An online automatic analysis method for the composition of steel based on X-ray fluorescence spectrometer

Abstract

The present invention relates to an online automatic analysis method for the composition of steel based on an X-ray fluorescence spectrometer. The method includes: M1. Detecting the steel online, obtaining the data information of the temperature of the steel in real time based on an infrared thermometer, obtaining the data information of the elemental analysis of the steel in real time based on the X-ray fluorescence spectrometer, and obtaining the data information of the measurement position of the X-ray fluorescence spectrometer in real time based on a laser sensor, and using an improved whale optimization-based lightweight gradient boosting machine classification prediction algorithm to predict the content of each element in the steel, so as to obtain the data information of the content of each element in the predicted steel. The present invention not only performs composition analysis operations automatically and standardizedly, is easy to use, realizes online automatic detection of the composition of steel, but also improves work efficiency, reduces labor costs, and improves the safety of operations.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic steel analysis, and in particular to an online automatic steel composition analysis method based on an X-ray fluorescence spectrometer. Background Art

[0002] In the ex-factory process of products such as bars, profiles, and building materials in steel enterprises, the identification of product grades is carried out, that is, the elemental composition of the steel itself is detected to avoid situations such as product confusion and unqualified components. At present, many enterprises use handheld X-ray fluorescence spectrometers for manual detection, which has the following disadvantages: there is no protection device, which will cause certain harm to the human body; it is prone to misoperation; handheld measurement is prone to measurement errors; the handheld device is heavy and inconvenient to operate; relying on manual data reading is prone to errors, and data storage, collection, transmission, and analysis are inconvenient. At the same time, the temperature of the steel itself and the change of the measurement position may affect the accuracy of XRF analysis. Summary of the Invention

[0003] In view of the above problems, the present invention provides an online automatic steel composition analysis method based on an X-ray fluorescence spectrometer, which not only automatically and standardly performs composition analysis operations, is easy to use, realizes online automatic detection of steel composition, but also improves work efficiency, reduces labor costs, and improves operation safety.

[0004] In order to achieve the above object and other related objects, the technical solution provided by the present invention is as follows:

[0005] An online automatic steel composition analysis method based on an X-ray fluorescence spectrometer, the method comprising:

[0006] M1. The steel is detected online, the data information of the temperature of the steel is obtained in real time based on an infrared thermometer, the data information of the elemental analysis of the steel is obtained in real time based on an X-ray fluorescence spectrometer, and the data information of the measurement position of the X-ray fluorescence spectrometer is obtained in real time based on a laser sensor;

[0007] M2. Based on the data information of the temperature of the steel, the data information of the elemental analysis of the steel, and the data information of the measurement position of the X-ray fluorescence spectrometer, an improved whale optimization lightweight gradient boosting machine classification prediction algorithm is used to predict the content of each element in the steel, and the data information of the content of each element in the predicted steel is obtained;

[0008] M3. Based on the data information of the content of each element in the predicted steel, an improved Archimedes optimization algorithm based on an attenuation factor and dynamic learning is used to optimize the content of each element in the steel, and the data information of the content of each element in the optimized steel is obtained;

[0009] M4. Based on the data information of the content of each element in the optimized steel, construct a steel quality evaluation function R, calculate the quality evaluation value of the steel, and obtain the data information of the quality evaluation value of the steel.

[0010] Further, the steel quality evaluation function R is

[0011] ,

[0012] where x i is the data information of the content of the i-th element in the optimized steel, n is the sample size, and α i is the weight coefficient of the content of the i-th element in the optimized steel.

[0013] Further, the constraint condition of the weight coefficient α i of the content of the i-th element in the optimized steel is

[0014] .

[0015] Further, the method further includes:

[0016] M5. Based on the data information of the quality evaluation value of the steel, set a preset threshold. If the quality evaluation value of the steel is less than the preset threshold, the steel does not meet the requirements and is a non-conforming product. If the quality evaluation value of the steel is greater than the preset threshold, the steel meets the requirements and is a qualified product.

[0017] Further, in step M2, the prediction of the content of each element in the steel by using the improved whale optimization lightweight gradient boosting machine classification prediction algorithm includes:

[0018] M21. Input the data information of the temperature of the steel, the data information of the elemental analysis of the steel, and the data information of the measurement position of the X-ray fluorescence spectrometer into the lightweight gradient boosting machine classification prediction model for training and learning, and initialize the hyperparameters of the model to obtain the data information of the hyperparameters of the initialized model;

[0019] M22. Based on the data information of the hyperparameters of the initialized model, initialize the whale population, determine the population parameters, and obtain the data information of the initialized whale population;

[0020] M23. Based on the data information of the initialized whale population, establish a fitness function Q for the population individuals,

[0021] ,

[0022] Among them, x is the data information of the whale population after initialization, α1, α2 and α3 are the dynamic adjustment factors of the fitness of the individuals in the whale population, and the fitness values ​​of the individuals in the whale population are calculated to obtain the data information of the fitness values ​​of the individuals in the whale population;

[0023] M24. Based on the data information of the fitness values ​​of the whale population individuals, establish the target optimization function W of the population,

[0024] ,

[0025] Among them, y is the data information of the fitness value of the individual whale population, β1, β2 and β3 are the target optimization factors, and the hyperparameters of the initialized model are optimized to obtain the data information of the hyperparameters of the optimized model;

[0026] M25. Inputting the data information of the hyperparameters of the optimized model into the lightweight gradient boosting machine classification prediction model for training to obtain a trained lightweight gradient boosting machine classification prediction model;

[0027] M26. Based on the trained lightweight gradient boosting machine classification prediction model, the data information of the temperature of the steel, the data information of the elemental analysis of the steel and the data information of the measurement position of the X-ray fluorescence spectrometer are input to predict the content of each element in the steel, and obtain the data information of the predicted content of each element in the steel.

[0028] Furthermore, the target optimization factors β1, β2 and β3 are,

[0029] ,

[0030] ,

[0031] ,

[0032] Among them, y is the data information of the fitness value of the individual whale population.

[0033] Furthermore, the fitness dynamic adjustment factors α1, α2 and α3 of the individual whale population are:

[0034] ,

[0035] ,

[0036] ,

[0037] Among them, x is the data information of the initialized whale population.

[0038] Further, in step M3, the optimization of the content of each element in the steel using the improved Archimedes optimization algorithm based on the attenuation factor and dynamic learning includes:

[0039] M31. Based on the data information of the content of each element in the predicted steel, initialize the Archimedes population, determine the population parameters and the maximum number of iterations K, and obtain the data information of the initialized Archimedes population;

[0040] M32. Based on the data information of the initialized Archimedes population, establish a fitness function G of the population individuals based on the attenuation factor,

[0041] ,

[0042] ,

[0043] ,

[0044] ,

[0045] where z is the data information of the initialized Archimedes population, δ1, δ2, and δ3 are attenuation factors, and the fitness values of the Archimedes population individuals are deduced to obtain the data information of the fitness values of the Archimedes population individuals;

[0046] M33. Based on the data information of the fitness values of the Archimedes population individuals, establish an optimization function H based on the dynamic learning factor,

[0047] ,

[0048] ,

[0049] ,

[0050] ,

[0051] where r is the data information of the fitness values of the Archimedes population individuals, λ1, λ2, and λ3 are dynamic learning factors, and the content of each element in the steel is optimized to obtain the data information of the content of each element in the optimized steel.

[0052] Further, the constraint conditions for the dynamic learning factors λ1, λ2, and λ3 are

[0053] 。

[0054] To achieve the above and other related objectives, the present invention also provides a system for implementing the online automatic analysis method of steel composition based on any one of the above X-ray fluorescence spectrometers. The system includes:

[0055] A data acquisition module, configured to obtain data information of the temperature of the steel in real time based on an infrared thermometer, obtain data information of the elemental analysis of the steel in real time based on an X-ray fluorescence spectrometer, and obtain data information of the measurement position of the X-ray fluorescence spectrometer in real time based on a laser sensor;

[0056] A prediction module for the elemental content of steel, connected to the data acquisition module, configured to predict the content of each element in the steel by using an improved whale optimization-based lightweight gradient boosting machine classification prediction algorithm, and obtain data information of the content of each element in the predicted steel;

[0057] An optimization module for the elemental content of steel, connected to the prediction module for the elemental content of steel, configured to optimize the content of each element in the steel by using an improved Archimedes optimization algorithm based on an attenuation factor and dynamic learning, and obtain data information of the content of each element in the optimized steel;

[0058] An evaluation module for the quality of steel, connected to the optimization module for the elemental content of steel, configured to construct a steel quality evaluation function R, calculate the quality evaluation value of the steel, and obtain data information of the quality evaluation value of the steel;

[0059] A discrimination module for the quality of steel, connected to the evaluation module for the quality of steel, configured to set a preset threshold. If the quality evaluation value of the steel is less than the preset threshold, the steel does not meet the requirements and is a non-conforming product. If the quality evaluation value of the steel is greater than the preset threshold, the steel meets the requirements and is a qualified product.

[0060] The present invention has the following positive effects:

[0061] 1. By using an improved whale optimization-based lightweight gradient boosting machine classification prediction algorithm to predict the content of each element in the steel and combining with an improved Archimedes optimization algorithm based on an attenuation factor and dynamic learning to optimize the content of each element in the steel, the present invention not only performs component analysis operations automatically and standardized, is easy to use, realizes online automatic detection of steel components, but also improves work efficiency, reduces labor costs, and improves the safety of operations.

[0062] 2. The present invention calculates the quality evaluation value of steel by constructing a steel quality evaluation function R, and combines the setting of a preset threshold to accurately and intelligently detect the quality of steel. This not only improves the reliability and security of data, avoids situations such as human judgment errors and data tampering, and ensures the quality of the produced products, but also realizes full-process automated operation, relies on computer terminals for control and automatic discrimination, and the data can be directly transmitted to the central control platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 is a schematic flowchart of the method of the present invention;

[0064] Figure 2 is a schematic flowchart of the improved whale optimization lightweight gradient boosting machine classification prediction algorithm of the present invention;

[0065] Figure 3 is a schematic flowchart of the improved Archimedes optimization algorithm based on the attenuation factor and dynamic learning of the present invention;

[0066] Figure 4 is a schematic diagram of the system framework of the present invention;

[0067] Figure 5 is a schematic structural diagram of the detection device of the present invention;

[0068] Figure 6 is an overall architecture diagram of the on-line automatic analysis system of the X-ray fluorescence spectrometer of the present invention.

[0069] Explanation of the reference numerals in the figure: 1 - infrared thermometer, 2 - X-ray fluorescence spectrometer, 3 - laser sensor. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0070] The following describes exemplary embodiments of the present disclosure, including various details of the embodiments of the present disclosure to facilitate understanding. It should be considered that they are merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted below.

[0071] Embodiment 1: As Figure 1 or Figure 5 shown, an on-line automatic analysis method for steel components based on an X-ray fluorescence spectrometer, the method comprising:

[0072] M1. The steel is detected online. Based on the infrared thermometer 1, the data information of the temperature of the steel is obtained in real time. Based on the X-ray fluorescence spectrometer 2, the data information of the elemental analysis of the steel is obtained in real time. And based on the laser sensor 3, the data information of the measurement position of the X-ray fluorescence spectrometer is obtained in real time;

[0073] M2. Based on the data information of the temperature of the steel, the data information of the elemental analysis of the steel, and the data information of the measurement position of the X-ray fluorescence spectrometer, an improved whale optimization lightweight gradient boosting machine classification prediction algorithm is used to predict the content of each element in the steel, and the data information of the content of each element in the predicted steel is obtained;

[0074] M3. Based on the data information of the content of each element in the predicted steel, an improved Archimedes optimization algorithm based on the attenuation factor and dynamic learning is used to optimize the content of each element in the steel, and the data information of the content of each element in the optimized steel is obtained;

[0075] M4. Based on the data information of the content of each element in the optimized steel, a steel quality evaluation function R is constructed to calculate the quality evaluation value of the steel, and the data information of the quality evaluation value of the steel is obtained.

[0076] In this embodiment, the steel quality evaluation function R is

[0077] ,

[0078] where x i is the data information of the content of the i-th element in the optimized steel, n is the sample size, and α i is the weight coefficient of the content of the i-th element in the optimized steel.

[0079] In this embodiment, the constraint condition of the weight coefficient α i of the content of the i-th element in the optimized steel is

[0080] .

[0081] In this embodiment, the method further includes:

[0082] M5. Based on the data information of the quality evaluation value of the steel, a preset threshold is set. If the quality evaluation value of the steel is less than the preset threshold, the steel does not meet the requirements and is a substandard product. If the quality evaluation value of the steel is greater than the preset threshold, the steel meets the requirements and is a qualified product.

[0083] In this embodiment, such as Figure 2As shown, in step M2, the prediction of the content of each element in the steel using the improved lightweight gradient boosting machine classification prediction algorithm based on whale optimization includes:

[0084] M21. Input the data information of the temperature of the steel, the data information of the elemental analysis of the steel, and the data information of the measurement position of the X-ray fluorescence spectrometer into the lightweight gradient boosting machine classification prediction model for training and learning, and initialize the hyperparameters of the model to obtain the data information of the hyperparameters of the initialized model;

[0085] M22. Based on the data information of the hyperparameters of the initialized model, initialize the whale population, determine the population parameters, and obtain the data information of the initialized whale population;

[0086] M23. Based on the data information of the initialized whale population, establish the fitness function Q of the population individuals,

[0087] ,

[0088] where x is the data information of the initialized whale population, and α1, α2, and α3 are the fitness dynamic adjustment factors of the whale population individuals, and the fitness values of the whale population individuals are calculated to obtain the data information of the fitness values of the whale population individuals;

[0089] M24. Based on the data information of the fitness values of the whale population individuals, establish the objective optimization function W of the population,

[0090] ,

[0091] where y is the data information of the fitness values of the whale population individuals, and β1, β2, and β3 are the objective optimization factors, and the hyperparameters of the initialized model are optimized to obtain the data information of the hyperparameters of the optimized model;

[0092] M25. Input the data information of the hyperparameters of the optimized model into the lightweight gradient boosting machine classification prediction model for training to obtain the trained lightweight gradient boosting machine classification prediction model;

[0093] M26. Based on the trained lightweight gradient boosting machine classification prediction model, input the data information of the temperature of the steel, the data information of the elemental analysis of the steel, and the data information of the measurement position of the X-ray fluorescence spectrometer, and predict the content of each element in the steel to obtain the data information of the content of each element in the predicted steel.

[0094] In this embodiment, the objective optimization factors β1, β2, and β3 are,

[0095] ,

[0096] ,

[0097] ,

[0098] Among them, y is the data information of the fitness value of the individual of the whale population.

[0099] In this embodiment, the fitness dynamic adjustment factors α1, α2, and α3 of the individual of the whale population are

[0100] ,

[0101] ,

[0102] ,

[0103] Among them, x is the data information of the initialized whale population.

[0104] Embodiment 2: On the basis of a method for automatic online analysis of steel components based on an X-ray fluorescence spectrometer in Embodiment 1, the present invention will be further described and described below.

[0105] As Figure 1 or Figure 5 shown, a method for automatic online analysis of steel components based on an X-ray fluorescence spectrometer, the method includes:

[0106] M1. The steel is detected online, the data information of the temperature of the steel is obtained in real time based on the infrared thermometer 1, the data information of the elemental analysis of the steel is obtained in real time based on the X-ray fluorescence spectrometer 2, and the measurement position data information of the X-ray fluorescence spectrometer is obtained in real time based on the laser sensor 3;

[0107] M2. Based on the data information of the temperature of the steel, the data information of the elemental analysis of the steel, and the measurement position data information of the X-ray fluorescence spectrometer, an improved whale optimization lightweight gradient boosting machine classification prediction algorithm is used to predict the content of each element in the steel, and the data information of the content of each element in the predicted steel is obtained;

[0108] M3. Based on the data information of the content of each element in the predicted steel, an improved Archimedes optimization algorithm based on the attenuation factor and dynamic learning is used to optimize the content of each element in the steel, and the data information of the content of each element in the optimized steel is obtained;

[0109] M4. Based on the data information of the content of each element in the optimized steel, construct a steel quality evaluation function R, calculate the quality evaluation value of the steel, and obtain the data information of the quality evaluation value of the steel.

[0110] In this embodiment, as Figure 3 shown, in step M3, the use of the improved Archimedes optimization algorithm based on the attenuation factor and dynamic learning to optimize the content of each element in the steel includes:

[0111] M31. Based on the data information of the content of each element in the predicted steel, initialize the Archimedes population, determine the population parameters and the maximum number of iterations K, and obtain the data information of the initialized Archimedes population;

[0112] M32. Based on the data information of the initialized Archimedes population, establish a fitness function G for the population individuals based on the attenuation factor,

[0113] ,

[0114] ,

[0115] ,

[0116] ,

[0117] where z is the data information of the initialized Archimedes population, δ1, δ2, and δ3 are attenuation factors, calculate the fitness values of the Archimedes population individuals, and obtain the data information of the fitness values of the Archimedes population individuals;

[0118] M33. Based on the data information of the fitness values of the Archimedes population individuals, establish an optimization function H based on the dynamic learning factor,

[0119] ,

[0120] ,

[0121] ,

[0122] ,

[0123] where r is the data information of the fitness values of the Archimedes population individuals, λ1, λ2, and λ3 are dynamic learning factors, optimize the content of each element in the steel, and obtain the data information of the content of each element in the optimized steel.

[0124] In this embodiment, the constraint conditions of the dynamic learning factors λ1, λ2, and λ3 are

[0125] 。

[0126] In this embodiment, as Figure 4 shown, the present invention provides a system for implementing the online automatic analysis method of steel composition based on any one of the X-ray fluorescence spectrometers. The system includes:

[0127] A data acquisition module, configured to obtain data information of the temperature of the steel in real time based on an infrared thermometer, obtain data information of the elemental analysis of the steel in real time based on an X-ray fluorescence spectrometer, and obtain data information of the measurement position of the X-ray fluorescence spectrometer in real time based on a laser sensor;

[0128] A prediction module for the content of steel elements, connected to the data acquisition module, configured to predict the content of each element in the steel by using an improved classification prediction algorithm of a lightweight gradient boosting machine based on whale optimization, and obtain data information of the content of each element in the predicted steel;

[0129] An optimization module for the content of steel elements, connected to the prediction module for the content of steel elements, configured to optimize the content of each element in the steel by using an improved Archimedes optimization algorithm based on an attenuation factor and dynamic learning, and obtain data information of the content of each element in the optimized steel;

[0130] An evaluation module for the quality of steel, connected to the optimization module for the content of steel elements, configured to construct a steel quality evaluation function R, calculate the quality evaluation value of the steel, and obtain data information of the quality evaluation value of the steel;

[0131] A discrimination module for the quality of steel, connected to the evaluation module for the quality of steel, configured to set a preset threshold. If the quality evaluation value of the steel is less than the preset threshold, the steel does not meet the requirements and is a non-conforming product. If the quality evaluation value of the steel is greater than the preset threshold, the steel meets the requirements and is a qualified product.

[0132] In this embodiment, as Figure 6 shown, the motion module: The entire motion module part is composed of one horizontal and two vertical modules. The horizontal motion module bears the horizontal movement of all functional parts, and the starting and ending points of the reciprocating stroke are determined by the limit mechanism; the vertical module is responsible for the up and down movement of the sample surface treatment device; the vertical module is responsible for the up and down movement of the pressing device, the ranging device, and the X-ray fluorescence spectrometer.

[0133] Cooling device: Since the temperature of the online bar is relatively high, it is necessary to cool it before measurement. The cooling device is composed of a blower and an air knife, and the bar is cooled through PLC control.

[0134] Temperature measuring device: The temperature measuring device uses infrared temperature measurement to measure the temperature of the online bars and determine whether the temperature has dropped to the set value.

[0135] Distance measuring device: The distance measuring device uses laser ranging to measure the surface position of bars of various sizes and specifications, and controls the grinding depth of surface treatment and the measurement position of the X-ray fluorescence spectrometer based on this data.

[0136] Pressing device: After being positioned at the bar to be measured through the cooperation of the distance measuring device and the horizontal movement module, the pressing device uses a cylinder to press the bar to avoid loosening of the bar during surface grinding.

[0137] Surface treatment device: The surface treatment device removes the oxide scale on the surface of the bar by grinding with a grinding wheel to make the measured data more accurate.

[0138] X-ray fluorescence spectrometer: The X-ray fluorescence spectrometer is handheld, remotely controlled by a PLC, and the measurement data is interacted through wireless transmission.

[0139] The present invention also provides a computer-readable storage medium, on which a computer program is stored that is programmed or configured to execute any one of the online steel component automatic analysis methods based on an X-ray fluorescence spectrometer.

[0140] In the embodiments provided in the present application, any reference to a memory, storage, database, or other medium may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0141] In summary, the present invention not only performs component analysis operations automatically and standardly, is easy to use, realizes automatic online detection of steel components, but also improves work efficiency, reduces labor costs, and improves the safety of operations.

[0142] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. An online automatic steel composition analysis method based on X-ray fluorescence spectrometer, characterized in that: The method comprises: M1. Steel is tested online, and data information of steel temperature is obtained in real time based on infrared thermometer, data information of elemental analysis of steel is obtained in real time based on X-ray fluorescence spectrometer, and data information of measurement position of X-ray fluorescence spectrometer is obtained in real time based on laser sensor; M2. Based on the data information of the temperature of the steel, the data information of the element analysis of the steel and the data information of the measurement position of the X-ray fluorescence spectrometer, the content of each element in the steel is predicted using an improved classification prediction algorithm based on whale optimization lightweight gradient boosting machine to obtain the data information of the content of each element in the steel after prediction; M3. Based on the predicted data information of the content of each element in the steel, the content of each element in the steel is optimized by using an improved Archimedean optimization algorithm based on attenuation factor and dynamic learning to obtain data information of the content of each element in the optimized steel; M4. Based on the data information of the content of each element in the optimized steel, a steel quality assessment function R is constructed to calculate the quality assessment value of the steel to obtain data information of the quality assessment value of the steel.

2. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 1 is characterized in that: The steel quality evaluation function R is: ， Among them, x i is the data information of the content of the i-th element in the optimized steel, n is the sample size, α i is the weight coefficient of the content of the i-th element in the optimized steel.

3. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 2 is characterized in that: The weight coefficient α of the content of the i-th element in the optimized steel is i The constraints are: 。 4. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 1, characterized in that: The method further comprises: M5. Based on the data information of the quality assessment value of the steel, a preset threshold is set. If the quality assessment value of the steel is less than the preset threshold, the steel does not meet the requirements and is an unqualified product. If the quality assessment value of the steel is greater than the preset threshold, the steel meets the requirements and is a qualified product.

5. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 1, characterized in that: In step M2, the use of the improved whale-optimized lightweight gradient boosting machine classification prediction algorithm to predict the content of each element in the steel includes: M21. Inputting the data information of the temperature of the steel, the data information of the elemental analysis of the steel, and the data information of the measurement position of the X-ray fluorescence spectrometer into the lightweight gradient boosting machine classification prediction model for training and learning, initializing the hyperparameters of the model, and obtaining the data information of the hyperparameters of the initialized model; M22. Based on the data information of the hyperparameters of the initialized model, the whale population is initialized, the population parameters are determined, and the data information of the initialized whale population is obtained; M23. Based on the data information of the initialized whale population, establish the fitness function Q of the population individuals, ， Among them, x is the data information of the whale population after initialization, α1, α2 and α3 are the dynamic adjustment factors of the fitness of the individuals in the whale population, and the fitness values ​​of the individuals in the whale population are calculated to obtain the data information of the fitness values ​​of the individuals in the whale population; M24. Based on the data information of the fitness values ​​of the whale population individuals, establish the target optimization function W of the population, ， Among them, y is the data information of the fitness value of the individual whale population, β1, β2 and β3 are the target optimization factors, and the hyperparameters of the initialized model are optimized to obtain the data information of the hyperparameters of the optimized model; M25. Inputting the data information of the hyperparameters of the optimized model into the lightweight gradient boosting machine classification prediction model for training to obtain a trained lightweight gradient boosting machine classification prediction model; M26. Based on the trained lightweight gradient boosting machine classification prediction model, the data information of the temperature of the steel, the data information of the elemental analysis of the steel and the data information of the measurement position of the X-ray fluorescence spectrometer are input to predict the content of each element in the steel, and obtain the data information of the predicted content of each element in the steel.

6. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 5 is characterized in that: The target optimization factors β1, β2 and β3 are, ， ， ， Among them, y is the data information of the fitness value of the individual whale population.

7. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 5 is characterized in that: The fitness dynamic adjustment factors α1, α2 and α3 of the individual whale population are: ， ， ， Among them, x is the data information of the initialized whale population.

8. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 1, characterized in that: In step M3, the optimization of the content of each element in the steel by using the improved Archimedean optimization algorithm based on attenuation factor and dynamic learning includes: M31. Based on the data information of the content of each element in the predicted steel, the Archimedean population is initialized, the population parameters and the maximum number of iterations K are determined, and the data information of the initialized Archimedean population is obtained; M32. Based on the data information of the initialized Archimedean population, establish a fitness function G of the population individuals based on the attenuation factor, ， ， ， ， Among them, z is the data information of the initialized Archimedes population, δ1, δ2 and δ3 are attenuation factors, and the fitness values ​​of the individuals in the Archimedes population are calculated to obtain the data information of the fitness values ​​of the individuals in the Archimedes population; M33. Based on the data information of the fitness values ​​of the individuals in the Archimedean population, an optimization function H based on a dynamic learning factor is established. ， ， ， ， Among them, r is the data information of the fitness value of the Archimedean population individual, λ1, λ2 and λ3 are dynamic learning factors, and the content of each element in the steel is optimized to obtain the data information of the content of each element in the optimized steel.

9. The method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to claim 8, characterized in that: The constraints of the dynamic learning factors λ1, λ2 and λ3 are: 。 10. A system for implementing the method for automatic online steel composition analysis based on X-ray fluorescence spectrometer according to any one of claims 1 to 9, characterized in that: The system comprises: A data acquisition module, used to acquire data information of the temperature of the steel in real time based on an infrared thermometer, acquire data information of elemental analysis of the steel in real time based on an X-ray fluorescence spectrometer, and acquire data information of the measurement position of the X-ray fluorescence spectrometer in real time based on a laser sensor; A prediction module for element content in steel, connected to the data acquisition module, is used to predict the content of each element in the steel using an improved classification prediction algorithm based on whale optimization lightweight gradient boosting machine to obtain data information on the content of each element in the steel after prediction; An optimization module for element content of steel, connected to the prediction module for element content of steel, is used to optimize the content of each element in the steel by using an improved Archimedean optimization algorithm based on attenuation factor and dynamic learning, and obtain data information on the content of each element in the optimized steel; A steel quality assessment module, connected to the steel element content optimization module, is used to construct a steel quality assessment function R, calculate the quality assessment value of the steel, and obtain data information of the quality assessment value of the steel; The steel quality identification module is connected to the steel quality evaluation module and is used to set a preset threshold. If the quality evaluation value of the steel is less than the preset threshold, the steel does not meet the requirements and is an unqualified product. If the quality evaluation value of the steel is greater than the preset threshold, the steel meets the requirements and is a qualified product.

Images