Low-energy-consumption smelting method for high-alloy building steel

Abstract

The invention belongs to the field of steel and iron metallurgy process innovation, and provides a low-energy-consumption smelting method for high-alloy building steel, which comprises the following steps: collecting and standardizing historical data, constructing a standard set and a time sequence, capturing a time sequence correlation coefficient by using an LSTM model, and constructing an intelligent control system to output a deoxidation slagging argon flow value. The molten steel component uniformity is monitored through an LIBS sensor array, an adjustment coefficient is calculated to judge whether adjustment is needed or not, if adjustment is needed, a correlation model is fitted to extract a flow component influence coefficient for adjustment, a flow temperature influence coefficient is calculated at the same time, and by analyzing the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient, the molten steel component uniformity is judged. Calculating a temperature argon flow value; accurate optimization of argon flow is achieved through multi-link cooperative regulation and control, cooperative homogenization of components and temperature under the condition of less electrification or no electrification is achieved, and the energy utilization efficiency is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of steel metallurgy process innovation, and particularly relates to a low-energy-consumption smelting method of high-alloy construction steel. BACKGROUND

[0002] With the development of smelting of high-alloy construction steel, how to realize low-energy-consumption smelting based on smelting of high-alloy construction steel has become a research hotspot in the field of steel metallurgy process innovation, but there are still significant technical bottlenecks in the LF refining process, therefore, it is of great significance to develop a low-energy-consumption smelting method of high-alloy construction steel. In the LF (ladle furnace) refining process, although the state of molten steel can be adjusted by argon flow, on the one hand, the flow is set based on static data at a single time point, and the argon flow value is set as a fixed flow threshold or by artificial experience, lacking systematic analysis of historical data of the deoxidization and slagging link, which leads to difficulty in accurately matching the actual working condition of the set argon flow, and easily causes fluctuation of the oxidizability of molten steel or uneven composition of the slag, for example, when a certain medium-sized steel enterprise produces 20CrMnTi gear steel, the argon flow is always 0.5 m3 / h, which leads to difficulty in floating inclusions due to insufficient flow in the deoxidization stage, and the slag is rolled in due to too high flow in the later stage of slagging, and the products of this batch are all unqualified in detection; On the other hand, the influence of the introduced argon on the composition and temperature of molten steel is not considered, and the feasibility of whether the influence of argon on the composition of molten steel and the temperature of molten steel can simultaneously meet the needs of both is not analyzed, which leads to the situation that when one is adjusted, the other also changes, and both cannot be simultaneously satisfied, causing adjustment conflicts and frequent work stoppage. Therefore, the application provides a low-energy-consumption smelting method of high-alloy construction steel. SUMMARY

[0003] In order to make up for the deficiencies of the prior art and solve at least one technical problem proposed in the background.

[0004] The technical scheme adopted by the application to solve the technical problems is as follows: a low-energy-consumption smelting method of high-alloy construction steel comprises the following steps: obtaining historical data of a deoxidization and slagging link, calculating time series correlation coefficients between standardized values of the historical data, and constructing an argon intelligent control system to output an argon flow value for deoxidization and slagging; inputting argon according to the argon flow value for deoxidization and slagging to perform deoxidization and slagging, and judging whether the argon flow value for deoxidization and slagging needs to be adjusted in the deoxidization and slagging process, if so, deducing a flow composition influence coefficient between the argon flow value and a composition uniformity analysis value, and calculating a composition argon flow value; The deoxidizing slagging argon flow value is first adjusted according to the component argon flow value, and the temperature uniformity analysis value is calculated in real time after the adjustment, and the flow temperature influence coefficient between the argon flow value and the temperature uniformity analysis value is derived; The consistency of the positive and negative of the flow component influence coefficient and the flow temperature influence coefficient is determined, if consistent, the temperature argon flow value is calculated, and the deoxidizing slagging argon flow value is second adjusted according to the temperature argon flow value.

[0005] Further, the process of calculating the time sequence correlation coefficient between the standardized values of the historical data is: Obtain the historical data of the deoxidizing slagging link, the historical data including the deoxidizer consumption, the slagging agent consumption, the argon flow value, the initial state of the oxidizing property of the molten steel and the slag conversion amount, and the historical data of the deoxidizing slagging link is standardized respectively; The standardized processing results are divided and selected to obtain a plurality of 3-step time sequence sequences; The plurality of 3-step time sequence sequences are input into the LSTM model, the time sequence correlation between the standardized values is captured through the hidden layer, and the time sequence correlation coefficient between the standardized values of the historical data is output.

[0006] Further, the process of dividing and selecting the standardized processing results to obtain a plurality of 3-step time sequence sequences is: The historical data of the deoxidizing slagging link is standardized; The deoxidizer consumption standardized value , the slagging agent consumption standardized value , the initial state of the oxidizing property of the molten steel standardized value , the slag conversion amount standardized value , and the argon flow value standardized value , and the five standardized values in the same time step are recorded as a standard set; Obtain the standard set in a plurality of time steps of the historical data, record three time-continuous time steps as a group, in the form of a 3x5 time sequence matrix, the 3x5 time sequence matrix is a 3-step time sequence, and a plurality of 3-step time sequence sequences are obtained by selecting a plurality of times.

[0007] Further, the process of constructing an argon intelligent control system to output the deoxidizing slagging argon flow value is: Input layer: the deoxidizer consumption standardized value, the slagging agent consumption standardized value, the initial state of the oxidizing property of the molten steel standardized value and the slag conversion amount standardized value in the same step of the historical data are taken as input; Hidden layer: the weight matrix between each standardized value output by the LSTM model and the argon flow value in the same step of the historical data; Output layer: output the predicted deoxidizing slagging argon flow value; obtaining the deoxidizer dosage and the slagging agent dosage in the current deoxidizing and slagging process, and calculating the corresponding deoxidizer dosage standardized value and the slagging agent dosage standardized value; setting the current value of the molten steel oxidizability initial state standardized value and the slag conversion amount standardized value to 1; inputting the current deoxidizer dosage standardized value, the slagging agent dosage standardized value, the molten steel oxidizability initial state standardized value and the slag conversion amount standardized value into the argon intelligent control model, and outputting the deoxidizing and slagging argon flow value.

[0008] Further, the process of determining whether the deoxidizing and slagging argon flow value needs to be adjusted is as follows: dividing the molten steel after adding the material into a plurality of unit regions, dividing different elements in the unit regions according to categories, and respectively calculating the composition uniformity analysis value of each unit region, if the composition uniformity analysis value of any unit region is greater than the composition uniformity analysis threshold value, it is determined that the element composition of the unit region is uneven; obtaining the number of unit regions with uneven element composition, and performing ratio processing with the total number of unit regions to obtain an adjustment coefficient, and comparing the adjustment coefficient with an adjustment coefficient threshold value: if the adjustment coefficient is greater than or equal to the adjustment coefficient threshold value, it is determined that the deoxidizing and slagging argon flow value needs to be adjusted; otherwise, it is determined that no adjustment is needed.

[0009] Further, the process of calculating the composition uniformity analysis value of each unit region is as follows: respectively performing standard deviation processing on different element sets in the unit region, and obtaining the average value of the standard deviations of all element sets, which is the composition uniformity analysis value.

[0010] Further, the process of deriving the flow composition influence coefficient of the argon flow value and the composition uniformity analysis value is as follows: comparing the calculated determination coefficient with a determination coefficient threshold value, and obtaining the flow composition influence coefficient according to the comparison result is as follows: if the determination coefficient is greater than or equal to the determination coefficient threshold value, it means that there is a linear correlation between the argon flow standardized value and the corresponding composition uniformity analysis value; otherwise, it means that there is no linear correlation.

[0011] if it is a linear correlation, the argon flow standardized value coefficient in the one-dimensional linear regression model formula is the flow composition influence coefficient; if it is a nonlinear correlation, a plurality of models are respectively constructed, and the determination coefficient of each model is calculated; taking the nonlinear model corresponding to the maximum determination coefficient as the nonlinear model between the argon flow standardized value and the corresponding composition uniformity analysis value standardized value; Based on the eigenvalue of the current argon flow standardization value of the model, the flow component influence coefficient is obtained. The flow temperature influence coefficient is the same as the flow temperature influence coefficient derivation process.

[0012] Further, the process of obtaining the flow component influence coefficient according to the comparison result is: If the determination coefficient is greater than or equal to the determination coefficient threshold value, it indicates that there is a linear correlation between the argon flow standardization value and the corresponding component uniform analysis value; On the contrary, it indicates that there is no linear correlation.

[0013] If it is linear correlation, the argon flow standardization value coefficient in the one-dimensional linear regression model formula is the flow component influence coefficient; If it is nonlinear correlation, a binary linear model, a high-order polynomial model, an exponential model, a power function model, and a trigonometric function model are constructed respectively, and the determination coefficients of each model are calculated; The nonlinear model corresponding to the maximum determination coefficient is taken as the nonlinear model between the argon flow standardization value and the corresponding component uniform analysis value standardization value; Based on the eigenvalue of the current argon flow standardization value of the model, the flow component influence coefficient is obtained.

[0014] Further, the process of judging the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient is: If the flow component influence coefficient and the flow temperature influence coefficient are the same or opposite, it indicates that the positive and negative consistency is consistent, otherwise, it is inconsistent.

[0015] Further, the process of adjusting the deoxidizing slag-making argon flow value according to the temperature argon flow value is: If the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient is consistent, the temperature uniform analysis value of all units is obtained, and the average processing is carried out to obtain the temperature uniform analysis average value; The temperature uniform analysis value of each unit is subtracted from the temperature uniform analysis average value respectively, and the maximum value is taken as the component uniform adjustment value; The component uniform adjustment value is processed by ratio with the flow temperature influence coefficient to obtain the component flow adjustment value, and is processed by addition with the component argon flow value to obtain the temperature argon flow value.

[0016] The beneficial effects of the present application are as follows: Effect one: improve the control precision and dynamic adaptability of argon flow, build a time sequence correlation model (LSTM network) of historical data of deoxidizing slag making link, quantify the dynamic relationship between deoxidizer dosage, slag making agent dosage, molten steel oxidizing property and argon flow, realize accurate matching of argon flow and process parameters, effectively reduce the fluctuation of molten steel oxidizing property and the problem of uneven composition of slag, and improve the refining stability; Effect two: under the condition of little or no power supply, dynamically adjust the argon flow to balance the composition uniformity and temperature uniformity, dynamically deduce the composition and temperature coefficient by unit argon amount control, realize accurate regulation of argon flow on composition uniformity and temperature uniformity, improve the control precision of uniformity, avoid excessive stirring or insufficient stirring, and significantly improve the uniformity of molten steel; Effect three: based on the dynamic flow adjustment mechanism of real-time working condition, reduce unnecessary excessive or insufficient use of argon, reduce energy consumption, at the same time, reduce the rework caused by quality failure through preventive adjustment, significantly shorten the refining period, and reduce the production cost. BRIEF DESCRIPTION OF DRAWINGS

[0017] The application will be further described below with reference to the drawings.

[0018] Figure 1 is a step flow schematic diagram of a low-energy smelting method of a high-alloy construction steel according to an embodiment of the application; Figure 2 is a logic judgment schematic diagram of a low-energy smelting method of a high-alloy construction steel according to an embodiment of the application. DETAILED DESCRIPTION

[0019] In order to make the technical means, creative features, purposes and effects realized by the application easy to understand, the application will be further described below with reference to the specific embodiments.

[0020] Embodiment 1: please refer to Figure 1 - Figure 2 A low-energy smelting method of a high-alloy construction steel according to an embodiment of the application includes the following steps: Step one: in the LF refining process, obtain the historical data of the deoxidizing slag making link, calculate the time sequence correlation coefficient between the standardized values of the historical data, and build an argon intelligent control model to output the deoxidizing slag making argon flow value; It should be noted that the refining is sequentially divided into: deoxidizing slag making link, argon fine adjustment link and stirring control link, and the next link is entered after the end of the previous link, and the argon fine adjustment link is for molten steel temperature uniformity fine adjustment after molten steel composition uniformity fine adjustment; In step one, the process of calculating the time sequence correlation coefficient between the standardized values of the historical data is: obtain the historical data of the deoxidizing slag making link; The historical data of the deoxidizing and slagging link includes: deoxidizer dosage, slagging agent dosage, argon flow value, initial state of oxidizing of molten steel, and slag conversion amount; The deoxidizer dosage in any time step in the historical data is standardized to obtain a deoxidizer dosage standardized value :

[0021] The deoxidizer dosage standard range is the minimum value of all historical deoxidizer dosage data and the maximum value , and the time step is set by experts according to actual conditions; Similarly, the slagging agent dosage standardized value in any time step in the historical data is calculated , the initial state of oxidizing of molten steel standardized value , the slag conversion amount standardized value , and the argon flow value standardized value ; The , , , , in the same time step are recorded as a standard set , , , , ; Exemplarily, the process of obtaining the standard set is as follows: The deoxidizer dosage of each ton of molten steel in a time step (0s-15s) is obtained , the deoxidizer dosage standard range is (0.2, 1.0) kg / t, the deoxidizer dosage standardized value is 0.375, and the slagging agent dosage standardized value is 0.467, is 0.556, is 0.779, is 0.837, and the argon flow value standardized value is 0.837, so the standard set is (0.375, 0.367, 0.456, 0.779, 0.837); The standard set in several time steps of the historical data is obtained; Exemplarily, the process of constructing a 3-step time sequence is as follows: Set the time step as 15 seconds, obtain the standard set in 3 adjacent time steps of historical data, Time step 1 (0-15 seconds) , Time step 2 (15-30 seconds) , Time step 3 (30-45 seconds) , Form a 3x5 time sequence matrix, that is, a 3-step time sequence; Input layer: input several 3-step time sequence as input features into the LSTM model; Hidden layer: learn long-term dependencies through the gating mechanism, use ReLU activation function to enhance non-linear expression ability and accelerate convergence, and capture the time sequence correlation between parameters; Output layer: output the time sequence correlation between the standardized values of historical data, in the form of a weight matrix between each standardized value; Build an argon intelligent control model, use the standardized values of deoxidizer dosage, slagging agent dosage, initial state of molten steel oxidation and slag conversion amount in the same step of historical data as input, combine the weight matrix between each standardized value output by the LSTM model and the argon flow value in the same step of historical data, and output the predicted deoxidizing and slagging argon flow value; It should be noted that the process of outputting the predicted deoxidizing and slagging argon flow value in combination with the weight matrix between the standardized values of historical data output by the LSTM model is as follows: Calculate the corresponding weight of the standardized value parameters of the input layer and the weight matrix to obtain the predicted argon flow value, calculate the mean square error between the predicted argon flow value and the argon flow value in the same step of historical data by using mean square error (MSE), and adjust the predicted argon flow value in combination with the flow value weight in the weight matrix. The final adjusted predicted argon flow value is output to obtain the predicted deoxidizing and slagging argon flow value; In step one, the process of obtaining the deoxidizing and slagging argon flow value is as follows: Obtain the deoxidizer dosage and slagging agent dosage of the current deoxidizing and slagging link, and calculate the corresponding deoxidizer dosage standardized value and slagging agent dosage standardized value; Set the standardized value of the initial state of molten steel oxidation and the standardized value of slag conversion amount as 1; It should be noted that the reason for setting the standardized value of the initial state of molten steel oxidation and the standardized value of slag conversion amount as 1 is that when the standardized value of the initial state of molten steel oxidation and the standardized value of slag conversion amount are set as 1, the initial state of molten steel oxidation and the slag conversion amount meet the standard of the initial state of oxidation and the standard of slag conversion amount; The deoxidizing agent dosage normalized value, the slagging agent dosage normalized value, the molten steel oxidizing initial state normalized value and the slag conversion amount normalized value are input into the argon intelligent control model, and the deoxidizing and slagging argon flow value is output; Step two: input argon according to the deoxidizing and slagging argon flow value to deoxidize and slag, and judge whether the deoxidizing and slagging argon flow value needs to be adjusted during the deoxidizing and slagging process. If so, derive the flow composition influence coefficient between the argon flow value and the composition uniformity analysis value, and calculate the composition argon flow value; In step two, the process of determining whether the deoxidizing and slagging argon flow value needs to be adjusted is as follows: By uniformly installing LIBS sensors at different heights and angles on the walls of the refining furnace, a laser-induced breakdown spectroscopy (LIBS) array is formed, and the molten steel after adding materials is divided into several unit regions; Based on all LIBS sensors in any unit region, different elements in the unit region are classified according to categories, and the element content of different categories in each monitoring point in the unit region is obtained; The element content of the same category is summarized as an element set, and the element set of each element category is obtained; Element composition uniformity analysis is performed on any unit region; The standard deviation of each element set is processed in the unit region, and the average value of the standard deviations of all element sets is obtained, which is the composition uniformity analysis value. The composition uniformity analysis value is compared with the composition uniformity analysis threshold value: If the composition uniformity analysis value is greater than the composition uniformity analysis threshold value, it is determined that the element composition of the unit region is not uniform; On the contrary, it is determined to be uniform; It should be noted that the composition uniformity analysis threshold value is set by experts according to current work needs; Similarly, several unit regions are processed, and the number of unit regions with non-uniform element composition is obtained; The number of unit regions with non-uniform element composition is processed by ratio with the total number of unit regions to obtain an adjustment coefficient; The adjustment coefficient is compared with the adjustment coefficient threshold value: If the adjustment coefficient is greater than or equal to the adjustment coefficient threshold value, it is determined that the deoxidizing and slagging argon flow value needs to be adjusted; If the adjustment coefficient is less than the adjustment coefficient threshold value, it is determined that the deoxidizing and slagging argon flow value does not need to be adjusted; It should be noted that the adjustment coefficient threshold value is set by experts according to current work needs; In step two, if the deoxidizing and slagging argon flow value needs to be adjusted, the process of calculating the flow composition influence coefficient of the argon flow value on the composition uniformity analysis value is as follows: Obtaining the fine-tuning argon flow values of different LF refining history, and calculating the corresponding composition uniformity analysis values; Standardizing all argon flow values and corresponding composition uniformity analysis values to obtain a plurality of argon flow standardized values and corresponding composition uniformity analysis value standardized values; Performing least square fitting processing on the argon flow standardized values and corresponding composition uniformity analysis values; Constructing a linear regression model between the argon flow standardized values and the corresponding composition uniformity analysis value standardized values ;

[0022] wherein, is a constant term, k is an influence factor, is a random error term; Introducing the determination coefficient , inputting the standardized value of the actual composition uniformity analysis value and the standardized value of the composition uniformity analysis value predicted by the linear regression model into the following formula, and calculating the determination coefficient to determine the fitting result:

[0023] wherein, is the mean value of the standardized value of the actual composition uniformity analysis value ; Comparing the calculated determination coefficient with the determination coefficient threshold value: If the determination coefficient is greater than or equal to the determination coefficient threshold value, it indicates that there is a linear correlation between the argon flow standardized value and the corresponding composition uniformity analysis value, and is the flow composition influence coefficient; Otherwise, it indicates that there is no linear correlation; If there is no linear correlation between the argon flow standardized value and the corresponding composition uniformity analysis value standardized value, constructing a binary linear model, a high-order polynomial model, an exponential model, a power function model, a trigonometric function model, etc. Through least square fitting processing, calculating the determination coefficient of each model; Taking the nonlinear model corresponding to the maximum determination coefficient as the nonlinear model between the argon flow standardized value and the corresponding composition uniformity analysis value standardized value; Based on the characteristic value (such as derivative, function value) of the current argon flow standardized value of the model, the control value reflecting the unit argon amount on the composition uniformity is obtained, and the characteristic value is the flow composition influence coefficient; Exemplary, the method for obtaining the flow composition influence coefficient based on the eigenvalue of the current argon flow standardization value of the model is as follows: If the maximum value of the coefficient of determination corresponds to a nonlinear model that is a quadratic function model (y=ax2+bx+c) , the influence factor is the derivative (2ax+b) of the current argon flow standardization value, which dynamically changes with the flow interval and reflects the regulation intensity at different stages, then the derivative (2ax+b) of the current argon flow standardization value is the flow composition influence coefficient; It should be noted that the physical meaning of obtaining the flow composition influence coefficient of the argon flow value on the composition uniform analysis value is as follows: The flow composition influence coefficient is obtained by least squares fitting processing of the argon flow standardization value and the standardization value of the corresponding composition uniform analysis value, which is essentially the average change amount of the composition uniform analysis value standardization value when the argon flow standardization value changes by 1 unit, directly reflecting the regulation rate of the molten steel stirring intensity of the argon flow on the composition uniformity; In step two, the process of calculating the composition argon flow value is as follows: Average the composition uniform analysis values of each unit to obtain the composition uniform analysis average value; Difference the composition uniform analysis value of each unit from the composition uniform analysis average value to obtain a plurality of to-be-determined composition uniform analysis adjustment values, and take the maximum value as the composition uniform analysis adjustment value; It should be noted that the reason for taking the maximum value as the deoxidizing slagging argon flow adjustment value is as follows: The to-be-determined deoxidizing slagging argon flow adjustment value is obtained by difference processing of the composition uniform analysis value of each unit from the composition uniform analysis average value, and the maximum value reflects the maximum degree. If the maximum to-be-determined deoxidizing slagging argon flow adjustment value is met, the remaining to-be-determined deoxidizing slagging argon flow adjustment values are also met; Ratio the composition uniform analysis adjustment value to the flow composition influence coefficient to obtain the deoxidizing slagging argon flow adjustment value; Add the deoxidizing slagging argon flow adjustment value to the deoxidizing slagging argon flow value to obtain the composition argon flow value; Step three: under the requirement of little or no power refining, adjust the deoxidizing slagging argon flow value according to the composition argon flow value for the first time, and calculate the temperature uniform analysis value in real time after adjustment, and derive the flow temperature influence coefficient between the argon flow value and the temperature uniform analysis value; In step three, the process of calculating the temperature uniform analysis value in real time after adjustment is as follows: Uniformly implant a plurality of thermocouple sensors at key positions such as the inner wall and bottom of the steelmaking furnace or the steel holding vessel to form a thermocouple sensor array; The molten steel after the component argon flow value is input is divided into a plurality of unit regions, and based on the thermocouple sensors in each unit region, real-time temperature values of a plurality of monitoring points in each unit region are obtained; Uniformity analysis is performed on the temperature of any unit region; Through the temperature values of the plurality of monitoring points in the unit region, standard deviation processing is performed to obtain a real-time temperature uniformity analysis value; All unit regions are processed to obtain a plurality of real-time temperature uniformity analysis values; In step three, the process of deriving the flow temperature influence coefficient of the argon flow value on the molten steel temperature uniformity value is as follows: The argon flow values and corresponding temperature uniformity analysis values in the fine-tuning link of historical LF refining are obtained; Through standardization processing of the argon flow values and corresponding temperature uniformity analysis values in the plurality of historical data, argon flow standardized values and standardized values of the corresponding temperature uniformity analysis values are obtained; In combination with the process of deriving the flow component influence coefficient of the argon flow value on the component uniformity analysis value in step two, the molten steel temperature uniformity value is replaced by the component uniformity analysis value, and the standardized value of the temperature uniformity analysis value is replaced by the standardized value of the component uniformity analysis value, and the flow temperature influence coefficient is derived according to the flow component influence coefficient derivation process; Step four: determining the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient, if consistent, calculating the temperature argon flow value, and performing secondary adjustment on the deoxidizing and slagging argon flow value according to the temperature argon flow value; In step four, the process of determining the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient is as follows: The flow component influence coefficient and the flow temperature influence coefficient are compared with the threshold value 0 respectively: If the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient is consistent, it is determined that the argon flow value can be adjusted, and the requirements of the molten steel temperature and component uniformity are met at the same time; It should be noted that if the positive and negative consistency is consistent, it is determined that the argon flow value can be adjusted, and the requirements of the molten steel temperature and component uniformity are met at the same time; The component argon flow value is adjusted according to the difference between the set component uniformity analysis threshold value and the component uniformity analysis value, wherein the deoxidizing and slagging argon flow value adjustment amount is the minimum value, if the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient is consistent, the temperature argon flow value is obtained by adjusting the component argon flow value, which meets the molten steel temperature uniformity, and the deoxidizing and slagging argon flow value adjustment amount is still within the range of meeting the component uniformity, so that the temperature argon flow value also meets the requirement of the molten steel component uniformity; If the positive and negative of the flow composition influence coefficient and the flow temperature influence coefficient are inconsistent, it is determined that the argon flow value cannot be adjusted to simultaneously meet the requirements of the steel temperature and composition uniformity; In step four, the temperature argon flow value is calculated, and the composition argon flow value is adjusted as follows: If the positive and negative of the flow composition influence coefficient and the unit flow temperature influence coefficient are consistent, the temperature uniformity analysis value of all units is obtained; The temperature uniformity analysis values of all units are averaged to obtain a temperature uniformity analysis average value; The temperature uniformity analysis value of each unit is subtracted from the temperature uniformity analysis average value to obtain a plurality of composition uniformity adjustment values, and the maximum value thereof is taken as the composition uniformity adjustment value; The composition uniformity adjustment value is divided by the flow temperature influence coefficient to obtain a composition flow adjustment value, which is added to the composition argon flow value to obtain a temperature argon flow value; If the positive and negative of the flow composition influence coefficient and the flow temperature influence coefficient are inconsistent, the composition argon flow value is adjusted by using other methods, such as flow field optimization design, etc. The technical scheme of the embodiment of the present application is as follows: historical data of a deoxidizing and slagging link is obtained, a time sequence correlation coefficient between standardized values of the historical data is calculated, an argon intelligent control system is constructed, a deoxidizing and slagging argon flow value is output, argon is input for deoxidizing and slagging according to the deoxidizing and slagging argon flow value, and it is determined whether the deoxidizing and slagging argon flow value needs to be adjusted during the deoxidizing and slagging process. If it needs to be adjusted, a flow composition influence coefficient between the argon flow value and the composition uniformity analysis value is derived, a composition argon flow value is calculated, the deoxidizing and slagging argon flow value is adjusted for the first time according to the composition argon flow value, a temperature uniformity analysis value is calculated in real time after the adjustment, a flow temperature influence coefficient between the argon flow value and the temperature uniformity analysis value is derived, the positive and negative consistency of the flow composition influence coefficient and the flow temperature influence coefficient is determined, if the positive and negative are consistent, a temperature argon flow value is calculated, and the deoxidizing and slagging argon flow value is adjusted for the second time according to the temperature argon flow value.

[0024] The basic principles, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A low-energy smelting method for high-alloy construction steel, characterized by: include: Obtain historical data of the deoxidation and slagging process, calculate the time series correlation coefficient between the standardized values ​​of the historical data, and build an argon intelligent control system to output the deoxidation and slagging argon flow value; Input argon gas according to the deoxidation slag forming argon gas flow rate value to perform deoxidation slag forming, and determine whether the deoxidation slag forming argon gas flow rate value needs to be adjusted during the deoxidation slag forming process. If necessary, derive the flow component influence coefficient between the argon gas flow rate value and the component uniformity analysis value, and calculate the component argon gas flow rate value; The deoxidation slag forming argon flow value is adjusted for the first time according to the component argon flow value, and the temperature uniformity analysis value is calculated in real time after the adjustment, and the flow temperature influence coefficient between the argon flow value and the temperature uniformity analysis value is derived; Determine the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient. If they are consistent, calculate the temperature argon flow value and make a secondary adjustment to the deoxidation and slagging argon flow value according to the temperature argon flow value.

2. The low-energy smelting method for high-alloy structural steel according to claim 1, characterized in that: The process of calculating the time series correlation coefficient between the standardized values ​​of historical data is: Obtain historical data of the deoxidation and slag-making process, including deoxidizer dosage, slag-making agent dosage, argon flow rate, initial state of molten steel oxidation, and slag conversion amount, and perform standardization on the historical data of the deoxidation and slag-making process; The normalization processing results are divided and selected to obtain several 3-step time series; Several three-step time series sequences are input into the LSTM model, the time series correlation between the standardized values ​​is captured through the hidden layer, and the time series correlation coefficient between the standardized values ​​of historical data is output.

3. The low-energy smelting method for high-alloy structural steel according to claim 2, characterized in that: The process of dividing and selecting the normalization results to obtain several 3-step time series is as follows: Standardize the historical data of the deoxidation and slagging process to obtain the standardized value of the deoxidizer dosage , standardized value of slagging agent dosage , Standardized value of initial state of molten steel oxidation , standardized value of slag conversion , argon flow rate standardization value , and record the five standardized values ​​in the same time step as a standard set; Obtain a standard set within several time steps of historical data, record three consecutive time steps as a group in the form of a 3×5 time series matrix, that is, a 3-step time series sequence, and perform several selections to obtain several 3-step time series sequences.

4. The low-energy smelting method for high-alloy structural steel according to claim 2, characterized in that: The process of constructing an intelligent argon control system and outputting the deoxidation and slagging argon flow value is as follows: Input layer: The standardized value of deoxidizer dosage, slag-forming agent dosage, initial state of molten steel oxidation, and slag conversion amount within the same step length in the historical data are taken as input; Hidden layer: The weight matrix between each standardized value output by the LSTM model and the argon flow value within the same step length in the historical data; Output layer: outputs the predicted argon flow value for deoxidation and slagging; Obtain the deoxidizer dosage and slag-forming agent dosage in the current deoxidation and slag-forming process, and calculate the corresponding deoxidizer dosage standardized value and slag-forming agent dosage standardized value; The normalized value of the initial state of molten steel oxidation and the normalized value of slag conversion amount are set to 1; The current standardized value of deoxidizer dosage, standardized value of slag-forming agent dosage, standardized value of initial state of molten steel oxidation and standardized value of slag conversion amount are input into the argon intelligent control model, and the deoxidation and slag-forming argon flow value is output.

5. The low-energy smelting method for high-alloy structural steel according to claim 1, characterized in that: The process of judging whether the deoxidation slagging argon gas flow rate needs to be adjusted is as follows: After the first adjustment of the deoxidation slag forming argon gas flow rate, the molten steel area is divided into several unit areas. The different elements in the unit areas are divided into categories. The element contents of the same category are summarized into element sets. The composition uniformity analysis value of each unit area is calculated separately. If the composition uniformity analysis value of any unit area is greater than the composition uniformity analysis threshold, it is determined that the element composition of the unit area is uneven. Get the number of uneven element compositions in the unit area and perform ratio processing on it with the total number of unit areas to obtain the adjustment coefficient, which is then compared with the adjustment coefficient threshold: If the adjustment coefficient is greater than or equal to the adjustment coefficient threshold, it is determined that the deoxidation slagging argon gas flow rate needs to be adjusted; Otherwise, it is determined that no adjustment is required.

6. The low-energy smelting method for high-alloy structural steel according to claim 5, characterized in that: The process of calculating the composition uniformity analysis value of each unit area is: The standard deviation of different element sets is processed in the unit area, and the average value of the standard deviation of all element sets is obtained, which is the component uniformity analysis value.

7. The low-energy smelting method for high-alloy structural steel according to claim 1, characterized in that: The process of deriving the flow composition influence coefficient of argon flow value and composition uniformity analysis value is as follows: Obtain the component uniformity analysis values ​​and argon flow values ​​of different historical LF refining processes, perform standardization on them respectively, obtain the standardized values ​​of the two, and perform least squares fitting on the two; Constructing standardized values ​​for argon flow Normalized value with the corresponding component uniformity analysis value The univariate linear regression model between Calculate the coefficient of determination of the univariate linear regression model fitting results; Comparing the calculated determination coefficient with the determination coefficient threshold, and obtaining the flow component influence coefficient based on the comparison result; The derivation process of flow rate temperature influence coefficient is the same as that of flow rate temperature influence coefficient.

8. The low-energy smelting method for high-alloy structural steel according to claim 6, characterized in that: The process of obtaining the flow component influence coefficient based on the comparison results is: If the coefficient of determination is greater than or equal to the threshold value of the coefficient of determination, it means that there is a linear correlation between the normalized value of the argon flow rate and the corresponding component uniformity analysis value; Otherwise, it means that there is no linear correlation; If it is a linear correlation, the coefficient of the standardized value of the argon flow rate in the univariate linear regression model formula is the flow component influence coefficient; If it is a nonlinear association, construct a bivariate linear model, a high-order polynomial model, an exponential model, a power function model, and a trigonometric function model, and calculate the coefficient of determination of each model; The nonlinear model corresponding to the maximum value of the determination coefficient is taken as the nonlinear model between the standardized value of the argon flow rate and the standardized value of the corresponding component uniform analysis value; The flow component influence coefficient is obtained based on the characteristic value of the normalized value of the current argon flow of the model.

9. The low-energy smelting method for high-alloy structural steel according to claim 1, characterized in that: The process of judging the positive and negative consistency of the flow component influence coefficient and the flow temperature influence coefficient is as follows: If the flow component influence coefficient and the flow temperature influence coefficient are the same positive or negative, it means that the positive and negative are consistent, otherwise they are inconsistent.

10. The low-energy smelting method for high-alloy structural steel according to claim 1, characterized in that: The process of secondary adjustment of the deoxidation slag forming argon flow value according to the temperature argon flow value is as follows: If the flow component influence coefficient and the flow temperature influence coefficient are consistent in positive and negative, then obtain the temperature uniformity analysis values ​​of all units and perform averaging to obtain the temperature uniformity analysis average value; Perform difference processing on the temperature uniformity analysis value of each unit and the average value of the temperature uniformity analysis, and take the maximum value as the composition uniformity adjustment value; The component uniformity adjustment value is ratioed with the flow temperature influence coefficient to obtain the component flow adjustment value, and then added to the component argon flow value to obtain the temperature argon flow value.

Images