Batching method, device, storage medium and computer equipment

Abstract

The present invention discloses a batching method, device, storage medium, and computer equipment, which relate to the technical field of molten steel smelting, and are primarily capable of improving the batching accuracy of alloys during the molten steel refining process and reducing batching costs. The method comprises: obtaining the current alloying element content, alloying element batching interval, and initial molten steel quality corresponding to the molten steel to be refined, furnace condition data of the refining furnace, and alloy property data of the alloy to be added; determining a predicted alloying element yield expected value and variance based on historical alloying element yields during historical smelting processes, and determining a predicted molten steel quality expected value and variance based on the initial molten steel quality; inputting the predicted alloying element yield expected value and variance, the predicted molten steel quality expected value and variance, the current alloying element content, alloying element batching interval, furnace condition data, and alloy property data into a preset alloy batching quantity prediction model to obtain a predicted alloy batching quantity; and batching the molten steel to be refined based on the predicted batching quantity.

Description

Technical Field

[0001] The present invention relates to the technical field of molten steel smelting, and in particular to a batching method, device, storage medium and computer equipment. Background Art

[0002] The LF furnace (Ladle Furnace) is the primary off-furnace refining equipment in steel production. Its primary smelting functions are to heat and maintain the temperature of molten steel, adjust its composition, and improve its purity. Refining in the LF furnace results in higher-quality molten steel. Composition adjustment is a key goal of LF furnace refining, primarily through the selective addition of alloying materials to ensure that the final alloying element composition of the molten steel, such as C, Mn, Si, and Cr, falls within the desired control range for the steel grade.

[0003] Currently, alloy formulations for various steelmaking processes are typically predicted using fixed element yields and molten steel quality. However, these parameters, such as element yield and molten steel quality, vary across different steelmaking processes and can change with changing smelting conditions. Consequently, these formulations can easily lead to excessive element content in the final molten steel, increasing alloy formulation costs and reducing alloy formulation accuracy. Summary of the Invention

[0004] The present invention provides a batching method, device, storage medium and computer equipment, which are mainly capable of improving the batching accuracy of alloys in the process of molten steel refining and reducing the batching cost of alloys in the process of molten steel refining.

[0005] According to a first aspect of the present invention, there is provided a batching method comprising:

[0006] Obtaining the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and obtaining the historical alloying element yield rate of the refining furnace during historical molten steel smelting processes;

[0007] Determining an expected value of a predicted alloying element yield and a predicted alloying element yield variance corresponding to the molten steel to be refined based on the historical alloying element yield, and determining an expected value of a predicted molten steel quality and a predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality;

[0008] Inputting the predicted alloying element yield expected value, the predicted alloying element yield variance, the predicted molten steel quality expected value, the predicted molten steel quality variance, the current alloying element content, the alloying element batching interval, the furnace condition data, and the alloy property data into a preset alloy batching amount prediction model to perform batching prediction, thereby obtaining the predicted batching amount of the alloy in the molten steel to be refined;

[0009] The molten steel to be refined is batched according to the predicted batching amount of the alloy.

[0010] Preferably, the alloy property data includes the alloy unit price of the alloy to be added and the element content in the alloy to be added.

[0011] Preferably, determining the expected value of the predicted alloying element yield and the predicted alloying element yield variance corresponding to the molten steel to be refined based on the historical alloying element yield comprises:

[0012] The current alloying element content and the historical alloying element yield are regressed and analyzed using a Gaussian process regression function to obtain an expected value and a variance of a predicted alloying element yield corresponding to the molten steel to be refined.

[0013] Preferably, determining the predicted molten steel quality expectation value and the predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality includes:

[0014] The Gaussian probability density function is used to perform density analysis on the probability distribution of the initial molten steel quality to obtain the expected value of the predicted molten steel quality and the predicted molten steel quality variance corresponding to the molten steel to be refined.

[0015] Preferably, the preset alloy batch quantity prediction model includes a batch quantity constraint layer and a batch value optimization layer, and the predicted alloy element yield expected value, predicted alloy element yield variance, predicted molten steel quality expected value, predicted molten steel quality variance, current alloy element content, alloy element batching interval, furnace condition data, and alloy property data are input into the preset alloy batch quantity prediction model to perform batching prediction, thereby obtaining the predicted batch quantity of the alloy in the molten steel to be refined, including:

[0016] Obtaining the alloy yield of the alloy to be added corresponding to the molten steel to be refined;

[0017] Input the predicted alloying element yield expected value, predicted alloying element yield variance, predicted molten steel quality expected value, predicted molten steel quality variance, current alloying element content, alloying element batching interval, element content in the alloy to be added, alloy yield and furnace condition data into the batching amount constraint layer to perform batching interval prediction, and obtain the predicted batching amount allowable interval corresponding to the alloy to be added;

[0018] Selecting an initial alloy batch quantity within the predicted batch quantity tolerance range according to an optimization algorithm, and determining a search step length corresponding to the initial alloy batch quantity;

[0019] Inputting the initial alloy batching amount and the alloy unit price into the batching value optimization layer for value prediction to obtain the batching value corresponding to the initial alloy batching amount;

[0020] The predicted alloy proportion in the molten steel to be refined is determined according to the proportion value corresponding to the initial alloy proportion and the search step length.

[0021] Preferably, determining the predicted alloy proportion in the molten steel to be refined according to the proportion value corresponding to the initial alloy proportion and the search step size includes:

[0022] According to the ingredient value corresponding to the initial alloy ingredient amount and the search step size, an iterative search is performed on the alloy ingredient amounts of each combination within the predicted ingredient amount allowable range, and during the iterative search process, it is determined whether the ingredient value corresponding to the alloy ingredient amount searched next is less than the ingredient value corresponding to the alloy ingredient amount searched previously;

[0023] If it is less than the value of the alloy ingredient amount corresponding to the last search, continue the iterative search and determine whether the search step is less than the preset convergence threshold;

[0024] If it is less than the preset convergence threshold, the iterative search is stopped, and the ingredient value corresponding to the alloy ingredient amount searched last time is determined;

[0025] The minimum proportioning value is determined between the proportioning value corresponding to the alloy proportioning quantity searched last time and the proportioning value corresponding to the alloy proportioning quantity searched last time, and the alloy proportioning quantity corresponding to the minimum proportioning value is determined as the predicted proportioning quantity of the alloy in the molten steel to be refined.

[0026] Preferably, it is characterized in that, before obtaining the current alloying element content corresponding to the molten steel to be refined, the method further comprises:

[0027] performing slagging and temperature-raising treatments on the molten steel to be refined to obtain treated molten steel to be refined;

[0028] The obtaining of the current alloying element content corresponding to the molten steel to be refined includes:

[0029] Obtain the current alloying element content of the processed molten steel to be refined.

[0030] Preferably, it is characterized in that the step of batching the molten steel to be refined according to the predicted batching amount of the alloy comprises:

[0031] According to the predicted proportion of the alloy, the alloy to be added is added to the molten steel to be refined.

[0032] According to a second aspect of the present invention, there is provided a batching device comprising:

[0033] an acquisition unit, configured to acquire the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning range corresponding to the type of molten steel to be refined, furnace condition data of the refining furnace corresponding to the molten steel to be refined, alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and historical alloying element yields of the refining furnace during historical molten steel smelting processes;

[0034] a determination unit, configured to determine an expected value of a predicted alloying element yield and a predicted alloying element yield variance corresponding to the molten steel to be refined based on the historical alloying element yield, and to determine an expected value of a predicted molten steel quality and a predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality;

[0035] a prediction unit, configured to input the predicted alloying element yield expected value, the predicted alloying element yield variance, the predicted molten steel quality expected value, the predicted molten steel quality variance, the current alloying element content, the alloying element batching interval, the furnace condition data, and the alloy property data into a preset alloy batching quantity prediction model to perform batching prediction, thereby obtaining a predicted batching quantity of the alloy in the molten steel to be refined;

[0036] A batching unit is used to batch the molten steel to be refined according to the predicted batching amount of the alloy.

[0037] According to a third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the above batching method is implemented.

[0038] According to a fourth aspect of the present invention, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above batching method when executing the program.

[0039] According to a batching method, device, storage medium and computer equipment provided by the present invention, compared with the current method of using the same fixed element yield and molten steel quality to predict the element batching scheme in each molten steel smelting process, the present invention obtains the current alloy element content corresponding to the molten steel to be refined, the alloy element batching range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and the historical alloy element yield of the refining furnace in the historical molten steel smelting process; and based on the historical alloy element yield , determine the expected value of the predicted alloy element yield and the predicted alloy element yield variance corresponding to the molten steel to be refined, and based on the initial molten steel quality, determine the expected value of the predicted molten steel quality and the predicted molten steel quality variance corresponding to the molten steel to be refined; then input the predicted alloy element yield expected value, the predicted alloy element yield variance, the predicted molten steel quality expected value, the predicted molten steel quality variance, the current alloy element content, the alloy element batching interval, the furnace condition data and the alloy property data into the preset alloy batching prediction model for batching prediction, and obtain the predicted batching amount of the alloy in the molten steel to be refined; finally, the molten steel to be refined is batched according to the predicted batching amount of the alloy. Therefore, by first predicting the expected value and variance of the element yield in the current molten steel refining process based on the historical alloying element yield, and at the same time predicting the expected value and variance of the molten steel quality based on the initial molten steel quality, the batching scheme of the molten steel to be refined is then determined based on the predicted expected value and variance of the element yield and the expected value and variance of the molten steel quality. This can overcome the parameter uncertainty problem caused by complex and changeable working conditions, and also avoid problems such as the excessive content of various elements in the final molten steel, thereby reducing the batching cost of the molten steel in the smelting process and improving the batching accuracy of the molten steel in the smelting process. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0041] Figure 1 A flow chart of a batching method provided by an embodiment of the present invention is shown;

[0042] Figure 2 A flow chart of another batching method provided by an embodiment of the present invention is shown;

[0043] Figure 3 A schematic structural diagram of a batching device provided by an embodiment of the present invention is shown;

[0044] Figure 4A schematic structural diagram of another batching device provided by an embodiment of the present invention is shown;

[0045] Figure 5 A schematic diagram of the physical structure of a computer device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0046] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0047] At present, the method of using the same fixed element yield and molten steel quality to predict the elemental proportioning scheme in each molten steel smelting process will lead to problems such as the content of each element in the final molten steel exceeding the limit, thereby resulting in higher proportioning costs for the molten steel during the smelting process, and also reducing the accuracy of the proportioning of the molten steel during the smelting process.

[0048] In order to solve the above problems, the embodiment of the present invention provides a batching method, such as Figure 1 As shown, the method includes:

[0049] 101. Obtain the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and obtain the historical alloying element recovery rate of the refining furnace during the historical smelting process of the molten steel.

[0050] Among them, the alloy property data include the alloy unit price of the alloy to be added and the element content in the alloy to be added; the furnace condition data include the temperature, material, wall thickness and other data of the refining furnace. The historical alloy element recovery rate refers to the element recovery rate of the refining furnace in the previous molten steel smelting process before this molten steel smelting; the alloy element recovery rate means that the alloy raw materials melted into the molten steel cannot convert all kinds of elements inside it into the molten steel 100%, and there will be some loss. Therefore, there is an alloy element recovery rate. The ratio of the part of the element that will be completely converted into the molten steel to the total amount of the element is called the alloy element recovery rate; the alloy element batching range refers to the range between the upper and lower limits of various alloy elements added to the molten steel; the alloy is composed of a variety of alloy elements, and the alloy element content refers to the proportion of various elements in the alloy to be added to the alloy to be added before the molten steel is added.

[0051] In the embodiment of the present invention, the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and the historical alloying element yield rate of the refining furnace in the historical molten steel smelting process are stored in the remote database. The above data can be obtained in the remote database, and the historical element yield rate can be used to predict the expected value and variance of the element yield rate in the current molten steel smelting process. The initial molten steel quality can be used to determine the expected value and variance of the predicted molten steel quality corresponding to the molten steel to be refined. Finally, the predicted expected value and plan of the alloying element yield rate, as well as the predicted expected value and variance of the molten steel quality, can be used to predict the proportioning plan for the molten steel to be refined. This can overcome the uncertainty problem of parameters such as element yield rate and molten steel quality caused by complex and changeable working conditions, thereby improving the proportioning accuracy of the molten steel during the smelting process and reducing the proportioning cost.

[0052] 102. Based on the historical alloying element yields, determine the expected value of the predicted alloying element yield and the predicted alloying element yield variance corresponding to the molten steel to be refined, and based on the initial molten steel quality, determine the expected value of the predicted molten steel quality and the predicted molten steel quality variance corresponding to the molten steel to be refined.

[0053] In an embodiment of the present invention, after obtaining the current alloying element content and historical alloying element yield of the molten steel to be refined, as well as the initial molten steel quality before smelting, a Gaussian process regression function is used to perform regression analysis on the current alloying element content and historical alloying element yield to obtain the expected value and variance of the predicted alloying element yield corresponding to the molten steel to be refined. Simultaneously, a Gaussian probability density function is used to perform density analysis on the probability distribution of the initial molten steel quality to obtain the expected value and variance of the predicted molten steel quality corresponding to the molten steel to be refined. This overcomes the uncertainty issues associated with parameters such as element yield and molten steel quality caused by complex and variable working conditions, thereby avoiding the problem of excessive alloying elements added to the molten steel, saving on steel batching costs during the smelting process, and improving the accuracy of steel batching during the smelting process.

[0054] 103. The expected value of the predicted alloy element yield, the predicted alloy element yield variance, the expected value of the predicted molten steel quality, the predicted molten steel quality variance, the current alloy element content, the alloy element batching interval, the furnace condition data and the alloy property data are input into the preset alloy batching prediction model to perform batching prediction, and obtain the predicted batching amount of the alloy in the molten steel to be refined.

[0055] Among them, the predicted batching amount refers to the weight of various alloys that need to be added to the molten steel during the steel smelting process, and the total batching cost is minimized.

[0056] For the embodiment of the present invention, after predicting the expected value and variance of the element yield rate and the expected value and variance of the molten steel quality during the smelting process, the predicted expected value and variance of the alloy element yield rate, the predicted expected value and variance of the molten steel quality, the current alloy element content, the alloy element batching interval, the furnace condition data and the alloy property data are input into a preset alloy batching quantity prediction model for batching prediction. The preset alloy batching quantity prediction model can predict the weight of various alloys that need to be added to the molten steel, and finally the molten steel is batched according to the weight of various alloys. Predicting the batching quantity of various alloys by the preset alloy batching quantity prediction model can improve the prediction efficiency of the alloy batching quantity and improve the prediction accuracy of the alloy batching quantity, thereby improving the batching accuracy and batching efficiency of the molten steel to be refined and reducing the batching cost.

[0057] 104. The molten steel to be refined is proportioned according to the predicted proportion of the alloy.

[0058] In the embodiments of the present invention, after predicting the amounts of various alloys required to be added to the molten steel, the amount of alloy to be added is determined from the alloys to be added, and the alloys are ultimately added to the molten steel in accordance with the alloy amounts. This method overcomes the difficulty in quantifying and evaluating the original parameter uncertainties by predicting the expected values ​​and variances of element yields and molten steel quality. Ultimately, the molten steel batching plan is determined based on the predicted expected values ​​and variances of element yields and molten steel quality, thus avoiding problems such as excessive element content in the molten steel and reducing the batching costs during the smelting process.

[0059] According to a batching method provided by the present invention, compared with the current method of using the same fixed element yield and molten steel quality to predict the element batching scheme in each molten steel smelting process, the present invention obtains the current alloying element content corresponding to the molten steel to be refined, the alloying element batching range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and the historical alloying element yield of the refining furnace in the historical molten steel smelting process; and based on the historical alloying element yield, determines the content of the molten steel to be refined. The expected value of the predicted alloying element yield and the predicted alloying element yield variance corresponding to water, as well as the predicted molten steel quality expected value and the predicted molten steel quality variance corresponding to the molten steel to be refined are determined based on the initial molten steel quality; then the predicted alloying element yield expected value, the predicted alloying element yield variance, the predicted molten steel quality expected value, the predicted molten steel quality variance, the current alloying element content, the alloying element batching interval, the furnace condition data and the alloy property data are input into the preset alloy batching quantity prediction model for batching prediction, and the predicted batching quantity of the alloy in the molten steel to be refined is obtained; finally, the molten steel to be refined is batched according to the predicted batching quantity of the alloy. Therefore, by first predicting the expected value and variance of the element yield in the current molten steel refining process based on the historical alloying element yield, and at the same time predicting the expected value and variance of the molten steel quality based on the initial molten steel quality, the batching scheme of the molten steel to be refined is then determined based on the predicted expected value and variance of the element yield and the expected value and variance of the molten steel quality. This can overcome the parameter uncertainty problem caused by complex and changeable working conditions, and also avoid problems such as the excessive content of various elements in the final molten steel, thereby reducing the batching cost of the molten steel in the smelting process and improving the batching accuracy of the molten steel in the smelting process.

[0060] Furthermore, in order to better illustrate the above batching process, as a refinement and extension of the above embodiment, the embodiment of the present invention provides another batching method, such as Figure 2 As shown, the method includes:

[0061] 201. Obtain the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and obtain the historical alloying element yield rate of the refining furnace during the historical smelting process of the molten steel.

[0062] For the embodiment of the present invention, in order to improve the batching accuracy of molten steel, the molten steel to be refined needs to be pre-treated. Based on this, before obtaining the molten steel property data of the molten steel to be refined, the method also includes: slagging and heating the molten steel to be refined to obtain the treated molten steel to be refined.

[0063] Specifically, in order to improve the batching accuracy of molten steel, it is first necessary to remove impurities in the molten steel to be refined. First, slag-forming agents and other steelmaking materials, such as calcium fluoride or calcium oxide, are added to the molten steel. The slag in the molten steel to be refined is removed by the slag-forming agent. At the same time, in order to improve the smelting efficiency of the molten steel, the molten steel to be refined can be appropriately heated in advance, thereby obtaining the treated molten steel to be refined.

[0064] Furthermore, after obtaining the processed molten steel to be refined, the current alloy element content corresponding to the molten steel to be refined, the alloy element ratio range corresponding to the type of molten steel to which the molten steel to be refined belongs, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined (including the alloy unit price of the alloy to be added, and the element content in the alloy to be added), the initial molten steel quality corresponding to the molten steel to be refined, and the historical alloy element recovery rate of the refining furnace in the historical smelting process of the molten steel are obtained in the remote database.

[0065] 202. The current alloying element content and the historical alloying element yield are regressed using the Gaussian process regression function to obtain the expected value and variance of the predicted alloying element yield corresponding to the molten steel to be refined.

[0066] In the embodiment of the present invention, after obtaining the current alloying element content, the historical alloying element yield, and the initial molten steel quality, a Gaussian process regression function can be used to perform regression analysis on the current alloying element content and the historical alloying element yield to obtain the expected value and variance of the predicted alloying element yield corresponding to the molten steel to be refined. Specifically, the expected value and variance of the predicted alloying element yield can be calculated using the following formula:

[0067] μ j =PR(X j )

[0068] Among them, μ j represents the expected value of the j-th alloying element recovery rate, GPR(*) represents the Gaussian process regression function, X j is the current alloying element content, and the alloying element yield variance is calculated using the following formula:

[0069]

[0070] in, represents the variance of the j-th predicted alloying element yield, y j represents the historical alloying element yield, from which the variance of the yield of various alloying elements can be calculated.

[0071] 203. The probability distribution of the initial molten steel quality is analyzed using the Gaussian probability density function to obtain the expected value and variance of the predicted molten steel quality corresponding to the molten steel to be refined.

[0072] In the embodiment of the present invention, after obtaining the initial molten steel quality, a probability density function can be used to perform regression analysis on the initial molten steel quality to obtain the expected value and variance of the predicted molten steel quality corresponding to the molten steel to be refined. Specifically, the expected value and variance of the predicted molten steel quality can be calculated using the following formula:

[0073]

[0074] Among them, μ Q represents the expected value of the predicted molten steel quality, f(X) represents the probability density function of the molten steel quality, X represents the current alloying element content, and the variance of the predicted molten steel quality is calculated by the following formula:

[0075]

[0076] In the formula Represents the predicted molten steel quality variance, E Q (*) represents the calculation function of molten steel quality, X Q The initial quality of the molten steel to be refined is represented by , and the expected value and variance of the predicted molten steel quality can be calculated using the probability density function.

[0077] 204. The expected value of the predicted alloy element yield, the predicted alloy element yield variance, the expected value of the predicted molten steel quality, the predicted molten steel quality variance, the current alloy element content, the alloy element batching interval, the furnace condition data and the alloy property data are input into a preset alloy batching prediction model to perform batching prediction, and obtain the predicted batching amount of the alloy in the molten steel to be refined.

[0078] In embodiments of the present invention, when using a preset alloy proportioning prediction model to predict the proportioning scheme for molten steel, in order to optimize the solution algorithm for the alloy proportioning model, ensure rapid convergence, and obtain a global optimal solution to the problem, after constructing the preset alloy proportioning prediction model, the uncertain variables in the model, namely, alloy element yield and molten steel quality, need to be separated from the function body and converted into a second-order cone form. The formula containing the uncertain variables alloy element yield and molten steel quality is converted into a formula represented by expected value and variance, thereby quickly solving the model and obtaining the optimal proportioning scheme. Based on this, step 204 specifically includes: obtaining the alloy yield of the alloy to be added corresponding to the molten steel to be refined; inputting the expected value of the predicted alloy element yield, the predicted alloy element yield variance, the predicted expected value of the molten steel quality, the predicted molten steel quality variance, the current alloy element content, the alloy element batching interval, the element content in the alloy to be added, the alloy yield and the furnace condition data into the batching constraint layer for batching interval prediction, and obtaining the predicted batching allowable interval corresponding to the alloy to be added; according to the optimization algorithm, selecting the initial alloy batching within the predicted batching allowable interval, and determining the search step corresponding to the initial alloy batching; inputting the initial alloy batching and the alloy unit price into the batching value optimization layer for value prediction, and obtaining the batching value corresponding to the initial alloy batching; determining the predicted batching amount of the alloy in the molten steel to be refined according to the batching value corresponding to the initial alloy batching and the search step.

[0079] Among them, the preset alloy ingredient quantity prediction model includes an ingredient quantity constraint layer and an ingredient value optimization layer; the alloy yield, that is, the alloy material yield, refers to the actual production process. Due to various physical and chemical reactions such as crystallization and precipitation in the refining furnace, the added alloy will be partially lost, and the added alloy raw materials cannot be completely dissolved in the molten steel. A part of it will form slag. The ratio of the remaining part of the alloy after removing the slag to the total amount of alloy is called the alloy yield.

[0080] Specifically, first, the expected value and variance of the predicted alloy element yield, the expected value and variance of the predicted molten steel quality, the current alloy element content, the alloy element batching interval, the element content in the alloy to be added, the alloy yield and the furnace condition data are input into the batching constraint layer, and the alloy batching is feasibility screened through the batching constraint layer to obtain the predicted batching allowable interval corresponding to the alloy to be added, and then a set of initial alloy batching is determined within the preset batching allowable interval, and a search step is determined at the same time, wherein the value of the search step can be set as needed, and the embodiment of the present invention does not specifically limit the search step. The initial alloy batching is then input into the batching value optimization layer, and the batching value (batch cost) corresponding to the initial alloy batching is output through the batching value optimization layer, and finally the predicted batching amount of the alloy in the molten steel to be refined is determined based on the batching value corresponding to the initial alloy batching and the search step. Based on this, the method includes: according to the ingredient value corresponding to the initial alloy ingredient quantity and the search step, iteratively searching the alloy ingredient quantities of each combination within the predicted ingredient quantity allowable range, and judging during the iterative search whether the ingredient value corresponding to the alloy ingredient quantity searched next time is less than the ingredient value corresponding to the alloy ingredient quantity searched last time; if it is less than the ingredient value corresponding to the alloy ingredient quantity searched last time, continuing the iterative search and judging whether the search step is less than a preset convergence threshold; if it is less than the preset convergence threshold, stopping the iterative search and determining the ingredient value corresponding to the alloy ingredient quantity searched last time; determining the minimum ingredient value between the ingredient value corresponding to the alloy ingredient quantity searched last time and the ingredient value corresponding to the alloy ingredient quantity searched last time, and determining the alloy ingredient quantity corresponding to the minimum ingredient value as the predicted ingredient quantity of the alloy in the molten steel to be refined.

[0081] Among them, the preset convergence threshold is a value set according to actual needs, and the embodiment of the present invention does not impose a specific limit on the numerical size of the preset convergence threshold. Specifically, first determine a group of initial alloy ingredient quantities within the predicted ingredient quantity allowable range, and determine the ingredient value corresponding to the initial alloy ingredient quantity. Then, according to the search step length, search for the first group of ingredient quantities within the predicted ingredient quantity allowable range, and calculate the ingredient value corresponding to the first group of ingredient quantities. If the ingredient value corresponding to the first group of ingredient quantities is less than the ingredient value corresponding to the initial alloy ingredient quantity, then use the first group of ingredient quantities as the starting point to redetermine the search step length, and search for the second group of alloy ingredient quantities within the predicted ingredient quantity allowable range according to the redetermined search step length. At the same time, determine whether the redetermined search step length is smaller than the initial alloy ingredient quantity. If the value of the alloy batches in the second group is less than the preset convergence threshold, the search is stopped and a determination is made as to whether the value of the alloy batches in the second group is less than the value of the alloy batches in the previous group (the first group). If so, the alloy batches in the second group are directly determined as the predicted amount of the alloy to be added to the molten steel to be refined. If the value of the alloy batches in the second group is greater than or equal to the value of the alloy batches in the previous group (the first group), the previous group (the first group) is directly determined as the predicted amount of the alloy to be added to the molten steel to be refined. At the same time, if the newly determined search step size is greater than or equal to the preset convergence threshold, the iterative search is continued in the above manner until the alloy batch with the lowest value is found, and the batch with the lowest value is determined as the predicted amount of the alloy to be added to the molten steel to be refined.

[0082] In another embodiment of the present invention, in the process of repeatedly iteratively optimizing the alloy amount within the allowable range of the predicted batching amount to obtain the global optimal solution, the solution can be specifically obtained by the following formula:

[0083]

[0084] Among them, A ij is the element content in the alloy to be added, s i is the alloy yield, S j To predict the expected value and variance of alloying element yield, x i Indicates furnace condition data, represents the incremental predicted value of the alloying element, n represents the number of alloy types, i represents the i-th alloy, and j represents the j-th alloying element.

[0085]

[0086] in, Indicates the predicted value of the alloy element content at the end point, is the predicted value of alloy element increment, Q is the initial mass of molten steel to be refined, E j is the current alloying element content.

[0087]

[0088] in, Indicates the predicted upper limit content of the end-point alloying elements, It is the upper limit of the alloying element ratio range.

[0089] The constraint equation for the upper limit of the endpoint alloying element content is as follows:

[0090]

[0091] in, is the predicted value of the upper limit content of the endpoint alloying elements, Pr{*} represents the probability of its internal establishment, and α represents the confidence level of the constraint condition.

[0092] The calculation formula for the predicted lower limit content of the end-point alloying elements is as follows:

[0093]

[0094] in, Indicates the predicted lower limit content of the end-point alloying element, It is the lower limit of the alloying element ratio range.

[0095] The constraint equation for the lower limit of the endpoint alloying element content is as follows:

[0096]

[0097] in, It is the predicted value of the lower limit content of the endpoint alloying elements.

[0098] Based on the above formula, it is necessary to perform modeling transformation on the above formula. The constraint conditions in the transformation process are: random constraint inequalities are transformed into deterministic analytical constraint equations.

[0099] The transformation process of the alloy in molten steel to meet the upper limit requirements is as follows:

[0100] The expected value of the difference between the alloying element increment and the upper limit of the alloying element increment is calculated as follows:

[0101]

[0102] Among them, E δ1 The expected value of the difference between the alloying element increment and the upper limit of the alloying element increment, μ j Represents the expected value of the predicted alloying element yield.

[0103] The calculation formula for the upper limit value of the allowable interval of the predicted ingredient amount corresponding to the alloy to be added is as follows:

[0104]

[0105] Among them, E Q1, Indicates the upper limit of the allowable interval of the predicted ingredient amount corresponding to the alloy to be added, μ Q Represents the expected value of predicted molten steel quality.

[0106] The formula for calculating the variance of alloying element increment is as follows:

[0107]

[0108] in, represents the variance of the alloying element increment, Represents the variance of the predicted alloying element yield.

[0109] The formula for calculating the variance of the difference between the current alloying element content and the upper limit of the current alloying element content is as follows:

[0110]

[0111] in, Indicates the variance of the difference between the current alloying element content and the upper limit of the current alloying element content, Represents the predicted variance of molten steel quality.

[0112] The analytical constraint equation for the upper limit of the endpoint alloy content (the upper limit of the allowable range of the predicted ingredient amount corresponding to the alloy to be added) is as follows:

[0113]

[0114] Among them, E δ1 represents the expected value of the difference between the alloying element increment and the upper limit of the alloying element content, E Q1, Represents the upper limit of the allowable interval of the predicted ingredient amount corresponding to the alloy to be added, represents the variance of alloying element increments, Represents the variance of the difference between the current alloying element content and the upper limit of the current alloying element content, Φ -1 (α) represents the inverse function of the standard normal distribution function.

[0115] The transformation process of molten steel composition meeting the lower limit requirements is as follows:

[0116] The calculation formula for the expected value of the difference between the alloying element increment and the lower limit of the alloying element increment is as follows:

[0117]

[0118] Among them, Eδ2 The expected value of the difference between the alloying element increment and the lower limit of the alloying element increment

[0119] The calculation formula for the lower limit of the allowable interval of the predicted ingredient amount corresponding to the alloy to be added is as follows:

[0120]

[0121] The calculation formula for the variance of the difference between the current alloying element content and the current lower limit of the alloying element content is as follows:

[0122]

[0123] The analytical constraint equation for the lower limit of the allowable interval of the predicted ingredient amount corresponding to the alloy to be added is as follows:

[0124]

[0125] Among them, E δ2 represents the expected value of the difference between the alloying element increment and the lower limit of the alloying element increment, E Q2, Represents the expected value of the lower limit of the allowable interval of the predicted ingredient amount corresponding to the alloy to be added, represents the variance of alloying element increments, The analytical constraint equation represents the lower limit of the alloying element content. The above formula can be used to determine the allowable interval of the predicted ingredient quantity corresponding to the alloy to be added. Then, the target alloy ingredient quantity within the allowable interval of the predicted ingredient quantity needs to be determined according to the interior point method. The target alloy ingredient quantity and the alloy unit price are input into the ingredient value optimization layer to obtain the ingredient value corresponding to the target alloy ingredient quantity. The ingredient quantity of each group within the allowable interval of the predicted ingredient quantity is iteratively calculated based on the ingredient value. Specifically, the iterative calculation can be performed using the following formula:

[0126]

[0127] x i ≥0; i=1,2,…,n; j=1,2,…,m

[0128] Among them, C is the value of each batch of ingredients, n is the number of alloy types, c i The unit price of alloy type i, x i is the batching quantity for the i-th alloy. The above formula can be used to determine the batching value corresponding to each batching quantity group within the predicted batching quantity tolerance range, and then determine the predicted batching quantity with the lowest batching value among each batching quantity group. The above formula can then be used to determine the predicted batching quantity of the alloy to be added to the molten steel to be refined.

[0129] 205. The molten steel to be refined is proportioned according to the predicted proportion of the alloy.

[0130] According to the embodiment of the present invention, after determining the predicted amount of alloy to be added to the molten steel, the molten steel to be refined needs to be dosed. Based on this, step 205 specifically includes: adding the alloy to be added to the molten steel to be refined according to the predicted amount of alloy.

[0131] Specifically, the predicted alloy dosage refers to the mass of alloy that needs to be added to the molten steel to be refined. According to the predicted alloy dosage, the corresponding mass of alloy is obtained from the alloy to be added and added to the molten steel being smelted.

[0132] According to another batching method provided by the present invention, compared with the current method of using the same fixed element yield and molten steel quality to predict the element batching scheme in each molten steel smelting process, the present invention obtains the current alloy element content corresponding to the molten steel to be refined, the alloy element batching range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and the historical alloy element yield of the refining furnace in the historical molten steel smelting process; and based on the historical alloy element yield, determines the content of the molten steel to be refined. The expected value of the predicted alloying element yield and the predicted alloying element yield variance corresponding to the molten steel, as well as the predicted molten steel quality expected value and the predicted molten steel quality variance corresponding to the molten steel to be refined are determined based on the initial molten steel quality; then the predicted alloying element yield expected value, the predicted alloying element yield variance, the predicted molten steel quality expected value, the predicted molten steel quality variance, the current alloying element content, the alloying element batching interval, the furnace condition data and the alloy property data are input into a preset alloy batching quantity prediction model for batching prediction, and the predicted batching quantity of the alloy in the molten steel to be refined is obtained; finally, the molten steel to be refined is batched according to the predicted batching quantity of the alloy. Therefore, by first predicting the expected value and variance of the element yield in the current molten steel refining process based on the historical alloying element yield, and at the same time predicting the expected value and variance of the molten steel quality based on the initial molten steel quality, the batching scheme of the molten steel to be refined is then determined based on the predicted expected value and variance of the element yield and the expected value and variance of the molten steel quality. This can overcome the parameter uncertainty problem caused by complex and changeable working conditions, and also avoid problems such as the excessive content of various elements in the final molten steel, thereby reducing the batching cost of the molten steel in the smelting process and improving the batching accuracy of the molten steel in the smelting process.

[0133] Further, as Figure 1 The specific implementation of the present invention provides a batching device, such as Figure 3 As shown, the device includes: an acquisition unit 31, a determination unit 32, a prediction unit 33 and a batching unit 34.

[0134] The acquisition unit 31 can be used to obtain the current alloy element content corresponding to the molten steel to be refined, the alloy element ratio range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and obtain the historical alloy element yield rate of the refining furnace in the historical smelting process of the molten steel.

[0135] The determination unit 32 can be used to determine the expected value of the predicted alloy element yield and the predicted alloy element yield variance corresponding to the molten steel to be refined based on the historical alloy element yield, and to determine the expected value of the predicted molten steel quality and the predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality.

[0136] The prediction unit 33 can be used to input the predicted expected value of alloy element yield, predicted alloy element yield variance, predicted expected value of molten steel quality, predicted molten steel quality variance, current alloy element content, alloy element batching interval, furnace condition data and alloy property data into a preset alloy batching quantity prediction model to perform batching prediction, and obtain the predicted batching quantity of the alloy in the molten steel to be refined.

[0137] The batching unit 34 can be used to batch the molten steel to be refined according to the predicted batching amount of the alloy.

[0138] In a specific application scenario, the alloy property data includes the alloy unit price of the alloy to be added and the element content in the alloy to be added.

[0139] In a specific application scenario, in order to predict the expected value and variance of the element yield in the steel smelting process, such as Figure 4 As shown, the determining unit 32 can be specifically used to use a Gaussian process regression function to perform regression analysis on the current alloy element content and the historical alloy element yield rate to obtain the expected value and variance of the predicted alloy element yield rate corresponding to the molten steel to be refined.

[0140] In a specific application scenario, in order to predict the expected value and variance of the molten steel quality, the determination unit 32 can also be used to perform density analysis on the probability distribution of the initial molten steel quality using a Gaussian probability density function to obtain the predicted expected value of the molten steel quality and the predicted variance of the molten steel quality corresponding to the molten steel to be refined.

[0141] In a specific application scenario, in order to predict the amount of alloy corresponding to the molten steel to be refined, the prediction unit 33 includes an acquisition module 331 , a prediction module 332 and a determination module 333 .

[0142] The acquisition module 331 may be used to acquire the alloy yield of the alloy to be added corresponding to the molten steel to be refined.

[0143] The prediction module 332 can be used to input the predicted expected value of alloy element yield, predicted alloy element yield variance, predicted expected value of molten steel quality, predicted molten steel quality variance, current alloy element content, alloy element batching range, element content in the alloy to be added, alloy yield and furnace condition data into the batching quantity constraint layer to perform batching interval prediction, and obtain the predicted batching quantity allowable range corresponding to the alloy to be added.

[0144] The determination module 333 may be configured to select an initial alloy batch quantity within the predicted batch quantity tolerance range according to an optimization algorithm, and determine a search step corresponding to the initial alloy batch quantity.

[0145] The prediction module 332 can be specifically used to input the initial alloy batch quantity and the alloy unit price into the batch value optimization layer to perform value prediction, and obtain the batch value corresponding to the initial alloy batch quantity.

[0146] The determination module 333 may be specifically configured to determine the predicted alloy proportion in the molten steel to be refined according to the proportion value corresponding to the initial alloy proportion and the search step length.

[0147] In a specific application scenario, in order to determine the predicted amount of alloy in the molten steel to be refined, the determination module 333 includes a search submodule, a judgment submodule and a determination submodule.

[0148] The search submodule can be used to iteratively search the alloy proportions of each combination within the predicted proportion tolerance range based on the proportion value corresponding to the initial alloy proportion and the search step size, and determine during the iterative search process whether the proportion value corresponding to the alloy proportion searched next time is less than the proportion value corresponding to the alloy proportion searched last time.

[0149] The judgment submodule can be used to continue the iterative search if the value of the alloy ingredient quantity is less than the ingredient value corresponding to the last search, and to judge whether the search step is less than a preset convergence threshold.

[0150] The determination submodule can be used to stop the iterative search if the value is less than the preset convergence threshold, and determine the ingredient value corresponding to the alloy ingredient amount searched for the last time.

[0151] The determination submodule can be specifically used to determine the minimum ingredient value between the ingredient value corresponding to the alloy ingredient quantity searched for the last time and the ingredient value corresponding to the alloy ingredient quantity searched for the previous time, and determine the alloy ingredient quantity corresponding to the minimum ingredient value as the predicted ingredient quantity of the alloy in the molten steel to be refined.

[0152] In a specific application scenario, before the molten steel to be refined is batched, in order to perform predictive processing on the molten steel to be refined, the device further includes: a processing unit 35 .

[0153] The processing unit 35 can be used to perform slagging and temperature-raising treatment on the molten steel to be refined to obtain processed molten steel to be refined.

[0154] The acquisition unit 31 may be specifically configured to acquire the current alloy element content corresponding to the processed molten steel to be refined.

[0155] It should be noted that for other corresponding descriptions of the functional modules involved in the batching device provided in the embodiment of the present invention, please refer to Figure 1 The corresponding description of the method shown will not be repeated here.

[0156] Based on the above Figure 1 The method shown, accordingly, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the following steps when executed by a processor: obtaining the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning interval corresponding to the type of molten steel to which the molten steel to be refined belongs, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and obtaining the historical alloying element yield rate of the refining furnace in the historical smelting process of the molten steel; based on the historical alloying element yield rate, determining the molten steel to be refined. The expected value of the predicted alloying element yield and the predicted alloying element yield variance corresponding to the molten steel, and the predicted molten steel quality expected value and the predicted molten steel quality variance corresponding to the molten steel to be refined are determined based on the initial molten steel quality; the predicted alloying element yield expected value, the predicted alloying element yield variance, the predicted molten steel quality expected value, the predicted molten steel quality variance, the current alloying element content, the alloying element batching interval, the furnace condition data and the alloy property data are input into a preset alloy batching quantity prediction model for batching prediction to obtain the predicted batching quantity of the alloy in the molten steel to be refined; and the molten steel to be refined is batched according to the predicted batching quantity of the alloy.

[0157] Based on the above Figure 1 The method shown and Figure 3 The embodiment of the device shown in the figure, the embodiment of the present invention also provides a physical structure diagram of a computer device, such as Figure 5As shown, the computer device includes: a processor 41, a memory 42, and a computer program stored in the memory 42 and executable on the processor, wherein the memory 42 and the processor 41 are both arranged on a bus 43 and when the processor 41 executes the program, the following steps are implemented: obtaining the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning interval corresponding to the type of molten steel to which the molten steel to be refined belongs, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and obtaining the historical alloying element yield rate of the refining furnace in the historical molten steel smelting process; based on the The method comprises the following steps: determining a historical alloying element yield, determining an expected value of a predicted alloying element yield and a predicted alloying element yield variance corresponding to the molten steel to be refined, and determining an expected value of a predicted molten steel quality and a predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality; inputting the predicted expected value of the alloying element yield, the predicted alloying element yield variance, the predicted expected value of the molten steel quality, the predicted molten steel quality variance, the current alloying element content, the alloying element batching interval, the furnace condition data and the alloy property data into a preset alloy batching quantity prediction model for batching prediction, and obtaining a predicted batching quantity of the alloy in the molten steel to be refined; and batching the molten steel to be refined according to the predicted batching quantity of the alloy.

[0158] Through the technical solution of the present invention, the present invention obtains the current alloy element content corresponding to the molten steel to be refined, the alloy element ratio range corresponding to the molten steel type to which the molten steel to be refined belongs, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and the historical alloy element yield rate of the refining furnace in the historical smelting process of the molten steel; and based on the historical alloy element yield rate, determines the expected value of the predicted alloy element yield rate and the predicted alloy element yield rate corresponding to the molten steel to be refined. The predicted yield variance, and based on the initial molten steel quality, the expected value of the predicted molten steel quality and the predicted molten steel quality variance corresponding to the molten steel to be refined are determined; then the predicted alloy element yield expected value, the predicted alloy element yield variance, the predicted molten steel quality expected value, the predicted molten steel quality variance, the current alloy element content, the alloy element batching interval, the furnace condition data and the alloy property data are input into the preset alloy batching prediction model for batching prediction, and the predicted batching amount of the alloy in the molten steel to be refined is obtained; finally, the molten steel to be refined is batched according to the predicted batching amount of the alloy. Therefore, by first predicting the expected value and variance of the element yield in the current molten steel refining process based on the historical alloying element yield, and at the same time predicting the expected value and variance of the molten steel quality based on the initial molten steel quality, the batching scheme of the molten steel to be refined is then determined based on the predicted expected value and variance of the element yield and the expected value and variance of the molten steel quality. This can overcome the parameter uncertainty problem caused by complex and changeable working conditions, and also avoid problems such as the excessive content of various elements in the final molten steel, thereby reducing the batching cost of the molten steel in the smelting process and improving the batching accuracy of the molten steel in the smelting process.

[0159] Obviously, those skilled in the art will appreciate that the various modules or steps of the present invention described above can be implemented using a general-purpose computing device, centralized on a single computing device, or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. In some cases, the steps shown or described can be performed in a different order than that shown, or can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0160] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A batching method, characterized in that: include: Obtaining the current alloying element content of the molten steel to be refined, the alloying element proportioning range corresponding to the type of molten steel to be refined, the furnace condition data of the refining furnace corresponding to the molten steel to be refined, the alloy property data of the alloy to be added to the molten steel to be refined, the initial molten steel quality of the molten steel to be refined, and obtaining the historical alloying element yield of the refining furnace during historical molten steel smelting processes; Determining an expected value of a predicted alloying element yield and a predicted alloying element yield variance corresponding to the molten steel to be refined based on the historical alloying element yield, and determining an expected value of a predicted molten steel quality and a predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality; Obtaining the alloy yield of the alloy to be added corresponding to the molten steel to be refined; inputting the predicted alloying element yield expected value, predicted alloying element yield variance, predicted molten steel quality expected value, predicted molten steel quality variance, current alloying element content, alloying element batching interval, element content in the alloy to be added, alloy yield and furnace condition data into the batching amount constraint layer of the preset alloy batching amount prediction model to perform batching interval prediction, and obtaining the predicted batching amount allowable interval corresponding to the alloy to be added; According to the optimization algorithm, an initial alloy batch quantity is selected within the predicted batch quantity allowable range, and a search step size corresponding to the initial alloy batch quantity is determined; the initial alloy batch quantity and the alloy unit price are input into the batch value optimization layer of the preset alloy batch quantity prediction model for value prediction to obtain the batch value corresponding to the initial alloy batch quantity; According to the ingredient value corresponding to the initial alloy ingredient quantity and the search step length, an iterative search is performed on the alloy ingredient quantities of each combination within the predicted ingredient quantity allowable range, and during the iterative search process, it is judged whether the ingredient value corresponding to the alloy ingredient quantity searched next time is less than the ingredient value corresponding to the alloy ingredient quantity searched last time; if it is less than the ingredient value corresponding to the alloy ingredient quantity searched last time, the iterative search is continued, and it is judged whether the search step length is less than a preset convergence threshold; if it is less than the preset convergence threshold, the iterative search is stopped, and the ingredient value corresponding to the alloy ingredient quantity searched last time is determined; the minimum ingredient value is determined between the ingredient value corresponding to the alloy ingredient quantity searched last time and the ingredient value corresponding to the alloy ingredient quantity searched last time, and the alloy ingredient quantity corresponding to the minimum ingredient value is determined as the predicted ingredient quantity of the alloy in the molten steel to be refined; The molten steel to be refined is batched according to the predicted batching amount of the alloy.

2. The method according to claim 1, characterized in that The alloy property data includes the alloy unit price of the alloy to be added and the element content in the alloy to be added.

3. The method according to claim 1, characterized in that The step of determining the expected value of the predicted alloying element yield and the predicted alloying element yield variance corresponding to the molten steel to be refined based on the historical alloying element yield comprises: The current alloying element content and the historical alloying element yield are regressed and analyzed using a Gaussian process regression function to obtain an expected value and a variance of a predicted alloying element yield corresponding to the molten steel to be refined.

4. The method according to claim 1, wherein The step of determining the predicted molten steel quality expectation value and the predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality comprises: The Gaussian probability density function is used to perform density analysis on the probability distribution of the initial molten steel quality to obtain the expected value of the predicted molten steel quality and the predicted molten steel quality variance corresponding to the molten steel to be refined.

5. The method according to claim 1, characterized in that Before obtaining the current alloying element content corresponding to the molten steel to be refined, the method further includes: performing slagging and temperature-raising treatments on the molten steel to be refined to obtain treated molten steel to be refined; The obtaining of the current alloying element content corresponding to the molten steel to be refined includes: Obtain the current alloying element content of the processed molten steel to be refined.

6. A batching device, characterized in that: include: an acquisition unit, configured to acquire the current alloying element content corresponding to the molten steel to be refined, the alloying element proportioning range corresponding to the type of molten steel to be refined, furnace condition data of the refining furnace corresponding to the molten steel to be refined, alloy property data of the alloy to be added corresponding to the molten steel to be refined, the initial molten steel quality corresponding to the molten steel to be refined, and historical alloying element yields of the refining furnace during historical molten steel smelting processes; a determination unit, configured to determine an expected value of a predicted alloying element yield and a predicted alloying element yield variance corresponding to the molten steel to be refined based on the historical alloying element yield, and to determine an expected value of a predicted molten steel quality and a predicted molten steel quality variance corresponding to the molten steel to be refined based on the initial molten steel quality; A prediction unit is configured to obtain an alloy yield of an alloy to be added corresponding to the molten steel to be refined; input the predicted alloying element yield expected value, predicted alloying element yield variance, predicted molten steel quality expected value, predicted molten steel quality variance, current alloying element content, alloying element batching interval, element content in the alloy to be added, alloy yield, and furnace condition data into a batching constraint layer in a preset alloy batching prediction model to perform batching interval prediction, thereby obtaining a predicted batching allowable interval corresponding to the alloy to be added; According to the optimization algorithm, an initial alloy batch quantity is selected within the predicted batch quantity allowable range, and a search step size corresponding to the initial alloy batch quantity is determined; the initial alloy batch quantity and the alloy unit price are input into the batch value optimization layer of the preset alloy batch quantity prediction model for value prediction to obtain the batch value corresponding to the initial alloy batch quantity; The prediction unit is used to iteratively search the alloy proportions of each combination within the predicted proportion allowable range according to the proportion value corresponding to the initial alloy proportion and the search step, and judge whether the proportion value corresponding to the alloy proportion searched next time is less than the proportion value corresponding to the alloy proportion searched last time during the iterative search process; if it is less than the proportion value corresponding to the alloy proportion searched last time, continue the iterative search and judge whether the search step is less than a preset convergence threshold; if it is less than the preset convergence threshold, stop the iterative search and determine the proportion value corresponding to the alloy proportion searched last time; determine the minimum proportion value between the proportion value corresponding to the alloy proportion searched last time and the proportion value corresponding to the alloy proportion searched last time, and determine the alloy proportion corresponding to the minimum proportion value as the predicted proportion of the alloy in the molten steel to be refined; A batching unit is used to batch the molten steel to be refined according to the predicted batching amount of the alloy.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images