Surface property interference under heat-resistant steel aging detection method and storage medium

Abstract

The application discloses a heat-resistant steel aging detection method under surface characteristic interference and a storage medium. The method comprises the following steps: obtaining a sample with a treated surface, performing LIBS detection on the sample, obtaining sample spectrum data and constructing a data set, and establishing an aging detection model by using the data set; obtaining a to-be-detected object with an untreated surface, performing LIBS detection on a plurality of measurement points of the to-be-detected object, and obtaining laser spectrum tomography data; performing screening, feature selection and intensity correction processing on the tomography data to obtain corrected spectrum data; and performing aging detection on the corrected spectrum data by using the heat-resistant steel aging detection model. The application can effectively obtain the matrix representative spectrum information of the heat-resistant steel under the surface characteristic interference, overcome the negative influence of the surface characteristics and the recondensation of the surface layer material of the to-be-detected object on the aging detection, ensure the effectiveness of the spectrum data, realize the rapid in-situ detection of the heat-resistant steel, and can be applied to the detection object without pretreatment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of heat-resistant steel aging detection, and particularly relates to a heat-resistant steel aging detection method under surface characteristic interference and a storage medium. BACKGROUND

[0002] High-temperature pressure-bearing equipment is made of high-temperature pressure-bearing materials, and the equipment can work in a high-temperature and high-pressure environment for a long time and is widely used in the industrial industry. In order to ensure the safe and stable operation of such equipment, it is necessary to regularly detect whether the high-temperature pressure-bearing material of the equipment is invalid and monitor the working state thereof, and to perform risk assessment. Heat-resistant steel is the most common high-temperature pressure-bearing material. The heat-resistant steel characteristic detection method based on the laser-induced breakdown spectroscopy (LIBS) is generally as follows: a nanosecond laser is used to ablate a detection object to form a laser-induced plasma, and the relationship between the characteristic spectral information reflecting the matrix characteristics emitted in the evolution process of the plasma and the microstructure or mechanical properties of the heat-resistant steel is established, so as to realize the detection of the characteristics of the heat-resistant steel. However, due to the long-term high-temperature and high-pressure environment of the pressure-bearing equipment, the surface of the heat-resistant steel of the equipment often appears to be oxidized, corroded and slagged due to environmental factors, resulting in impurities such as slag mixture and oxide layer generated on the surface of the heat-resistant steel, so that the thickness of the surface layer is thickened, and it is impossible to ensure that the laser can completely break through the surface layer of the heat-resistant steel, thereby causing a certain degree of negative impact on the subsequent spectral collection step, and it is impossible to detect the effective matrix information of the monitoring object, thereby affecting the operation and maintenance monitoring work of the high-temperature pressure-bearing equipment.

[0003] In order to effectively reduce or eliminate the influence of the impurities in the surface layer, some existing technologies optimize the parameters such as laser energy and focal point position; in the aspect of data analysis, some existing technologies attempt to use spectral intensity normalization, multivariate scatter correction and other methods to improve the spectral quality, combine the spectral feature selection method, select the spectral feature variables with higher correlation with the matrix characteristics of the material, and use a machine learning algorithm to establish a heat-resistant steel characteristic analysis model, and then realize the characteristic analysis. These methods have obtained good accuracy and stability.

[0004] However, the prior art has the following defects: first, the detection object of the prior art is usually a heat-resistant pipe sample that has been fully polished and polished, and the sample completely removes the interference of the surface characteristics on the detection, so that the prior art cannot effectively obtain representative spectral information of the heat-resistant steel base under the influence of the surface state, thereby reducing the accuracy of the operation and maintenance monitoring of the high-temperature pressure-bearing equipment; second, the premise of spectral analysis is that the laser-induced breakdown completely breaks through the surface layer of the detection object, however, the prior art only determines the breakdown of the surface layer based on the intensity change of the typical characteristic spectrum, which cannot accurately determine the breakdown of the surface layer and lacks a fundamental basis; third, after obtaining the spectral data of the broken surface layer, the existing spectral data optimization method only corrects the spectral line noise and baseline drift caused by instrument fluctuations, environmental interference and surface state, and cannot correct the negative effects of pit depth variation and surface layer impurity recondensation on the spectrum. In actual industrial applications, whether it is remote online monitoring or portable rapid detection of high-temperature pressure-bearing materials, the detection object needs to be detected without pretreatment, and if a complicated pretreatment process is added during the detection process, the detection difficulty will be greatly increased and the detection efficiency will be reduced. It is impossible to realize real-time detection.

[0005] Therefore, how to overcome the negative effects of the impurities in the surface layer on the heat-resistant steel aging detection and realize in-situ detection of high-temperature pressure-bearing materials, reduce the detection difficulty and improve the detection efficiency have become problems to be solved in the field. SUMMARY

[0006] The purpose of the present application is to provide a heat-resistant steel aging detection method under the interference of surface characteristics and a storage medium to solve one or more technical problems in the prior art and at least provide a beneficial choice or create conditions.

[0007] The solution to the technical problem of the present application is: in a first aspect, the present application provides a heat-resistant steel aging detection method under the interference of surface characteristics, comprising the following steps:

[0008] A sample heat-resistant steel after surface pretreatment is obtained, LIBS technology is used to ablate and laser tomography the sample heat-resistant steel, sample spectral data and its corresponding aging grade label are obtained, a spectral feature data set is constructed, and a heat-resistant steel aging detection model is established using the spectral feature data set;

[0009] A to-be-detected heat-resistant steel without surface pretreatment is obtained, LIBS technology is used to ablate and laser tomography a plurality of measurement points of the to-be-detected heat-resistant steel, and laser spectral tomography data of a plurality of the measurement points are obtained;

[0010] The laser spectral tomography data are screened for representative spectra to obtain base spectral data;

[0011] performing feature selection and intensity correction on the base body spectral data to obtain modified spectral data;

[0012] performing aging detection on the modified spectral data by using the aging detection model of the heat-resistant steel to obtain an aging detection result.

[0013] As a further improvement of the above technical solution, the filtering of the laser spectrum tomography data to obtain base body spectral data comprises:

[0014] The filtering step is performed once for the laser spectrum tomography data of each measurement point:

[0015] According to the laser spectrum tomography data of the measurement point, the breakdown pulse number and the plasma temperature of the surface layer of the measurement point are calculated;

[0016] It is judged whether the breakdown pulse number of the measurement point is greater than a breakdown pulse reference value; if yes, the next step is entered; if no, the laser spectrum tomography data of the measurement point is discarded;

[0017] It is judged whether the difference between the plasma temperature of the measurement point and a preset plasma reference temperature is less than a first value; if yes, the laser spectrum tomography data of the measurement point is taken as base body measurement spectral data; if no, the laser spectrum tomography data of the measurement point is discarded;

[0018] When the filtering steps of all measurement points are completed, the base body measurement spectral data of all measurement points is output as base body spectral data.

[0019] As a further improvement of the above technical solution, the first value is 10%.

[0020] As a further improvement of the above technical solution, the calculation of the breakdown pulse number of the surface layer of the measurement point according to the laser spectrum tomography data of the measurement point comprises:

[0021] The laser spectrum tomography data of the measurement point is divided into several data groups, and the first mean data and the second mean data of each data group are calculated according to the laser spectrum tomography data; wherein the ablation time value of the first mean data and the second mean data of each data group is continuous, and the ablation time value is defined as the time value of ablation of the measurement point;

[0022] The first mean data is the mean value of the characteristic spectral line intensity of the surface layer element of the measurement point of the heat-resistant steel sample, and the second mean data is the mean value of the characteristic spectral line intensity of the base body element of the measurement point of the heat-resistant steel sample;

[0023] Determine whether the first mean data of the current data group is continuously increasing and whether the second mean data is continuously decreasing; if yes, analyze the stability of the characteristic spectral intensity of the surface layer element and the matrix element corresponding to the current data group and proceed to the next step; if no, skip the current data group and determine the next data group.

[0024] When the stability is greater than the stability threshold, calculate the first difference between two adjacent first mean data within the current data group, and calculate the second difference between two adjacent second mean data within the current data group.

[0025] When both the first difference and the second difference are less than the second value, the starting pulse number of the laser spectral tomography data of the current data group is taken as the breakdown pulse number of the surface layer at this measurement point.

[0026] As a further improvement to the above technical solution, the step of calculating the plasma temperature of the surface layer at the measurement point based on the laser spectral tomography data at the measurement point includes:

[0027] Based on the laser tomography data, the intensity of the plasma spectral line of the surface layer corresponding to the measurement point is calculated using the following formula:

[0028]

[0029] Among them, I ij λ represents the spectral line intensity of the laser tomography data. ij A is the wavelength of the spectral line. ij Let g be the transition probability from energy level i to energy level j. i E represents the statistical weight of the energy level. i For the transition to the upper energy level, k is the Boltzmann constant, T is the plasma temperature, N(T) is the density function, and U(T) is the partition function;

[0030] Several spectral lines of the same ionization energy level of the same element are selected as the analysis object, and the plasma spectral intensity of each line is used as the criterion. Using the vertical axis as the ordinate, the energy levels E of the transitions of each spectral line are... i Using the horizontal axis as the abscissa, a Boltzmann curve is obtained after fitting, and the plasma temperature of the surface layer is obtained from the Boltzmann curve.

[0031] As a further improvement to the above technical solution, the feature selection of the matrix spectral data includes: screening out spectral features within a 5nm range near six specific feature spectral lines in the matrix spectral data to obtain the screened matrix spectral data.

[0032] As a further improvement to the above technical solution, the six specific characteristic spectral lines are Na I 589.00nm, Na I 589.59nm, KI 766.53nm, KI 769.90nm, Ca I 393.37nm, Ca I 396.85nm, and Ca I 422.67nm.

[0033] As a further improvement to the above technical solution, the step of intensity correction of the matrix spectral data includes:

[0034] Calculate the average intensity of the data in the spectral feature dataset;

[0035] Based on the average intensity, the matrix spectral data after feature selection is corrected using the following formula:

[0036]

[0037] Where S1 represents the corrected spectral data, S0 represents the matrix spectral data after feature selection, I0 represents the spectral average intensity of the matrix spectral data after feature selection, and I mean This represents the average intensity of the data in the spectral feature dataset.

[0038] As a further improvement to the above technical solution, the step of establishing an aging detection model for heat-resistant steel using the spectral feature dataset includes:

[0039] The regression feature elimination method based on support vector machine is used to select features from the spectral feature dataset to obtain the optimal feature subset;

[0040] Determine the Gaussian kernel function as the kernel function for the support vector machine model, and establish the support vector machine model; wherein the Gaussian kernel function satisfies:

[0041] k(x i x j )=exp(-g||x i -x j ||);

[0042] Where, x i Let x be the number of spectra of the i-th sample in the spectral feature dataset. i Let g be the number of spectra of the j-th sample in the spectral feature dataset, and g be a constant used to determine the Gaussian distribution of the Gaussian kernel function;

[0043] A support vector machine model is trained using the optimal feature subset to obtain an aging detection model for heat-resistant steel.

[0044] In a second aspect, the present application also provides a storage medium, wherein the storage medium stores processor-executable instructions, and the processor-executable instructions, when executed by a processor, are configured to implement the surface characteristic interference-based aging detection method for heat-resistant steel.

[0045] The present application has the following beneficial effects: the surface characteristic interference-based aging detection method for heat-resistant steel and the storage medium are provided, the spectral evolution trend characteristics obtained by spectral tomography are combined with the surface layer and the matrix difference analysis to determine the spectral data affected by the surface characteristics, and the plasma temperature is combined to determine the matrix representative spectral information of the heat-resistant steel under the influence of the surface state, so as to ensure the acquisition of the matrix representative spectrum of the detection object; in addition, the present application adopts the targeted spectral feature selection method, reduces the influence of the recondensation of the surface layer material on the laser ablation and the spectral data, and combines the spectral intensity correction method to reduce the influence of the surface layer property difference and the environmental fluctuation on the identification degree and the process stability of the actual detection object, so as to ensure the effectiveness of the spectral data representing the matrix and the accuracy of the to-be-detected data source; the aging detection method of the present application can be applied to the detection object without pretreatment, can realize rapid in-situ detection, and can be coupled with a remote or portable detection device to complete the aging detection of the high-temperature pressure-resistant material.

[0046] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application can be realized and achieved by the structures specifically pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 A flowchart of the surface characteristic interference-based aging detection method for heat-resistant steel provided by the present application is shown in the figure;

[0048] Figure 2 A flowchart of the method for obtaining the matrix spectral data provided by the present application is shown in the figure;

[0049] Figure 3 A flowchart of the method for calculating the breakdown pulse number of the surface layer provided by the present application is shown in the figure;

[0050] Figure 4 A schematic diagram of the change of the spectral line intensity of the typical matrix element, alloy element and surface layer component of the heat-resistant steel with the pulse number provided by the present application is shown in the figure. DETAILED DESCRIPTION

[0051] In order to make the objects, technical solutions and advantages of the present application more clear, the present application is further described in detail below with reference to the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application, and are not used to limit the present application.

[0052] The application will be further described below in conjunction with the accompanying drawings and specific embodiments. The described embodiments should not be regarded as limiting the application, and all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the application.

[0053] In the following description, "some embodiments" are related to a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the application belongs. The terms used herein are only for the purpose of describing the embodiments of the application and are not intended to limit the application.

[0055] With the advancement of industrialization, high-temperature pressure equipment is widely used in power, metallurgy, power machinery, petroleum and chemical industry and other industries. In order to ensure the safe and stable operation of such equipment, it is necessary to regularly monitor the working state of high-temperature pressure materials and assess the risk. At present, the working state monitoring of high-temperature pressure materials is mainly realized by preventive planned maintenance, inspection of surface state and mechanical properties of the detected components during work, and evaluation of material failure state. The existing high-temperature pressure material failure detection methods mainly include non-destructive testing and destructive evaluation. The destructive evaluation method needs to cut the pipe for sampling, which affects the continuous operation of the equipment, the detection period is long, and the representativeness of the sampling needs to be improved. The traditional non-destructive testing method has the disadvantages of radiation risk, only macroscopic defects can be detected or the detection depth is limited.

[0056] Laser-induced breakdown spectroscopy (LIBS) is a new non-destructive testing method. By micro-ablation of the detection object, the breakdown spectroscopy information characterizing the matrix characteristics of the substance can be obtained, which can realize characteristic detection without damaging the detection object. Moreover, the measurement device of laser-induced breakdown spectroscopy technology can be integrated into a portable measurement device, which can realize remote measurement in extreme environments, thereby meeting the requirements of online monitoring of high-temperature pressure equipment and manual convenient detection during maintenance, and further ensuring the safe operation of high-temperature pressure major equipment. Laser-induced breakdown spectroscopy technology is widely used in the fields of high-temperature pressure material failure and aging detection.

[0057] Heat-resistant steel is a common high-temperature pressure-bearing material. The heat-resistant steel characteristic detection method based on LIBS technology is generally as follows: a laser-induced plasma is formed by ablation of a nanosecond laser on a detection object, and the relationship between the characteristic spectral information reflecting the matrix characteristics emitted during the evolution of the plasma and the microstructure or mechanical properties of the heat-resistant steel is established to realize the detection of the characteristics of the heat-resistant steel. However, the working environment of high-temperature pressure-bearing equipment is often very harsh, and the equipment is often in a high-temperature and high-pressure environment for a long time. The surface of the heat-resistant steel of the equipment will inevitably be oxidized, corroded and slagged due to environmental factors, resulting in the generation of impurities such as slag mixture and oxide layer on the surface of the metal pipeline. The chemical composition and physical properties of the impurities in the surface layer are often significantly different from those of the matrix of the heat-resistant steel. When detecting the aging of the heat-resistant steel of the equipment, laser-induced breakdown is required for the heat-resistant steel. Due to the presence of other impurities such as slag mixture and oxide layer on the surface layer of the heat-resistant steel, the thickness of the surface layer is thickened, and it is impossible to ensure that the laser can completely break through the surface layer of the heat-resistant steel, thereby causing a certain degree of negative impact on the subsequent spectral collection step, and it is impossible to detect the effective matrix information of the monitoring object, thereby affecting the operation and maintenance monitoring of the high-temperature pressure-bearing equipment.

[0058] At present, in order to effectively reduce or eliminate the influence of the impurities in the surface layer, some existing technologies optimize the parameters such as laser energy, focal point position, light collection delay, etc.; in the aspect of data analysis, some existing technologies try to use spectral intensity normalization, wavelet threshold denoising, multivariate scatter correction, standard normal transformation and other methods to improve the spectral quality, and combine the innovation and use of spectral feature screening method to select spectral feature variables with higher correlation with the matrix characteristics of the material, use machine learning algorithm to establish a heat-resistant steel characteristic analysis model, and then realize characteristic analysis. These methods have obtained good accuracy and stability.

[0059] However, the above existing technologies have the following defects:

[0060] (1) The detection object is a heat-resistant steel sample that has been fully polished and polished, which is a sample completely free of surface characteristics interference. This method has certain limitations and cannot effectively obtain the representative spectral information of the matrix of the heat-resistant steel under the influence of the surface state, thereby affecting the accuracy of the operation and maintenance monitoring of the high-temperature pressure-bearing equipment;

[0061] (2) The premise of spectral analysis is that the laser-induced breakdown completely breaks through the surface layer of the detection object. However, the existing technology only determines the breakdown of the surface layer based on the intensity change of the typical characteristic spectral line, and cannot accurately determine the breakdown of the surface layer, and lacks a basis.

[0062] (3) After obtaining the spectral data of the broken surface layer, the existing spectral data optimization method only corrects the spectral line noise and baseline drift caused by instrument fluctuations, environmental interference and surface state, and cannot correct the negative impact of pit depth variation and surface layer impurity recondensation on the spectrum.

[0063] In actual industrial applications, whether it is remote online monitoring of high-temperature pressure-bearing materials or portable rapid detection, detection needs to be performed without pretreatment of the detection object. If a complex pretreatment process is added during detection, the detection difficulty will be greatly increased and the detection efficiency will be reduced. True in-situ rapid detection cannot be achieved.

[0064] To solve the problems in the prior art, the present application provides a heat-resistant steel aging detection method under surface characteristic interference and a storage medium. First, based on the sample data and aging grade labels of a plurality of sample heat-resistant steels of known aging grades, a spectral feature data set is established, a support vector machine is trained through the spectral feature data set, and a heat-resistant steel aging detection model is obtained. Then, a to-be-detected heat-resistant steel with surface characteristic interference is obtained, laser ablation and laser tomography are performed on a plurality of measurement points of the to-be-detected heat-resistant steel through LIBS technology, corresponding laser spectral tomography data are obtained, and representative spectra are selected. Specifically, based on the spectral evolution trend features obtained by spectral tomography, in combination with surface layer and matrix difference analysis, spectral data affected by surface characteristics are determined, the to-be-detected heat-resistant steel matrix representative spectrum is obtained by combining the spectral data with plasma temperature discrimination, and then the aging detection model is obtained. After that, a targeted spectral feature selection method is used to reduce the influence of surface layer material recondensation on laser ablation and spectral data, ensure the effectiveness of the spectral representation of the matrix characteristics, and at the same time, a spectral correction method is used to reduce the influence of surface layer property differences and environmental fluctuations on the identification degree and process stability of the actual detection object, and feature selection is completed. At the same time, the intensity of the matrix spectral data after feature selection is corrected by using the spectral feature data set, and the influence of intensity difference on the model discrimination process is reduced. Finally, the aging detection model is used to realize the aging detection of the heat-resistant steel in combination with the corrected spectral data.

[0065] Referring to Figure 1 , Figure 1 The flowchart of the heat-resistant steel aging detection method under surface characteristic interference provided by the embodiment of the present application is shown. In one embodiment of the present application, the heat-resistant steel aging detection method under surface characteristic interference will be described and explained below. The method can include but is not limited to the following steps.

[0066] S100, obtaining a sample heat-resistant steel after surface pretreatment, performing ablation and laser tomography on the sample heat-resistant steel by using LIBS technology, obtaining sample spectral data and corresponding aging grade labels, constructing a spectral feature data set, and establishing a heat-resistant steel aging detection model by using the spectral feature data set.

[0067] It should be noted that the surface pretreatment generally refers to polishing and polishing and other operations on the surface of the heat-resistant steel, so that the surface is smooth and flat. The sample heat-resistant steel after surface pretreatment does not exist surface characteristic interference.

[0068] S200, obtaining the heat-resistant steel to be measured with surface characteristic interference, and using LIBS technology to ablate and laser tomography on multiple measurement points of the heat-resistant steel to be measured, to obtain laser spectral tomography data of the multiple measurement points.

[0069] It should be noted that the laser spectral tomography data carries the ablation time value. The heat-resistant steel to be measured without surface pretreatment has surface characteristic interference, and the surface characteristic interference specifically refers to the phenomenon of uneven surface, surface oxidation, surface dust and other surface characteristic interference on the surface of the heat-resistant steel, and these surface characteristic interference phenomena will have a negative impact on the ablation detection, spectrum collection and other processes of the LIBS technology.

[0070] In this step, the heat-resistant steel sample with surface characteristic interference is detected by LIBS technology. The laser generally selects a nanometer laser to obtain the best signal-to-noise ratio of the characteristic spectrum as the benchmark, and the orthogonal experiment method is used to optimize and adjust the laser energy, spot diameter and light delay. At the same time, different measurement positions are selected for laser ablation for each sample to improve the detection representativeness of the sample, each measurement point is an ablation point, and continuous laser pulses are used for multiple ablation and laser tomography for each ablation point to obtain the evolution trend of the spectral characteristics along the radial direction, and then the laser spectral tomography data corresponding to the continuous laser pulses of different measurement points of the sample is obtained, that is, the laser spectral tomography data, to determine the acquisition scheme of the matrix representative spectrum.

[0071] S300, screening the representative spectrum of the laser spectral tomography data to obtain the matrix spectrum data;

[0072] S400, performing feature selection and intensity correction on the matrix spectrum data to obtain corrected spectrum data;

[0073] S500, using the heat-resistant steel aging detection model to perform aging detection on the corrected spectrum data to obtain an aging detection result.

[0074] Optionally, the aging detection result includes the aging grade of the heat-resistant steel to be measured.

[0075] Referring to Figure 2 , as shown in Figure 2The flowchart for obtaining the substrate spectrum data is shown. In one embodiment of the present application, step S300 will be further described and explained. In order to ensure that the obtained spectrum data corresponds to the substrate which has been effectively ablated by the pulse, and to improve the effectiveness of the spectrum data, the present application screens the obtained spectrum tomography data for representative spectrum through step 300, and then obtains effective spectrum data which can map the ablation of the substrate. Step S300 can include but is not limited to the following steps.

[0076] S310, the following screening step is performed once for each measurement point of the laser spectrum tomography data:

[0077] S311, the breakdown pulse number and the plasma temperature of the surface layer of the measurement point are calculated according to the laser spectrum tomography data of the measurement point;

[0078] S312, it is judged whether the breakdown pulse number of the measurement point is greater than the breakdown pulse reference value; if yes, go to S313; if no, discard the laser spectrum tomography data of the measurement point.

[0079] S313, it is judged whether the difference between the plasma temperature of the measurement point and the preset plasma reference temperature is less than the first value; if yes, the laser spectrum tomography data of the measurement point is taken as the substrate measurement spectrum data; if no, the laser spectrum tomography data of the measurement point is discarded.

[0080] S320, when the screening steps of all measurement points are completed, the substrate measurement spectrum data of all measurement points is output as the substrate spectrum data.

[0081] Optionally, the first value is 10%.

[0082] In the specific embodiment, first, by obtaining the variation rule of the spectrum characteristics along the radial direction, the variation trend of the characteristic spectrum intensity of the main sample typical substrate element, alloy element and surface layer component with the increase of the laser pulse number is explored, and the principle analysis of the sample working environment and the structure and properties of the surface layer is combined to determine the pulse number for breaking through the surface layer. The pulse number for breaking through the surface layer is compared with the pulse reference value to complete the first data screening.

[0083] In the art, there is a large difference between the surface layer and the base of the heat-resistant steel in physical and chemical properties, which affects the laser-substance interaction process and causes changes in plasma characteristics. Therefore, the breakdown of the surface layer can be determined based on the plasma temperature. After the first screening of the data is completed, the plasma temperature of the heat-resistant steel sample after surface pretreatment is calculated as the reference value of the plasma temperature of the heat-resistant steel base. Then, the plasma temperature corresponding to the laser spectroscopy tomography data is calculated, and when the temperature difference with the reference value of the plasma temperature of the heat-resistant steel base is less than 10%, i.e. the first value, it is considered that the current pulse has effectively ablated the heat-resistant steel base. In this way, the second screening of the data is completed. Through the screening of the data twice, the base spectral data within the pulse number range of the representative spectrum is determined.

[0084] Referring to Figure 3 , Figure 3 The flowchart for calculating the number of breakdown pulses of the surface layer provided by the embodiments of the present application is shown. Further, in step S311, the number of breakdown pulses of the surface layer corresponding to the measurement point is calculated according to the laser spectroscopy tomography data of the measurement point, including:

[0085] The laser spectroscopy tomography data of the measurement point is divided into several data groups, and the first mean data and the second mean data of each data group are calculated according to the laser spectroscopy tomography data.

[0086] It should be noted that the first mean data is the mean value of the characteristic spectral line intensity of the surface layer element corresponding to all measurement points of the heat-resistant steel sample.

[0087] It should be noted that the second mean data is the mean value of the characteristic spectral line intensity of the base element corresponding to all measurement points of the heat-resistant steel sample.

[0088] It should be noted that the ablation time value of the first mean data and the second mean data of each data group is continuous, and the ablation time value refers to the time value of the ablation of the measurement point. The purpose of setting the ablation time value is to facilitate the subsequent determination of whether the mean values of the characteristic spectral line intensities of the base element and the surface layer element continue to rise or continue to fall.

[0089] It is determined whether the first mean data of the current data group is continuously rising and the second mean data is continuously falling. If so, the stability of the characteristic spectral line intensity of the surface layer element and the base element corresponding to the current data group is analyzed, and the next step is performed. If not, the current data group is skipped, and the next data group is judged until the judgment of all data groups is completed.

[0090] When the stability is greater than the stability threshold, the first difference between the adjacent two first mean data in the current data group is calculated, and the second difference between the adjacent two second mean data in the current data group is calculated;

[0091] When both the first difference and the second difference are less than the second value, the start pulse number of the laser spectroscopy tomography data of the current data set is taken as the breakdown pulse number of the surface layer of the measurement point.

[0092] Optionally, the second value is 5%.

[0093] In the embodiment, the breakdown pulse number of the surface layer is determined based on the evolution characteristics of the characteristic spectral lines. Referring to Figure 4 illustrated in FIG. 2, Figure 4 illustrated in FIG. 2 is a schematic diagram of the variation of the characteristic spectral line intensity of the typical base elements, alloy elements and surface layer components of the heat-resistant steel with the pulse number. The spectral line intensity can be generally considered to be proportional to the element content. Therefore, the gradual breakdown of the surface impurity layer is accompanied by the decrease of the characteristic spectral line intensity of the surface layer elements such as Ca, Na and K. Specifically, due to the influence of the surface layer thickness, the spectral line intensity can be enhanced in a certain pulse range at the beginning, but the intensity will show a downward trend with the ablation of the surface layer. At the same time, the characteristic spectral line intensity of the base elements such as Fe, Mn, V and Cr is increased. After complete breakdown, the variation of each characteristic spectral line intensity tends to be stable. Therefore, based on the evolution characteristics, the determination steps of the above surface layer breakdown pulse number can be obtained.

[0094] The above steps will be described below with an example. It is assumed that after the ablation and laser tomography of the heat-resistant steel, laser spectroscopy tomography data of a plurality of measurement points are obtained. Each measurement point has 100 laser spectroscopy tomography data corresponding to 100 laser pulses. The determination steps of the surface layer breakdown pulse number are as follows:

[0095] For each measurement point, first, the one hundred laser spectroscopy tomography data corresponding thereto are divided into 20 groups, and each group of data contains 5 laser spectroscopy tomography data. Then, for each group of data, the average value of the characteristic spectral line intensity of the 5 base elements, i.e., the first average data, and the average value of the characteristic spectral line intensity of the 5 surface layer elements, i.e., the second average data, are calculated according to the 5 laser spectroscopy tomography data contained in the group. These average values also carry the same ablation time values as the laser spectroscopy tomography data, and the ablation time values of the first average data and the second average data are continuous to ensure that the first average data and the second average data in each group of data are continuous.

[0096] Afterwards, the mean of the characteristic spectral line intensity of the 5 base elements and the mean of the characteristic spectral line intensity of the 5 surface layer elements contained in each group of data are judged and analyzed in turn. When the mean of the characteristic spectral line intensity of the base elements in the current data group appears for 5 times in succession, and the mean of the characteristic spectral line intensity of the surface layer elements decreases for 5 times in succession, the stability of the characteristic spectral line intensity is analyzed. When the difference between the mean of the spectral line intensity of the base elements in the current data group for 2 times in succession is less than 5%, and the difference between the mean of the spectral line intensity of the surface layer elements for 2 times in succession is also less than 5%, the starting pulse number of the spectral data used for calculating the mean is taken as the surface layer breakdown pulse number of the measuring point.

[0097] Optionally, the breakdown pulse reference value is determined by the following steps:

[0098] The heat-resistant steel sample with surface characteristic interference is pre-processed by ablation and laser tomography based on LIBS technology, and then the surface layer breakdown pulse number of each measuring point of the sample is confirmed based on the laser spectral tomography data. The maximum value of the surface layer breakdown pulse number of all measuring points is selected as the surface layer breakdown pulse reference value of the detection object, i.e. the breakdown pulse reference value.

[0099] Optionally, the plasma temperature reference value is determined by the following steps:

[0100] The heat-resistant steel sample with surface characteristic interference is pre-processed by ablation and laser tomography based on LIBS technology, and then the surface layer breakdown pulse number of each measuring point of the sample is confirmed based on the laser spectral tomography data. The maximum value of the surface layer breakdown pulse number of all measuring points is selected as the surface layer breakdown pulse reference value of the detection object, i.e. the breakdown pulse reference value.

[0101] Based on the above embodiment, there is a large difference between the surface layer and the base physicochemical characteristics of the heat-resistant steel, which affects the laser-substance interaction process and causes changes in plasma characteristics. Therefore, the breakdown of the surface layer can be distinguished based on the plasma temperature. The plasma temperature is one of the most important parameters of the laser-induced plasma and is an important physical quantity for understanding the processes of dissociation, ionization and reaction in the plasma. The plasma temperature includes electron temperature, excitation temperature and ionization temperature, which are consistent when the plasma is in a local thermal equilibrium state. Meanwhile, the particles in the plasma satisfy the Boltzmann distribution, and the plasma temperature can be obtained by using the spectral intensity of multiple characteristic spectral lines of an element and the Boltzmann method. When the laser-induced plasma is in a local thermal equilibrium state, the Boltzmann curve is established by using the spectral intensity of multiple spectral lines of an element, and the points corresponding to the unused spectral lines will also fall on the fitted straight line.

[0102] Further, in step S310, the step of calculating the plasma temperature of the surface layer of the measurement point according to the laser spectroscopy tomography data of the measurement point, specifically includes:

[0103] First, the plasma spectral line intensity of the surface layer corresponding to the measurement point is calculated according to the laser spectroscopy tomography data.

[0104] It should be noted that the plasma spectral line intensity is calculated by the following formula:

[0105]

[0106] Where I ij represents the spectral line intensity of the laser spectroscopy tomography data, λ ij represents the spectral line wavelength, A ij represents the transition probability from i energy level to j energy level, g i represents the statistical weight of the energy level, E i represents the upper energy level of the transition, k represents the Boltzmann constant, T represents the plasma temperature, N(T) represents the density function, and U(T) represents the partition function determined by the statistical weight of the atomic ground state. For a spectral line, its g i A ij and E i can be obtained by querying the database.

[0107] Then, several spectral lines of the same ionization energy level of the same element are selected as the analysis object, and the plasma spectral line intensity of each spectral line is taken as the ordinate, and the upper energy level E i of each spectral line is taken as the abscissa, and the Boltzmann curve is obtained after fitting, and the plasma temperature of the surface layer is obtained according to the Boltzmann curve.

[0108] In the prior art, the intensity ratio of two different spectral lines of an element is usually calculated to calculate the plasma temperature T. However, the plasma temperature calculated by this method has a large error, and therefore, more characteristic spectral lines are used to improve the calculation accuracy of the plasma temperature. The plasma spectral line intensity can be calculated by the above formula, and when the plasma is in local thermal equilibrium, the Boltzmann curve is established using this formula, and the slope of the straight line is The plasma temperature can be obtained by the Boltzmann curve.

[0109] Alternatively, the density function N(T) satisfies:

[0110] N(T)=hcN o ;

[0111] Where h is the Planck constant, c is the speed of light, and N ois the total particle number density.

[0112] As an optional embodiment, the spectrum of the heat-resistant steel after surface pre-treatment, without the interference of surface characteristics, is taken as a typical reference spectrum of the sample matrix. The reference spectrum is correlated with the matrix spectrum data selected according to the discrimination result, and the matrix spectrum data effectively ablated to the sample matrix is further selected as the further analysis data. The calculation formula of the correlation coefficient is as follows:

[0113]

[0114] wherein y i represents the spectral intensity of the i-th pixel point of the typical reference spectrum, represents the average value of the spectral intensity of the typical reference spectrum, x i ' is the spectral intensity of the i-th pixel point of the matrix spectrum data of a single measurement pulse, is the average value of the spectral intensity of the i-th pixel point of the matrix spectrum data.

[0115] When r = 1.0, the matrix spectrum data is completely correlated with the reference spectrum. When r > 0.8, the matrix spectrum data is highly correlated with the reference spectrum. When 0.4 ≤ r ≤ 0.8, the matrix spectrum data is moderately correlated with the reference spectrum. When r < 0.4, the matrix spectrum data is lowly linearly correlated with the reference spectrum. Optionally, the data of the matrix spectrum data that meets the moderate correlation and low linear correlation is discarded.

[0116] In an embodiment of the present application, the following will further illustrate and describe the feature selection and intensity correction of the matrix spectrum data in step S400 to obtain the corrected spectrum data. After the surface layer of the heat-resistant steel is ablated and broken by the continuous laser pulses, a small amount of molten or gasified surface layer material will still produce recondensation phenomenon, and the recondensation product will adhere to the ablation area. The composition of the surface layer material is greatly different from the matrix, which will affect the characterization of the real state of the matrix by the obtained spectrum. For the heat-resistant steel, the surface layer contains oxides and coal combustion residues, among which the characteristic spectrum of the residual alkali metal Na, K and Ca elements is easier to be excited, and the peak intensity is high and the half-width is wide, which will have a great influence on the adjacent other element spectrum, and further affect the final analysis result of the characteristics of the heat-resistant steel matrix. Therefore, according to the characteristics of the impurity elements in the surface layer and the wavelength position of the characteristic spectrum, the spectral features in a certain wavelength range near the corresponding typical characteristic spectrum are screened out. The feature selection of the matrix spectrum data can include but is not limited to the following steps.

[0117] The spectral features in a range of 5 nm near the six specific characteristic spectrum lines in the matrix spectrum data are screened out to obtain the screened matrix spectrum data, and a spectral feature data set is constructed.

[0118] It should be noted that the six specific characteristic spectral lines are Na I 589.00 nm spectral line, Na I 589.59 nm spectral line, K I 766.53 nm spectral line, K I 769.90 nm spectral line, Ca I 393.37 nm spectral line, Ca I 396.85 nm spectral line, and Ca I 422.67 nm spectral line.

[0119] Based on the above embodiment, the spectral feature data set established by the model and the data of the detection object may be affected by factors such as changes in the collection environment, fluctuations in the LIBS device, and slight differences in the surface layer characteristics, so that the overall intensity of the data spectrum is affected. Therefore, based on the spectral feature data set, the average intensity of the average spectrum of the matrix spectral data after feature selection is corrected, the influence of the intensity difference on the model discrimination process is reduced, and the correction process is as follows:

[0120] The average intensity of the data in the spectral feature data set is calculated.

[0121] According to the average intensity, the matrix spectral data after feature selection is corrected by the following formula:

[0122]

[0123] wherein S1 represents the corrected spectral data, S0 represents the matrix spectral data after feature selection, I0 represents the spectral average intensity of the matrix spectral data after feature selection, and I mean represents the average intensity of the data in the spectral feature data set.

[0124] As an optional embodiment, the data of the constructed spectral feature data set is subjected to spectral quality optimization and interference correction. The mechanical properties of the heat-resistant steel will cause differences in laser ablation samples, and factors such as slight differences in the surface curvature of the heat-resistant steel pipe and different surface oxide layer structures will affect the interaction between the laser and the sample, thereby interfering with the correlation between the spectrum and the characteristics of the heat-resistant steel.

[0125] First, the standard normal variate transform (SNV) is used to eliminate the influence of the surface roughness of the detection sample, the different sizes of solid particles, and surface scattering. The specific calculation formula is as follows:

[0126]

[0127] wherein represents the average value of the spectral intensity of the data in the spectral feature data set, N is the number of wavelength points measured by the spectrometer, and x SNV represents the spectral data after SNV transformation.

[0128] Then, the baseline drift is corrected by using the multivariate scatter correction, and the scattering effect caused by the uneven metal surface is eliminated. The multivariate scatter correction is one of the common algorithms for hyperspectral data preprocessing. The multivariate scatter correction can effectively eliminate the spectral differences caused by different scattering levels, thereby enhancing the correlation between the spectrum and the data. The method corrects the baseline translation and offset of the spectral data through the ideal spectrum. In practice, we cannot obtain the real ideal spectral data, so we often assume the average value of all spectral data as the "ideal spectrum". The specific process is as follows:

[0129] Firstly, the average value of all spectral data in the spectral feature data set is taken as the ideal spectrum, and the ideal spectrum is taken as the target of each spectrum Then, the spectrum of each data in the spectral feature data set is linearly regressed with the ideal spectrum, and the baseline translation and offset of each data, i.e. the slope k and the intercept b0, are obtained by solving the least square method: Finally, the spectrum of each data is corrected, and the corrected data x is obtained by subtracting the baseline translation and dividing by the offset of each data spectrum MSC :

[0130] In an embodiment of the present application, the process of establishing the aging detection model of the heat-resistant steel in step S100 will be described and explained below. The model establishment process is the process of establishing the relationship between the spectrum and the aging grade. The basic process is to divide the sample spectral data based on the known aging grade into a training set and a test set, use the training set to establish the model, and then use the test set to evaluate the model, and output the final heat-resistant steel aging detection model. The modeling data source is the spectral feature data set, which includes sample spectral data of a plurality of sample heat-resistant steels, and the sample spectral data carries aging grade labels. The sample heat-resistant steel is an artificial aging heat-resistant steel with known aging grade and a polished heat-resistant steel with surface characteristics eliminated.

[0131] The heat-resistant steel aging detection model is established by using the spectral feature data set, including:

[0132] S110, the regression feature elimination method based on support vector machine is used to select features of the spectral feature data set, and an optimal feature subset is obtained.

[0133] It should be noted that, because the spectral data contains rich spectral characteristics, it is easy to cause dimension disaster, that is, the number of characteristic variables is much larger than the number of samples, causing the sparsity of sample data in high-dimensional space. The classification model is easy to find a complex hyperplane in the high-dimensional space due to the high sparsity of the sample, resulting in overfitting of the classification problem. Therefore, in order to improve the prediction and generalization ability of the classification model and better understand the contribution of each variable to the classification model, it is necessary to select the spectral characteristic variables in the spectral feature data set. Based on recursive feature elimination (Recursive Feature Elimination, RFE), the contribution of variables to the model is represented by the weight vector coefficient of the learner, and the elimination of invalid and redundant variables is realized.

[0134] In S120, the Gaussian kernel function is determined as the kernel function of the support vector machine model, and the support vector machine model is established.

[0135] It should be noted that, because the spectral data contains rich spectral characteristics, it is easy to cause dimension disaster, that is, the number of characteristic variables is much larger than the number of samples, causing the sparsity of sample data in high-dimensional space. The classification model is easy to find a complex hyperplane in the high-dimensional space due to the high sparsity of the sample, resulting in overfitting of the classification problem. Therefore, in order to improve the prediction and generalization ability of the classification model and better understand the contribution of each variable to the classification model, it is necessary to select the spectral characteristic variables in the spectral feature data set. Based on recursive feature elimination (Recursive Feature Elimination, RFE), the contribution of variables to the model is represented by the weight vector coefficient of the learner, and the elimination of invalid and redundant variables is realized.

[0136] k(x i , x j )=exp(-g||x i -x j ||);

[0137] wherein x i is the spectrum number of the i th sample in the spectral feature data set, x j is the spectrum number of the j th sample in the spectral feature data set, and g is a constant used to determine the Gaussian distribution of the Gaussian kernel function.

[0138] In S130, the optimal feature subset is used to train the support vector machine model to obtain the heat-resistant steel aging detection model.

[0139] Specifically, the optimal feature subset is divided into a training set and a test set at a certain ratio, the support vector machine model is trained through the training set, and the performance of the trained support vector machine model is tested through the test set. If the performance meets the expectation, the model output is output as the heat-resistant steel aging detection model. If the performance does not meet the expectation, the training parameters of the model are adjusted, and the support vector machine model is retrained through the training set.

[0140] In addition, the embodiment of the present application also provides a storage medium, wherein the storage medium stores processor-executable instructions, and the processor-executable instructions are used to execute a heat-resistant steel aging detection method under surface characteristic interference when executed by a processor.

[0141] The application can realize rapid detection and in-situ detection of the characteristics of the actual service heat-resistant steel, and provide a reliable analysis and testing scheme for remote online monitoring and real-time in-situ detection of the heat-resistant steel. With reference to the scheme and steps, a detection scheme for different types and working state high-temperature pressure-bearing materials can be customized, and then a monitoring data management and remote analysis and diagnosis, equipment maintenance scheme based on intelligent photoelectric diagnosis can be developed, thereby providing guidance for safe operation and production of high-temperature pressure-bearing equipment. The application can be widely applied in the power, metallurgy, power machinery, petroleum and chemical industries.

[0142] The following takes a typical heat-resistant steel T91 as a detection object to illustrate the technical scheme of the application.

[0143] In the first step, a typical heat-resistant steel T91 is taken as an object, and a T91 steel used in a coal-fired boiler with unknown aging grade is selected as a detection experimental sample. The LIBS detection is performed without surface pretreatment, and the spectral data corresponding to the continuous laser pulses of each measurement point of the sample, i.e., the laser spectral tomography data of the measurement point, are obtained.

[0144] In the second step, the variation law of the laser tomography data of the actual service T91 steel sample with the pulse number is explored, and the characteristic spectral lines of the matrix elements and alloy elements are mainly focused on. The principle analysis of the surface effects such as slagging and surface oxidation of the T91 steel is combined to obtain a pulse number reference value 1 of the surface layer breakdown, and the accuracy is improved through analysis of different measurement points.

[0145] In the third step, the plasma temperature of the spectral data of the actual service T91 steel sample under different pulses is calculated, and compared with the plasma temperature of a reference sample whose surface characteristics are eliminated by polishing and polishing. When the difference between the plasma temperature of the detection sample and the reference sample is within 10%, the pulse number reference value 2 of the surface layer breakdown is determined. The maximum value in the reference values 1 and 2 is taken as a reference to determine the pulse number range corresponding to the surface spectrum, and the matrix spectral data within the pulse number range is selected.

[0146] In the fourth step, the representative spectral data obtained after determining the pulse range are subjected to spectral feature selection. Based on the characteristics of the impurity elements in the surface layer and the excitation characteristics of the characteristic spectral lines, the spectral features within 5nm around the characteristic spectral lines of Na I 589.00nm, Na I 589.59nm, K I 766.53nm, K I 769.90nm, Ca I 393.37nm, Ca I 396.85nm, and Ca I 422.67nm are screened out, and the spectral intensity is corrected with reference to the average intensity of the average spectrum of the modeling data set.

[0147] In the fifth step, the corrected spectral data is analyzed by using a pre-constructed aging detection model of heat-resistant steel to obtain the aging grade of the T91 object.

[0148] The present application has the following technical effects:

[0149] Based on the spectral evolution trend features obtained by spectral tomography, combined with surface layer and matrix difference analysis, the spectral data affected by surface characteristics is determined, combined with plasma temperature discrimination, the matrix representative spectral information of heat-resistant steel under the influence of surface state is effectively obtained, and the acquisition of the matrix representative spectrum of the detection object is ensured. Moreover, the present application adopts a targeted spectral feature selection method, reduces the influence of surface layer material recondensation on laser ablation and spectral data, and simultaneously combines a spectral intensity correction method to reduce the influence of surface layer property difference and environmental fluctuation on the recognition degree and process stability of the actual detection object, ensure the effectiveness of the spectral data representing the matrix, and ensure the accuracy of the to-be-measured data source.

[0150] The detection scheme provided by the present application overcomes the interference of the surface characteristics of the to-be-measured object, simultaneously overcomes the negative influence of pit depth variation and surface layer impurity recondensation on the spectrum, supports a non-preprocessed detection object, can realize rapid in-situ detection, and can be coupled with a remote or portable detection device to complete the aging detection of high-temperature pressure-resistant materials.

[0151] The terms "first", "second", "third", "fourth" and the like used in the description of the present application and the above drawings, if any, are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0152] It should be understood that, in the application, "at least one" refers to one or more, and "multiple" refers to two or more. "And / or" is used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, "A and / or B" can represent three cases of only A, only B, and A and B existing at the same time, wherein A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after it. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can represent a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, and c can be single or multiple.

[0153] In several embodiments provided in the application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the above-described device embodiments are only schematic. For example, the division of the units is only a logical function division, and actual implementation can have another division manner. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0154] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0155] In addition, each functional unit in each embodiment of the application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0156] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0157] For the step numbers in the above method embodiments, they are only set for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

Claims

1. A method for detecting aging of heat-resistant steel under surface property interference, characterized by, The method comprises the following steps: obtaining a sample of heat-resistant steel after surface pretreatment, ablation and laser tomography of the sample of heat-resistant steel by using LIBS technology to obtain sample spectral data and corresponding aging grade labels, constructing a spectral feature data set, and establishing a heat-resistant steel aging detection model by using the spectral feature data set; obtaining a to-be-tested heat-resistant steel without surface pretreatment, ablation and laser tomography of multiple measurement points of the to-be-tested heat-resistant steel by using LIBS technology to obtain laser spectral tomography data of the multiple measurement points; screening representative spectra from the laser spectral tomography data to obtain matrix spectral data; performing feature selection and intensity correction on the matrix spectral data to obtain corrected spectral data; performing aging detection on the corrected spectral data by using the heat-resistant steel aging detection model to obtain an aging detection result; wherein the screening of representative spectra from the laser spectral tomography data to obtain matrix spectral data comprises: for the laser spectral tomography data of each measurement point, performing the following screening steps once: calculating the breakdown pulse number and plasma temperature of the surface layer of the measurement point according to the laser spectral tomography data of the measurement point; determining whether the breakdown pulse number of the measurement point is greater than a breakdown pulse reference value; if yes, proceeding to the next step; if no, discarding the laser spectral tomography data of the measurement point; determining whether the difference between the plasma temperature of the measurement point and a preset plasma reference temperature is less than a first value; if yes, taking the laser spectral tomography data of the measurement point as matrix measurement spectral data; if no, discarding the laser spectral tomography data of the measurement point; when the screening steps of all measurement points are completed, outputting the matrix measurement spectral data of all measurement points as matrix spectral data; wherein the calculation of the breakdown pulse number of the surface layer of the measurement point according to the laser spectral tomography data of the measurement point comprises: dividing the laser spectral tomography data of the measurement point into several data groups, and calculating first mean data and second mean data of each data group according to the laser spectral tomography data; wherein the ablation time values of the first mean data and the second mean data of each data group are continuous, and the ablation time value is defined as the time value of ablation of the measurement point; wherein the first mean data is the mean value of the characteristic spectral line intensity of the surface layer element of the measurement point of the heat-resistant steel sample, and the second mean data is the mean value of the characteristic spectral line intensity of the matrix element of the measurement point of the heat-resistant steel sample; determining whether the first mean data of the current data group is continuously rising and the second mean data is continuously falling; if yes, analyzing the stability of the characteristic spectral line intensity of the surface layer element and the matrix element corresponding to the current data group, and proceeding to the next step; if no, skipping the current data group and determining the next data group. When the stability is greater than the stability threshold, a first difference between the first mean data of the current data group and the first mean data of its adjacent data group is calculated, and a second difference between the second mean data of the current data group and the second mean data of its adjacent data group is calculated; When the first difference and the second difference are both less than the second numerical value, the starting pulse number of the laser spectroscopy tomography data of the current data group is taken as the breakdown pulse number of the surface layer of the measurement point; The step of calculating the plasma temperature of the surface layer of the measurement point according to the laser spectroscopy tomography data of the measurement point comprises: The plasma spectral line intensity of the surface layer corresponding to the measurement point is calculated according to the laser spectroscopy tomography data by the following formula: wherein, is the spectral line intensity of the laser spectroscopy tomography data, is the spectral line wavelength, is the transition probability from energy level i to energy level j, is the statistical weight of the energy level, is the upper energy level of the transition, k is the Boltzmann constant, and T is the plasma temperature, is the density function, is the partition function; Several spectral lines of the same ionization energy level of the same element are selected as analysis objects, and the plasma spectral line intensity of each spectral line is taken as the ordinate The transition upper energy level of each spectral line is taken as the abscissa as the ordinate After fitting, a Boltzmann curve is obtained, and the surface layer plasma temperature is obtained according to the Boltzmann curve.

2. The method of claim 1, wherein the method is characterized by: The first numerical value is 10%.

3. The method of claim 1, wherein the method is characterized by: The feature selection on the base body spectral data comprises: screening out the spectral features in the range of 5nm around six specific characteristic spectral lines in the base body spectral data, to obtain the screened base body spectral data.

4. The method of claim 1, wherein the method is characterized by: The step of performing intensity correction on the base body spectral data comprises: The average intensity of the data in the spectral feature data set is calculated; The base body spectral data after feature selection is corrected according to the average intensity by the following formula: wherein is the corrected spectral data, is the base spectral data after feature selection, is the spectral average intensity of the base spectral data after feature selection, is the average intensity of the data in the spectral feature data set.

5. The method of claim 1, wherein the method is characterized by: The establishment of the heat-resistant steel aging detection model by using the spectral feature data set comprises: The spectral feature data set is selected by a regression feature elimination method based on a support vector machine, to obtain an optimal feature subset; A Gaussian kernel function is determined as the kernel function of the support vector machine model, and the support vector machine model is established; wherein the Gaussian kernel function satisfies: wherein, is the number of spectra of the i-th sample in the spectral feature dataset, is the number of spectra of the j-th sample in the spectral feature dataset, and g is a constant used to determine the Gaussian distribution of the Gaussian kernel function. The optimal feature subset is used to train the support vector machine model, to obtain the heat-resistant steel aging detection model.

6. A storage medium having stored therein instructions executable by a processor, the instructions causing the processor to perform the method of any one of claims 1-5. The instructions executable by the processor, when executed by the processor, are used to perform the heat-resistant steel aging detection method under the interference of surface characteristics.

Images