Method and system for detecting flame retardant rating of glass fiber reinforced unsaturated polyester molding compound

Abstract

The invention discloses a method and a system for detecting the flame retardant grade of a glass fiber reinforced unsaturated polyester molding compound. The method comprises the following steps: collecting glass fiber reinforced unsaturated polyester molding compound samples with different flame retardant grades; the method comprises the following steps: acquiring spectral data of a glass fiber reinforced unsaturated polyester molding compound sample through a laser-induced breakdown spectroscopy technology; processing the spectroscopic data by adopting a principal component analysis method, and extracting characteristic dimensions of the spectroscopic data; based on the spectral data feature dimension, establishing a flame retardant grade prediction model, and training and verifying the model; and inputting the spectral data of the new sample into the trained flame retardant grade prediction model to obtain a flame retardant grade prediction result. According to the method, the relationship between the spectral data and the flame retardant rating of the SMC sample is established, so that the flame retardant rating of the SMC material can be rapidly detected by utilizing the LIBS technology.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring the flame retardant grade of materials for electric energy metering boxes, and particularly to a method and system for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compounds. Background Art

[0002] With the construction of a new power system and the large-scale use of smart meters, the State Grid Corporation has put forward higher requirements for the collection and management of power information, and it is necessary to quickly detect the status information of power equipment such as electric energy metering boxes. Aging of electrical circuits, overloading of electricity consumption, humidity, incorrect installation, unauthorized wiring, etc. may all cause the electric energy metering box to catch fire, posing a major threat to personal and property safety. Therefore, it is necessary to detect the flame retardant performance of the meter box to reduce the risk of fire.

[0003] The shell of the electric energy metering box mostly uses glass fiber reinforced unsaturated polyester molding compound, which has the advantages of corrosion resistance, good insulation performance, light weight, etc. However, during use, it will be affected by many environmental factors such as ultraviolet rays, oxygen, ozone, water, temperature, humidity, microorganisms, etc., and a series of complex changes that are not conducive to product use will occur. The aging of non-metallic materials is mainly manifested as color change of appearance, surface cracking and warping deformation, decline of physical properties and flame retardant properties, etc. The aging of materials will seriously affect the appearance and performance indicators of products and greatly shorten the service life of products. The material targeted by the present invention is SMC (glass fiber reinforced unsaturated polyester molding compound), a typical material for non-metallic electric energy metering boxes. At present, the conventional method for detecting the flame retardant grade is to measure through a vertical burning test, which has strict requirements for sample size, experimental conditions, etc., and is time-consuming and laborious, and cannot meet the needs of rapid detection. Therefore, it is very necessary to propose a rapid detection method to complete the rapid detection of the flame retardant grade of the materials for electric energy metering boxes, improve the detection efficiency, and enrich the quality control means for electric energy metering boxes. Summary of the Invention

[0004] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this part, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this part, the abstract, and the title, and such simplifications or omissions shall not be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a method and system for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound to solve the problems of long detection period and insufficient manpower and material resources in the conventional method for detecting the flame retardant grade.

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound, comprising:

[0009] Collecting glass fiber reinforced unsaturated polyester molding compound samples with different flame retardant grades;

[0010] Based on the glass fiber reinforced unsaturated polyester molding compound samples, obtaining spectral data of the samples by laser induced breakdown spectroscopy;

[0011] Processing the spectral data by principal component analysis to extract the characteristic dimensions of the spectral data;

[0012] Based on the characteristic dimensions of the spectral data, establishing a flame retardant grade prediction model, and training and validating the model;

[0013] Inputting the spectral data of the new sample into the trained flame retardant grade prediction model to obtain the flame retardant grade prediction result.

[0014] As a preferred scheme of the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to the present invention, wherein:

[0015] The laser induced breakdown spectroscopy includes the following steps:

[0016] Focusing the laser energy on the glass fiber reinforced unsaturated polyester molding compound through an optical path to ablate the material surface to generate plasma;

[0017] The plasma radiates atomic and ionic line spectra during the cooling process;

[0018] Collecting spectral signals through an optical fiber, and then analyzing the elemental composition and state characteristics of the glass fiber reinforced unsaturated polyester molding compound according to the collected characteristic spectral information.

[0019] As a preferred scheme of the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to the present invention, wherein:

[0020] The processing of the spectral data by principal component analysis includes the following steps:

[0021] The original data set X = {x1, x2,..., x n}, where x i is a vector with a length of m;

[0022] Normalizing the data;

[0023] Calculating the covariance matrix;

[0024] Solving the eigenvalues and eigenvectors of the covariance matrix;

[0025] Sort according to the eigenvalue magnitudes, select the largest k eigenvalues, and form an m×k dimensional matrix P with the eigenvectors corresponding to the eigenvalues as column vectors;

[0026] Multiply the original dataset X by the eigenvector matrix P to obtain the principal component matrix of the original data, and retain the principal components corresponding to the cumulative variance of more than 95%.

[0027] As a preferred embodiment of the method for detecting the flame retardant grade of the glass fiber reinforced unsaturated polyester molding compound described in the present invention, wherein:

[0028] The standardization processing of the data includes implementing mean centering and unit variance scaling on the original spectral data.

[0029] As a preferred embodiment of the method for detecting the flame retardant grade of the glass fiber reinforced unsaturated polyester molding compound described in the present invention, wherein:

[0030] The calculation of the covariance matrix is expressed as:

[0031]

[0032] The principal component matrix of the original data is expressed as:

[0033] Y = XP.

[0034] As a preferred embodiment of the method for detecting the flame retardant grade of the glass fiber reinforced unsaturated polyester molding compound described in the present invention, wherein:

[0035] The establishment of the flame retardant grade prediction model includes the following steps:

[0036] The number of input layer nodes is determined by the dimensionality of the spectral data features after principal component processing;

[0037] The first hidden layer is set to 50 nodes, and the second hidden layer is 30 nodes;

[0038] The number of output layer nodes corresponds to the number of categories of the flame retardant grade;

[0039] Adopt the cross-entropy loss function to measure the difference between the model output probability distribution and the true label distribution;

[0040] Select the stochastic gradient descent as the optimization algorithm, update the model parameters by calculating the gradient on each small batch of data, and in the parameter settings, the maximum number of iterations is 150 and the initial learning rate is 0.02.

[0041] As a preferred embodiment of the method for detecting the flame retardant grade of the glass fiber reinforced unsaturated polyester molding compound described in the present invention, wherein:

[0042] The training of the model includes training a neural network model using training set data, continuously adjusting the weights through multiple iterations, and learning how to map spectral features to corresponding flame retardant grades.

[0043] In a second aspect, the present invention provides a flame retardant grade detection system for glass fiber reinforced unsaturated polyester molding compound, including:

[0044] A sample collection module for collecting samples of glass fiber reinforced unsaturated polyester molding compound with different flame retardant grades;

[0045] A spectral data collection module for obtaining spectral data of the sample based on the glass fiber reinforced unsaturated polyester molding compound sample through laser induced breakdown spectroscopy;

[0046] A feature extraction module for processing the spectral data by using principal component analysis method to extract the spectral data feature dimension;

[0047] A model establishment, training and verification module for establishing a flame retardant grade prediction model based on the spectral data feature dimension, and training and verifying the model;

[0048] An acquisition module for inputting the spectral data of a new sample into the trained flame retardant grade prediction model to obtain the flame retardant grade prediction result.

[0049] In a third aspect, the present invention provides a computing device, including:

[0050] A memory for storing programs;

[0051] A processor for executing the computer executable instructions, and when the computer executable instructions are executed by the processor, the steps of the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound are implemented.

[0052] In a fourth aspect, the present invention provides a computer readable storage medium, including: when the program is executed by the processor, the steps of the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound are implemented.

[0053] The beneficial effects of the present invention: The present invention focuses laser energy on the material surface through an optical path, ablates the material surface to vaporize a small amount of the sample to form a plasma with high temperature and high electron density. The plasma radiates a large number of atomic and ionic line spectra during the cooling process. The spectral signal is collected through an optical fiber, and then the elemental composition and state characteristics of the ablated material are analyzed based on the collected characteristic spectral information. By analyzing the relationship between the LIBS spectral characteristics and the flame retardant grade of the sample, a model for predicting the flame retardant grade based on spectral data is established, which can realize the rapid detection of the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound. Description of the Drawings

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Among them:

[0055] Figure 1 It is a schematic diagram of the basic process of a method for detecting the flame retardant grade of a glass fiber reinforced unsaturated polyester molding compound provided by an embodiment of the present invention;

[0056] Figure 2 It is a diagram of the LIBS system of a method for detecting the flame retardant grade of a glass fiber reinforced unsaturated polyester molding compound provided by an embodiment of the present invention;

[0057] Figure 3 It is a flow chart for predicting the flame retardant grade of an SMC sample of a method for detecting the flame retardant grade of a glass fiber reinforced unsaturated polyester molding compound provided by an embodiment of the present invention;

[0058] Figure 4 It is a result diagram of training the spectral data of an SMC sample and predicting the flame retardant grade using a neural network for a method for detecting the flame retardant grade of a glass fiber reinforced unsaturated polyester molding compound provided by an embodiment of the present invention; Detailed implementation manners

[0059] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific implementation manners of the present invention with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0060] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0061] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selectively exclusive embodiment from other embodiments.

[0062] The present invention will be described in detail with reference to the schematic diagrams. When describing the embodiments of the present invention in detail, for the convenience of explanation, the cross-sectional views showing the device structure will be enlarged locally in a non-general proportion, and the schematic diagrams are only examples and should not limit the scope of protection of the present invention herein. In addition, in actual production, three-dimensional spatial dimensions including length, width, and depth should be included.

[0063] At the same time, in the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "upper, lower, inner, and outer" is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation to the present invention. In addition, the terms "first, second, or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0064] Unless otherwise clearly defined and limited in the present invention, the terms "mounted, connected, and coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can also be a mechanical connection, an electrical connection, or a direct connection, and can also be indirectly connected through an intermediate medium, or can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0065] Embodiment 1

[0066] Referring to Figures 1-3 , an embodiment of the present invention provides a method for detecting the flame retardant grade of glass fiber-reinforced unsaturated polyester molding compound. As Figure 1 , shown in FIG. 3, it includes the following steps:

[0067] S1: Collect glass fiber-reinforced unsaturated polyester molding compound samples with different flame retardant grades;

[0068] S2: Based on the glass fiber-reinforced unsaturated polyester molding compound samples, obtain the spectral data of the samples through laser-induced breakdown spectroscopy technology;

[0069] In the embodiment of the present application, as Figure 2 , shown in FIG., the LIBS system mainly consists of four parts: a laser, an optical path system, a controller, and a spectrometer. By selecting appropriate laser energy, light collection angle, and spectrometer delay time, spectral signals with high signal-to-noise ratio and signal-to-background ratio can be obtained.

[0070] In the embodiments of the present application, during the determination of laser-induced breakdown spectroscopy (LIBS), restricted by sample properties (surface roughness and hardness), environmental differences (temperature, air pressure, and light), and hardware device performance (stability of laser output energy and resolution of spectrometer light collection), the original spectra collected by LIBS usually have a relatively high baseline and noise, which affects the stability of spectral signals and is not conducive to the extraction of effective spectral signals and subsequent analysis. Additionally, due to differences in the response characteristics of different channels of the spectrometer, there are obvious demarcation points between the spectral data of different channels. Therefore, before analyzing the spectral lines, it is necessary to perform baseline correction and smoothing denoising on the original spectra to improve the quality of the spectra.

[0071] S3: Process the spectral data using the principal component analysis method to extract the characteristic dimensions of the spectral data;

[0072] In the embodiments of the present application, the principal component analysis method is used to process the spectral data. Its basic idea is to transform a set of variables that may be correlated through an orthogonal transformation into a set of linearly uncorrelated variables. The transformed variables are called principal components. Through principal component analysis, correlated variables can be transformed into linearly uncorrelated variables, and the amount of information contained in the transformed variables is reflected by the variance. The larger the variance, the more information is contained. By sorting the variances of the transformed variables, it can be found that a few variables with the largest variances can already contain most of the information of the original data. Therefore, more variables with very small variances can be discarded without significantly affecting the final data analysis results.

[0073] The main process of using principal component analysis for data dimensionality reduction is as follows:

[0074] The original data set X = {x1, x2,..., x n}, where x i is a vector of length m.

[0075] (1) First, standardize the data; Since glass fiber-reinforced unsaturated polyester molding compounds may exhibit different spectral characteristics due to factors such as manufacturing processes and types of additives, special attention needs to be paid to the standardization process to ensure comparability between different samples. Considering the possible non-uniformity and aging phenomena on the material surface, perform mean centering and unit variance scaling on the original spectral data to eliminate the influence of these factors.

[0076] (2) Calculate the covariance matrix:

[0077]

[0078] (3) Calculate the eigenvalues and eigenvectors of the covariance matrix. For SMC materials, the presence and concentration of certain specific elements (such as halogens, phosphorus, nitrogen, etc.) have a direct impact on the flame retardancy of the materials. Therefore, when solving the eigenvalues and eigenvectors, special attention should be paid to the wavelength regions related to these key elements to ensure that they are fully considered in subsequent analyses.

[0079] (4) Select the largest k eigenvalues, and use the eigenvectors corresponding to these eigenvalues as column vectors to form an m×k dimensional matrix P;

[0080] (5) Multiply the original data set X by the eigenvector matrix P to obtain the principal component matrix of the original data:

[0081] Y = XP

[0082] The obtained principal component matrix Y will be used to train machine learning models, especially those classifiers aimed at predicting the flame retardancy grade of SMC materials. In this way, the data dimension can be effectively reduced while retaining the information that best represents the flame retardant characteristics of the materials, thereby improving the efficiency and accuracy of the model. At least retain the principal components corresponding to more than 95% of the cumulative variance as the objects of subsequent data analysis.

[0083] S4: Based on the spectral data feature dimension, establish a flame retardancy grade prediction model, and train and validate the model;

[0084] The number of input layer nodes of the three-layer neural network is determined according to the spectral data feature dimension after principal component processing. The first hidden layer has 50 nodes, the second hidden layer has 30 nodes, and the number of output layer nodes corresponds to the number of categories of the flame retardancy grade. The cross-entropy loss function is used in training, which is suitable for classification tasks and can measure the difference between the model output probability distribution and the true label distribution. The optimization algorithm selects stochastic gradient descent, which updates the model parameters by calculating the gradient on each small batch of data. The maximum number of iterations in the parameter settings is 150, and the initial learning rate is 0.02. These selections are because the cross-entropy loss is sensitive to the classification effect, stochastic gradient descent has high computational efficiency and is easy to implement, while the number of iterations and learning rate are initially set according to experience and data characteristics, and can be adjusted later according to the performance on the validation set.

[0085] Use the results of the validation set to adjust hyperparameters such as the learning rate and batch size. If the model performance on the validation set improves slowly or overfitting occurs, the learning rate can be appropriately reduced to make the model update more cautiously; if the performance improves but the convergence speed is slow, the batch size can be appropriately increased to improve the stability of gradient estimation. By continuous trial and verification, find the combination of hyperparameters that makes the model perform optimally on the validation set, thereby optimizing the model performance and improving its generalization ability on unknown data.

[0086] In fact, the machine learning model will automatically learn from a large amount of training data which spectral features can most effectively distinguish different flame retardant grades. Moreover, due to the complexity and diversity of materials, even materials of the same category may exhibit different spectral features, so the learning process of the model is crucial.

[0087] To accurately link spectral features with flame retardant grades, actual data analysis must be relied on. By collecting a large number of samples verified by UL94 tests and using LIBS technology to obtain their spectral information, and then using machine learning algorithms for pattern recognition and classification, a reliable prediction model can be established. After the model training is completed, its accuracy and reliability can be tested by evaluating the performance of the model on unknown samples.

[0088] S5: Input the spectral data of the new sample into the trained flame retardant grade prediction model to obtain the flame retardant grade prediction result.

[0089] This embodiment also provides a detection system for the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound, including:

[0090] A sample collection module for collecting samples of glass fiber reinforced unsaturated polyester molding compound with different flame retardant grades;

[0091] A spectral data collection module for obtaining the spectral data of the sample based on the glass fiber reinforced unsaturated polyester molding compound sample through laser induced breakdown spectroscopy technology;

[0092] A feature extraction module for processing the spectral data by using the principal component analysis method to extract the spectral data feature dimension;

[0093] A model establishment and training verification module for establishing a flame retardant grade prediction model based on the spectral data feature dimension and training and verifying the model;

[0094] An acquisition module for inputting the spectral data of the new sample into the trained flame retardant grade prediction model to obtain the flame retardant grade prediction result.

[0095] Furthermore, it also includes:

[0096] A memory for storing programs;

[0097] A processor for loading the program to execute the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound.

[0098] This embodiment also provides a computer-readable storage medium, which stores a program, and when the program is executed by a processor, the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound is implemented.

[0099] The storage medium proposed in this embodiment and the flame retardant grade detection method of the glass fiber reinforced unsaturated polyester molding compound proposed in the above embodiment belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0100] From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general-purpose hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that makes a contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disc of a computer, etc., including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0101] Embodiment 2

[0102] Referring to Figure 4 , which is an embodiment of the present invention, provides a method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound. In order to verify its beneficial effects, scientific demonstration is carried out through specific implementation methods and implementation effects.

[0103] The specific content of this embodiment is as follows:

[0104] Training set:

[0105] Among the samples with an actual flame retardant grade of V-0, 1092 were correctly predicted as V-0, and 33 were incorrectly predicted as V-1.

[0106] Among the samples with an actual flame retardant grade of V-1, 40 were incorrectly predicted as V-0, 660 were correctly predicted as V-1, and 50 were incorrectly predicted as V-2.

[0107] Among the samples with an actual flame retardant grade of V-2, 1 was incorrectly predicted as V-0, 50 were incorrectly predicted as V-1, and 699 were correctly predicted as V-2.

[0108] Validation set:

[0109] Among the samples with an actual flame retardant grade of V-0, 219 were correctly predicted as V-0, and 6 were incorrectly predicted as V-1.

[0110] Among the samples with an actual flame retardancy rating of V-1, 6 were mispredicted as V-0, 135 were correctly predicted as V-1, and 9 were mispredicted as V-2.

[0111] Among the samples with an actual flame retardancy rating of V-2, 1 was mispredicted as V-0, 10 were mispredicted as V-1, and 139 were correctly predicted as V-2.

[0112] Test set:

[0113] Among the samples with an actual flame retardancy rating of V-0, 148 were correctly predicted as V-0, and 2 were mispredicted as V-1.

[0114] Among the samples with an actual flame retardancy rating of V-1, 9 were mispredicted as V-0, 88 were correctly predicted as V-1, and 3 were mispredicted as V-2.

[0115] Among the samples with an actual flame retardancy rating of V-2, 5 were mispredicted as V-1, and 95 were correctly predicted as V-2.

[0116] Total dataset:

[0117] Among the samples with an actual flame retardancy rating of V-0, 1459 were correctly predicted as V-0, and 41 were mispredicted as V-1.

[0118] Among the samples with an actual flame retardancy rating of V-1, 55 were mispredicted as V-0, 883 were correctly predicted as V-1, and 62 were mispredicted as V-2.

[0119] Among the samples with an actual flame retardancy rating of V-2, 2 were mispredicted as V-0, 65 were mispredicted as V-1, and 933 were correctly predicted as V-2.

[0120] These confusion matrices show the prediction performance of the model on different datasets, including accuracy and other evaluation metrics. From these data, the prediction ability and generalization ability of the model can be evaluated.

[0121] From Figure 4 it can be concluded that the prediction accuracy of the neural network for the flame retardancy rating of SMC samples is about 90%, and the generalization performance on the test set is good.

[0122] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound, characterized in that: include: Collect samples of glass fiber reinforced unsaturated polyester molding compounds with different flame retardant grades; Based on the glass fiber reinforced unsaturated polyester molding compound sample, the spectral data of the sample was obtained by laser induced breakdown spectroscopy technology; The principal component analysis method is used to process the spectral data and extract the characteristic dimensions of the spectral data; Based on the characteristic dimensions of spectral data, a flame retardant grade prediction model is established, and the model is trained and verified; The spectral data of the new sample is input into the trained flame retardant grade prediction model to obtain the flame retardant grade prediction result.

2. The method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to claim 1, characterized in that: The laser induced breakdown spectroscopy technique comprises the following steps: Focus the laser energy on the glass fiber reinforced unsaturated polyester molding compound through the optical path, ablate the material surface to generate plasma; The plasma radiates a line spectrum of atoms and ions as it cools; The spectral signals are collected through optical fibers, and then the elemental composition and state characteristics of the glass fiber reinforced unsaturated polyester molding compound are analyzed based on the collected characteristic spectral information.

3. The method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to claim 1 or 2, characterized in that: The method of processing the spectral data by principal component analysis comprises the following steps: The original data set X = {x1, x2, ..., x n }, where x i is a vector of length m; Standardize the data; Calculate the covariance matrix; Solve for the eigenvalues ​​and eigenvectors of the covariance matrix; Sort by eigenvalue, select the largest k eigenvalues, and use the eigenvectors corresponding to the eigenvalues ​​as column vectors to form an m×k dimensional matrix P; The original data set X is multiplied by the eigenvector matrix P to obtain the principal component matrix of the original data, retaining the principal components corresponding to more than 95% of the cumulative variance.

4. The method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to claim 3, characterized in that: The data standardization process includes performing mean centering and unit variance scaling on the original spectral data.

5. The method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to claim 4, characterized in that: The calculated covariance matrix is ​​expressed as: The principal component matrix of the original data is expressed as: Y=XP.

6. The method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to claim 5, characterized in that: The flame retardant grade prediction model is established, comprising the following steps: The number of nodes in the input layer is determined by the characteristic dimension of the spectral data after principal component processing; The first hidden layer is set to 50 nodes and the second hidden layer is set to 30 nodes; The number of nodes in the output layer corresponds to the number of categories of flame retardancy levels; The cross entropy loss function is used to measure the difference between the model output probability distribution and the true label distribution; Stochastic gradient descent is selected as the optimization algorithm. The model parameters are updated by calculating the gradient on each small batch of data. The maximum number of iterations in the parameter setting is 150 and the initial learning rate is 0.

02.

7. The method for detecting the flame retardancy grade of glass fiber reinforced unsaturated polyester molding compound according to claim 6, characterized in that: The training of the model includes training the neural network model using the training set data, continuously adjusting the weights through multiple iterations, and learning how to map the spectral features to the corresponding flame retardant levels.

8. A system based on the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound according to claim 1, characterized in that: Sample collection module, used to collect glass fiber reinforced unsaturated polyester molding compound samples with different flame retardant grades; Spectral data acquisition module, used to obtain spectral data of samples based on glass fiber reinforced unsaturated polyester molding compound samples through laser induced breakdown spectroscopy technology; A feature extraction module is used to process the spectral data using the principal component analysis method to extract the characteristic dimensions of the spectral data; Model building and training verification module, which is used to build a flame retardant grade prediction model based on the characteristic dimensions of spectral data, and to train and verify the model; The acquisition module is used to input the spectral data of the new sample into the trained flame retardant grade prediction model to obtain the flame retardant grade prediction result.

9. A computing device, characterized in that include: Memory, used to store programs; A processor is used to load the program to execute the steps of the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a program, characterized in that: When the program is executed by a processor, the steps of the method for detecting the flame retardant grade of glass fiber reinforced unsaturated polyester molding compound as described in any one of claims 1 to 7 are implemented.

Images