Memory, metal pitting process identification method, device and equipment

Abstract

This invention discloses a method and apparatus for identifying the pitting corrosion process in metals. The method includes the following steps: A. Acquiring surface images of pitted pits during the pitting corrosion process of a specimen to obtain the surface deformation and strain distribution of the metal pitting corrosion process, and simultaneously acquiring corresponding acoustic emission signals; B. Identifying the images of surface deformation and strain distribution; C. Based on the identified images, segmenting the synchronously acquired acoustic emission signals into stages, and corroding the acoustic emission signals of the three stages of bubble generation, film rupture, and pit growth with the images; D. Extracting characteristic parameters of the acoustic emission signals under different temperatures, corrosion intensities, and stress levels, using them as input variables for a pitting corrosion acoustic emission identification neural network model and training the model; and identifying the metal pitting corrosion process through the acoustic emission output variables of the model for bubble generation, film rupture, and pit growth. This invention can simultaneously acquire surface image data of pitted pits and acoustic emission signals, and by analyzing the acoustic emission signals, can provide timely identification and early warning when metal materials are damaged by pitting corrosion.

Description

Technical Field

[0001] This invention relates to the field of industrial nondestructive testing technology, and in particular to a method and apparatus for identifying metal pitting processes. Background Technology

[0002] With the continuous improvement of the industrial system, industrial production has driven the ever-increasing demand for industrial equipment. In many heavy industry enterprises, industrial equipment and pipelines often operate in high-temperature and highly corrosive environments, inevitably leading to various safety hazards, the most common of which is pitting corrosion. While images of pitting corrosion can provide a direct view of its state, images of pitting corrosion on operating equipment are difficult to obtain. Acoustic emission methods can monitor pitting corrosion on operating equipment and pipelines; however, the acquired acoustic emission signals cannot yet be correlated with different stages of pitting corrosion, especially under the combined effects of high temperature and stress. Further research is needed to understand how the initiation and propagation of pitting corrosion under material tension and the changes and interrelationships of acoustic emission signals occur.

[0003] Chinese patent application CN108088746A discloses a method for testing and analyzing the deformation mechanics and acoustic emission characteristics of metals under combined tension and torsion. This method includes the design of a testing system, a description of the signal features to be extracted, the calculation and derivation process of the failure stress and failure surface direction of different material specimens under tension and torsion, and a method for determining the material failure type and failure stress. The method includes the following steps: loading and acquiring mechanical signals such as axial force, torque, axial deformation, and torsion angle of the specimen using a tension-torsion electronic testing machine system; acquiring acoustic emission signals using an acoustic emission instrument and acoustic sensors; deriving the calculation formula for the combined tension-torsion failure stress and its direction based on the theory of mechanics of materials; and using the acquired signals and features to determine the failure type and calculate the failure stress and failure surface direction of the specimen. This prior art can obtain and compare the combined tension-torsion failure mechanics and acoustic emission characteristics, failure type, failure stress, and failure surface of different metallic materials under different loading rates.

[0004] Chinese patent application CN111398057A discloses a method for calculating the stress intensity factor of cracks in heterogeneous materials using DIC technology. The method includes: preparing standard compression or tension specimens of the heterogeneous material and subjecting them to uniaxial compression or tension to obtain the elastic modulus and Poisson's ratio; preparing standard fracture specimens of the heterogeneous material with cracks and preparing speckle patterns for DIC testing on the surface of the area surrounding the cracks; conducting fracture experiments on the specimens to obtain the displacement field and crack size in the region near the crack tip under different loads; selecting different integration paths, dividing the integration region into sub-elements, and calculating the J-integral value of each sub-element; filtering the J-integral values ​​and then superimposing them to obtain the J-integral values ​​on the integration path, converting the J-integral values ​​into a stress intensity factor K; repeating the above steps to calculate the J-integral values ​​under different loads / displacements to obtain the stress intensity factor K at each loading moment.

[0005] No significant technological advancements have yet been found that can acquire acoustic emission signals and surface image information of metals under different temperatures, corrosion intensities, stress magnitudes, and pit shapes during the pitting corrosion propagation process, using both imaging and acoustic methods, to analyze the relationship between the propagation patterns of pitting damage and acoustic emission signals. Therefore, it is necessary to design an in-situ visualized, multi-field coupled acoustic signal analysis method for metal pitting corrosion to study the macroscopic and microscopic mechanisms of pitting corrosion under high-temperature tensile conditions. This will provide a deeper understanding of the relationship between pitting corrosion propagation in metallic materials and various factors such as temperature, corrosion intensity, stress magnitude, and pit shape, offering a theoretical basis for real-time monitoring signal analysis of equipment operating under high-temperature tensile conditions, and proposing specific indicators for the safe operating range of high temperatures, corrosion intensities, and stress magnitudes experienced by equipment.

[0006] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide a method and apparatus for identifying the pitting process of metals. It can simultaneously collect surface image data and acoustic emission signals of pitting pits and analyze the signal characteristics of pitting expansion, so as to better analyze the acoustic emission signals of each stage of the pitting process and provide timely identification and early warning when metal materials are damaged by pitting.

[0008] To achieve the above objectives, according to a first aspect of the present invention, the present invention provides a method for identifying the pitting process of metal, comprising the following steps: A. Acquiring surface images of pits in the pitting process of a specimen to obtain the surface deformation and strain distribution of the metal pitting process, and simultaneously acquiring corresponding acoustic emission signals; B. Identifying the images of surface deformation and strain distribution; C. Based on the identified images, segmenting the synchronously acquired acoustic emission signals into stages, and corresponding the acoustic emission signals of the three stages of bubble generation, film rupture, and pit growth with the images; D. Extracting feature parameters of the acoustic emission signals under different temperatures, corrosion intensities, and stress levels, using them as input variables for a pitting acoustic emission identification neural network model and training the model; and identifying the metal pitting process through the acoustic emission output variables of the model for bubble generation, film rupture, and pit growth.

[0009] Furthermore, in the above technical solution, the specimen in step A can be a tensile specimen made of metal. The specimen has groove-like structures on multiple different surfaces. The groove-like structures are placed in different temperature and corrosion test environments to simulate the pitting process of the material in actual operation.

[0010] Furthermore, in the above technical solution, the image recognition in step B can adopt a two-dimensional convolutional neural network model. Through model training and testing, the image can be recognized as at least three categories of pitting corrosion phenomena, including bubble generation, film rupture, and pit growth.

[0011] Furthermore, in the above technical solution, step C, which involves segmenting the acquired acoustic emission signal in stages, can specifically include: calibrating the start time of acoustic emission signal acquisition and the start time of image acquisition; recording the times of bubble generation, film rupture, and pit growth based on the identified image, and segmenting the acoustic emission waveform between a first preset time before that time and a second preset time after that time; setting an amplitude threshold for the segmented acoustic emission waveform, and performing a second segmentation of the acoustic emission waveform based on the amplitude threshold. Both the first and second preset times can be set to 1 second.

[0012] Furthermore, in the above technical solution, the acoustic emission waveform signal after secondary interception corresponds to the corresponding images of bubble generation, film rupture, and pit growth.

[0013] Furthermore, the above technical solution may include a noise reduction step after the acquired acoustic emission signal is segmented in stages. The noise reduction process may specifically involve: performing a four-level decomposition of the original acoustic emission signal using the db4 wavelet function in the MATLAB wavelet toolbox, decomposing it into acoustic signals of different frequency bands; fitting the acoustic signals of different frequency bands and filtering out high-frequency noise interference; and reconstructing the decomposition function with high correlation to improve the signal-to-noise ratio.

[0014] Furthermore, in the above technical solution, the characteristic parameters of the acoustic emission signal in step D may include the amplitude, energy, duration, ring count, and derived ratios of rise time to amplitude, ring count to duration, and rise time to duration extracted from the acoustic emission signal after secondary interception.

[0015] Furthermore, in the above technical solution, the pitting corrosion acoustic emission recognition neural network model can be an algorithm model based on a BP neural network. This model adopts a single hidden layer topology, and the connection relationship nlm between the input layer, hidden layer, and output layer of this topology can specifically be 7-6-3; the algorithm of the model can specifically include:

[0016]

[0017] Where g(x) is the activation function, which is the Sigmoid function;

[0018]

[0019] Among them, H j The output of the hidden layer; wij The weights from the input layer to the hidden layer; x i The feature parameters of the input layer; a j The bias from the input layer to the hidden layer;

[0020]

[0021] Among them, O k For the output of the output layer; w jk b represents the weights from the hidden layer to the output layer. k This is the bias from the hidden layer to the output layer; i = 1…n, j = 1…l, k = 1…m.

[0022] Furthermore, in the above technical solution, the error calculation in the model algorithm may specifically include:

[0023]

[0024] Where E is the error; Y k The expected output;

[0025]

[0026] Formula (5) is the weight update formula, which is used to correct the weights and minimize the error E.

[0027]

[0028] Formula (6) is the bias update formula, which is used to correct the bias and minimize the error E.

[0029] To achieve the above objectives, according to a second aspect of the present invention, a metal pitting process identification device is provided, comprising: an image acquisition module for acquiring images of pitted pit surfaces during the pitting process of a specimen, and obtaining surface deformation and strain distribution during the metal pitting process; an image recognition module for recognizing the images of surface deformation and strain distribution; an acoustic signal interception module for intercepting synchronously acquired acoustic emission signals in stages based on the recognized images, and corresponding the acoustic emission signals of the three stages of bubble generation, film rupture, and pit growth with the images; and a pitting process identification module for extracting characteristic parameters of acoustic emission signals under different temperatures, corrosion intensities, and stress levels as input variables of a pitting acoustic emission identification neural network model; and identifying the metal pitting process through the acoustic emission output variables of the model for bubble generation, film rupture, and pit growth.

[0030] Furthermore, in the above technical solution, the pitting corrosion process identification module may include: a feature extraction submodule, which is used to extract feature parameters of acoustic emission signals under different temperatures, corrosion intensities, and stress magnitudes; a sample training submodule, which is used to input the feature parameters of the acoustic emission signals in the sample database into the pitting corrosion acoustic emission identification neural network model for model training; and an identification submodule, which uses the trained model to identify the metal pitting corrosion process through the acoustic emission output variables of bubble generation, film rupture, and pit growth.

[0031] To achieve the above objectives, according to a third aspect of the present invention, a memory is provided, including an instruction set adapted for a processor to execute steps as described in the aforementioned method for identifying metal pitting processes.

[0032] To achieve the above objectives, according to a fourth aspect of the present invention, a metal pitting process identification device is provided, 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 enable interaction with a user; and the processor is used to execute the instruction set in the memory.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1) The method of the present invention can be applied to damage monitoring and analysis during the operation of high-temperature equipment and pipelines;

[0035] 2) By analyzing the acoustic emission signals at each stage of the pitting process, timely identification and early warning can be provided for equipment when it suffers pitting damage in actual operation;

[0036] 3) This invention utilizes acoustic emission technology and microscopic image acquisition technology to simultaneously acquire, extract, and correlate acoustic emission signals and image information during the initiation and propagation of metal pitting corrosion under high temperature stress environment, thereby distinguishing the acoustic emission signals at each stage of pitting corrosion.

[0037] 4) After the acoustic emission signal is intercepted, wavelet denoising can be used to reduce the noise of the acoustic signal;

[0038] 5) This invention can collect pitting acoustic emission signals under different stresses, temperatures, pitting pit shapes, and corrosion intensities, establish a sample database, and, through the training and testing of an acoustic emission recognition neural network, can more effectively identify pitting acoustic emission signals, thereby solving the problem of pitting signals not being processed and warned in a timely manner during actual equipment operation.

[0039] Other features and aspects of the invention will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0040] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0041] Figure 1 This is a flowchart illustrating the entire process of experimentation, signal acquisition and processing, and signal recognition involved in this invention.

[0042] Figure 2 This is a schematic diagram of the metal pitting process identification method of Embodiment 1 of the present invention;

[0043] Figure 3 This is a schematic diagram showing the corresponding two-dimensional convolutional image recognition and pitting acoustic emission signal extraction in Embodiment 1 of the present invention;

[0044] Figure 4 This is a schematic diagram of the metal pitting process identification device according to Embodiment 2 of the present invention;

[0045] Figure 5 This is a schematic diagram of the metal pitting process identification device according to Embodiment 4 of the present invention. Detailed Implementation

[0046] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific embodiments.

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprising of," etc., will be understood to include the stated elements or components, and does not exclude other elements or other components.

[0048] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0049] Furthermore, to better illustrate the present invention, numerous specific details are provided in the following detailed embodiments. Those skilled in the art should understand that the present invention can be practiced even without certain specific details. In some instances, methods, means, and elements well-known to those skilled in the art have not been described in detail, in order to highlight the spirit of the present invention.

[0050] This invention combines digital image correlation technology with acoustic emission technology to analyze the pitting corrosion propagation phenomenon in metals operating under conditions where direct image observation is not possible. The method includes extracting acoustic emission signals from pitting corrosion, denoising these signals, extracting data from microscopic images, and establishing correlations between acoustic, optical, and mechanical signals. This allows for a deeper understanding of the relationship between metal pitting corrosion propagation and acoustic emission signals under different temperatures, corrosion intensities, pit shapes, and stress levels, providing a basis for monitoring equipment operation in high-temperature pitting corrosion environments. The entire experimental process, signal acquisition and processing, and signal recognition can be referenced. Figure 1 .

[0051] Example 1

[0052] like Figure 2 As shown, Embodiment 1 of the metal pitting corrosion process identification method of the present invention can accurately identify the metal pitting corrosion process by simultaneously acquiring and analyzing the signal characteristics of crack propagation through microscopic images and acoustic emission signals. Specifically, it includes the following steps:

[0053] Step S101: Acquire surface images of pitting pits during the pitting process of the specimen to obtain the surface deformation and strain distribution of the metal pitting expansion. Specifically, first, prepare a tensile specimen made of metal, and prefabricate four groove-shaped structures of different shapes at the center of the four surfaces of the tensile specimen. Treat the parts other than the groove-shaped structures with high-temperature corrosion resistance to effectively observe the metal pitting process at the groove-shaped structures. Second, place the specimen on an in-situ tensile testing machine and clamp it with tensile fixtures. Then, use electrochemical corrosion to corrode the groove-shaped structures on the surface of the specimen to simulate the pitting behavior of metal materials. The corrosion solution can be changed to alter the corrosion intensity. Finally, use an existing high-frequency induction heater to heat the specimen, simulating different temperature environments during material operation.

[0054] Tensile tests were conducted on the specimens, with stress applied slowly. The corrosion rate increased with increasing load, and the number of pits and cracks on the alloy surface increased after corrosion. An industrial camera was used to capture images of material deformation on the pitted surface during the dynamic experiment, obtaining information on surface deformation and strain distribution during pitting expansion. Simultaneously, an acoustic emission detection system was used to acquire acoustic emission signals from the pitting process. This system, using a piezoelectric sensor, converts the mechanical waves emitted by microscopic deformation of the material surface and the generation of bubbles, film rupture, and pit growth during corrosion into electrical signals, thus providing a visual representation of the acoustic emission signals in waveform form.

[0055] Step S101 creates pitting corrosion conditions for specimens under different environments, thereby obtaining microscopic images and acoustic emission signal data of pitting corrosion under different temperatures, different corrosive liquid concentrations, and different stress intensities. This provides richer sample data for establishing a neural network model for pitting corrosion acoustic emission recognition in subsequent steps, thereby improving the accuracy of the recognition model.

[0056] Step S102 involves recognizing the images of surface deformation and strain distribution obtained in step S101. Specifically, the microscopic images of the pitting surface acquired in real time by an industrial camera are input into a two-dimensional convolutional neural network. Preferably, but not limitingly, during the continuous pitting process, 500 microscopic images each of three scenarios (bubble generation, film rupture, and pit growth) can be selected and input into the input layer of the two-dimensional convolutional neural network for training, establishing an image recognition model for each stage of pitting. After model training and testing, pitting experiments are conducted again. At this point, the image recognition model can identify the three different pitting phenomena in the pitting process, that is, it can be identified as including at least three pitting phenomenon categories: bubble generation, film rupture, and pit growth.

[0057] Step S103, as follows Figure 3 As shown, based on the image identified in step S102, the acquired acoustic emission signals are segmented into stages, and the acoustic emission signals of the three stages—bubble generation, film rupture, and pit growth—are correlated with the corresponding identified images. Specifically, when the image recognition model based on a two-dimensional convolutional neural network identifies a certain pitting corrosion phenomenon (i.e., bubble generation, film rupture, or pit growth), that is... Figure 3 The microscopic image on the left side of the image captures a segment of acoustic emission signal at the moment the phenomenon occurs. This allows us to correlate the images of the three characteristics of bubble generation, film rupture, and pit growth with the acoustic emission signals, and establish separate sample databases for the three cases. This achieves the goal of extracting the real pitting acoustic emission signals for each stage.

[0058] To achieve accurate correspondence between the image and the acoustic emission signal, preferably but not limitingly, the acoustic emission signal interception in this embodiment can be carried out in the following manner: First, the start time of the acoustic emission signal acquisition and the start time of the image acquisition need to be calibrated to ensure that the time of the image and the acoustic emission signal acquired separately are completely synchronized; Second, based on the identified image, the times when bubble generation, film rupture and pit growth occur are recorded, and the acoustic emission waveform between the first preset time before the time and the second preset time after the time is intercepted. Specifically, when the convolutional neural network identifies a certain pitting corrosion phenomenon (bubble formation, film rupture, or pit growth), it records the moment the phenomenon occurs. Considering the delay in signal transmission and the duration of the signal itself, the acoustic emission waveform is captured in the acoustic emission acquisition system for one second before (i.e., the first preset time) and one second after (i.e., the second preset time), totaling two seconds. This acoustic emission waveform is used as an initial pitting corrosion phenomenon waveform data file. A fixed amplitude threshold (i.e., waveform amplitude threshold) is set for the captured acoustic emission waveform. Based on this amplitude threshold, the acoustic emission waveform is captured a second time to finally obtain the complete waveform of the acoustic emission signal corresponding to a certain pitting corrosion phenomenon. At this point, the association between the pitting corrosion microscopic image and the acoustic emission signal is completed, and the complete waveform of the acoustic emission signal can be exported and saved in the corresponding sample database.

[0059] It should be noted that the acoustic emission signal may contain a lot of noise unrelated to pitting corrosion. To remove this noise as much as possible, a noise reduction step can be included after the acquired acoustic emission signal is segmented in stages. This noise reduction step can be done as follows: First, the original pitting corrosion acoustic signal is decomposed into four levels using the db4 wavelet function in the MATLAB wavelet toolbox, decomposing the original signal into four signals of different frequency bands, such as a1, d1, d2, and d3. Second, the four decomposed signals are processed by setting fitting coefficients. The maximum coefficient is selected in the toolbox to filter out all signals in the d1, d2, and d3 frequency bands, removing high-frequency noise interference. Finally, the decomposition function with high correlation is reconstructed to reconstruct a new pitting corrosion signal, thereby improving the signal-to-noise ratio.

[0060] Step S104: Extract feature parameters of acoustic emission signals under different temperatures, corrosion intensities, and stress levels from the sample database, and use them as input variables for the pitting corrosion acoustic emission recognition neural network model. Specifically, these feature parameters are extracted from the aforementioned wavelet-denoised acoustic emission signals, including but not limited to: four feature parameters of acoustic emission signal amplitude, energy, duration, and ring count, and three derived acoustic emission feature parameters: the ratio of rise time to amplitude, the ratio of ring count to duration, and the ratio of rise time to duration. Use these seven feature parameters as the input layer of the acoustic emission recognition model, and the three types of pitting corrosion acoustic emission sources (i.e., bubble generation, film rupture, and pit growth) as the output layer to establish a single hidden layer acoustic emission recognition neural network model.

[0061] Step S105: Input the data from the sample database into the acoustic emission recognition model for network model training and testing. This acoustic emission recognition neural network model is a BP neural network-based algorithm model. This model can adopt a single hidden layer topology, and the connection relationship (nlm) between the input layer, hidden layer, and output layer of this topology is specifically 7-6-3. The specific algorithm of the model is as follows:

[0062]

[0063] Where g(x) is the activation function, which is the Sigmoid function;

[0064]

[0065] Among them, H j The output of the hidden layer; w ij The weights from the input layer to the hidden layer; x i These are the feature parameters of the input layer (i.e., the seven feature parameters in step S104); a j The bias from the input layer to the hidden layer;

[0066]

[0067] Among them, O k The output of the output layer (i.e., the three types of pitting acoustic emission sources: bubble generation, film rupture, and pit growth); w jk b represents the weights from the hidden layer to the output layer. k This is the bias from the hidden layer to the output layer; i = 1…n, j = 1…l, k = 1…m.

[0068] Preferably, but not restrictively, in order to gradually reduce errors during model training and bring the actual output as close as possible to the desired output, error calculation in the model algorithm can be performed in the following way:

[0069]

[0070] Where E is the error; Y k The expected output;

[0071]

[0072] Formula (5) is the weight update formula, which is used to correct the weights and minimize the error function E during the backpropagation of the error.

[0073]

[0074] Formula (6) is the bias update formula, which is used to correct the bias and minimize the error function E during the backpropagation of the error.

[0075] Step S106: After model training and testing, the metal pitting corrosion process can be actually identified by the acoustic emission output variables of bubble generation, film rupture and pit growth in the model, thereby realizing signal identification and early warning at each stage of the pitting corrosion process.

[0076] The method of this invention is applicable to damage monitoring and analysis during the operation of high-temperature equipment and pipelines. The metal pitting process identification method provided in Embodiment 1 of this invention can better analyze the acoustic emission signals at each stage of the pitting process, enabling timely identification and early warning of pitting damage to equipment and pipelines during actual operation. This invention utilizes acoustic emission technology and microscopic image acquisition technology to extract and correlate acoustic emission signals and image information during the initiation and propagation of metal pitting under high-temperature stress environments. It distinguishes the acoustic emission signals at each stage of pitting and collects pitting acoustic emission signals under different stresses, temperatures, pit shapes, and corrosion intensities to establish a sample database. Through training and testing of an acoustic emission recognition neural network, it can effectively identify pitting acoustic emission signals. This solves the problem of the inability to process and warn of pitting signals in a timely manner during actual equipment operation.

[0077] Example 2

[0078] like Figure 4 As shown, the metal pitting process identification device in this embodiment is a device corresponding to the method in Embodiment 1. That is, the method in Embodiment 1 is implemented by means of a virtual device. Each virtual module constituting the metal pitting process identification device can be executed by an electronic device, such as a network device, a terminal device, or a server.

[0079] The metal pitting corrosion process identification device provided in this embodiment specifically includes: an image acquisition module 201, an image recognition module 202, an acoustic signal interception module 203, and a pitting corrosion process identification module 204. The image acquisition module 201 processes images of pitted pits collected during the pitting corrosion process of the specimen to obtain images of the three stages of the metal pitting corrosion process, as well as the corresponding surface deformation and strain distribution. The image recognition module 202 identifies the images of surface deformation and strain distribution. The acoustic signal interception module 203, based on the identified images, intercepts the collected acoustic emission signals in stages, corresponding the acoustic emission signals of the three stages—bubble generation, film rupture, and pit growth—to the images. The pitting corrosion process identification module 204 extracts characteristic parameters of the acoustic emission signals under different temperatures, corrosion intensities, and stress levels as input variables for the pitting corrosion acoustic emission identification neural network model; and identifies the metal pitting corrosion process through the acoustic emission output variables of the model for bubble generation, film rupture, and pit growth.

[0080] Furthermore, the pitting corrosion process identification module 204 in the metal pitting corrosion process identification device of this embodiment may specifically include: a feature extraction submodule, a sample training submodule, and an identification submodule. The feature extraction submodule can be used to extract feature parameters of the acoustic emission signals under different temperatures, corrosion intensities, and stress levels; the sample training submodule can be used to input the feature parameters of the acoustic emission signals in the sample database into the pitting corrosion acoustic emission identification neural network model for model training; the identification submodule can use the trained model to identify the metal pitting corrosion process through the acoustic emission output variables of bubble generation, film rupture, and pit growth.

[0081] Example 3

[0082] This embodiment provides a memory, which can be a non-transitory (non-volatile) computer storage medium. The computer storage medium stores computer-executable instructions that can execute each step of the metal pitting process identification method in any of the above method embodiments and achieve the same technical effect.

[0083] Example 4

[0084] This embodiment provides a metal pitting process identification device. 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 metal pitting process identification method described in the above aspects and achieve the same technical effect.

[0085] Figure 5 This is a schematic diagram of the hardware structure of the electronic device in this embodiment, as shown below. Figure 5As shown, the device includes one or more processors 610 and a memory 620. Taking one processor 610 as an example, the device may also include an input device 630 and an output device 640.

[0086] The processor 610, memory 620, input device 630, and output device 640 can be connected via a bus or other means. Figure 3 Taking the example of a connection between China and Israel via a bus.

[0087] The 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. The processor 610 executes various functional applications and data processing of the electronic device by running the non-transitory software programs, instructions, and modules stored in the memory 620, thereby implementing the processing method of the above-described method embodiments.

[0088] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; the data storage area may store data, etc. Furthermore, the memory 620 may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory 620 may optionally include memory remotely located relative to the processor 610, and these remote memories may be connected to the processing device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0089] Input device 630 can receive input digital or character information and generate signal input. Output device 640 may include display devices such as a display screen.

[0090] The one or more modules are stored in the memory 620, and when executed by the one or more processors 610, they perform the metal pitting process identification method of the present invention. The above product can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying the pitting process of metals, characterized in that, Includes the following steps: A. Acquire images of the pitted surfaces of the specimen during the pitting process to obtain the surface deformation and strain distribution of the metal during the pitting process, and simultaneously acquire the corresponding acoustic emission signals. B. Identify the images of the surface deformation and strain distribution; C. Based on the identified image, the synchronously acquired acoustic emission signals are extracted in stages, and the acoustic emission signals of the three stages of bubble generation, film rupture and pit growth are correlated with the image. The process of segmenting the acquired acoustic emission signals in stages specifically includes calibrating the start time of the acoustic emission signal acquisition and the start time of the image acquisition. Based on the identified image, the times when the bubble generation, film rupture, and pit growth phenomena occur are recorded, and the acoustic emission waveform between a first preset time before the time and a second preset time after the time is extracted; an amplitude threshold is set for the extracted acoustic emission waveform, and the acoustic emission waveform is extracted a second time according to the amplitude threshold; D. Extract the characteristic parameters of the acoustic emission signal under different temperatures, corrosion intensities, and stress levels, and use them as input variables for the pitting corrosion acoustic emission recognition neural network model and train the model; identify the metal pitting corrosion process through the acoustic emission output variables of the model, such as bubble generation, film rupture, and pit growth; the characteristic parameters of the acoustic emission signal include the amplitude, energy, duration, ring count, and derived rise time to amplitude, ring count to duration, and rise time to duration ratio extracted from the acoustic emission signal after the second interception.

2. The method for identifying metal pitting corrosion process according to claim 1, characterized in that, The specimen in step A is a tensile specimen made of metal. The specimen has groove-like structures on multiple different surfaces. These groove-like structures are placed in different temperature and corrosion test environments to simulate the pitting process of the material in actual operation.

3. The method for identifying metal pitting corrosion process according to claim 1, characterized in that, The image recognition in step B uses a two-dimensional convolutional neural network model. Through model training and testing, the image is identified as including at least three pitting corrosion phenomena: bubble generation, film rupture, and pit growth.

4. The method for identifying metal pitting corrosion process according to claim 1, characterized in that, The first preset time and the second preset time are both 1 second.

5. The method for identifying metal pitting corrosion process according to claim 1, characterized in that, The acoustic emission waveform signal after secondary interception corresponds to the corresponding images of bubble generation, film rupture, and pit growth.

6. The method for identifying metal pitting corrosion process according to claim 1, characterized in that, The process of segmenting the acquired acoustic emission signal into stages also includes a noise reduction step.

7. The method for identifying metal pitting corrosion process according to claim 6, characterized in that, The noise reduction process specifically includes: The original acoustic emission signal was decomposed into four levels using the db4 wavelet function in the MATLAB wavelet toolbox, resulting in acoustic signals in different frequency bands. The acoustic signals of different frequency bands are fitted and high-frequency noise interference is filtered out. Reconstruction of decomposition functions with high correlation is performed to improve the signal-to-noise ratio.

8. The method for identifying metal pitting corrosion process according to claim 1, characterized in that, The described pitting corrosion acoustic emission recognition neural network model is an algorithm model based on a backpropagation (BP) neural network. This model adopts a single hidden layer topology, and the connection relationship between the input layer, hidden layer, and output layer of this topology is as follows: nlm Specifically, it is 7-6-3; the algorithm of the model specifically includes: Formula (1); Where g(x) is the activation function, which is the Sigmoid function; Formula (2); Among them, H j For the output of the hidden layer; w ij The weights from the input layer to the hidden layer; x i The feature parameters of the input layer; a j The bias from the input layer to the hidden layer; Formula (3); Among them, O k For the output of the output layer; w jk b represents the weights from the hidden layer to the output layer. k The bias from the hidden layer to the output layer; above .

9. The method for identifying metal pitting corrosion process according to claim 8, characterized in that, The error calculation in the model algorithm specifically includes: , remember Formula (4); Where E is the error; Y k The expected output; Formula (5); Formula (5) is a weight update formula, used to correct the weight and minimize the error E; Formula (6): Formula (6) is a bias update formula, which is used to correct the bias and minimize the error E.

10. A device for identifying the pitting process of metal, characterized in that, include: The image acquisition module is used to acquire images of the pitted pit surface of the specimen during the pitting process, and to obtain the surface deformation and strain distribution of the metal during the pitting process. An image recognition module is used to recognize images of the surface deformation and strain distribution. The acoustic signal interception module is used to intercept the synchronously acquired acoustic emission signals in stages according to the identified image, and to correlate the acoustic emission signals of the three stages of bubble generation, film rupture and pit growth with the image. The process of segmenting the acquired acoustic emission signals in stages specifically includes calibrating the start time of the acoustic emission signal acquisition and the start time of the image acquisition. Based on the identified image, the times when the bubble generation, film rupture, and pit growth phenomena occur are recorded, and the acoustic emission waveform between a first preset time before the time and a second preset time after the time is extracted; an amplitude threshold is set for the extracted acoustic emission waveform, and the acoustic emission waveform is extracted a second time according to the amplitude threshold; The pitting corrosion process identification module is used to extract characteristic parameters of the acoustic emission signals under different temperatures, corrosion intensities, and stress levels, as input variables of the pitting corrosion acoustic emission identification neural network model; the metal pitting corrosion process is identified by the acoustic emission output variables of the model, such as bubble generation, film rupture, and pit growth; the characteristic parameters of the acoustic emission signals include amplitude, energy, duration, ring count, and derived rise time to amplitude, ring count to duration, and rise time to duration ratios extracted from the acoustic emission signals after secondary interception.

11. The metal pitting process identification device according to claim 10, characterized in that, The pitting process identification module includes: The feature extraction submodule is used to extract feature parameters of the acoustic emission signal under different temperatures, corrosion intensities, and stress levels. The sample training submodule is used to input the feature parameters of the acoustic emission signals in the sample database into the pitting acoustic emission recognition neural network model for model training. The identification submodule uses the trained model to identify the metal pitting corrosion process through acoustic emission output variables of bubble generation, film rupture, and pit growth.

12. A memory, characterized in that, It includes an instruction set adapted for a processor to perform the steps in the metal pitting process identification method as described in any one of claims 1 to 9.

13. A device for identifying the pitting process of metal, characterized in that, It includes a bus, an input device, an output device, a processor, and a memory as described in claim 12; The bus is used to connect the memory, input device, output device, and processor; The input and output devices are used to enable interaction with the user; The processor is used to execute the instruction set in the memory.

Images