Online regulation and control method and system for molten steel temperature of steel plant

Abstract

The invention discloses an online regulation and control method and system for the molten steel temperature of a steel plant, and belongs to the field of intelligent smelting technologies and equipments.The online regulation and control method comprises the steps that relevant information of smelting equipment and steel ladles of main working procedures of the steel plant is obtained, and a data set corresponding to each working procedure is obtained after screening is conducted according to historical production data of the main working procedures; based on the steel ladle lining temperature, the steel ladles are classified, and the temperature drop condition of molten steel in all types of steel ladles is obtained; based on the data set and the steel ladle temperature drop condition, a molten steel temperature presetting model is established through a case-based reasoning algorithm, a molten steel temperature prediction model is established through an intelligent algorithm, and before and during smelting of each furnace of molten steel and after scheduling time information is updated, all subsequent procedure presetting temperatures and the next procedure prediction temperature are obtained; and the production plan is adjusted according to the preset molten steel temperature, the predicted molten steel temperature and the actually measured molten steel temperature. The method assists in achieving accurate control over the temperature of the molten steel, and the phenomenon of high-temperature tapping is reduced advantageously.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent smelting technology and equipment, and in particular to an online control method and system for molten steel temperature in a steelmaking plant. Background Art

[0002] For the existing long-process steel manufacturing process, the smelting process from converter, refining to continuous casting is the key link to obtain qualified molten steel in the steelmaking plant. In this process, molten steel temperature is a key parameter, and its control level affects factors such as production rhythm, billet quality and energy cost. Accurate control should ensure that each process, especially the terminal process - continuous casting, can obtain molten steel with a temperature that meets the standard, while reducing the terminal molten steel temperature to reduce energy consumption.

[0003] To achieve precise temperature control, first of all, a reasonable temperature system must be established before smelting, that is, requirements are set for the temperature of the molten steel at the beginning and end of each process. The current mainstream temperature system is to evaluate the average temperature drop of the molten steel based on the standard process smelting time, and roughly compensate for the heat storage of the ladle or continuous casting tundish that is insufficient. This system has a relatively rough consideration of the impact of the thermal state of the ladle, a wide range of regulations for the reasonable temperature range, and lacks dynamic consideration of the target temperature of each furnace of molten steel at each process node under different production conditions.

[0004] In smelting, some scholars predict the temperature of molten steel within each process or between processes in order to know the temperature of molten steel in advance and adjust the process control to meet the temperature system requirements. Prediction models include methods based on mechanism models, data-driven models or multi-model combinations. The application of data-driven models has reduced the impact of complex molten pool reaction conditions and the nonlinear relationship between many influencing factors on molten steel temperature prediction to a certain extent, but the algorithms commonly used in the prior art are prone to fall into local optimality and it is difficult to achieve classified optimization processing for different data conditions. In addition, existing models rarely consider the impact of the thermal state of the ladle on the molten steel temperature, or simply use the relevant parameters of the thermal state of the ladle as one of the input conditions of the algorithm.

[0005] The rational formulation of the temperature system and the accurate prediction of the molten steel temperature are complementary to each other. At present, there is a lack of molten steel temperature control system that can be applied to online production and fully considers the temperature connection and matching within and between processes, so as to truly guide on-site operators. At the same time, in online production, time is an important influencing factor of molten steel temperature, and the regulation of molten steel temperature and the scheduling of production time are inseparable. The stability of molten steel temperature depends on the dynamic scheduling methods of steel mills, including planned scheduling and overhead crane scheduling, and is also an important goal of dynamic scheduling. In the existing technology, the regulation of molten steel temperature in steel mills is relatively isolated, lacks information association with planned scheduling, and it is difficult to achieve coordinated matching optimization of temperature and time. Summary of the invention

[0006] In order to solve the above problems, the purpose of the present invention is to provide a technology for online control of molten steel temperature in a steel plant. Based on the simulation analysis of the thermal state of the ladle, this technology considers the influencing factors of the molten steel temperature in each section, establishes a data-driven molten steel temperature prediction model and prediction model, and combines them to form an online molten steel temperature control system. Before and during smelting, the production operators are provided with the predetermined and predicted temperatures of the molten steel in each process to assist them in taking control measures.

[0007] In order to achieve the above technical objectives, the present application provides an online control method for molten steel temperature in a steelmaking plant, comprising the following steps:

[0008] Obtain relevant information about the smelting equipment and ladles in the main process of the steel plant, and filter the historical production data of the main process to obtain the data set for model training in each process;

[0009] Based on the ladle lining temperature, ladles with different thermal states are classified to obtain the temperature drop of molten steel in each category of ladles;

[0010] Based on the data set, the historical furnaces were screened, and according to the temperature drop of the ladle, the case-based reasoning algorithm was used to establish a molten steel temperature prediction model, and the K-means and BPNN algorithms were used to establish a molten steel temperature prediction model;

[0011] Based on the molten steel temperature booking and prediction model, the scheduled temperatures of all subsequent processes and the predicted temperature of the next process are obtained before and during the smelting of each batch of molten steel and after the scheduling time information is updated. The production plan is adjusted according to the scheduled molten steel temperature, the predicted molten steel temperature and the actual molten steel temperature.

[0012] Preferably, in the process of obtaining relevant information about the smelting equipment and ladle in the main process of the steelmaking plant, relevant information about the steelmaking workshop is obtained, including: overall layout, equipment parameters, production steel types and process paths, and relevant information about the ladle, including: ladle masonry diagram, brick masonry thermophysical parameters, and ladle turnover status, among which the relevant information about the steelmaking workshop is used for the selection of factors affecting molten steel temperature, the classification of models, the determination of standard smelting time and standard molten steel temperature drop, and the relevant information about the ladle is used for numerical simulation and classification of the ladle.

[0013] Preferably, in the process of obtaining the temperature drop of molten steel in ladles of various categories, the ladle age, empty ladle time and baking time are used as the thermal state of the ladle according to the ladle turnover situation, and numerical simulation is performed on each category of ladle to obtain the temperature drop rate of each category of ladles in the static time period, and correction is made according to historical production data to obtain the compensation value of the molten steel temperature in different thermal states of ladles in each section. Among them, for the empty ladle time, according to the overall production rhythm and ladle turnover procedures, the ladles with empty times less than 90 minutes, between 90-150 minutes, and greater than 150 minutes are marked respectively.

[0014] Preferably, in the process of building the molten steel temperature prediction model, the historical heats with better performance are selected, all factors affecting the molten steel temperature from the converter to the continuous casting process are taken as input, and the temperature of each process node is taken as output to establish a case-based reasoning data set;

[0015] Input a new heat, select a historical heat in the data set with the same ladle thermal state category and the highest similarity of other influencing factors, and use the temperature of this historical heat at each process node as the target temperature of the new heat.

[0016] Preferably, in the process of obtaining the predicted molten steel temperature, the influencing factors of the molten steel temperature in each process section are analyzed, and based on the data set, variables with a larger Pearson correlation coefficient and in line with the mechanism analysis are selected as the dimensions of subsequent modeling, except for the ladle thermal state;

[0017] Using the K-means cluster analysis method, we tried to increase the number of cluster centers through enumeration, calculated the number of clusters when the rate of decrease of the sum of squared errors slowed down significantly, and used this number of clusters to divide the data set.

[0018] Based on the BPNN model, according to the divided data set, the model of each heat of molten steel is called to make predictions according to its own data set;

[0019] For heats with abnormal ladle thermal conditions, the corresponding temperature correction value is subtracted from the model predicted temperature to obtain the corrected predicted molten steel temperature.

[0020] Preferably, in the process of obtaining the predetermined temperature of all subsequent processes and the predicted temperature of the next process, for each furnace of molten steel, it is triggered before the end of converter smelting, and the liquidus temperature of the steel grade is used as the starting point. The temperature drop in each stage is gradually reversed according to the superheat and production plan, and an optimized temperature system for each node is formulated. Then, compensation is made according to the thermal state of the ladle to obtain the target temperature of each node and update it in time. Among them, the temperature prediction model is triggered in the smelting process from converter to continuous casting. Based on the previous process node, according to the currently known production conditions, the molten steel temperature of the next process node is calculated, providing a basis for determining the temperature control parameters of each process.

[0021] Preferably, in the process of adjusting the production plan, the steelmaking production plan issued by the superior system is received, and an operation plan for the production site is generated according to the on-site production performance, logistics information and abnormal operating conditions; and based on the received operation plan, basic information on the steel type and process path of the specific furnace, as well as information on the planned smelting time in each process and the transportation time between processes are obtained, and corresponding adjustments are made when the operation plan changes.

[0022] The present invention discloses an online control system for the temperature of molten steel in a steelmaking plant, which is used to implement the above-mentioned online control method for the temperature of molten steel in a steelmaking plant. The system comprises:

[0023] The data acquisition module is used to obtain relevant information about the smelting equipment and ladles in the main process of the steel plant. According to the historical production data of the main process, the data set for model training of each process is obtained after screening;

[0024] The data analysis module is used to classify ladles of different thermal states based on the ladle lining temperature and obtain the temperature drop of molten steel in each category of ladles;

[0025] The model building module is used to screen historical heats based on the data set, establish a molten steel temperature prediction model using the case-based reasoning algorithm according to the ladle temperature drop, and establish a molten steel temperature prediction model using the K-means and BPNN algorithms;

[0026] The online control module is used to obtain the scheduled temperature of all subsequent processes and the predicted temperature of the next process before and during the smelting of each furnace of molten steel and after the scheduling time information is updated based on the molten steel temperature booking and molten steel temperature prediction model, and adjust the production plan according to the scheduled molten steel temperature, predicted molten steel temperature and actual molten steel temperature.

[0027] The present invention discloses the following technical effects:

[0028] The present invention establishes an online molten steel temperature control system for a steel plant based on the pre-setting and prediction of the molten steel temperature, and integrates the system with the dynamic scheduling system of the steel plant to provide accurate molten steel temperature recommendation and perception during the production process, assist in achieving precise control of the molten steel temperature, and help reduce the high-temperature steelmaking phenomenon. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 It is a structural diagram of the molten steel temperature prediction model of the present invention;

[0031] Figure 2 It is a trigger flow chart of the online molten steel temperature control system described in the present invention. DETAILED DESCRIPTION

[0032] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0033] like Figure 1-2 As shown, the present invention provides a method for online control of molten steel temperature in a steelmaking plant. Based on the simulation analysis of the thermal state of the ladle, the method considers the influencing factors of the molten steel temperature in each section, establishes a molten steel temperature predetermined model and a prediction model based on data drive, and combines them to form an online molten steel temperature control system, which provides the production operators with the predetermined and predicted temperatures of the molten steel in each process before and during smelting, and assists them in taking control measures.

[0034] Among them, the predetermined temperature is the optimized temperature system and the target molten steel temperature, and the predicted temperature is the currently predicted molten steel temperature at the end of each process. The operator adjusts the operation according to the predicted temperature so that the actual molten steel temperature finally reaches the target of the predetermined temperature. For the molten steel temperature prediction model, an overall prediction method based on case reasoning is proposed. Compared with the existing step-by-step reverse forward prediction method, the prediction result is more in line with the actual requirements; for the molten steel temperature prediction model, a prediction method based on K-means clustering algorithm (K-means), back propagation neural network (BPNN) and ladle thermal state correction is proposed; compared with the prediction model without clustering analysis or without detailed consideration of the ladle thermal state, the prediction hit rate is significantly improved, which can better guide on-site production.

[0035] On this basis, the relationship between the online molten steel temperature control system and the dynamic scheduling system of the steel plant is established, the planned time and scheduling situation provided by the dynamic scheduling are received, and the control results of the molten steel temperature are returned, providing decision support at the temperature level for optimizing the production process scheduling.

[0036] Specifically, the present invention discloses a method for online temperature control of molten steel in a steelmaking plant, comprising the following steps:

[0037] Step S1, obtaining information such as parameters, structure, and operation status of the smelting equipment and ladle of the main process of the steel plant, obtaining historical production data of the main process, and obtaining a data set for model training for each process after screening;

[0038] Step S2, using numerical simulation technology combined with the on-site measured ladle lining temperature, classifying ladles in different thermal states to obtain the temperature drop of molten steel in each category of ladles;

[0039] Step S3, based on the training data set, select the historical heats with better performance, and establish a molten steel temperature prediction model using a case-based reasoning algorithm in combination with the ladle temperature drop;

[0040] Step S4, based on the training data set, clustering is performed using the K-means algorithm, a BPNN model is established for each type of clustered data set, and a molten steel temperature prediction model is established in combination with the ladle temperature drop;

[0041] Step S5, forming an online control system for molten steel temperature, obtaining time information provided by the dynamic scheduling system of the steel plant, calling the molten steel temperature prediction model to display the predicted temperature of all subsequent processes before and during the smelting of each furnace of molten steel and after the scheduling time information is updated, and calling the molten steel temperature prediction model to display the predicted temperature of the next process in turn;

[0042] Step S6, the scheduled molten steel temperature, predicted molten steel temperature and measured molten steel temperature are returned to the steel plant dynamic scheduling system, which adjusts the production plan and performs overhead crane scheduling based on the difference between the measured temperature and the target temperature, and returns the time information to the molten steel temperature online control system.

[0043] The process of establishing the predetermined model of molten steel temperature in step S3 includes:

[0044] Step S31, select the historical heats with better performance, take all factors affecting the temperature of molten steel from converter to continuous casting as input, and take the temperature of each process node as output, to establish a case-based reasoning data set;

[0045] Step S32, input a new heat, select a historical heat in the data set with the same ladle thermal state category and the highest similarity of other influencing factors, and use the temperature of this historical heat at each process node as the target temperature of the new heat.

[0046] The process of establishing the molten steel temperature prediction model in step S4 includes:

[0047] Step S41, analyzing the influencing factors of the molten steel temperature in each process section, and based on the data set, selecting variables with a larger Pearson correlation coefficient and consistent with the mechanism analysis, except for the ladle thermal state, as the dimensions for subsequent modeling;

[0048] Step S42, using the K-means cluster analysis method, trying to increase the number of cluster centers by enumeration method, calculating the number of clusters when the rate of decrease of the sum of squared errors slows down significantly, and using the number of clusters to divide the data set;

[0049] Step S43, establishing BPNN models for the divided data sets respectively, and calling the model of the data set to which each heat of molten steel belongs for prediction;

[0050] Step S44, for the heats with abnormal ladle thermal state, subtract the corresponding temperature correction value from the model predicted temperature to obtain the corrected predicted molten steel temperature.

[0051] The operation logic of the online molten steel temperature control system described in step S5 in practical application includes:

[0052] Step S51, the system obtains the production plan and ladle thermal state provided by the dynamic scheduling system of the steel plant, and pre-sets the molten steel temperature. Starting from the converter smelting, the scheduling personnel and the operators of each process can see the recommended target temperature provided by the system in advance;

[0053] Step S52, if a new scheduling plan is generated during the smelting process, the target temperature will be dynamically adjusted according to the change of rhythm;

[0054] Step S53, after the specific operation of each process (such as adding materials, turning on and off the oxygen supply, turning on and off the power supply, etc.) and at regular intervals, the molten steel temperature prediction model will automatically predict the temperature of the molten steel after it reaches the target node;

[0055] In step S54, on-site personnel determine subsequent intra-process operations and inter-process scheduling based on the degree of deviation between the predicted temperature and the predetermined temperature, and ultimately make the predicted molten steel temperature at each process node close to the predetermined molten steel temperature.

[0056] Embodiment: The embodiment of the present invention provides a method for online temperature control of molten steel in a steelmaking plant, comprising the following steps:

[0057] Step S1, obtain the parameters, structure, operation status and other information of the smelting equipment and ladle of the main process of the steel plant, obtain the historical production data of the main process, and obtain the data set for model training of each process after screening.

[0058] In this step, the information that needs to be obtained includes relevant information of the steelmaking workshop, such as the overall layout, equipment parameters, production steel types, process paths, etc., as well as ladle-related ladle diagrams, bricklaying thermophysical parameters, ladle turnover, etc. Among them, the relevant information of the steelmaking workshop is used for the selection of factors affecting molten steel temperature, the classification of models, the determination of standard smelting time and standard molten steel temperature drop, and the relevant information of the ladle is mainly used for numerical simulation and classification of the ladle. The screening of historical production data mainly includes the elimination of blank values ​​and abnormal values, data normalization, etc.

[0059] Step S2, using numerical simulation technology combined with the on-site measured ladle lining temperature, ladles in different thermal states are classified to obtain the temperature drop of molten steel in each category of ladles.

[0060] In this step, since the ladle is used as a container for holding, transporting molten steel and refining outside the furnace, the influence of its thermal state on the temperature of the molten steel runs through the entire process. It is necessary to classify the ladle according to its thermal state to facilitate subsequent analysis and modeling. According to the turnover of the ladle, the thermal state of the ladle is mainly composed of ladle age, empty ladle time and baking time. Taking a steel mill as an example, the ladle of the steel mill needs to be repaired more than 20 times a week, which is divided into minor repairs, medium repairs and major repairs, and different maintenance operations are carried out until the lining bricks are completely removed and newly built. Considering that baking is required before the repair is completed and the ladle goes online again, and the 4 furnace ladles before going online have a large heat storage space, the ladle is divided into the first 4 furnaces of the new ladle going online, the 5th furnace of the new ladle going online to maintenance, the first 4 furnaces after minor repairs, and the 5th furnace after minor repairs going online to maintenance. For the empty ladle time, according to the overall production rhythm and ladle turnover procedures, the ladles with empty ladle time less than 90 minutes, between 90-150 minutes, and greater than 150 minutes are marked as A, B, and C. Then, numerical simulation is performed on each category of ladles to obtain the temperature drop rate of each category of ladles during the static time period, and corrections are made based on historical production data to obtain the compensation value of the molten steel temperature in different ladle thermal states in each section.

[0061] Step S3, based on the training data set, select the historical heats with better performance, and use the case-based reasoning algorithm to establish a molten steel temperature prediction model in combination with the ladle temperature drop.

[0062] In this step, based on the filtered historical production data set obtained in step S1 and the ladle classification result obtained in step S2, a molten steel temperature predetermined model is established by using a case-based reasoning method. The specific establishment process includes the following steps:

[0063] Step S31, screen out historical heats with better performance, take all factors that affect the temperature of molten steel from converter to continuous casting as input, and use the temperature of each process node as output to establish a case-based reasoning data set.

[0064] The screening criteria are: continuous casting superheat meets the standard, the temperature of each process node meets the temperature system, and the smelting time or transfer time of each section is short. Among these heats, the heats with LF power supply between the middle 50% (25%-75%) in the converter-LF refining-continuous casting path and the heats without RH oxygen blowing operation in the converter-RH refining-continuous casting path are retained. The factors that affect the temperature of molten steel that can be known before the process smelting mainly include: ladle heat state, pouring sequence, planned processing time in each process, planned transfer time between each process, refining heating power, converter tapping times, alloy addition amount, etc. The process nodes mainly include the converter end point, refining start, refining end, continuous casting start, etc.

[0065] Step S32, input a new heat, select a historical heat with the same ladle thermal state category and the highest similarity of other influencing factors in the data set, and use the temperature of the historical heat at each process node as the target temperature of the new heat. The similarity can be calculated using Euclidean distance similarity, the Euclidean distance calculation formula is shown in formula (1), and the similarity calculation formula is shown in formula (2).

[0066]

[0067] Where d(X,Y) represents the Euclidean distance between the new heat and the heat in the case library, S sim (X,Y) is the Euclidean distance similarity, m is the number of influencing factors in a heat, xj represents the jth influencing factor of the new heat, y j represents the jth influencing factor of the heat in the case library, w j Represents the weight of the jth influencing factor.

[0068] This embodiment uses a total of 3667 data from a steel plant for illustration. Through step S31, 1622 furnaces are screened out, 1000 of which are randomly selected as the data set for case reasoning, and the remaining 622 furnaces are used as test sets for verification. The model predetermined results and the actual process node molten steel temperature of the test set furnace are compared, and the hit rate within the allowable error range is obtained as shown in Table 1. In order to ensure the actual application effect, the error range of the molten steel temperature at the refining end node is selected as [-5℃, 5℃], and the error range of the molten steel temperature at the other front process nodes is selected as [-7℃, 7℃].

[0069] Table 1

[0070] Converter end point Steelmaking Argon Station Refining begins Refining is finished Full node hit 91.48％ 92.12％ 93.57％ 94.86％ 90.03％

[0071] Step S4, based on the training data set, clustering is performed using the K-means algorithm, a BPNN model is established for each type of clustered data set, and a molten steel temperature prediction model is established in combination with the ladle temperature drop.

[0072] In this step, based on the filtered historical production data set obtained in step S1 and the ladle classification result obtained in step S2, a molten steel temperature prediction model is established. The model structure is as follows: Figure 1 As shown, the specific establishment process includes the following steps:

[0073] Step S41, analyzing the influencing factors of the molten steel temperature in each process section, and based on the data set, selecting variables with a larger Pearson correlation coefficient and consistent with the mechanism analysis, except for the thermal state of the ladle, as the dimensions for subsequent modeling.

[0074] Step S42, using the K-means cluster analysis method, trying to increase the number of cluster centers by enumeration method, calculating the number of clusters when the rate of decrease of the sum of squared errors slows down significantly, and using the number of clusters to divide the data set;

[0075] Step S43, establishing BPNN models for the divided data sets respectively, and calling the model of the data set to which each heat of molten steel belongs for prediction;

[0076] Step S44, for the heats with abnormal ladle thermal state, subtract the corresponding temperature correction value from the model predicted temperature to obtain the corrected predicted molten steel temperature.

[0077] This embodiment uses a total of 3667 data from a steel plant for illustration. Taking the process from the start of LF refining to the end of LF refining as an example, the factors affecting the molten steel temperature with a larger Pearson correlation coefficient are selected as shown in Table 2.

[0078] Table 2

[0079]

[0080]

[0081] The factors in Table 2 are used as the dimensions of Kmeans cluster analysis. The number of cluster centers is increased by enumeration method. The sum of squared errors is calculated. The number of clusters is selected as 4, and the data set is divided into 4 categories. For each type of data set, the influencing factors other than the thermal state of the ladle in Table 2 are used as the input of the BP neural network model. The BP neural network model is established. For the furnaces of ladles whose empty ladle time mark is not A or whose turnover is less than 5 times, the corresponding temperature correction value is subtracted from the temperature predicted by the BP model to obtain the corrected predicted molten steel temperature. 2500 furnaces are used as training data, and the remaining 1167 furnaces are used as test data for verification. The prediction hit rate of molten steel temperature under different error ranges of this method is shown in Table 3.

[0082] Table 3

[0083] [-5℃,5℃] [-7℃,7℃] [-10℃,10℃] 87.23％ 93.14％ 97.86％

[0084] Step S5, forms an online control system for molten steel temperature, obtains the time information provided by the dynamic scheduling system of the steel plant, and calls the molten steel temperature prediction model to display the predicted temperature of all subsequent processes before and during the smelting of each batch of molten steel and after the scheduling time information is updated, and calls the molten steel temperature prediction model in turn to display the predicted temperature of the next process.

[0085] In this step, for each furnace of molten steel, the triggering process of the molten steel temperature online control system is as follows: Figure 2 As shown, the temperature prediction model is triggered before the end of converter smelting. It takes the liquidus temperature of the steel grade as the starting point, and gradually reverses the temperature drop in each stage according to the superheat, production plan, etc., and formulates the optimized temperature system for each node, and then compensates according to the thermal state of the ladle to obtain the target temperature of each node and update it in time. The temperature prediction model is triggered during the smelting process from converter to continuous casting. It is based on the previous process node and calculates the molten steel temperature of the next process node according to the currently known production conditions, providing a basis for determining the temperature control parameters of each process and providing guidance for operators to perform temperature control operations. It specifically includes the following steps:

[0086] Step S51, the system obtains the production plan and ladle thermal state provided by the dynamic scheduling system of the steel plant, and pre-sets the molten steel temperature. Starting from the converter smelting, the scheduling personnel and the operators of each process can see the recommended target temperature provided by the system in advance;

[0087] Step S52, if a new scheduling plan is generated during the smelting process, the target temperature will be dynamically adjusted according to the change of rhythm;

[0088] Step S53, after the specific operation of each process (such as adding materials, turning on and off the oxygen supply, turning on and off the power supply, etc.) and at regular intervals, the molten steel temperature prediction model will automatically predict the temperature of the molten steel after it reaches the target node;

[0089] In step S54, on-site personnel determine subsequent intra-process operations and inter-process scheduling based on the degree of deviation between the predicted temperature and the predetermined temperature, and ultimately make the predicted molten steel temperature at each process node close to the predetermined molten steel temperature.

[0090] Step S6, the scheduled molten steel temperature, predicted molten steel temperature and measured molten steel temperature are returned to the steel plant dynamic scheduling system, which adjusts the production plan and performs overhead crane scheduling based on the difference between the measured temperature and the target temperature, and returns the time information to the molten steel temperature online control system.

[0091] In this step, the dynamic scheduling system of the steel plant receives the steelmaking production plan issued by the superior system, and generates the operation plan for the production site based on the actual production performance, logistics information and abnormal working conditions. The molten steel temperature control system will receive the operation plan, obtain basic information such as the steel type and process path of the specific furnace, as well as time information such as the planned smelting time in each process and the transportation time between processes. When the operation plan changes, the molten steel temperature control system will recalculate accordingly. After the calculation is completed, the molten steel temperature control system will transmit the results and the actual temperature measurement values ​​of each furnace tracked to the dynamic scheduling system for analysis and adjustment. The temperature reservation and prediction results can be presented separately or integrated into the Gantt chart interface of the dynamic scheduling system.

[0092] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0093] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0094] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for online control of molten steel temperature in a steelmaking plant, characterized in that: The following steps are involved: Obtain relevant information about the smelting equipment and ladles in the main process of the steel plant, and filter the historical production data of the main process to obtain the data set for model training in each process; Based on the ladle lining temperature, ladles with different thermal states are classified to obtain the temperature drop of molten steel in each category of ladles; Based on the data set, historical heats are screened, and according to the temperature drop of the ladle, a molten steel temperature prediction model is established using a case-based reasoning algorithm, and a molten steel temperature prediction model is established using K-means and BPNN algorithms; Based on the molten steel temperature booking model and the molten steel temperature prediction model, before and during the smelting of each furnace of molten steel and after the scheduling time information is updated, the scheduled temperatures of all subsequent processes and the predicted temperature of the next process are obtained respectively, and the production plan is adjusted according to the scheduled molten steel temperature, the predicted molten steel temperature and the actual molten steel temperature.

2. The method for online control of molten steel temperature in a steelmaking plant according to claim 1, characterized in that: In the process of obtaining relevant information about the smelting equipment and ladle in the main process of the steelmaking plant, relevant information of the steelmaking workshop is obtained, including: overall layout, equipment parameters, production steel types and process paths, and relevant information of the ladle, including: ladle masonry diagram, brick masonry thermophysical parameters, and ladle turnover status. Among them, the relevant information of the steelmaking workshop is used for the selection of factors affecting molten steel temperature, classification of models, determination of standard smelting time and standard molten steel temperature drop, and the relevant information of the ladle is used for numerical simulation and classification of the ladle.

3. The method for online control of molten steel temperature in a steelmaking plant according to claim 2, characterized in that: In the process of obtaining the temperature drop of molten steel in each category of ladles, the ladle age, empty ladle time and baking time are used as the thermal state of the ladle according to the ladle turnover situation, and numerical simulation is carried out on each category of ladle to obtain the temperature drop rate of each category of ladles in the static time period, and correction is made according to the historical production data to obtain the compensation value of the molten steel temperature in different thermal states of ladles in each section. Among them, for the empty ladle time, according to the overall production rhythm and ladle turnover procedures, the ladles with empty ladle time less than 90min, between 90-150min, and greater than 150min are marked respectively.

4. The method for online control of molten steel temperature in a steelmaking plant according to claim 3, characterized in that: In the process of building the molten steel temperature prediction model, we screened out the historical heats with better performance, took all the factors that affect the molten steel temperature from converter to continuous casting as input, and used the temperature of each process node as output to establish a case-based reasoning data set. Input a new heat, select a historical heat in the data set with the same ladle thermal state category and the highest similarity of other influencing factors, and use the temperature of this historical heat at each process node as the target temperature of the new heat.

5. The method for online control of molten steel temperature in a steelmaking plant according to claim 4, characterized in that: In the process of obtaining the predicted molten steel temperature, the influencing factors of the molten steel temperature in each process section are analyzed, and based on the data set, variables with a larger Pearson correlation coefficient and consistent with the mechanism analysis are selected as the dimensions for subsequent modeling, except for the ladle thermal state; Using the K-means cluster analysis method, we tried to increase the number of cluster centers through enumeration, calculated the number of clusters when the rate of decrease of the sum of squared errors slowed down significantly, and used this number of clusters to divide the data set. Based on the BPNN model, according to the divided data set, for each heat of molten steel, the model of the data set to which it belongs is called for prediction; For heats with abnormal ladle thermal conditions, the corresponding temperature correction value is subtracted from the model predicted temperature to obtain the corrected predicted molten steel temperature.

6. The method for online control of molten steel temperature in a steelmaking plant according to claim 5, characterized in that: In the process of obtaining the predetermined temperature of all subsequent processes and the predicted temperature of the next process, for each furnace of molten steel, it is triggered before the end of converter smelting, and the liquidus temperature of the steel grade is used as the starting point. The temperature drop in each stage is gradually reversed according to the superheat and production plan, and the optimized temperature system for each node is formulated. Then, compensation is made according to the thermal state of the ladle to obtain the target temperature of each node and update it in time. Among them, the temperature prediction model is triggered in the smelting process from converter to continuous casting. Based on the previous process node and according to the currently known production conditions, the molten steel temperature of the next process node is calculated, providing a basis for determining the temperature control parameters of each process.

7. The method for online control of molten steel temperature in a steelmaking plant according to claim 6, characterized in that: In the process of adjusting the production plan, receive the steelmaking production plan issued by the superior system, and generate the operation plan for the production site based on the actual production performance, logistics information and abnormal working conditions; Based on the received operation plan, we obtain basic information about the steel type and process path of the specific furnace, as well as the planned smelting time in each process and the transportation time between processes. When the operation plan changes, we make corresponding adjustments.

8. An online control system for molten steel temperature in a steelmaking plant, characterized in that: The system is used to implement an online control method for molten steel temperature in a steelmaking plant as claimed in any one of claims 1 to 7, and the system comprises: The data acquisition module is used to obtain relevant information about the smelting equipment and ladles in the main process of the steel plant. According to the historical production data of the main process, the data set for model training of each process is obtained after screening; The data analysis module is used to classify ladles of different thermal states based on the ladle lining temperature and obtain the temperature drop of molten steel in each category of ladles; A model building module is used to screen historical heats based on the data set, establish a molten steel temperature prediction model using a case-based reasoning algorithm according to the ladle temperature drop, and establish a molten steel temperature prediction model using a K-means and BPNN algorithm; The online control module is used to obtain the scheduled temperature of all subsequent processes and the predicted temperature of the next process based on the molten steel temperature booking model and the molten steel temperature prediction model before and during the smelting of each furnace of molten steel and after the scheduling time information is updated, and adjust the production plan according to the scheduled molten steel temperature, the predicted molten steel temperature and the actual molten steel temperature.

Images