Converter blowing process slagging period and copper making period end point judgment and intelligent regulation and control method

Abstract

The invention discloses an end point judgment and intelligent regulation and control method for a slagging period and a copper making period in a converter blowing process, which comprises the following steps: acquiring real-time flue gas parameters and an in-furnace flame image through an improved CEMS (Continuous Emission Monitoring System) device, and adding an industrial high-temperature-resistant camera probe and a photoelectric detector in the CEMS device based on the improved CEMS device; the in-furnace flame image is processed, the end point of the slagging period in the converter blowing process is intelligently judged based on the processed in-furnace flame image, and a first judgment result is obtained; based on the real-time flue gas parameters, a flue gas concentration time sequence change rule is obtained; based on the smoke concentration time sequence change rule, the terminal point of the copper making period in the converter blowing process is intelligently judged, and a second judgment result is obtained; and based on the judgment result and the real-time flue gas parameters, the parameters in the converter blowing process are intelligently regulated and controlled. The method is suitable for intelligent regulation and control of the molten pool internal temperature field image in the molten pool smelting process, and can be applied to the metallurgical industry field, the chemical industry field, the pharmacy field and other fields.

Description

Technical Field

[0001] The present invention belongs to the technical field of metallurgy, and in particular relates to a method for determining and intelligently controlling the endpoints of a slagging period and a copper-making period in a converter blowing process. Background Art

[0002] Currently, molten pool smelting is the mainstream nonferrous pyrometallurgical smelting technology, offering advantages such as high smelting intensity, smoke-free operation, and a high degree of autothermal smelting. As a core process in pyrometallurgy, it is primarily applicable to the smelting of nonferrous sulfide ores such as copper or oxide ores. Converter (BOF) blowing is a crucial step in pyrometallurgical copper smelting, directly impacting the converter's production efficiency. The product of matte blowing is blister copper, and both under- and over-blowing can result in substandard blister copper quality. Therefore, preventing under- and over-blowing is crucial to the entire blowing process. Determining the end point of BOF blowing has become a crucial step in production control, necessitating the development of a model for determining the end point of blowing. Based on the aforementioned findings, new technologies, such as intelligent methods for determining and controlling the end points of the slagging and copper-making phases of the BOF blowing process, can improve energy efficiency and reduce carbon emissions in nonferrous metal smelting, conserving material and energy resources and reducing emissions such as carbon dioxide. This research and development of energy-saving and carbon-reducing technologies for molten pool smelting has significant theoretical and practical value.

[0003] The existing methods for determining the endpoints of the slag-making and copper-making periods in the converter blowing process mainly include: methods based on product color, methods based on flue gas flame color, etc. However, these methods have technical defects such as insufficient accuracy, narrow evaluation range, and insufficient reliability. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes a method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process. The method is suitable for intelligently controlling the temperature field image inside the molten pool during the molten pool smelting process, and can be applied to many fields besides the metallurgical industry, such as chemical industry and pharmaceutical industry.

[0005] To achieve the above objectives, the present invention provides a method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in a converter blowing process, comprising:

[0006] Acquiring real-time flue gas parameters and furnace flame images through an improved CEMS device, wherein the improved CEMS device is based on adding an industrial high-temperature resistant camera probe and a photoelectric detector to the CEMS device;

[0007] processing the furnace flame image, and intelligently determining an end point of a slagging period in a converter blowing process based on the processed furnace flame image to obtain a first determination result;

[0008] Based on the real-time flue gas parameters, obtaining a temporal variation pattern of flue gas concentration;

[0009] Based on the time-series variation pattern of the flue gas concentration, an intelligent judgment is made on the end point of the copper-making period in the converter blowing process to obtain a second judgment result;

[0010] Based on the judgment results and real-time flue gas parameters, the parameters of the converter blowing process are intelligently controlled.

[0011] Optionally, the improved CEMS device includes: a smoke detection subsystem, a smoke parameter monitoring subsystem, a flame image acquisition subsystem, and a system control and data acquisition and processing subsystem;

[0012] The smoke detection subsystem is used to monitor the concentration of particulate matter in the exhaust gas;

[0013] The flue gas parameter monitoring subsystem is used to monitor the basic physical parameters of the flue gas;

[0014] The flame image acquisition subsystem is used to collect flame images and changes in illumination and brightness in the furnace;

[0015] The system control and data acquisition processing subsystem is used to receive the collected furnace information and control the information collection work of the camera probe and various sensors.

[0016] Optionally, the flue gas parameters include: SO2 concentration, flue gas temperature, CO concentration and NO x concentration.

[0017] Optionally, processing the image of the flame in the furnace includes: processing the image of the flame in the furnace using a YOLOv8 intelligent recognition model, wherein the YOLOv8 intelligent recognition model adopts an improved CSPDarknet as the backbone network, introduces PAN as the feature fusion part, and uses the Detect head to perform final target detection and regression, wherein the improved CSPDarknet uses a 3×3 convolution layer instead of a Focus layer.

[0018] Optionally, intelligently determining the end point of the slagging period in the converter blowing process based on the processed furnace flame image, and obtaining a first determination result includes:

[0019] When the flame color of the processed furnace flame image tends to be white and the flame brightness detected by the photoelectric detector is within a preset range, it is determined that the slagging period has reached its end.

[0020] Optionally, obtaining a time-series variation pattern of flue gas concentration based on the real-time flue gas parameter includes:

[0021] Processing the real-time flue gas parameters to obtain a time series data set of flue gas concentration;

[0022] The time series data set is input into the intelligent prediction model to obtain the time series change law of the flue gas concentration, wherein the intelligent prediction model includes a signal decomposition module and a recurrent neural network. The signal decomposition module is used to decompose the time series data set to reduce the complexity of the time series data set, and the recurrent neural network is used to predict the time series change law.

[0023] Optionally, the signal decomposition module includes: a fully integrated empirical mode decomposition CEEMDAN unit and a variational mode decomposition VMD unit;

[0024] The CEEMDAN unit is used to add white noise during the decomposition process to obtain a decomposition result;

[0025] The VMD unit is used to adaptively decompose the signal in the frequency domain using a variational model.

[0026] Optionally, based on the time-series variation pattern of the flue gas concentration, intelligently judging the end point of the copper-making period in the converter blowing process, and obtaining the second judgment result includes:

[0027] When SO2 concentration rises rapidly, flue gas temperature rises, and CO concentration rises, it is judged to be the early stage of copper making;

[0028] When the SO2 concentration rises to the maximum value and then begins to decrease, the flue gas temperature rises slowly, and the CO concentration decreases, it is judged to be the middle stage of copper making;

[0029] When the SO2 concentration decreases and tends to be stable, the flue gas temperature decreases, and the CO concentration tends to be 0, it is judged that the copper making end point has been reached.

[0030] Optionally, based on the judgment results and real-time flue gas parameters, intelligent control of the parameters of the converter blowing process includes:

[0031] Based on the judgment result and the real-time flue gas parameters, the PID control model is used to intelligently control the parameters of the converter blowing process.

[0032] Compared with the prior art, the present invention has the following advantages and technical effects:

[0033] On the one hand, the present invention uses an improved CEMS device to monitor the flue gas parameter changes in the copper making period and the slag making period of the converter blowing process in real time, solving the problem of lack of visualization in high temperature environments and reducing manual operation costs. On the other hand, the present invention uses the YOLOv8 flame image intelligent recognition system and the CEEMDAN-VMD-BILSTM-MATT flue gas concentration intelligent prediction system based on the combustion image, SO2 concentration, NO xThe method and system can be used for intelligent, real-time and accurate judgment and parameter control of the end of the slagging period and the copper-making period in the converter blowing process, and can be applied to intelligent control of the temperature field image in the molten pool in the molten pool smelting process, and can be applied to many fields such as chemical industry and pharmaceutical industry. BRIEF DESCRIPTION OF DRAWINGS

[0034] The drawings constituting a part of this application are used to provide further understanding of the application, the illustrative embodiments of the application and the description thereof are used to explain the application, and do not constitute improper limitation on the application. In the drawings:

[0035] Figure 1 is a flow chart of a method for judging and intelligently controlling the end of the slagging period and the copper-making period in the converter blowing process according to an embodiment of the application;

[0036] Figure 2 is a system framework diagram for real-time intelligent judgment of the end of the slagging period and the copper-making period in the converter blowing process according to an embodiment of the application;

[0037] Figure 3 is a whole framework diagram of the intelligent judgment system for the end of the slagging period and the copper-making period according to an embodiment of the application. DETAILED DESCRIPTION

[0038] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0039] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a group of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0040] The present embodiment proposes a method for judging and intelligently controlling the end of the slagging period and the copper-making period in the converter blowing process, as shown in Figure 1 The method specifically comprises the following steps:

[0041] Real-time flue gas parameters and in-furnace flame images are obtained through the improved CEMS device, and the improved CEMS device is based on the addition of an industrial high-temperature camera probe and a photoelectric detector in the CEMS device;

[0042] The in-furnace flame image is processed, the end of the slagging period in the converter blowing process is intelligently judged based on the processed in-furnace flame image, and a first judgment result is obtained;

[0043] Based on real-time flue gas parameters, obtain the time series variation pattern of flue gas concentration;

[0044] Based on the time series variation pattern of flue gas concentration, the end point of the copper-making period in the converter blowing process is intelligently judged to obtain a second judgment result;

[0045] Based on the judgment results and real-time flue gas parameters, the parameters of the converter blowing process are intelligently controlled.

[0046] Specifically, such as Figure 1 As shown, this embodiment specifically includes: S1, using the improved CEMS device to monitor the SO2 concentration and NO x Concentration, flue gas temperature and other flue gas parameters and collect flame images inside the furnace and save the data in the cloud data center;

[0047] S2. Use the YOLOv8 intelligent recognition model to process and classify the collected flame images in the furnace. When the system identifies the flame color as milky white, it determines that the slagging period has reached the end and feeds the judgment result back to the operation console in real time. This enables intelligent judgment of the slagging period end point of the converter blowing process.

[0048] S3. Input the processed flue gas concentration time series data set into the CEEMDAN-VMD-BiLSTM-MATT intelligent prediction model to obtain the time series variation pattern of flue gas concentration during the converter blowing process, thereby realizing intelligent judgment of the end point of the copper making period of the converter blowing process;

[0049] S4. Using the PID control system, the feedback of flue gas concentration changes and flame temperature images can achieve precise control of the chemical reaction rate, furnace temperature, copper production efficiency, etc. during the converter blowing process, thereby completing the intelligent regulation of various operating parameters of the converter blowing process.

[0050] Furthermore, the improved CEMS device includes: a smoke detection subsystem, a smoke parameter monitoring subsystem, a flame image acquisition subsystem, and a system control and data acquisition and processing subsystem;

[0051] a smoke detection subsystem for monitoring the concentration of particulate matter in exhaust gases;

[0052] Flue gas parameter monitoring subsystem, used to monitor the basic physical parameters of flue gas;

[0053] Flame image acquisition subsystem, used to collect flame images and changes in illumination and brightness inside the furnace;

[0054] The system control and data acquisition and processing subsystem is used to receive the collected furnace information and control the information collection work of the camera probe and various sensors.

[0055] Specifically, the improved CEMS device consists of four basic parts: smoke detection subsystem, flue gas parameter monitoring subsystem, flame image acquisition system, system control and data acquisition and processing subsystem. It receives external signals through sensors to monitor SO2 concentration, NO x The flue gas parameters such as concentration and flue gas temperature are recorded and uploaded to the cloud server to save the data set.

[0056] Furthermore, the flue gas parameters include: SO2 concentration, flue gas temperature, CO concentration and NO x concentration.

[0057] Specifically, the converter sensor receives external signals through extraction condensation method, extraction heat and moisture method, in-situ method, in-situ method, etc. to monitor SO2 concentration, NO x The flame image acquisition system is composed of 3-6 industrial CCD cameras and photoelectric detectors placed in the CEMS device to collect real-time images of the flame temperature in the furnace from multiple angles.

[0058] Furthermore, processing the flame image inside the furnace includes: processing the flame image inside the furnace using the YOLOv8 intelligent recognition model, wherein the YOLOv8 intelligent recognition model adopts the improved CSPDarknet as the backbone network, introduces PAN as the feature fusion part, and uses the Detect head to perform final target detection and regression, wherein the improved CSPDarknet uses a 3×3 convolutional layer instead of a Focus layer to simplify the model structure and improve versatility and compatibility.

[0059] Specifically, the YOLOv8 intelligent recognition model is used to process and classify the collected flame images from the furnace. When the system identifies the flame as milky white, it determines that the slagging period has ended and feeds the judgment result back to the operation console in real time. This enables intelligent judgment of the slagging period endpoint during the converter blowing process.

[0060] like Figure 2 As shown in the figure, the YOLOv8 intelligent recognition model is used to process and classify the collected furnace flame images. YOLOv8 uses the improved CSPDarknet as the backbone network, introduces PAN (Path Aggregation Network) as the feature fusion part, and uses the Detect head to perform the final target detection and regression. Its mathematical expression is:

[0061]

[0062] in, is the predicted probability of target category j, s jare logits output by the neural network, s j ) is a sigmoid activation function.

[0063] Further, based on the processed in-furnace flame image, the end point of the slagging period of the converter blowing process is intelligently judged, and a first judgment result is obtained, which includes:

[0064] When the flame color of the processed in-furnace flame image tends to be white and the flame brightness detected by the photoelectric detector is within a preset range, it is judged that the end point of the slagging period is reached.

[0065] Specifically, the YOLOv8 intelligent recognition model is used to process and classify the collected in-furnace flame image. When the system identifies the flame color as milky white, it is determined that the end point of the slagging period is reached and the judgment result is fed back to the operation platform in real time. The YOLOv8 intelligent image recognition system is used to identify and classify the collected flame image dataset according to the different flame colors. When the flame color tends to be white and the flame brightness detected by the photoelectric detector is within the range of 1000-3000 cd / m 2 , it is approximately determined that the end point of the slagging period is reached and the judgment result is fed back to the operation platform in real time.

[0066] Further, based on the real-time flue gas parameters, the time sequence variation law of the flue gas concentration is obtained, which includes:

[0067] The real-time flue gas parameters are processed to obtain the time sequence dataset of the flue gas concentration;

[0068] The time sequence dataset is input into an intelligent prediction model to obtain the time sequence variation law of the flue gas concentration, wherein the intelligent prediction model includes a signal decomposition module and a recurrent neural network. The signal decomposition module is used to decompose the time sequence dataset to reduce the complexity of the time sequence dataset. The recurrent neural network is used to predict the time sequence variation law.

[0069] Specifically, Figure 2 The overall framework diagram of the intelligent judgment method for the end point of the slagging period and the copper-making period is shown. The processed time sequence dataset of the flue gas concentration is input into the CEEMDAN-VMD-BiLSTM-MATT intelligent prediction model to obtain the time sequence variation law of the flue gas concentration in the converter blowing process, thereby realizing intelligent judgment of the end point of the copper-making period in the converter blowing process.

[0070] The constructed CEEMDAN-VMD-BILSTM-MATT intelligent prediction model combines the signal decomposition algorithm CEEMDAN-VMD with the intelligent prediction model BILSTM-MATT. By decomposing the high-dimensional complex flue gas concentration time sequence data signal multiple times to reduce the complexity of the signal, the prediction accuracy of the intelligent prediction model is improved, and the error is further reduced.

[0071] Furthermore, the signal decomposition module includes: a fully integrated empirical mode decomposition CEEMDAN unit and a variational mode decomposition VMD unit;

[0072] CEEMDAN unit, used to add white noise during the decomposition process to obtain the decomposition results;

[0073] The VMD unit is used to adaptively decompose the signal in the frequency domain using a variational model.

[0074] Specifically, CEEMDAN effectively reduces the impact of noise and ensures the decomposition quality of the signal by adding different white noises in each decomposition process and performing weighted averaging of all decomposition results in an integrated manner. Its mathematical model is expressed as:

[0075]

[0076] Among them, the IMF i (a) is the result of the i-th mode in the n-th iteration, where N is the number of noises.

[0077] VMD (Variational Mode Decomposition) is a signal decomposition algorithm based on the variational principle. It decomposes the signal into a series of modal components (IMFs) with specific bandwidths. VMD introduces a variational model to adaptively decompose the signal in the frequency domain, making each mode have a different center frequency and bandwidth, thus avoiding the common modal aliasing and over-decomposition problems in EMD algorithms. Its mathematical model is:

[0078]

[0079] Where C is the number of decomposed modes, is the cth mode.

[0080] Bidirectional Long Short-Term Memory Network (BiLSTM) is a commonly used recurrent neural network (RNN) variant, primarily used for processing and predicting time series data. By combining two LSTM networks (one forward and one backward), BiLSTM can simultaneously capture both forward and backward information in the sequence, thereby better understanding the dependencies between positions in the sequence. Its mathematical model is:

[0081]

[0082] in, is the hidden state of the forward LSTM at time t, is the hidden state of the reverse LSTM at time t, is the cell function of the forward LSTM, is the cell function of the reverse LSTM.

[0083] Furthermore, based on the time-series variation pattern of the flue gas concentration, an intelligent judgment is made on the end point of the copper-making period in the converter blowing process, and the second judgment result is obtained, including:

[0084] When SO2 concentration rises rapidly, flue gas temperature rises, and CO concentration rises, it is judged to be the early stage of copper making;

[0085] When the SO2 concentration rises to the maximum value and then begins to decrease, the flue gas temperature rises slowly, and the CO concentration decreases, it is judged to be the middle stage of copper making;

[0086] When the SO2 concentration decreases and tends to be stable, the flue gas temperature decreases, and the CO concentration tends to be 0, it is judged that the copper making end point has been reached.

[0087] Specifically, the processed flue gas concentration time series data set is input into the CEEMDAN-VMD-BILSTM-MATT intelligent prediction model to obtain the time series variation law of flue gas concentration in the converter blowing process, and the collected SO2 concentration, NO x The historical time series of data such as concentration and flue gas temperature are used as input to the CEEMDAN-VMD-BILSTM-MATT intelligent prediction model to predict the changing trend of flue gas concentration in the future. By comparing the error of the prediction results, the prediction accuracy of the CEEMDAN-VMD-BILSTM-MATT intelligent prediction model for different flue gas concentrations is analyzed, and one of them is selected as the judgment indicator for the end point of the copper making period.

[0088] The processed flue gas concentration time series dataset is input into the CEEMDAN-VMD-BILSTM-MATT intelligent prediction model to obtain the time series variation pattern of flue gas concentration during the converter blowing process, thereby realizing intelligent judgment of the end point of the copper-making period of the converter blowing process. When the SO2 concentration rises rapidly, the flue gas temperature increases, and the CO concentration increases, it indicates that the redox reaction in the furnace is vigorous and it is in the early stage of copper-making. When the SO2 concentration rises to the maximum value and begins to decrease, the flue gas temperature rises slowly, and the CO concentration decreases, it indicates that the reaction in the furnace is stabilizing and it is in the middle stage of copper-making. When the SO2 concentration decreases and stabilizes, the flue gas temperature decreases, and the CO concentration approaches 0, it indicates that the reaction rate in the furnace is slowing down and the copper-making end point has been reached.

[0089] Furthermore, based on the judgment results and real-time flue gas parameters, the parameters of the converter blowing process are intelligently controlled, including:

[0090] Based on the judgment results and real-time flue gas parameters, the PID control model is used to intelligently control the parameters of the converter blowing process.

[0091] Specifically, the PID control system is used to feedback the flue gas concentration changes and flame temperature images to achieve precise control of the chemical reaction rate, furnace temperature, copper making efficiency, etc. during the converter blowing process, thereby completing the intelligent regulation of various operating parameters of the converter blowing process.

[0092] PID control system is widely used in industrial control systems. It is used to adjust the system output (such as temperature, speed, position, etc.) to the required target value. Its mathematical model is expressed as:

[0093]

[0094] Where u(q) is the output of the controller (control signal), e(q) represents the error signal, and K p Represents the proportional gain, K l Indicates the integral gain, K r represents the differential gain.

[0095] This embodiment also discloses a real-time intelligent judgment system for the end points of the slagging period and the copper-making period in the converter blowing process, which consists of a YOLOv8 flame image intelligent recognition system, a CEEMDAN-VMD-BILSTM-MATT flue gas concentration intelligent prediction system, and a PID control system. The YOLOv8 flame image intelligent recognition system achieves accurate judgment of the end point of the slagging period by accurately identifying flame color and brightness, and the CEEMDAN-VMD-BILSTM-MATT flue gas concentration intelligent prediction system achieves accurate judgment of the end point of the copper-making period in the converter blowing process by intelligently predicting the concentrations of various types of flue gases.

[0096] An intelligent judgment and real-time control system for the end point of the copper-making period in a converter blowing process includes a communication-connected storage and an actuator. The storage stores a program for realizing an intelligent judgment and real-time control method for the end point of the copper-making period in a converter blowing process when the actuator executes the program.

[0097] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for determining and intelligently controlling the endpoints of the slagging and copper-making periods in a converter blowing process, characterized in that: include: Acquiring real-time flue gas parameters and furnace flame images through an improved CEMS device, wherein the improved CEMS device is based on adding an industrial high-temperature resistant camera probe and a photoelectric detector to the CEMS device; processing the furnace flame image, and intelligently determining an end point of a slagging period in a converter blowing process based on the processed furnace flame image to obtain a first determination result; Based on the real-time flue gas parameters, obtaining a temporal variation pattern of flue gas concentration; Based on the time-series variation pattern of the flue gas concentration, an intelligent judgment is made on the end point of the copper-making period in the converter blowing process to obtain a second judgment result; Based on the judgment results and real-time flue gas parameters, the parameters of the converter blowing process are intelligently controlled.

2. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 1, characterized in that: The improved CEMS device includes: a smoke detection subsystem, a smoke parameter monitoring subsystem, a flame image acquisition subsystem, and a system control and data acquisition and processing subsystem; The smoke detection subsystem is used to monitor the concentration of particulate matter in the exhaust gas; The flue gas parameter monitoring subsystem is used to monitor the basic physical parameters of the flue gas; The flame image acquisition subsystem is used to collect flame images and changes in illumination and brightness in the furnace; The system control and data acquisition processing subsystem is used to receive the collected furnace information and control the information collection work of the camera probe and various sensors.

3. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 1, characterized in that: The flue gas parameters include: SO2 concentration, flue gas temperature, CO concentration and NO x concentration.

4. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 1, wherein: Processing the flame image in the furnace includes: processing the flame image in the furnace using a YOLOv8 intelligent recognition model, wherein the YOLOv8 intelligent recognition model adopts an improved CSPDarknet as a backbone network, introduces PAN as a feature fusion part, and uses a Detect head to perform final target detection and regression, wherein the improved CSPDarknet uses a 3×3 convolutional layer instead of a Focus layer.

5. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 1, characterized in that: Based on the processed flame image in the furnace, the endpoint of the slagging period in the converter blowing process is intelligently judged, and the first judgment result is obtained, including: When the flame color of the processed furnace flame image tends to be white and the flame brightness detected by the photoelectric detector is within a preset range, it is determined that the slagging period has reached its end.

6. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 1, wherein: Based on the real-time flue gas parameters, obtaining a time-series variation pattern of flue gas concentration includes: Processing the real-time flue gas parameters to obtain a time series data set of flue gas concentration; The time series data set is input into the intelligent prediction model to obtain the time series change law of the flue gas concentration, wherein the intelligent prediction model includes a signal decomposition module and a recurrent neural network. The signal decomposition module is used to decompose the time series data set to reduce the complexity of the time series data set, and the recurrent neural network is used to predict the time series change law.

7. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 6, characterized in that: The signal decomposition module includes: a fully integrated empirical mode decomposition CEEMDAN unit and a variational mode decomposition VMD unit; The CEEMDAN unit is used to add white noise during the decomposition process to obtain a decomposition result; The VMD unit is used to adaptively decompose the signal in the frequency domain using a variational model.

8. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 3, wherein: Based on the time-series variation pattern of the flue gas concentration, an intelligent judgment is made on the end point of the copper-making period in the converter blowing process, and obtaining a second judgment result includes: When SO2 concentration rises rapidly, flue gas temperature rises, and CO concentration rises, it is judged to be the early stage of copper making; When the SO2 concentration rises to the maximum value and then begins to decrease, the flue gas temperature rises slowly, and the CO concentration decreases, it is judged to be the middle stage of copper making; When the SO2 concentration decreases and tends to be stable, the flue gas temperature decreases, and the CO concentration tends to be 0, it is judged that the copper making end point has been reached.

9. The method for determining and intelligently controlling the endpoints of the slag-making period and the copper-making period in the converter blowing process according to claim 1, wherein: Based on the judgment results and real-time flue gas parameters, the parameters of the converter blowing process are intelligently controlled, including: Based on the judgment result and the real-time flue gas parameters, the PID control model is used to intelligently control the parameters of the converter blowing process.

Images