Laser induced breakdown spectroscopy quantitative detection model establishment method and medium and controller

Abstract

The invention relates to the technical field of laser induced breakdown spectroscopy, and discloses a method for establishing a quantitative detection model of laser induced breakdown spectroscopy, a medium, and a controller. The method comprises: obtaining a sample data set, sample uncertainty, and system uncertainty, and dividing the sample data set into a sample training set and a sample test set; using the original spectral data in the sample training set and its corresponding sample measured value, establishing an original quantitative detection model of the ratio of the spectral line intensity of the measured element to the internal standard element with respect to the sample measured value; using the original quantitative detection model to calculate the absolute deviation corresponding to each test sample in the sample test set, recording the test sample whose absolute deviation is greater than the sample uncertainty as an uncertainty set, and using the uncertainty set to fit the original quantitative detection model according to the sample uncertainty and the system uncertainty, and determining the target quantitative detection model. The laser induced breakdown spectroscopy quantitative detection model established by the method fits the actual application of industrial materials.

Description

Technical Field

[0001] The invention relates to the technical field of laser induced breakdown spectroscopy, and in particular to a method for establishing a quantitative detection model of laser induced breakdown spectroscopy, a medium and a controller. Background Art

[0002] The material composition of metallurgical and other industries is the core parameter of process control and the evaluation index of product quality. Composition detection plays a very important role in various technical and economic indicators such as the smelting degree, quality of smelting products and metal recovery rate of smelting products. Its online real-time detection is one of the difficult problems in the intelligent improvement and upgrading of smelting, especially the online detection of high-temperature melt composition.

[0003] The material composition detection in the relevant smelting process adopts offline laboratory detection, which has problems such as insufficient real-time performance, insufficient reliability, insufficient guidance, and insufficient safety. It is not conducive to the real-time monitoring, fine control and process improvement of material composition, and cannot support the intelligent upgrade of corresponding process nodes. The component analyzer based on laser induced breakdown spectroscopy (LIBS) can detect the composition of process materials. It has the characteristics of no sampling, no sample preparation, no contact, and no radiation. It helps to realize closed-loop control based on real-time composition guidance process regulation, thereby realizing precise process control, optimizing process connection efficiency, and supporting intelligent construction.

[0004] In actual use, the laser induced breakdown spectroscopy component analyzer will produce system uncertainty due to equipment drift, and sample uncertainty due to sample inhomogeneity. Both system uncertainty and sample uncertainty have an impact on the quantitative detection model and cannot be ignored. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems in the related art to a certain extent. To this end, one purpose of the present invention is to propose a method for establishing a quantitative detection model of laser induced breakdown spectroscopy, which is based on system uncertainty and sample uncertainty to establish a quantitative detection model of laser induced breakdown spectroscopy, which is more suitable for the actual application of industrial materials.

[0006] A second object of the present invention is to provide a computer-readable storage medium.

[0007] The third object of the present invention is to provide a controller.

[0008] To achieve the above-mentioned purpose, the first aspect of the present invention proposes a method for establishing a quantitative detection model of laser induced breakdown spectroscopy, the method comprising: obtaining a sample data set, sample uncertainty and system uncertainty, and dividing the sample data set into a sample training set and a sample test set, wherein the sample data set includes original spectral data and its corresponding sample measured values; using the original spectral data in the sample training set and its corresponding sample measured values, establishing an original quantitative detection model of the ratio of the spectral line intensity of the measured element to the internal standard element with respect to the sample measured value; using the original quantitative detection model to calculate the absolute deviation corresponding to each test sample in the sample test set, recording the test samples whose absolute deviation is greater than the sample uncertainty as an uncertainty set, using the uncertainty set to fit the original quantitative detection model according to the sample uncertainty and the system uncertainty, and determining the target quantitative detection model.

[0009] According to the method for establishing a laser induced breakdown spectroscopy quantitative detection model in an embodiment of the present invention, a laser induced breakdown spectroscopy quantitative detection model is established based on system uncertainty and sample uncertainty, and the influencing factors of uncertainty are added to the training process of the laser induced breakdown spectroscopy quantitative detection model, which is more in line with the actual application of industrial materials and achieves better modeling effects.

[0010] In addition, the method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to the above embodiment of the present invention may also have the following additional technical features:

[0011] According to one embodiment of the present invention, the acquisition of the sample data set includes: using a laser induced breakdown spectroscopy component analyzer to emit lasers to multiple detection points of industrial materials to collect multiple original spectral data, wherein the number of the original spectral data is greater than a preset threshold number; sampling the industrial material at each detection point position, and performing chemical testing on the sampled material to obtain multiple sample measured values, wherein the sample measured values ​​correspond one-to-one to the original spectral data.

[0012] According to one embodiment of the present invention, it is characterized in that obtaining the sample uncertainty includes: repeatedly sampling n portions of industrial materials within a preset time, and testing the n portions of sampled materials to determine the single uncertainty, wherein n>1; repeating the step of repeatedly sampling n portions of industrial materials within a preset time, and testing the n portions of sampled materials to determine the single uncertainty m times, wherein m>1; determining the sample uncertainty based on the m single uncertainties.

[0013] According to one embodiment of the present invention, it is characterized in that the determining the single uncertainty by testing n portions of sampled materials includes: calculating the average value of the measured values ​​obtained by testing n portions of sampled materials to obtain the measured average value; calculating the deviation between each of the measured values ​​and the measured average value, and recording the maximum deviation as the single uncertainty; determining the sample uncertainty based on m single uncertainties includes: determining the maximum value among the m single uncertainties, and recording the maximum single uncertainty as the sample uncertainty.

[0014] According to one embodiment of the present invention, it is characterized in that obtaining the system uncertainty includes: obtaining a plurality of measured spectral data obtained by continuously measuring a standard sample of an industrial material by the same operator at the same time; determining a single characteristic spectral line intensity corresponding to each of the measured spectral data, and calculating an average characteristic spectral line intensity of the plurality of single characteristic spectral line intensities; and calculating the system uncertainty based on the single characteristic spectral line intensities and the average characteristic spectral line intensity.

[0015] According to one embodiment of the present invention, before dividing the sample data set into a sample training set and a sample test set, the method also includes: preprocessing the original spectral data for each piece of original spectral data in the sample training set; performing feature extraction on the preprocessed original spectral data and comparing it with a standard spectral line database; recording spectral line peaks that meet the standard spectral line range as characteristic peaks, and recording spectral line peaks that do not meet the standard spectral line range as non-characteristic peaks, thereby obtaining characteristic peaks and non-characteristic peaks of the original spectral data.

[0016] According to one embodiment of the present invention, the use of the original quantitative detection model to calculate the absolute deviation corresponding to each test sample in the sample test set includes: using the original quantitative detection model to calculate the model prediction value corresponding to the original spectral data of each test sample in the sample test set; calculating the absolute deviation based on the model prediction value corresponding to the original spectral data of each test sample in the sample test set and the sample measured value.

[0017] According to one embodiment of the present invention, the uncertainty set is used to fit the original quantitative detection model according to the sample uncertainty and the system uncertainty to determine the target quantitative detection model, including: determining the horizontal coordinate of each uncertain sample in the uncertainty set according to the system uncertainty; and determining the vertical coordinate of each uncertain sample in the uncertainty set according to the sample uncertainty to obtain a plurality of adjusted coordinates corresponding to each of the uncertain samples; fitting the original quantitative detection model according to the plurality of adjusted coordinates corresponding to each of the uncertain samples, and recording the one with a goodness of fit close to 1 as the target quantitative detection model.

[0018] To achieve the above-mentioned purpose, the second aspect of the present invention proposes a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for establishing a quantitative detection model of laser induced breakdown spectroscopy proposed in the second aspect of the present invention is implemented.

[0019] To achieve the above-mentioned purpose, the third aspect of the present invention proposes a controller, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the method for establishing a quantitative detection model of laser induced breakdown spectroscopy proposed in the second aspect of the present invention is implemented.

[0020] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a flow chart of a method for establishing a quantitative detection model according to an embodiment of the present invention;

[0022] Figure 2 is a flow chart of obtaining a sample data set according to an embodiment of the present invention;

[0023] Figure 3 It is a technical principle diagram of an embodiment of the present invention;

[0024] Figure 4 is a flow chart of obtaining sample uncertainty according to an embodiment of the present invention;

[0025] Figure 5 is a flow chart of obtaining system uncertainty according to an embodiment of the present invention;

[0026] Figure 6 is a flow chart of model training with uncertainty according to an embodiment of the present invention;

[0027] Figure 7 is a flow chart of preprocessing and feature extraction according to an embodiment of the present invention;

[0028] Figure 8 is a flow chart of determining absolute deviation according to one embodiment of the present invention;

[0029] Fig. 9 is a curve diagram of an original quantitative detection model of an embodiment of the present invention;

[0030] Fig.10 is a curve diagram of a target quantitative detection model established in one embodiment of the present invention;

[0031] Fig.11is a schematic diagram of a quantitative detection model building device according to an embodiment of the present invention;

[0032] Fig.12 4 is a structural block diagram of a controller according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0034] The laser is emitted by a laser. After the laser is focused on the sample surface, the material is peeled off the sample surface and heated to generate plasma. The original spectrum data is collected using a laser-induced breakdown spectroscopy component analyzer. The intensity of the plasma emission spectrum line can be calculated using the Roman King-Syber formula (1):

[0035] (1)

[0036] in, is the plasma emission line intensity, i and j are the upper and lower energy levels of electron transition, F is the experimental parameter, is the transition particle concentration, is the transition probability, is the degeneracy of energy level i, is the partition function, is the upper energy level, is the Boltzmann constant, and T is the plasma temperature.

[0037] Based on the above principle, the internal standard method can be used for quantitative analysis. A matrix element with a relatively stable element content is determined from the sample to be tested as the internal standard element. According to the above formula (1), the following is obtained:

[0038] Spectral line intensity of the measured element

[0039] Internal standard element spectral line intensity

[0040] According to the above formula, it can be simplified to:

[0041] (2)

[0042] in, is the spectral line intensity of the measured element, The concentration of the spectral line selected for the element being measured, is the upper energy level degeneracy of the measured element, is the transition probability of the measured element, is the partition function of the element being measured, is the upper energy level energy of the element being measured, is the spectral line intensity of the internal standard element, is the concentration of the spectral line selected for the internal standard element, is the upper energy level degeneracy of the internal standard element, is the transition probability of the internal standard element, is the partition function of the internal standard element, is the upper level energy of the internal standard element.

[0043] According to the above formula, the ratio of the spectral line intensity of the measured element and the internal standard element can be And the sample measured value Perform linear regression to establish the model curve, which is the internal standard method:

[0044] (3)

[0045] The internal standard method is widely used in the field of LIBS spectral analysis because it introduces matrix element spectral lines with relatively stable concentrations and is less affected by environmental disturbances, matrix effects, and self-absorption effects.

[0046] The above is a common application method, but in actual use, the laser induced breakdown spectroscopy component analyzer will produce system uncertainty due to equipment drift, and sample uncertainty due to sample inhomogeneity. If the factors affecting the uncertainty are not considered, the modeling effect is often not ideal.

[0047] For example, it is generally believed that the samples tested in a short period of time during a melt discharge process are uniform and consistent in composition. However, sampling experiments have shown that different elements have uneven sample distribution. If the unevenness reaches 0.2 (less than the preset deviation threshold), a deviation of 0.2 between two test results is still reasonable.

[0048] To solve the above problems, the present invention provides a method for establishing a quantitative detection model of laser induced breakdown spectroscopy, a medium, and a controller. The following describes the method for establishing a quantitative detection model of laser induced breakdown spectroscopy, a medium, and a controller in detail in conjunction with the accompanying drawings and specific implementations of the present invention.

[0049] Figure 1 FIG. 1 is a flow chart of a method for establishing a quantitative detection model according to an embodiment of the present invention. Figure 1 The method for establishing a quantitative detection model of laser induced breakdown spectroscopy may include:

[0050] S101, obtaining a sample data set, sample uncertainty and system uncertainty, and dividing the sample data set into a sample training set and a sample test set, wherein the sample data set includes original spectral data and its corresponding sample measured values;

[0051] S102, using the original spectral data in the sample training set and the corresponding sample measured values, establish an original quantitative detection model of the ratio of the spectral line intensity of the measured element to the internal standard element with respect to the sample measured value;

[0052] S103, using the original quantitative detection model to calculate the corresponding absolute deviation of each test sample in the sample test set, recording the test samples whose absolute deviation is greater than the sample uncertainty as the uncertainty set, and using the uncertainty set to fit the original quantitative detection model according to the sample uncertainty and system uncertainty to determine the target quantitative detection model.

[0053] In order to improve the accuracy of the laser induced breakdown spectroscopy quantitative detection model, the embodiment of the present invention establishes a laser induced breakdown spectroscopy quantitative detection model based on system uncertainty and sample uncertainty, and adds the influencing factors of uncertainty in the training process of the laser induced breakdown spectroscopy quantitative detection model to better fit the actual application of industrial materials and achieve better modeling effects.

[0054] Specifically, when obtaining the sample data set, sample uncertainty and system uncertainty, data accumulation can be performed in advance by means of experiments to accumulate the sample data set, sample uncertainty and system uncertainty. For example, three experiments can be used to accumulate the original spectral data and its corresponding sample measured value data, sample uncertainty data and system uncertainty data.

[0055] It should be noted that before obtaining the sample measured value in the sample data set, the measured element needs to be determined so as to obtain the concentration value of the measured element, that is, the sample measured value, through experiments.

[0056] The sample data set is divided into a sample training set and a sample test set. The sample training set is used to determine the original quantitative detection model, and the sample test set is used to adjust the original quantitative detection model to achieve model training with uncertainty for the laser induced breakdown spectroscopy quantitative detection model.

[0057] Specifically, the original spectral data in the sample training set are used to calculate the spectral line intensity of the measured element and the spectral line intensity of the internal standard element. Then the ratio of the spectral line intensity of the measured element to the spectral line intensity of the internal standard element is calculated, and a linear fit is performed based on the ratio of the spectral line intensity of the measured element and the internal standard element corresponding to the original spectral data in the sample training set and the actual measured value of the sample. The original quantitative detection model obtained by fitting uses the ratio of the spectral line intensity of the measured element and the internal standard element as the original horizontal coordinate, and the actual measured value of the sample obtained by the laboratory as the original vertical coordinate. The model curve obtained by linear regression is:

[0058]

[0059] in, is the measured value of the sample, is the spectral line intensity of the measured element, is the spectral line intensity of the internal standard element, K is a constant, and B is a constant.

[0060] Evaluate the original quantitative detection model, and use the original quantitative detection model to calculate the absolute deviation corresponding to each test sample in the sample test set. Determine whether the absolute deviation is greater than the sample uncertainty. If there is no absolute deviation greater than the sample uncertainty, the original quantitative detection model is recorded as the target quantitative detection model. If there is an absolute deviation greater than the sample uncertainty, the test sample with an absolute deviation greater than the sample uncertainty is recorded as an uncertainty set, and the uncertainty set is used to fit the original quantitative detection model within the uncertainty range according to the sample uncertainty and system uncertainty to obtain the target quantitative detection model.

[0061] In one embodiment of the present invention, Figure 2 As shown, obtaining a sample data set may include:

[0062] S201, using a laser to emit lasers to multiple detection points of industrial materials, and using a laser induced breakdown spectroscopy component analyzer to collect multiple raw spectral data, wherein the number of raw spectral data is greater than a preset threshold number;

[0063] S202, sampling industrial materials at each detection point, and performing chemical tests on the sampled materials to obtain a plurality of sample measured values, wherein the number of the sample measured values ​​is the same as the number of the original spectral data.

[0064] The embodiment of the present invention accumulates original spectral data and its corresponding sample measured values ​​through experiments.

[0065] In order to reduce the error of the accumulated raw spectral data and its corresponding sample measured value, sampling can be carried out during the stable period of industrial material discharge. Figure 3As shown, during the stable period of industrial material discharge, a laser is used to emit laser light, which is emitted to multiple detection points of industrial materials through an optical system, and a laser induced breakdown spectroscopy component analyzer is used to collect the original spectrum data of each detection point. The original spectrum data collected by the laser induced breakdown spectroscopy component analyzer is obtained by a computer. Among them, the computer is used to control the laser to emit laser light through a timing controller.

[0066] Sampling of industrial materials is performed at each detection point. Two samples can be taken in parallel each time, one for laboratory testing and one for standard retention. The industrial materials sent to the laboratory for testing are tested and tested to obtain multiple sample measured values ​​(concentration values ​​of the measured elements). It should be noted that when two samples are taken in parallel, it is necessary to ensure that each sample meets the laboratory testing requirements. Among them, standard retention is for subsequent sample inspection and retesting.

[0067] In an embodiment of the present invention, the preset threshold number is at least 60.

[0068] Exemplarily, when obtaining 80 data pairs (raw spectral data and its corresponding sample measured values), 80 detection points can be set on the surface of the industrial material at one time, and the raw spectral data and its corresponding sample measured values ​​can be obtained in the above manner. 10 laser detection points can also be set, and the raw spectral data and its corresponding sample measured values ​​can be obtained in the above manner, and the above process can be repeated 8 times. It should be noted that the embodiment of the present invention does not limit the number of preset thresholds and the sampling method of the raw spectral data and its corresponding sample measured values, and data accumulation can be performed according to actual needs.

[0069] In one embodiment of the present invention, Figure 4 As shown, obtaining sample uncertainty may include:

[0070] S301, repeatedly sampling n portions of industrial materials within a preset time, and performing chemical tests on the n portions of sampled materials to determine a single uncertainty, wherein n>1;

[0071] S302, repeating the step of sampling n portions of industrial materials within a preset time, and performing chemical testing on the n portions of sampled materials to determine the single uncertainty m times, wherein m>1;

[0072] S303, determining the sample uncertainty according to the m single uncertainties.

[0073] The embodiment of the present invention accumulates sample uncertainty data through experiments, and evaluates the accumulated sample uncertainty data to determine the sample uncertainty.

[0074] For example, sampling is performed during the stable period of industrial material discharge, such as 5 repeated samples of industrial materials within 5 minutes, and sent to the laboratory for testing, and the corresponding measured values ​​of each sampled material are obtained. The measured average value is calculated based on the corresponding measured values ​​of each sampled material, and the deviation between each measured value and the measured average value is counted, and the maximum deviation is recorded as the single uncertainty. Repeat the above process 6 times, and determine the sample uncertainty based on the 6 single uncertainties.

[0075] In one embodiment of the present invention, the single uncertainty is determined by performing a test on n portions of sampled materials, including:

[0076] Calculate the average of the measured values ​​obtained by testing n portions of sampled materials to obtain a measured average value;

[0077] Calculate the deviation between each measured value and the measured average value, and record the maximum deviation as the single uncertainty;

[0078] Determine the sample uncertainty based on m individual uncertainties, including:

[0079] Determine the maximum value among the m single uncertainties and record the maximum single uncertainty as the sample uncertainty.

[0080] Specifically, calculate the average value of the measured values ​​(concentration values ​​of the measured elements) obtained by the analysis of n samples of material to obtain the measured average value. Calculate each measured value With the determination of the average The maximum deviation is recorded as the single uncertainty. The single-shot uncertainty can be calculated using the following formula (4): :

[0081] (4)

[0082] in, is the nth measured value of a certain element in the reference sample by the laboratory, It is the average value of n measurements of a certain element in the reference sample by the laboratory.

[0083] Specifically, by comparing the uncertainty of each single , determine the maximum single uncertainty, and record the maximum single uncertainty as the sample uncertainty The sample uncertainty E of a certain element of the reference sample is calculated according to the following formula (5):

[0084] (5)

[0085] in, The number of reference samples for calculating sampling error, For the The sample uncertainty of an element in a reference sample.

[0086] It should be noted that a certain element in the reference sample is the same as the element being measured.

[0087] In one embodiment of the present invention, Figure 5 As shown, the system uncertainty is obtained, including:

[0088] S401, obtaining a plurality of measured spectral data obtained by the same operator measuring a standard sample of an industrial material for multiple consecutive times at the same time;

[0089] S402, determining the single characteristic spectral line intensity corresponding to each measured spectrum data, and calculating the average characteristic spectral line intensity of multiple single characteristic spectral line intensities;

[0090] S403, calculating the system uncertainty according to the single characteristic spectral line intensity and the average characteristic spectral line intensity.

[0091] The embodiment of the present invention accumulates system uncertainty data through experiments, and evaluates the accumulated system uncertainty data to determine the system uncertainty.

[0092] For example, at the same time, the same operator measures the industrial material standard sample for 11 consecutive times, and saves the 11 measured spectral data obtained by the measurement. Perform feature extraction on each measured spectral data obtained to obtain the single characteristic spectral line intensity (characteristic spectral line intensity of the measured element) corresponding to each measured spectral data, and calculate the average characteristic spectral line intensity of the 11 single characteristic spectral line intensities. And according to the following formula (6), the system uncertainty SD is calculated based on the single characteristic spectral line intensity and the average characteristic spectral line intensity.

[0093] (6)

[0094] in, To measure the intensity of a characteristic spectral line in a single measurement, To measure the average characteristic line intensity, n is the number of measurements.

[0095] When training a model with uncertainty in the embodiment of the present invention, the sample data set is divided into a sample training set and a sample test set in a ratio of 7:3 or 6:4, wherein the sample training set is used for model training and the sample test set is used for model evaluation.

[0096] Specifically, Figure 6 As shown, the original spectral data is feature extracted, the peak height or peak area of ​​the extracted spectral characteristic peak is calculated as the intensity, and the spectral line intensity of the measured element is calculated , spectral line intensity of internal standard element The ratio of the spectral line intensity of the measured element and the internal standard element is the original horizontal axis, with the actual measured value of the sample obtained in the laboratory Perform linear regression on the original ordinate and establish the original quantitative detection model curve of the ratio of the spectral line intensity of the measured element to the internal standard element with respect to the actual measured value of the sample:

[0097]

[0098] in, is the measured value of the sample, is the spectral line intensity of the measured element, is the spectral line intensity of the internal standard element, K is a constant, and B is a constant.

[0099] In one embodiment of the present invention, Figure 7 As shown, before dividing the sample data set into a sample training set and a sample test set, the method for establishing a laser induced breakdown spectroscopy quantitative detection model further includes:

[0100] For each piece of original spectral data in the sample training set, preprocessing the original spectral data;

[0101] The preprocessed raw spectral data is subjected to feature extraction and compared with the standard spectral line database. The spectral line peaks that meet the standard spectral line range are recorded as characteristic peaks, and the spectral line peaks that do not meet the standard spectral line range are recorded as non-characteristic peaks, thus obtaining the characteristic peaks and non-characteristic peaks of the raw spectral data.

[0102] Specifically, the raw spectral data in the sample data set is preprocessed, wherein the preprocessing includes at least one of spectral denoising, baseline correction and abnormal spectrum removal to eliminate noise, improve the signal-to-noise ratio and highlight the spectral features. Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA) and other methods can be used to extract features from the preprocessed raw spectral data, and the extracted characteristic spectral line intensity is compared with the standard spectral line database (National Institute of Standards and Technology nIST database). If it meets the standard spectral line range, it is determined as a characteristic peak and the positioning is successful; if it does not meet the standard spectral line range, it is determined as a non-characteristic peak and the result is discarded. By determining the characteristic peaks and non-characteristic peaks of each raw spectral data in the sample data set, it is convenient to calculate the spectral line intensity of the measured element and the internal standard element. That is, the peak height or peak area of ​​the characteristic peak is used as the intensity.

[0103] In one embodiment of the present invention, Figure 8 As shown, the original quantitative detection model is used to calculate the absolute deviation corresponding to each test sample in the sample test set, including:

[0104] S501, using the original quantitative detection model, calculating the model prediction value corresponding to the original spectral data of each test sample in the sample test set;

[0105] S502, calculating the absolute deviation according to the model prediction value and the sample measured value corresponding to the original spectral data of each test sample in the sample test set.

[0106] Evaluate the original quantitative detection model, and use the obtained original quantitative detection model to calculate the model prediction value corresponding to the original spectral data of each test sample in the sample test set .

[0107] The original quantitative detection model was evaluated according to the following absolute deviation formula (7).

[0108] (7)

[0109] in, is the absolute deviation, is the measured value of the sample, is the model's predicted value.

[0110] If there is an absolute deviation For test samples that exceed the sample uncertainty E, it is determined that the original quantitative detection model needs to be optimized and all absolute deviations are output. The set of samples m (uncertainty set) that exceeds the sample uncertainty E. If there is no absolute deviation For test samples with uncertainty greater than the sample uncertainty E, the original quantitative detection model is recorded as the target quantitative detection model.

[0111] In one embodiment of the present invention, Figure 6 As shown, the uncertainty set is used to fit the original quantitative detection model according to the sample uncertainty and system uncertainty to determine the target quantitative detection model, including:

[0112] According to the system uncertainty, the horizontal coordinate of each uncertain sample in the uncertainty set is determined, and according to the sample uncertainty, the vertical coordinate of each uncertain sample in the uncertainty set is determined to obtain a plurality of adjusted coordinates corresponding to each uncertain sample;

[0113] The original quantitative detection model is fitted according to the multiple adjustment coordinates corresponding to each uncertain sample, and the model with a goodness of fit close to 1 is recorded as the target quantitative detection model.

[0114] It should be noted that the original horizontal coordinate corresponding to the original quantitative detection model Recorded as x, the original ordinate corresponding to the original quantitative detection model Denoted as y.

[0115] Specifically, taking a single point (a sample point in the uncertainty set) as an example, calculate the ratio of the spectral line intensity of the measured element and the internal standard element at this single point , recorded as the horizontal coordinate x0 of the single point. Considering the uncertainty, the horizontal coordinate of the single point becomes [x0-SD, x0+SD]. Similarly, the vertical coordinate of the single point becomes [y0-E, y0+E]. The values ​​are taken within the horizontal and vertical coordinate ranges to determine the multiple adjustment coordinates corresponding to the single point.

[0116] All points are processed as above, the original quantitative detection model is trained within the uncertainty range, and all goodness of fit are output. , as shown in formula (8):

[0117] (8)

[0118] in, is the goodness of fit, is the model prediction value, It is the measured value of the sample.

[0119] The calculated goodness of fit R2 is compared, and the model training result with the goodness of fit R2 closest to 1 is selected as the target quantitative detection model.

[0120] The following is a specific example to illustrate the target quantitative detection model:

[0121] There are ten test samples in the sample test set. For example, there are original points 1 to 10. After the original quantitative model is established and evaluated, it is found that the absolute deviations of points 3, 4, and 5 are greater than the sample uncertainty. Keep the other points in the sample test set unchanged, expand the horizontal and vertical coordinates of points 3, 4, and 5 to the average points between [x0-SD, x0+SD] and [y0-E, y0+E] (this is to reduce the amount of calculation) and then fit. For example, there are 9 combinations (9 average points) of point 3 from x0, y0 to [x0-SD, x0+SD], [y0-E, y0+E], then there are 9*9*9 combinations of points 3, 4, and 5. Each outputs a goodness of fit. Compare the above goodness of fit and select the model training result closest to 1.

[0122] Fig. 9 This is a curve diagram of the original quantitative detection model of an embodiment of the present invention. Before implementation, due to system drift and sample inhomogeneity, the model training goodness of fit was 96.84%. The quantitative detection model with uncertainty was established using the method of the present invention, and the effect was as follows: Fig.10 The goodness of fit is 99.09%, and the modeling effect is significantly improved.

[0123] The method for establishing a quantitative detection model of laser induced breakdown spectroscopy in an embodiment of the present invention adds uncertainty influencing factors to the quantitative detection model of a laser induced breakdown spectroscopy (LIBS) component analyzer to better fit the actual application of industrial materials, achieve better modeling effects, and improve detection accuracy.

[0124] The present invention provides a device for establishing a quantitative detection model of laser induced breakdown spectroscopy. The device may include:

[0125] like Fig.11 As shown, the device for establishing a quantitative detection model of laser induced breakdown spectroscopy may include a data accumulation module, an uncertainty evaluation module and a model training module with uncertainty.

[0126] The data accumulation module is used to obtain a sample data set, wherein the sample data set includes original spectral data and its corresponding sample measured values; the uncertainty evaluation module allows the user to obtain sample uncertainty and system uncertainty; the model training module with uncertainty is used to divide the sample data set into a sample training set and a sample test set; the original spectral data in the sample training set and its corresponding sample measured values ​​are used to establish an original quantitative detection model of the ratio of the spectral line intensity of the measured element to the internal standard element with respect to the sample measured value; the sample test set is used to calculate the absolute deviation of the original quantitative detection model, and the test samples with absolute deviations greater than the sample uncertainty are recorded as uncertainty sets; the uncertainty set is used to fit the original quantitative detection model according to the sample uncertainty and system uncertainty to determine the target quantitative detection model.

[0127] like Fig.11 As shown, the device for establishing a quantitative detection model of laser induced breakdown spectroscopy may also include a spectrum preprocessing module and a feature extraction module.

[0128] The spectrum preprocessing module is used to preprocess the raw spectrum data in the sample training set. The feature extraction module extracts features from the preprocessed raw spectrum data and compares it with the standard spectrum line database. If it meets the standard spectrum line range, it is determined to be a characteristic peak; if it does not meet the standard spectrum line range, it is determined to be a non-characteristic peak.

[0129] The present invention provides a computer-readable storage medium.

[0130] In this embodiment, a computer program is stored on a computer-readable storage medium. When the computer program is executed by a processor, the above-mentioned method for establishing a quantitative detection model of laser induced breakdown spectroscopy is implemented.

[0131] The invention provides a controller.

[0132] In this embodiment, the controller may include a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the above-mentioned method for establishing a quantitative detection model of laser induced breakdown spectroscopy is implemented.

[0133] Fig.12 4 is a structural block diagram of a controller according to an embodiment of the present invention.

[0134] like Fig.12 As shown, the controller 500 includes: a processor 501 and a memory 503. The processor 501 and the memory 503 are connected, such as through a bus 502. Optionally, the controller 500 may also include a transceiver 504. It should be noted that in actual applications, the transceiver 504 is not limited to one, and the structure of the controller 500 does not constitute a limitation on the embodiments of the present invention.

[0135] The processor 501 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of the present invention. The processor 501 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0136] The bus 502 may include a path to transmit information between the above components. The bus 502 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 502 may be divided into an address bus, a data bus, a control bus, etc. For convenience, Fig.12 There is only one thick line in the diagram, but it does not mean that there is only one bus or one type of bus.

[0137] The memory 503 is used to store a computer program corresponding to the method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to the above embodiment of the present invention, and the computer program is controlled and executed by the processor 501. The processor 501 is used to execute the computer program stored in the memory 503 to implement the contents shown in the above method embodiment. Fig.12 The controller 500 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0138] The computer-readable storage medium and controller of the embodiment of the present invention utilize the above-mentioned laser induced breakdown spectroscopy quantitative detection model establishment method to increase the influencing factors of uncertainty. The established target quantitative detection model is more in line with the actual application of industrial materials, achieves better modeling effects, and improves detection accuracy.

[0139] It should be noted that the logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, device or equipment (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or equipment and execute instructions), or in combination with these instruction execution systems, devices or equipment. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or equipment or in combination with these instruction execution systems, devices or equipment. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAm), a read-only memory (ROm), an erasable and programmable read-only memory (EPROm or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROm). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0140] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0141] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0142] In the description of the present invention, it is to be understood that the terms “center”, “longitudinal”, “lateral”, “length”, “width”, “thickness”, “up”, “down”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside”, “outside”, “clockwise”, “counterclockwise”, “axial”, “radial”, “circumferential”, etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0143] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0144] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0145] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0146] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A method for establishing a quantitative detection model of laser induced breakdown spectroscopy, characterized in that: The method comprises: Acquire a sample data set, sample uncertainty and system uncertainty, and divide the sample data set into a sample training set and a sample test set, wherein the sample data set includes original spectral data and its corresponding sample measured values; Using the original spectral data in the sample training set and the corresponding sample measured values, an original quantitative detection model of the ratio of the spectral line intensity of the measured element to the internal standard element with respect to the sample measured value is established; The original quantitative detection model is used to calculate the absolute deviation corresponding to each test sample in the sample test set, and the test samples whose absolute deviation is greater than the sample uncertainty are recorded as an uncertainty set. The uncertainty set is used to fit the original quantitative detection model according to the sample uncertainty and the system uncertainty to determine the target quantitative detection model, wherein the uncertainty set is used to fit the original quantitative detection model according to the sample uncertainty and the system uncertainty to determine the target quantitative detection model, including: Determine the horizontal coordinate of each uncertain sample in the uncertainty set according to the system uncertainty, and determine the vertical coordinate of each uncertain sample in the uncertainty set according to the sample uncertainty, to obtain a plurality of adjusted coordinates corresponding to each uncertain sample; The original quantitative detection model is fitted according to the multiple adjustment coordinates corresponding to each of the uncertain samples, and the model with a goodness of fit close to 1 is recorded as the target quantitative detection model.

2. The method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to claim 1, characterized in that: The step of obtaining a sample data set comprises: Using a laser to emit lasers to multiple detection points of industrial materials, and using a laser induced breakdown spectroscopy component analyzer to collect a plurality of the raw spectral data, wherein the number of the raw spectral data is greater than a preset threshold number; The industrial material is sampled at each detection point, and the sampled material is tested to obtain a plurality of sample measured values, wherein the sample measured values ​​correspond one-to-one to the original spectral data.

3. The method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to claim 1, characterized in that: Obtain sample uncertainty, including: Repeatedly sample n portions of industrial materials within a preset time, and perform laboratory tests on the n portions of sampled materials to determine the single uncertainty, where n>1; Repeat the steps of sampling the industrial material n times within a preset time, and testing the n samples to determine the single uncertainty m times, wherein m>1; The sample uncertainty is determined based on the m single uncertainties.

4. The method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to claim 3, characterized in that: The method of performing chemical testing on n portions of sampled materials to determine a single uncertainty includes: Calculate the average of the measured values ​​obtained by testing n portions of sampled materials to obtain a measured average value; Calculate the deviation between each of the measured values ​​and the measured average value, and record the maximum deviation as the single uncertainty; The determining the sample uncertainty according to the m single uncertainties comprises: Determine the maximum value among the m single uncertainties, and record the maximum single uncertainty as the sample uncertainty.

5. The method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to claim 1, characterized in that: Obtain system uncertainties, including: Acquire multiple measured spectral data obtained by the same operator measuring the standard sample of industrial materials multiple times in succession at the same time; Determine the single characteristic spectral line intensity corresponding to each of the measured spectrum data, and calculate the average characteristic spectral line intensity of a plurality of the single characteristic spectral line intensities; The system uncertainty is calculated according to the single characteristic spectral line intensity and the average characteristic spectral line intensity.

6. The method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to claim 1, characterized in that: Before dividing the sample data set into a sample training set and a sample test set, the method further comprises: For each piece of original spectral data in the sample data set, preprocessing the original spectral data; The preprocessed raw spectral data is subjected to feature extraction and compared with the standard spectral line database, the spectral line peaks that meet the standard spectral line range are recorded as characteristic peaks, and the spectral line peaks that do not meet the standard spectral line range are recorded as non-characteristic peaks, thereby obtaining the characteristic peaks and non-characteristic peaks of the raw spectral data.

7. The method for establishing a quantitative detection model of laser induced breakdown spectroscopy according to claim 1, characterized in that: The calculating the absolute deviation corresponding to each test sample in the sample test set by using the original quantitative detection model includes: Using the original quantitative detection model, calculating the model prediction value corresponding to the original spectral data of each test sample in the sample test set; The absolute deviation is calculated based on the model prediction value and the sample measured value corresponding to the original spectral data of each test sample in the sample test set.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for establishing a laser induced breakdown spectroscopy quantitative detection model according to any one of claims 1 to 7 is implemented.

9. A controller, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that: When the computer program is executed by the processor, the method for establishing a laser induced breakdown spectroscopy quantitative detection model according to any one of claims 1 to 7 is implemented.

Images