Intelligent management method and system for smelting of refining furnace based on edge calculation

Abstract

The invention discloses a refining furnace smelting intelligent management method and system based on edge calculation, relates to the technical field of data processing, and comprises the steps of analyzing an association relationship between smelting information in each smelting stage, and providing a reliable basis for establishment and early warning of a subsequent abnormal event standard state. The multiple components of each abnormal event are determined through the incidence relation between the smelting information, abnormity evaluation is conducted on all dimensions, and the smelting early warning reliability and accuracy of the refining furnace are improved. The early warning period of edge calculation and analysis of each smelting stage is set according to the fluctuation condition of the smelting information, and an adaptive early warning period is set according to each smelting stage in combination with the real-time data processing and analysis capability of edge calculation, so that the real-time performance and timeliness of early warning are ensured. The possible abnormal events are analyzed through the state transition probability between the smelting real-time state and the standard state of the abnormal events, management optimization of smelting of the refining furnace is achieved, and efficient and safe operation of the smelting process of the refining furnace is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an intelligent management method and system for refining furnace smelting based on edge computing. Background Art

[0002] With the rapid development of the steel industry, the complexity and real-time requirements of the refining furnace smelting process are increasing. Traditional management methods can no longer meet the needs of efficient and accurate production. Edge computing, as an emerging computing model, achieves low latency, high bandwidth utilization and enhanced privacy protection by placing computing resources and data storage on edge devices close to where data is generated. In the field of refining furnace smelting, edge computing technology can process and analyze massive amounts of data in the smelting process in real time, providing strong support for intelligent management. By real-time monitoring of equipment status, smelting process parameters and output product quality, edge computing technology can detect anomalies and issue warnings in a timely manner to ensure the safe, stable and efficient operation of the smelting process.

[0003] In the prior art, early warning is often only based on a single type of data. However, relying solely on a single type of data, the correlation between multiple types of data is not considered, resulting in poor reliability and low accuracy of refining furnace smelting early warning, which is not conducive to the intelligent management of subsequent early warning measures.

[0004] Therefore, how to improve the reliability and accuracy of refining furnace smelting early warning is a technical problem that needs to be solved. Summary of the invention

[0005] The purpose of the present invention is to solve the problem of poor reliability and low accuracy of refining furnace smelting early warning in the prior art due to early warning based only on single type of data, and propose an intelligent management method for refining furnace smelting based on edge computing, the method comprising: Define the various smelting stages of the refining furnace smelting process and analyze the correlation between the smelting information in each smelting stage; Collect the smelting information of all abnormal events in each smelting stage, determine the multiple components of each abnormal event through the correlation between the smelting information, and integrate the multiple components to establish the standard state of each abnormal event in each stage; Analyze the fluctuation of smelting information at each smelting stage, and set the early warning cycle of edge computing analysis at each smelting stage according to the fluctuation of smelting information; Monitor the smelting information at each smelting stage in real time, define the real-time status of smelting, and regularly update the real-time status of smelting based on the early warning cycle. Generate early warning information through the real-time status of smelting and the standard status of abnormal events, and optimize the management of refining furnace smelting based on the early warning information.

[0006] In some embodiments of the present application, the correlation between the smelting information in each smelting stage is analyzed, including: Smelting information includes equipment status data, smelting process data and smelting output data; Constructing equipment status data sequence, smelting process data sequence and smelting output data sequence according to equipment status data, smelting process data and smelting output data, and processing the equipment status data sequence, smelting process data sequence and smelting output data sequence according to timestamps; The equipment status data sequence, the smelting process data sequence and the smelting output data sequence are assigned to the corresponding smelting stages. For the same smelting stage, the cross-correlation function among the equipment status data sequence, the smelting process data sequence and the smelting output data sequence is calculated, and a cross-correlation diagram is drawn. The correlation relationship among the equipment status data, the smelting process data and the smelting output data in each smelting stage is obtained through the cross-correlation diagram among the equipment status data, the smelting process data and the smelting output data. Among them, the horizontal axis of the cross-correlation diagram represents the time lag between any two of the equipment status data, the smelting process data and the smelting output data, and the vertical axis represents the correlation coefficient between any two of the equipment status data, the smelting process data and the smelting output data.

[0007] In some embodiments of the present application, the correlation relationship between the equipment status data, the smelting process data and the smelting output data in each smelting stage is obtained through the cross-correlation diagram between the equipment status data, the smelting process data and the smelting output data, including: According to the correlation coefficient of the highest point in the cross-correlation diagram, the data category combination with higher correlation is screened out. In the cross-correlation diagram under the data category combination with higher correlation, the cross-correlation diagram is divided into multiple curve segments according to the horizontal axis range value, and the average value of the slope change of each curve segment is calculated. After integration, the overall slope change average value is obtained. The data category combination is divided into two categories: stable relationship and unstable relationship through the overall slope change average value, and the trend component in the cross-correlation diagram is analyzed; For the data category combination with stable relationship, the horizontal coordinate value of the highest point in its cross-correlation diagram is taken as the standard time lag; For the data category combination of non-stable relationship, determine the curve segment with the highest frequency in the cross-correlation diagram, and use the abscissa value of the curve segment and the highest point as the standard time lag; The standard time lag, correlation coefficient and trend component were input into the multiple regression model to determine the correlation between equipment status data, smelting process data and smelting output data.

[0008] In some embodiments of the present application, multiple components of each abnormal event are determined by the association relationship between smelting information, including: For the same abnormal event, the abnormal data parameter ranges in the three types of data, namely, equipment status data, smelting process data and smelting output data, are screened out to confirm the changing trend before and after the data abnormality. The abnormal correlation is determined through the correlation between the equipment status data, smelting process data and smelting output data and the abnormal data; The constituent elements include key elements and evaluation elements. The abnormal data parameter range, the changing trend before and after the data anomaly, and the abnormal correlation are taken as key elements. Each key element is evaluated, and the event abnormality degree level is generated after comprehensive analysis. The abnormality degree level is used as an evaluation element.

[0009] In some embodiments of the present application, each key factor is evaluated and the event abnormality level is generated after integration, including: Determine the evaluation features corresponding to each key factor, generate the evaluation index of each key factor through the evaluation features, and generate the abnormality level of the event by combining the evaluation index of each key factor; ; in, For the The abnormality level of each abnormal event, , , They are the combined weights of the three key factors: abnormal data parameter range, change trend before and after data anomaly, and abnormal correlation. , , Respectively The evaluation indicators of the three key factors of abnormal data parameter range, data change trend before and after abnormality, and abnormal correlation are as follows: for The maximum of the three, , are preset constants, is the rounding symbol.

[0010] In some embodiments of the present application, the fluctuation of smelting information in each smelting stage is analyzed, and an early warning cycle of edge computing analysis in each smelting stage is set according to the fluctuation of smelting information, including: Calculate the coefficient of variation of three types of data: equipment status data, smelting process data, and smelting output data, collect the abnormality levels of all abnormal events in the smelting stage, and determine the early warning cycle; ; in, For the The early warning cycle in each smelting stage, Indicates The next smelting stage The coefficient of variation of the class data, For the The initial warning cycle under the smelting stage, Indicated by The next smelting stage The initial warning period of the data of the first category is obtained by mapping the coefficient of variation of the data of the first category. For the The number of types of abnormal events in each smelting stage, For the The next smelting stage The abnormality level of each abnormal event, For the The next smelting stage The abnormal conversion coefficient is determined by the frequency of occurrence of abnormal events. is a preset constant.

[0011] In some embodiments of the present application, both the standard state of abnormal events and the real-time state of smelting have key elements and evaluation elements corresponding to the abnormal data parameter range, the change trend before and after the data abnormality, and the abnormal correlation.

[0012] In some embodiments of the present application, early warning information is generated by using the real-time status of smelting and the standard status of abnormal events, including: According to the abnormality level of events under the real-time smelting state, the standard state of abnormal events under the smelting stage is screened, and abnormal events with the same abnormality level are screened out. The real-time smelting state and the standard state are matched for the screened abnormal events, and the state transition probability is calculated to generate early warning information.

[0013] Correspondingly, the present application also provides an intelligent management system for refining furnace smelting based on edge computing, including: The first module is used to define the various smelting stages of the refining furnace smelting process and analyze the correlation between the smelting information in each smelting stage; The second module is used to collect the smelting information of all abnormal events in each smelting stage, determine the multiple components of each abnormal event through the correlation between the smelting information, and integrate the multiple components to establish the standard state of each abnormal event in each stage; The third module is used to analyze the fluctuation of smelting information at each smelting stage, and set the early warning cycle of edge computing analysis at each smelting stage according to the fluctuation of smelting information; The fourth module is used to monitor the smelting information in each smelting stage in real time, define the real-time status of smelting, and regularly update the real-time status of smelting based on the early warning cycle. Early warning information is generated through the real-time status of smelting and the standard status of abnormal events, and the management optimization of refining furnace smelting is realized based on the early warning information.

[0014] By applying the above technical solutions, the various smelting stages of the refining furnace smelting process are defined, the correlation between the smelting information in each smelting stage is analyzed, and the correlation between the equipment status data, smelting process data and smelting output data is analyzed, providing a reliable basis for the establishment and early warning of the subsequent standard state of abnormal events. The multiple components of each abnormal event are determined by the correlation between the smelting information, and the abnormal conditions of the abnormal events in multiple dimensions are described by multiple components. The abnormal evaluation of each dimension is performed to improve the reliability and accuracy of the smelting early warning of the refining furnace. According to the fluctuation of smelting information, the early warning cycle of edge computing analysis is set for each smelting stage. Combined with the real-time data processing and analysis capabilities of edge computing, an adaptive early warning cycle is set for each smelting stage to ensure the real-time and timeliness of the early warning. The smelting information in each smelting stage is monitored in real time, the real-time state of smelting is defined, and the real-time state of smelting is regularly updated based on the early warning cycle. The possible abnormal events are analyzed by the state transition probability between the real-time state of smelting and the standard state of abnormal events, thereby generating early warning information, realizing the management optimization of refining furnace smelting according to the early warning information, and ensuring the efficient and safe operation of the refining furnace smelting process. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of the process of the intelligent management method for refining furnace smelting based on edge computing proposed by the present invention; Figure 2 This is a structural schematic diagram of the refining furnace smelting intelligent management system based on edge computing proposed in the present invention. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0017] Reference Figure 1 , the intelligent management method of refining furnace smelting based on edge computing includes the following steps: Step S101, defining each smelting stage of the refining furnace smelting process, and analyzing the correlation between the smelting information in each smelting stage.

[0018] In this embodiment, the refining furnace smelting process can generally be divided into the following stages: Preheating stage: Preheat the charge in preparation for the smelting stage.

[0019] Melting stage: The charge is melted at high temperature to form a molten metal pool.

[0020] Refining stage: remove impurities in the molten pool by adding refining agents, adjusting temperature and stirring.

[0021] Alloying stage: Add alloying elements as needed to adjust the metal composition.

[0022] Steel tapping stage: pouring out the refined molten metal in preparation for the next process.

[0023] In this embodiment, the smelting information includes equipment status data, including parameters such as the operating status, temperature, pressure, and flow rate of the smelting equipment. These data are used to monitor the health of the equipment and the stability of the smelting process. The smelting process data includes real-time information on oxygen content measurement, real-time information on temperature measurement, power consumption, argon consumption, soft blowing time, soft blowing argon consumption, wire feeding length, real-time wire feeding speed, etc. These data reflect the real-time status and process parameters of the smelting process, and the smelting output data include flue gas emissions, sewage discharge, slag volume, output, etc. There are complex correlations between these three types of data. For example, changes in furnace temperature in the equipment status data will affect the amount of refining agent in the smelting process data, and then affect the flue gas emissions in the smelting output data. Through data analysis and mining, these correlations can be revealed, providing a basis for abnormal event detection.

[0024] In some embodiments of the present application, the correlation between the smelting information in each smelting stage is analyzed, including: Smelting information includes equipment status data, smelting process data and smelting output data; Constructing equipment status data sequence, smelting process data sequence and smelting output data sequence according to equipment status data, smelting process data and smelting output data, and processing the equipment status data sequence, smelting process data sequence and smelting output data sequence according to timestamps; The equipment status data sequence, the smelting process data sequence and the smelting output data sequence are assigned to the corresponding smelting stages. For the same smelting stage, the cross-correlation function among the equipment status data sequence, the smelting process data sequence and the smelting output data sequence is calculated, and a cross-correlation diagram is drawn. The correlation relationship among the equipment status data, the smelting process data and the smelting output data in each smelting stage is obtained through the cross-correlation diagram among the equipment status data, the smelting process data and the smelting output data. Among them, the horizontal axis of the cross-correlation diagram represents the time lag between any two of the equipment status data, the smelting process data and the smelting output data, and the vertical axis represents the correlation coefficient between any two of the equipment status data, the smelting process data and the smelting output data.

[0025] In this embodiment, there is a time lag between the three types of data, namely, equipment status data, smelting process data and smelting output data, so they are converted into the form of data series for correlation calculation. The cross-correlation function between equipment status data, smelting process data and smelting output data is calculated to analyze the dynamic correlation between different data series. Through the cross-correlation diagram, the time lag effect and mutual dependence between data series are observed. The cross-correlation function is a statistic that measures the degree of correlation between two time series at different time lags. By calculating the cross-correlation function, we can obtain the correlation coefficient under different time lags, so as to understand how the change of one data series affects another data series. In order to observe the time lag effect and mutual dependence more intuitively, we can draw the results of the cross-correlation function into a cross-correlation diagram. In the cross-correlation diagram, the horizontal axis represents the time lag and the vertical axis represents the correlation coefficient. By observing the cross-correlation diagram, we can find the time lag point with the largest correlation coefficient, which is the time lag with the strongest correlation between the two data series (for the cross-correlation diagram with small changes).

[0026] In some embodiments of the present application, the correlation relationship between the equipment status data, the smelting process data and the smelting output data in each smelting stage is obtained through the cross-correlation diagram between the equipment status data, the smelting process data and the smelting output data, including: According to the correlation coefficient of the highest point in the cross-correlation diagram, the data category combination with higher correlation is screened out. In the cross-correlation diagram under the data category combination with higher correlation, the cross-correlation diagram is divided into multiple curve segments according to the horizontal axis range value, and the average value of the slope change of each curve segment is calculated. After integration, the overall slope change average value is obtained. The data category combination is divided into two categories: stable relationship and unstable relationship through the overall slope change average value, and the trend component in the cross-correlation diagram is analyzed; For the data category combination with stable relationship, the horizontal coordinate value of the highest point in its cross-correlation diagram is taken as the standard time lag; For the data category combination of non-stable relationship, determine the curve segment with the highest frequency in the cross-correlation diagram, and use the abscissa value of the curve segment and the highest point as the standard time lag; The standard time lag, correlation coefficient and trend component were input into the multiple regression model to determine the correlation between equipment status data, smelting process data and smelting output data.

[0027] In this embodiment, time lag effect: observe which data sequence changes lag behind the changes in another data sequence, and the length of the lag. The lag effect may be caused by a variety of factors, such as equipment response speed, material transfer time, chemical reaction rate, etc. This helps to understand the time delays and interactions between different links in the smelting process. Interdependence: Analyze the size and sign of the correlation coefficient between equipment status data, smelting process data and smelting output data. The size of the correlation coefficient reflects the degree of association between the two data sequences, while the sign indicates the positive or negative correlation between them. Dynamic changes: Observe the changes in the cross-correlation function over time to understand whether the dynamic correlation is stable and whether there are periodic or trend changes. This helps to identify potential problems and optimization points in the smelting process.

[0028] In this embodiment, for data category combinations with small changes, that is, stable, we look for the closest connection between the two time series, that is, the strongest correlation between them. By finding the point with the largest correlation coefficient, we can determine when (that is, which time lag) the two time series reach the strongest correlation, which is crucial to understanding the interaction and causal relationship between them. For data category combinations with large changes, that is, non-stable, in addition to the point with the largest correlation coefficient, we can also consider the points in the mode range of the correlation coefficient. The mode range refers to the range of values ​​that appear the most or are relatively concentrated in a set of data. Selecting points in the mode range as time lags can take into account the strong correlation under multiple time lags and increase the robustness of the results. Based on the point with the largest correlation coefficient and the point in the mode range of the correlation coefficient, we can synthesize a standard time lag. The standard time lag, correlation coefficient and trend component are used as input variables. Use these input variables to build a multivariate regression model (time lag terms can be added) to determine the specific correlation relationship (functional relationship) between each data category combination with high correlation.

[0029] Step S102, collecting smelting information of all abnormal events in each smelting stage, determining multiple components of each abnormal event through the correlation between the smelting information, and integrating the multiple components to establish a standard state of each abnormal event in each stage.

[0030] In this embodiment, historical data is reviewed to collect equipment status data, smelting process data, and smelting output data of abnormal events occurring at each stage. For each abnormal event, its parameters (such as abnormal value, duration), change trend (such as sudden increase, gradual decrease), and the correlation between the three types of data are analyzed as constituent elements. The standard state of each abnormal event at each stage is established.

[0031] In some embodiments of the present application, multiple components of each abnormal event are determined by the association relationship between smelting information, including: For the same abnormal event, the abnormal data parameter ranges in the three types of data, namely, equipment status data, smelting process data and smelting output data, are screened out to confirm the changing trend before and after the data abnormality. The abnormal correlation is determined through the correlation between the equipment status data, smelting process data and smelting output data and the abnormal data; The constituent elements include key elements and evaluation elements. The abnormal data parameter range, the changing trend before and after the data anomaly, and the abnormal correlation are taken as key elements. Each key element is evaluated, and the event abnormality degree level is generated after comprehensive analysis. The abnormality degree level is used as an evaluation element.

[0032] In this embodiment, for each abnormal event, its parameters are first analyzed, such as the range of abnormal values, the length of duration, etc. These parameters are the basis for describing the characteristics of abnormal events. Analyzing the changing trend of abnormal event data, such as the sudden increase or gradual decrease of equipment status data, the fluctuation of smelting process data, etc., these trends help to understand the development process of abnormal events. Abnormal association is the matching of three types of data association relationships. For example, one type of data is abnormal, causing the other two types of data to be abnormal: this is an abnormal association that conforms to the three types of data association relationships. For example, the temperature in the equipment status data increases abnormally, resulting in an accelerated reaction speed in the smelting process data, which in turn causes a decrease in product quality-related parameters in the smelting output data. This abnormal situation shows that the abnormality of the equipment status directly affects the smelting process and output results. For example, one type of data is abnormal, but the other two types of data with which it is associated are normal: this situation is relatively rare, but equally important. For example, the pressure in the equipment status data increases abnormally, but the smelting process data and smelting output data remain normal. This may mean that there are potential faults or hidden dangers inside the equipment, but it has not yet had a direct impact on the smelting process and output results. Because this does not conform to the correlation between the three types of data, it will make the anomaly more complex or unpredictable.

[0033] In some embodiments of the present application, each key factor is evaluated and the event abnormality level is generated after integration, including: Determine the evaluation features corresponding to each key factor, generate the evaluation index of each key factor through the evaluation features, and generate the abnormality level of the event by combining the evaluation index of each key factor; ; in, For the The abnormality level of each abnormal event, , , They are the combined weights of the three key factors: abnormal data parameter range, change trend before and after data anomaly, and abnormal correlation. , , Respectively The evaluation indicators of the three key factors of abnormal data parameter range, data change trend before and after abnormality, and abnormal correlation are as follows: for The maximum of the three, , are preset constants, is the rounding symbol.

[0034] In this embodiment, the evaluation features corresponding to each key factor are determined. For example, for the parameter range of abnormal data, the evaluation features are determined by the range of abnormal values, the amplitude beyond the normal range, the frequency and duration of abnormal values. The trend of changes before and after the data anomaly (considering the speed of data change, the amplitude of change, and the direction of change, etc.), for example, a sudden large change may be regarded as highly abnormal, while a slow small change may be regarded as normal or slightly abnormal. Abnormal association, statistics on whether the abnormal data of the three types of data match the association relationship. If it matches, it means that the abnormal situation is more controllable and traceable. If it does not match, it means that the abnormal situation is more complex or more serious, which is not conducive to abnormal location. Integrate all evaluation features and all parameters to determine the evaluation indicators of each type of data in the three types of data, and jointly explain the abnormal situation from three dimensions (parameter range, change trend and abnormal association).

[0035] In this embodiment, It means that the average values ​​of the three dimensions are corrected by the maximum evaluation index among the three types of data. The evaluation index is an indicator that describes the degree of abnormality. Different abnormal events may correspond to different serious situations or abnormal situations.

[0036] Step S103: Analyze the fluctuation of smelting information in each smelting stage, and set an early warning cycle of edge computing analysis in each smelting stage according to the fluctuation of smelting information.

[0037] In this embodiment, by taking advantage of the fact that edge computing can perform data analysis and processing on the device side, different warning cycles are set for different smelting stages. When the time cycle is met, the real-time smelting status is updated in real time, and the state transition probability is calculated.

[0038] In some embodiments of the present application, the fluctuation of smelting information in each smelting stage is analyzed, and an early warning cycle of edge computing analysis in each smelting stage is set according to the fluctuation of smelting information, including: Calculate the coefficient of variation of three types of data: equipment status data, smelting process data, and smelting output data, collect the abnormality levels of all abnormal events in the smelting stage, and determine the early warning cycle; ; in, For the The early warning cycle in each smelting stage, Indicates The next smelting stage The coefficient of variation of the class data, For the The initial warning cycle under the smelting stage, Indicated by The next smelting stage The initial warning period of the data of the first category is obtained by mapping the coefficient of variation of the data of the first category. For the The number of types of abnormal events in each smelting stage, For the The next smelting stage The abnormality level of each abnormal event, For the The next smelting stage The abnormal conversion coefficient is determined by the frequency of occurrence of abnormal events. is a preset constant.

[0039] In this embodiment, the coefficient of variation can describe the fluctuation of each parameter under the three types of data. The coefficient of variation of the three types of data is generated by combining all parameters. The 3 in the summation formula represents the three types of data: equipment status data, smelting process data, and smelting output data. It means that the average value of the initial warning cycle is corrected by the severity of all abnormal events in the smelting stage. The higher the abnormality, the shorter the cycle, and vice versa.

[0040] Step S104, real-time monitoring of smelting information in each smelting stage, definition of real-time smelting status, and regular update of real-time smelting status based on an early warning cycle, generating early warning information through the real-time smelting status and the standard status of abnormal events, and optimizing the management of refining furnace smelting based on the early warning information.

[0041] In some embodiments of the present application, both the standard state of abnormal events and the real-time state of smelting have key elements and evaluation elements corresponding to the abnormal data parameter range, the change trend before and after the data abnormality, and the abnormal correlation.

[0042] In this embodiment, if there is no abnormality in a key factor under the real-time smelting state, the corresponding part is 0.

[0043] In some embodiments of the present application, early warning information is generated by using the real-time status of smelting and the standard status of abnormal events, including: According to the abnormality level of events under the real-time smelting state, the standard state of abnormal events under the smelting stage is screened, and abnormal events with the same abnormality level are screened out. The real-time smelting state and the standard state are matched for the screened abnormal events, and the state transition probability is calculated to generate early warning information.

[0044] In this embodiment, the three key elements are integrated into one state, the state transition probability between the real-time state of smelting and the standard state is calculated, and the probability of transferring from the current state to the standard state of abnormal events is calculated using probability models such as Markov chains. When the state transition probability of a certain abnormal event is detected to exceed the warning threshold, the corresponding warning information is generated according to the nature, scope of impact and possible risk consequences of the abnormal event. The warning information should include the description of the abnormal event, the level of abnormality, the possible risk consequences and the recommended countermeasures, etc., to assist in the subsequent smelting process management.

[0045] Correspondingly, this application also provides an intelligent management system for refining furnace smelting based on edge computing, such as Figure 2 As shown, including, The first module is used to define the various smelting stages of the refining furnace smelting process and analyze the correlation between the smelting information in each smelting stage; The second module is used to collect the smelting information of all abnormal events in each smelting stage, determine the multiple components of each abnormal event through the correlation between the smelting information, and integrate the multiple components to establish the standard state of each abnormal event in each stage; The third module is used to analyze the fluctuation of smelting information at each smelting stage, and set the early warning cycle of edge computing analysis at each smelting stage according to the fluctuation of smelting information; The fourth module is used to monitor the smelting information in each smelting stage in real time, define the real-time status of smelting, and regularly update the real-time status of smelting based on the early warning cycle. Early warning information is generated through the real-time status of smelting and the standard status of abnormal events, and the management optimization of refining furnace smelting is realized based on the early warning information.

[0046] By applying the above technical solutions, the various smelting stages of the refining furnace smelting process are defined, the correlation between the smelting information in each smelting stage is analyzed, and the correlation between the equipment status data, smelting process data and smelting output data is analyzed, providing a reliable basis for the establishment and early warning of the subsequent standard state of abnormal events. The multiple components of each abnormal event are determined by the correlation between the smelting information, and the abnormal conditions of the abnormal events in multiple dimensions are described by multiple components. The abnormal evaluation of each dimension is performed to improve the reliability and accuracy of the smelting early warning of the refining furnace. According to the fluctuation of smelting information, the early warning cycle of edge computing analysis is set for each smelting stage. Combined with the real-time data processing and analysis capabilities of edge computing, an adaptive early warning cycle is set for each smelting stage to ensure the real-time and timeliness of the early warning. The smelting information in each smelting stage is monitored in real time, the real-time state of smelting is defined, and the real-time state of smelting is regularly updated based on the early warning cycle. The possible abnormal events are analyzed by the state transition probability between the real-time state of smelting and the standard state of abnormal events, thereby generating early warning information, realizing the management optimization of refining furnace smelting according to the early warning information, and ensuring the efficient and safe operation of the refining furnace smelting process.

[0047] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present invention can be implemented by hardware, or by software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present invention.

[0048] Those skilled in the art will appreciate that the accompanying drawings are merely schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily required for implementing the present invention.

[0049] Those skilled in the art will appreciate that the modules in the system in the implementation scenario can be distributed in the system of the implementation scenario according to the implementation scenario description, or can be changed accordingly and located in one or more systems different from the implementation scenario. The modules in the above implementation scenario can be combined into one module, or can be further split into multiple submodules.

[0050] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. The intelligent management method of refining furnace smelting based on edge computing is characterized by: include, Define the various smelting stages of the refining furnace smelting process and analyze the correlation between the smelting information in each smelting stage; Collect the smelting information of all abnormal events in each smelting stage, determine the multiple components of each abnormal event through the correlation between the smelting information, and integrate the multiple components to establish the standard state of each abnormal event in each stage; Analyze the fluctuation of smelting information at each smelting stage, and set the early warning cycle of edge computing analysis at each smelting stage according to the fluctuation of smelting information; Monitor the smelting information at each smelting stage in real time, define the real-time status of smelting, and regularly update the real-time status of smelting based on the early warning cycle. Generate early warning information through the real-time status of smelting and the standard status of abnormal events, and optimize the management of refining furnace smelting based on the early warning information.

2. The intelligent management method for refining furnace smelting based on edge computing according to claim 1 is characterized in that: Analyze the correlation between smelting information at each smelting stage, including: Smelting information includes equipment status data, smelting process data and smelting output data; Constructing equipment status data sequence, smelting process data sequence and smelting output data sequence according to equipment status data, smelting process data and smelting output data, and processing the equipment status data sequence, smelting process data sequence and smelting output data sequence according to timestamps; The equipment status data sequence, the smelting process data sequence and the smelting output data sequence are assigned to the corresponding smelting stages. For the same smelting stage, the cross-correlation function among the equipment status data sequence, the smelting process data sequence and the smelting output data sequence is calculated, and a cross-correlation diagram is drawn. The correlation relationship among the equipment status data, the smelting process data and the smelting output data in each smelting stage is obtained through the cross-correlation diagram among the equipment status data, the smelting process data and the smelting output data. Among them, the horizontal axis of the cross-correlation diagram represents the time lag between any two of the equipment status data, the smelting process data and the smelting output data, and the vertical axis represents the correlation coefficient between any two of the equipment status data, the smelting process data and the smelting output data.

3. The intelligent management method for refining furnace smelting based on edge computing according to claim 2 is characterized in that: The correlation between the equipment status data, smelting process data and smelting output data in each smelting stage is obtained through the cross-correlation diagram between the equipment status data, smelting process data and smelting output data, including: According to the correlation coefficient of the highest point in the cross-correlation diagram, the data category combination with higher correlation is screened out. In the cross-correlation diagram under the data category combination with higher correlation, the cross-correlation diagram is divided into multiple curve segments according to the horizontal axis range value, and the average value of the slope change of each curve segment is calculated. After integration, the overall slope change average value is obtained. The data category combination is divided into two categories: stable relationship and unstable relationship through the overall slope change average value, and the trend component in the cross-correlation diagram is analyzed; For the data category combination with stable relationship, the horizontal coordinate value of the highest point in its cross-correlation diagram is taken as the standard time lag; For the data category combination of non-stable relationship, determine the curve segment with the highest frequency in the cross-correlation diagram, and use the abscissa value of the curve segment and the highest point as the standard time lag; The standard time lag, correlation coefficient and trend component were input into the multiple regression model to determine the correlation between equipment status data, smelting process data and smelting output data.

4. The intelligent management method for refining furnace smelting based on edge computing according to claim 2 is characterized in that: The multiple components of each abnormal event are determined through the correlation between smelting information, including: For the same abnormal event, the abnormal data parameter ranges in the three types of data, namely, equipment status data, smelting process data and smelting output data, are screened out to confirm the changing trend before and after the data abnormality. The abnormal correlation is determined through the correlation between the equipment status data, smelting process data and smelting output data and the abnormal data; The constituent elements include key elements and evaluation elements. The abnormal data parameter range, the changing trend before and after the data anomaly, and the abnormal correlation are taken as key elements. Each key element is evaluated, and the event abnormality degree level is generated after comprehensive analysis. The abnormality degree level is used as an evaluation element.

5. The intelligent management method for refining furnace smelting based on edge computing according to claim 4 is characterized in that: Each key factor is evaluated and the abnormality level of the event is generated after integration, including: Determine the evaluation features corresponding to each key factor, generate the evaluation index of each key factor through the evaluation features, and generate the abnormality level of the event by combining the evaluation index of each key factor; ； in, For the The abnormality level of each abnormal event, , , They are the combined weights of the three key factors: abnormal data parameter range, change trend before and after data anomaly, and abnormal correlation. , , Respectively The evaluation indicators of the three key factors of abnormal data parameter range, data change trend before and after abnormality, and abnormal correlation are as follows: for The maximum of the three, , are preset constants, is the rounding symbol.

6. The intelligent management method for refining furnace smelting based on edge computing according to claim 4 is characterized in that: Analyze the fluctuation of smelting information at each smelting stage, and set the early warning cycle of edge computing analysis at each smelting stage according to the fluctuation of smelting information. include, Calculate the coefficient of variation of three types of data: equipment status data, smelting process data, and smelting output data, collect the abnormality levels of all abnormal events in the smelting stage, and determine the early warning cycle; ； in, For the The early warning cycle in each smelting stage, Indicates The next smelting stage The coefficient of variation of the class data, For the The initial warning cycle under the smelting stage, Indicated by The next smelting stage The initial warning period of the data of the first category is obtained by mapping the coefficient of variation of the data of the first category. For the The number of types of abnormal events in each smelting stage, For the The next smelting stage The abnormality level of each abnormal event, For the The next smelting stage The abnormal conversion coefficient is determined by the frequency of occurrence of abnormal events. is a preset constant.

7. The intelligent management method for refining furnace smelting based on edge computing according to claim 4 is characterized in that: Both the standard state of abnormal events and the real-time state of smelting have key elements and evaluation elements corresponding to the abnormal data parameter range, the change trend before and after the data abnormality, and the abnormal correlation.

8. The intelligent management method for refining furnace smelting based on edge computing according to claim 7 is characterized in that: Generate early warning information through the real-time status of smelting and the standard status of abnormal events. include, According to the abnormality level of events under the real-time smelting state, the standard state of abnormal events under the smelting stage is screened, and abnormal events with the same abnormality level are screened out. The real-time smelting state and the standard state are matched for the screened abnormal events, and the state transition probability is calculated to generate early warning information.

9. The intelligent management system for refining furnace smelting based on edge computing is characterized by: include, The first module is used to define the various smelting stages of the refining furnace smelting process and analyze the correlation between the smelting information in each smelting stage; The second module is used to collect the smelting information of all abnormal events in each smelting stage, determine the multiple components of each abnormal event through the correlation between the smelting information, and integrate the multiple components to establish the standard state of each abnormal event in each stage; The third module is used to analyze the fluctuation of smelting information at each smelting stage, and set the early warning cycle of edge computing analysis at each smelting stage according to the fluctuation of smelting information; The fourth module is used to monitor the smelting information in each smelting stage in real time, define the real-time status of smelting, and regularly update the real-time status of smelting based on the early warning cycle. Early warning information is generated through the real-time status of smelting and the standard status of abnormal events, and the management optimization of refining furnace smelting is realized based on the early warning information.

Images