identifying hydrogen-induced cracking processes, characterized in that: comprising a bus, an input device, an output device, a processor and a memory as claimed in claim 9; The bus is used to connect the memory, input device, output device and processor; The input device and the output device are used to implement interaction with the user; The processor is configured to execute an instruction set in the memory.

Images