Data and mechanism combined driving converter steelmaking oxygen consumption calculation method and system

Abstract

The invention discloses a data and mechanism combined driving converter steelmaking oxygen consumption calculation method and system, and belongs to the technical field of converter steelmaking. Dividing the data into field measurable data and parameters which cannot be measured in real time (including key parameters which cannot be measured in real time and non-key parameters which cannot be measured in real time); inputting field measurable data into the parameter prediction model to obtain a key parameter prediction value which cannot be measured in real time and an oxygen consumption prediction value; an oxygen consumption mechanism calculation model is adopted, and the oxygen consumption is calculated according to the converter steelmaking production data; wherein the key parameters which cannot be measured in real time adopt model prediction values; and weighting the predicted value of the oxygen consumption and the calculated value of the mechanism calculation model to obtain an oxygen consumption calculation result. According to the method, the oxygen consumption in the converter steelmaking process can be accurately predicted, and the actual requirement for accurately controlling the oxygen consumption according to the input material and the control target is met.

Description

Technical Field

[0001] The present invention relates to the technical field of converter steelmaking, and particularly to a method and system for calculating the oxygen consumption of converter steelmaking jointly driven by data and mechanism. Background Art

[0002] Converter steelmaking is the most important steelmaking method in current steel production. Its core task is to reduce the content of carbon and other impurity elements in molten iron by blowing oxygen into the molten iron and increase the temperature of the molten steel. In this process, oxygen reacts with elements such as carbon, silicon, manganese, and phosphorus in the molten iron to convert these impurity elements into gases or slag and discharge them, and finally obtain molten steel with qualified composition and temperature. Therefore, accurately controlling the input amount of oxygen is crucial for achieving efficient and precise steelmaking production. Too much or too little oxygen injection will cause inappropriate effects. If the oxygen injection is insufficient, excessive carbon and other impurity elements will remain in the molten iron, making the molten steel unable to meet the expected chemical composition requirements, thereby affecting the performance of the steel. If too much oxygen is blown in, it will lead to energy waste and increased production costs, and may even cause safety problems during the operation process. In addition, excessive oxygen will also cause an increase in the oxygen content in the molten steel, forming oxide inclusions, seriously affecting the quality of the molten steel and the performance of the steel. Therefore, in the actual converter steelmaking production process, accurately predicting and controlling the amount of oxygen blown in is the key to ensuring the quality of the molten steel.

[0003] Currently, there are mainly two types of methods for predicting the oxygen consumption in the converter steelmaking process, namely: the method based on mechanism calculation and the method based on data-driven. The method based on mechanism calculation relies on the physical and chemical reaction principles in the metallurgical field and describes the oxidation behavior of each element and its required oxygen consumption by establishing a mathematical model. This method has good interpretability, but the mechanism calculation process involves some parameters that cannot be measured in real time, such as slag ratio, oxygen utilization rate, etc. These parameters will change dynamically with the furnace conditions of each furnace and cannot be measured in real time. Most of them are given a fixed value according to experience and cannot adapt to the dynamic changes of the furnace conditions. The data-driven method uses machine learning algorithms to mine the complex implicit correlation relationships between data for oxygen consumption prediction. This method can overcome the problem that parameters are difficult to measure in real time in mechanism calculation, but due to the high dependence of the data-driven method on data quality and the lack of mechanism constraints, the prediction accuracy and robustness are not high and the method lacks interpretability.

[0004] In summary, it can be seen that the prior art is difficult to accurately predict the oxygen consumption in the converter steelmaking process. Summary of the Invention

[0005] The present invention provides a method and system for calculating the oxygen consumption of converter steelmaking jointly driven by data and mechanism to solve the technical problem that the prior art is difficult to accurately predict the oxygen consumption in the converter steelmaking process.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] On the one hand, the present invention provides a method for calculating oxygen consumption in converter steelmaking driven jointly by data and mechanism. The method for calculating oxygen consumption in converter steelmaking driven jointly by data and mechanism includes:

[0008] Obtain converter steelmaking production data;

[0009] Divide the converter steelmaking production data into on-site measurable data and non-real-time measurable parameters; among them, the non-real-time measurable parameters include key non-real-time measurable parameters and non-key non-real-time measurable parameters;

[0010] Input the on-site measurable data into a preset parameter prediction model, and output the predicted values of key non-real-time measurable parameters and the predicted value of oxygen consumption through the parameter prediction model;

[0011] Adopt a preset oxygen consumption mechanism calculation model to calculate oxygen consumption according to the converter steelmaking production data; among them, the on-site measurable data in the converter steelmaking production data adopts the measured value, the non-key non-real-time measurable parameters adopt the empirical values set manually, and the key non-real-time measurable parameters adopt the predicted values output by the parameter prediction model;

[0012] Weight the predicted value of oxygen consumption output by the parameter prediction model and the oxygen consumption calculated by the oxygen consumption mechanism calculation model to obtain the calculation result of oxygen consumption in converter steelmaking.

[0013] Further, the on-site measurable data includes: molten iron element content, end-point molten steel element content, molten steel weight, molten iron weight, scrap steel weight, pig iron weight, tapping temperature, gas recovery amount, quicklime weight, light-burned dolomite consumption, and magnesium ball consumption.

[0014] Further, the non-key non-real-time measurable parameters include: the proportion of furnace lining erosion in molten iron amount, dust ratio, iron content in slag, and oxygen purity.

[0015] Further, the key non-real-time measurable parameters include: incomplete combustion rate of carbon, slag ratio, carbon content in scrap steel, and oxygen utilization rate.

[0016] Further, the parameter prediction model is a multi-layer neural network model.

[0017] Further, the training process of the parameter prediction model includes:

[0018] Collect historical converter steelmaking production data and historical oxygen consumption, and preprocess the collected data. Use the preprocessed historical converter steelmaking production data and historical oxygen consumption to construct a sample data set;

[0019] Train the parameter prediction model using the sample data set; wherein, the input of the parameter prediction model is on-site measurable data, and the output is key non-real-time measurable parameters and oxygen consumption.

[0020] Further, the preprocessing of the collected data includes:

[0021] Use the box plot method to identify and remove outlier data points in the data.

[0022] Further, the loss function of the parameter prediction model is the sum of the mechanism loss and the data loss, expressed as:

[0023] Loss = Loss M + Loss D

[0024] where Loss is the loss function of the parameter prediction model; Loss M is the mechanism loss; Loss D is the data loss;

[0025] The data loss Loss D The calculation formula is:

[0026]

[0027] where N is the total number of training data; O prei is the oxygen consumption predicted based on the i-th training data; O i is the true data of the collected oxygen consumption corresponding to the i-th training data;

[0028] The mechanism loss Loss M The calculation formula is:

[0029]

[0030] where O cali is the oxygen consumption calculated after substituting the key non-real-time measurable parameters predicted by the parameter prediction model based on the i-th training data into the oxygen consumption mechanism calculation model.

[0031] Further, the expression of the oxygen consumption mechanism calculation model is:

[0032]

[0033] where M represents the calculated oxygen consumption; O2_ratio represents the oxygen utilization rate;

[0034]

[0035]

[0036] Among them, W iron represents the weight of hot metal; W hc , W hsi , W hmn , W hp , W hs respectively represent the contents of elements C, Si, Mn, P, and S in hot metal; W scrap represents the weight of scrap; W sc , W ssi , W smn , W sp , W ss respectively represent the contents of elements C, Si, Mn, P, and S in scrap; W pig_iron represents the weight of pig iron; W pc , W psi , W pmn , W pp , W ps respectively represent the contents of elements C, Si, Mn, P, and S in pig iron; W steel represents the weight of the produced molten steel, W tc , W tsi , W tmn , W tp , W ts respectively represent the contents of elements C, Si, Mn, P, and S in molten steel; per_C represents the proportion of carbon monoxide generated; per_S represents the proportion of sulfur element oxidized to sulfur dioxide; Slag_ratio represents the slag ratio; Slag_TFe represents the iron content in the slag; Dust_ratio represents the proportion of dust in the hot metal; Dust_FeO represents the proportion of FeO generated; Dust_Fe2O3 represents the proportion of Fe2O3 generated; per_lnr represents the proportion of lining erosion in the hot metal; lnr_C represents the carbon content in the lining.

[0037] On the other hand, the present invention also provides a converter steelmaking oxygen consumption calculation system jointly driven by data and mechanism. The converter steelmaking oxygen consumption calculation system jointly driven by data and mechanism includes:

[0038] A data acquisition module, used for:

[0039] Obtain converter steelmaking production data;

[0040] Divide the converter steelmaking production data into on-site measurable data and non-real-time measurable parameters; among them, the non-real-time measurable parameters include key non-real-time measurable parameters and non-key non-real-time measurable parameters;

[0041] A data processing module, used for:

[0042] Input on-site measurable data into a preset parameter prediction model, and output the predicted values of key non-real-time measurable parameters and the predicted value of oxygen consumption through the parameter prediction model;

[0043] Adopt a preset oxygen consumption mechanism calculation model to calculate the oxygen consumption according to the converter steelmaking production data; among them, the on-site measurable data in the converter steelmaking production data adopts the measured value, the non-key non-real-time measurable parameters adopt the empirical values set manually, and the key non-real-time measurable parameters adopt the predicted values output by the parameter prediction model;

[0044] A result output module, which is used to weight the predicted value of oxygen consumption output by the parameter prediction model and the oxygen consumption calculated by the oxygen consumption mechanism calculation model to obtain the calculation result of the oxygen consumption in converter steelmaking.

[0045] On the other hand, the present invention also provides an electronic device, which includes a processor and a memory; wherein, at least one instruction is stored in the memory, and the instruction is loaded and executed by the processor to implement the above method.

[0046] On another aspect, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored, and the instruction is loaded and executed by the processor to implement the above method.

[0047] The beneficial effects brought by the technical solution provided by the present invention at least include:

[0048] The present invention realizes the dynamic identification of non-real-time measurable parameters in the oxygen consumption calculation process by integrating the mechanism equation into the neural network model. This method combines the advantages of data-driven and mechanism models to predict the oxygen consumption in the converter steelmaking process. While improving the accuracy and robustness of the prediction, the model has better interpretability, which helps to deeply understand the oxygen consumption mechanism in the converter steelmaking process and provides strong support for actual industrial applications. Experimental results show that compared with a single data-driven or mechanism model, the combined drive model used in this method has significant improvements in both accuracy and robustness. Especially when noise is added to the data, this model shows stronger stability, proving its superior robustness.

[0049] From a computational perspective, integrating physical knowledge into the machine learning framework to enhance the generalization ability is a successful attempt, which shows the potential of using domain knowledge to extract more informative features and enhance the model interpretability. For complex smelting scenarios, this model with strong generalization ability and mechanism supervision can help discover more accurate mapping relationships and provide more robust prediction performance. Description of the Drawings

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 It is a schematic diagram of the execution flow of the method for calculating oxygen consumption of converter steelmaking driven by data and mechanism provided in an embodiment of the present invention;

[0052] Figure 2 Schematic diagram of oxygen input and consumption in a converter steelmaking process provided by an embodiment of the present invention;

[0053] Figure 3 is a schematic diagram of the structure of a parameter prediction model provided by an embodiment of the present invention;

[0054] Figure 4 is a flow chart of identifying parameters that cannot be measured in real time provided by an embodiment of the present invention;

[0055] Figure 5 It is a system block diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0057] First of all, it should be noted that in the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "exemplarily" is intended to present the concept in a concrete way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0058] First embodiment

[0059] In order to accurately predict the oxygen consumption in the converter steelmaking process, this embodiment provides a method for calculating the oxygen consumption in the converter steelmaking process driven by data and mechanism. The method can be implemented by an electronic device, which can be a terminal or a server. The execution flow of the method is as follows: Figure 1 As shown, the following steps are included:

[0060] S1, obtaining converter steelmaking production data;

[0061] S2. Divide the converter steelmaking production data into on-site measurable data and non-real-time measurable parameters. Among them, the non-real-time measurable parameters include key non-real-time measurable parameters and non-key non-real-time measurable parameters.

[0062] Among them, it should be noted that according to the calculation process of the oxygen consumption mechanism and combined with on-site data collection, the converter steelmaking production data can be divided into: on-site measurable data and non-real-time measurable parameters. Among them, the on-site measurable data mainly includes hot metal element content, hot metal weight, molten steel element content, molten steel weight, scrap weight, quicklime, dolomite weight, etc. The non-real-time measurable parameters mainly include the incomplete combustion rate of carbon, scrap element content, proportion of lining erosion to hot metal volume, dust proportion, iron content in slag, oxygen purity, oxygen utilization rate, etc. After multiple calculations, four parameters, namely the incomplete combustion rate of carbon, slag ratio, carbon content in scrap, and oxygen utilization rate, have a greater impact on the oxygen consumption calculation result. Therefore, they can be used as key non-real-time measurable parameters, and subsequent machine learning methods need to be used for dynamic identification, that is, dynamic optimization according to the actual input materials and control parameters of each smelting heat. The others are non-key non-real-time measurable parameters, and empirical values can be taken during calculation. The oxygen consumption mechanism calculation can be abstracted as: M = f(on-site measurable data, non-real-time measurable parameters).

[0063] S3. Input the on-site measurable data into a preset parameter prediction model, and output the predicted values of the key non-real-time measurable parameters and the predicted value of the oxygen consumption through the parameter prediction model.

[0064] S4. Adopt a preset oxygen consumption mechanism calculation model to calculate the oxygen consumption according to the converter steelmaking production data. Among them, the measured values are used for the on-site measurable data in the production data, the empirical values set manually are used for the non-key non-real-time measurable parameters, and the predicted values output by the parameter prediction model are used for the key non-real-time measurable parameters.

[0065] Among them, it should be noted that based on the mechanism knowledge in the metallurgical field, by analyzing the physical and chemical reactions during the converter steelmaking process, especially the oxygen consumption mechanism of the oxidation reactions of elements (such as carbon, silicon, manganese, etc.) in the hot metal, an oxygen consumption mechanism calculation equation is constructed. Then, the oxygen consumption is calculated one by one according to the reaction mechanism equation and summed up as the total oxygen consumption. The mechanism calculation process takes into account the influence of key parameters such as oxygen purity, oxygen utilization rate, slag ratio, and carbon content in scrap. The oxygen input and consumption during the converter steelmaking process are as Figure 2 shown.

[0066] According to the oxygen consumption reaction mechanism, the total oxygen consumption is calculated as follows:

[0067]

[0068] Among them, M represents the oxygen consumption calculated by the mechanism model; 0.99 is the oxygen ratio in the blown gas (99% oxygen, 1% nitrogen), and O2_ratio is the oxygen utilization rate.

[0069] The calculation formulas for the oxygen consumption of the reactions of carbon, silicon, manganese, phosphorus, and sulfur elements in the converter are as follows:

[0070]

[0071] Among them, M C_O 、M Si_O 、M Mn_O 、M P_O 、M S_O are the oxygen consumptions of the oxidation reactions of C, Si, Mn, P, and S elements respectively in converter steelmaking production; W iron is the weight of hot metal; W hc 、W hsi 、W hmn 、W hp 、W hs are the contents of C, Si, Mn, P, and S elements in hot metal respectively; W scrap is the weight of scrap steel; W sc 、W ssi 、W smn 、W sp 、W ss are the contents of C, Si, Mn, P, and S elements in scrap steel respectively; W pig_iron is the weight of pig iron; W pc 、W psi 、W pmn 、W pp 、W ps are the contents of C, Si, Mn, P, and S elements in pig iron respectively; W steel is the weight of the produced molten steel; W tc 、W tsi 、W tmn 、W tp 、W ts are the contents of C, Si, Mn, P, and S elements in molten steel respectively. The oxidation reaction of carbon element generates carbon monoxide and carbon dioxide. per_C represents the ratio of carbon monoxide generated, that is, the incomplete combustion rate of carbon, and (1 - per_C) is the ratio of carbon dioxide generated. Sulfur element reacts with calcium oxide and oxygen respectively to generate calcium sulfide and sulfur dioxide. per_S represents the ratio of sulfur element oxidized to sulfur dioxide.

[0072] The oxidation of iron element contained in the converter will generate Fe2O3 and FeO. Generally, it is assumed that all the iron in the slag is oxidized to FeO. The calculation formula for the oxidation of iron in the slag is as follows:

[0073]

[0074] In the formula, Slag_ratio is the slag ratio, and lag_TFe is the iron content in the slag.

[0075] The calculation formula for the oxygen consumption of iron oxidation in the fume and dust is as follows:

[0076]

[0077] In the formula, Dust_ratio is the proportion of fume and dust in the hot metal; Dust_FeO is the proportion of generated FeO; Dust_Fe2O3 is the proportion of generated Fe2O3. Generally, the above parameters take empirical values according to expert experience.

[0078] The converter lining refers to the refractory layer built inside the converter metal shell. During the smelting process, a part of the carbon element in the lining will undergo an oxidation reaction at high temperature. The formula is as follows:

[0079]

[0080] In the formula, per_lnr is the proportion of lining erosion in the hot metal; lnr_C refers to the carbon content in the lining.

[0081] For the prediction model, in this embodiment, a multi-layer fully connected neural network as shown in Figure 3 is selected as the backbone model. The key non-real-time parameters and oxygen consumption in the mechanism calculation process are used as the outputs of the neural network. This network can identify parameters dynamically and also realize data-driven oxygen consumption prediction. During the training process, first, 19 relevant input data are input into the neural network, and the identification results of non-real-time measurable parameters are obtained through training. At the same time, the mechanism prediction value of oxygen consumption is calculated according to the total oxygen consumption calculation formula. In addition, the neural network directly learns based on historical data to generate a data-driven oxygen consumption prediction result. Finally, the two prediction values are combined by weighted summation to obtain a comprehensive oxygen consumption prediction result. The loss function is designed as the weighted sum of the mechanism loss and the data loss, ensuring that the model can not only learn the patterns in the data but also comply with physical laws. By adjusting the weights of the mechanism loss and the data loss, the influence of both on the total loss function is balanced, enabling the model to capture the potential patterns in the data and conform to physical laws. While improving the prediction accuracy, generalization ability, and robustness of the model for non-predictable parameters of oxygen consumption, the interpretability of the model is ensured.

[0082] Specifically, the method includes four parts: feature extraction, parameter identification, data-driven output prediction, and mechanism and data joint loss. The 19 inputs pass through the feature extraction network and can be used for the identification of 4 key non-real-time measurable parameters or for data-driven output prediction. The loss function Loss uses the mechanism and data joint loss and is designed as the mechanism loss Loss M and the data loss LossD The weighted sum is expressed as follows:

[0083] Loss = Loss M + Loss D

[0084] The data loss method is calculated as follows:

[0085]

[0086] where N is the total number of training data, is the oxygen consumption predicted based on the i-th training data, and O i is the true oxygen consumption data collected corresponding to the i-th training data.

[0087] The mechanism loss method is calculated as follows:

[0088]

[0089] where, is the oxygen consumption calculated after substituting the key non-real-time measurable parameters predicted by the parameter prediction model based on the i-th training data into the oxygen consumption mechanism calculation model.

[0090] By training and optimizing the joint drive model based on historical production data, a high-precision oxygen consumption prediction result can be obtained and the interpretability of the model can be ensured. The training process of this parameter prediction model is as follows:

[0091] Collect the historical production data and historical oxygen consumption of converter steelmaking, preprocess the collected data, and construct a sample data set using the preprocessed historical production data and historical oxygen consumption of converter steelmaking;

[0092] It should be noted that the converter steelmaking process is carried out in a high-temperature and complex container. In such a complex production environment, when collecting some data, abnormal situations such as data drift and missing will occur. Before using this part of the data, it is necessary to preprocess the collected data to obtain accurate data. For this reason, in the data preprocessing stage, this embodiment adopts the box plot method to identify and remove outlier data points.

[0093] Specifically, this embodiment selects the data set of a certain steel plant for experiments. This data set contains 3169 samples. After the data preprocessing operation, 3084 effective samples are finally obtained. Before model training, the data is randomly shuffled and then divided into a training set and a test set according to a ratio of 8:2.

[0094] Use the sample data set to train the parameter prediction model; where the input of the parameter prediction model is on-site measurable data, and the output is key non-real-time measurable parameters and oxygen consumption.

[0095] Among them, it should be noted that before predicting the oxygen consumption using the network model, the ranges of on-site measurable data and non-real-time measurable parameters need to be determined. For the selection of input data, in this embodiment, the Pearson algorithm is adopted and combined with expert experience to determine the input factors. Thus, for the oxygen consumption prediction of the converter, 19 input factors are determined, including: the contents of molten iron elements (carbon, silicon, manganese, phosphorus, sulfur), the contents of end-point molten steel elements (carbon, silicon, manganese, phosphorus, sulfur), the weight of molten steel, the weight of molten iron, the weight of scrap steel, the weight of pig iron, the tapping temperature, the gas recovery amount, the weight of quicklime, the consumption of lightly burned dolomite, and the consumption of magnesium balls.

[0096] Based on the above, this method first screens the data set, and then designs a neural network that combines the oxygen consumption mechanism to conduct supervised training. Through network training, 4 key non-real-time measurable parameters (Per_C, Slag_ratio, Scrap_C, O2_ratio) in the mechanism formula and the oxygen consumption O directly predicted by the network model are obtained. pre According to the mechanism formula calculated based on the furnace condition identification of each furnace, O is obtained. cal After that, the data loss Loss is calculated. D and the mechanism loss Loss M are obtained to get the final total loss value Loss. After training, according to the directly predicted oxygen consumption O pre and the mechanism formula calculated based on the furnace condition identification of each furnace, O is obtained. cal The final oxygen consumption O is jointly calculated. end Finally, its hit rate is evaluated. Among them, the identification process of non-real-time measurable parameters is as Figure 4 shown.

[0097] S5, weight the predicted value of the oxygen consumption output by the parameter prediction model and the oxygen consumption calculated by the oxygen consumption mechanism calculation model to obtain the oxygen consumption calculation result of converter steelmaking.

[0098] Among them, it should be noted that after the model training is completed, in this embodiment, based on data-driven, the neural network is used to obtain the oxygen consumption prediction result O pre ; and based on the total oxygen consumption calculation formula, by inputting relevant data for calculation, the mechanism calculation result O is obtained. cal ; finally, the two are weighted and summed to obtain the oxygen consumption prediction result O end :

[0099] O end = 0.5×O cal + 0.5×O pre

[0100] Next, the hit rate commonly used in the prediction of the oxygen consumption of a converter is adopted as the performance evaluation index to evaluate the effect of the present invention. The hit rate is calculated within the error ranges of ±5%, ±7.5% and ±10% respectively.

[0101] For a dataset of 3084 groups of data from a certain factory, the ratio of training data to test data is 8:2.

[0102] Within the error ranges of ±5%, ±7.5% and ±10%, the predicted hit rates are 36.3%, 72.93% and 91.09% respectively. Compared with the single mechanism model, the hit rate is increased by about 1.28%, 3.08% and 3.57%. Compared with the single data-driven model, the hit rate is increased by about 0.78%, 3.59% and 3.75%.

[0103] Select some actual data as a case for display. Table 1 shows the identification results of the non-real-time measurable parameters and the final oxygen consumption prediction results. It can be seen that for each heat, these non-real-time measurable parameters will change to a certain extent, indicating that this parameter identification method can dynamically adapt to the actual furnace conditions of each heat.

[0104] Table 1 Identification results of non-real-time measurable parameters and predicted values of oxygen consumption

[0105]

[0106]

[0107] Furthermore, in order to evaluate the robustness of the model, in this embodiment, a method of adding noise to the training set is designed, and the change degree of the model accuracy is evaluated. Specifically, the influencing factors of oxygen consumption are analyzed, and the input factor with the greatest influence: the carbon content of hot metal, is selected, and 5%, 10% and 20% of noise are added to it respectively. Finally, the degradation of the model prediction accuracy under different noise ratios is evaluated to analyze the influence of noise data on the model performance. The results are shown in Table 2.

[0108] It can be observed from the data in Table 2 that as the noise data increases, the prediction accuracies of the three models gradually decrease. However, the data and mechanism jointly driven model shows a higher hit rate than the other two models at different noise levels. In addition, within different hit rate ranges, the prediction accuracy of the data-driven model drops most significantly, while the drop degree of the jointly driven model is relatively small. The results show that when facing data noise, the jointly driven model has stronger robustness than the data-driven model.

[0109] Table 2 Accuracy degradation of the model under different noise ratios

[0110]

[0111]

[0112] (In the table, the difference represents the largest data difference among the three groups of data, which can measure the model degradation situation)

[0113] In summary, this embodiment provides a method for calculating the oxygen consumption in converter steelmaking jointly driven by data and mechanism. By constructing a physics-informed neural network, the identification of non-real-time measurable parameters in oxygen consumption calculation is realized, and then the oxygen consumption prediction jointly driven by data and mechanism is realized. The experimental results show that this method for calculating the oxygen consumption in converter steelmaking jointly driven by data and mechanism significantly improves the oxygen consumption hit rate within a certain error range. Compared with the pure data-driven method, this method has significant improvements in the accuracy and robustness of the prediction task, and the training process has good interpretability due to the mechanism supervision.

[0114] For the oxygen consumption prediction task, this jointly driven method uses mechanism supervision to drive the data-driven training process, thus ensuring the interpretability of the method. At the same time, it dynamically identifies the key non-real-time measurable parameters in the mechanism calculation process through data driving, thus ensuring the dynamic adaptability and accuracy of the mechanism calculation. This method helps to optimize the production process, improve product quality, reduce production costs, and promote the sustainable development of the steel industry.

[0115] Second Embodiment

[0116] This embodiment provides a system for calculating the oxygen consumption in converter steelmaking jointly driven by data and mechanism. The system for calculating the oxygen consumption in converter steelmaking jointly driven by data and mechanism includes the following modules:

[0117] Data acquisition module, for:

[0118] Obtain the converter steelmaking production data;

[0119] Divide the converter steelmaking production data into on-site measurable data and non-real-time measurable parameters; among them, the non-real-time measurable parameters include key non-real-time measurable parameters and non-key non-real-time measurable parameters;

[0120] Data processing module, for:

[0121] Input the on-site measurable data into a preset parameter prediction model, and output the predicted values of the key non-real-time measurable parameters and the predicted value of the oxygen consumption through the parameter prediction model;

[0122] Adopt a preset oxygen consumption mechanism calculation model to calculate the oxygen consumption according to the converter steelmaking production data; among them, the on-site measurable data in the converter steelmaking production data adopts the measured value, the non-key non-real-time measurable parameters adopt the empirical values set manually, and the key non-real-time measurable parameters adopt the predicted values output by the parameter prediction model;

[0123] A result output module, configured to weight the predicted value of oxygen consumption output by the parameter prediction model and the oxygen consumption calculated by the oxygen consumption mechanism calculation model to obtain the calculation result of oxygen consumption for converter steelmaking.

[0124] It should be noted that, for the convenience of description, the data and mechanism jointly driven converter steelmaking oxygen consumption calculation system in this embodiment corresponds to the data and mechanism jointly driven converter steelmaking oxygen consumption calculation method in the above first embodiment; among them, the functions implemented by each functional module in the data and mechanism jointly driven converter steelmaking oxygen consumption calculation system in this embodiment correspond one by one to each process step in the data and mechanism jointly driven converter steelmaking oxygen consumption calculation method in the above first embodiment; therefore, it will not be elaborated here.

[0125] The third embodiment

[0126] This embodiment provides an electronic device, as Figure 5 shown, the electronic device includes: a processor and a memory; wherein, the processor and the memory can be connected through a communication bus; at least one instruction is stored in the memory, and the instruction is loaded and executed by the processor to implement the method of the above first embodiment. In addition, the electronic device may further include a transceiver, the processor and the transceiver can be connected through a communication bus, and the transceiver is used for communicating with other devices.

[0127] Next, in combination with Figure 5 specific introductions will be made to the respective components of the electronic device:

[0128] Among them, the processor is the control center of the electronic device. The electronic device may include multiple processors, and each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may be a single processor or a collective term for multiple processing elements. For example, the processor may be one or more central processing units (CPUs), or other general-purpose processors, application specific integrated circuits (ASICs), or one or more integrated circuits configured to implement the embodiments of the present invention. For example: one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor can execute various functions of the electronic device by running or executing software programs stored in the memory and calling data stored in the memory.

[0129] In a specific implementation, as an embodiment, the processor may include one or more CPUs. For example Figure 5 CPU0 and CPU1 shown in, of course, this is only an exemplary illustration.

[0130] The memory is used to store the software program for implementing the solution of the present invention and is controlled by the processor for execution. The specific implementation manner may refer to the above method embodiments and will not be elaborated here.

[0131] Optionally, the memory may be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, or may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may be integrated with the processor or may exist independently and be coupled to the processor through the interface circuit ( Figure 5 not shown) of the electronic device. The embodiments of the present invention do not make specific limitations in this regard.

[0132] The transceiver may include a receiver and a transmitter ( Figure 5 not shown separately). Among them, the receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function. The transceiver may be integrated with the processor or may exist independently and be coupled to the processor through the interface circuit ( Figure 5 not shown) of the electronic device. The embodiments of the present invention do not make specific limitations in this regard.

[0133] In addition, it should be noted that Figure 5 the structure of the electronic device shown in does not constitute a limitation on the device. The actual device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements. In addition, the technical effects achieved by the electronic device when executing the method of the first embodiment above may refer to the technical effects described in the first embodiment above. Therefore, they will not be elaborated here.

[0134] Fourth Embodiment

[0135] This embodiment provides a computer-readable storage medium in which at least one instruction is stored. The instruction is loaded and executed by the processor to implement the method of the first embodiment above. Among them, the computer-readable storage medium may be ROM, random access memory, CD-ROM, magnetic tape, floppy disk, and optical data storage devices, etc. The instructions stored therein can be loaded and executed by the processor in the terminal to implement the above method.

[0136] In addition, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, the embodiments of the present invention can take the form of all or part of a hardware embodiment, all or part of a software embodiment, or an embodiment combining software and hardware aspects. Moreover, when implemented in software, the embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present invention are fully or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center containing one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0137] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing terminal device generate a device for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0138] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1the functions specified in one or more boxes. These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable terminal device provide steps for implementing the functions specified in one or more processes and / or boxes Figure 1 one process or more processes and / or boxes Figure 1 the steps of the functions specified in one or more boxes.

[0139] It should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising the said element. In addition, the term "and / or" is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. Among them, A and B can be singular or plural. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood with reference to the context. "At least one" means one or more, and "a plurality" means two or more. "At least one of the following (items)" or similar expressions refer to any combination of these items, including any combination of single items (items) or plural items (items). For example, at least one of a, b or c can mean: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, c can be single or multiple.

[0140] In addition, it can be understood that in various embodiments of the present invention, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0141] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0142] In several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of functional modules / units is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in each embodiment of the present invention, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0143] If the method is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0144] Finally, it should be noted that the above description is only the preferred embodiment of the present invention. It should be pointed out that although the preferred embodiments of the present invention have been described, for those of ordinary skill in the art, once the basic creative concept of the present invention is known, several improvements and refinements can be made without departing from the principle described in the present invention. These improvements and refinements should also be regarded as the protection scope of the present invention. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A method for calculating oxygen consumption of converter steelmaking driven by data and mechanism, characterized in that: include: Obtain converter steelmaking production data; The converter steelmaking production data is divided into on-site measurable data and parameters that cannot be measured in real time; wherein the parameters that cannot be measured in real time include key parameters that cannot be measured in real time and non-key parameters that cannot be measured in real time; Inputting the on-site measurable data into a preset parameter prediction model, and outputting the predicted values ​​of key parameters that cannot be measured in real time and the predicted values ​​of oxygen consumption through the parameter prediction model; The preset oxygen consumption mechanism calculation model is used to calculate the oxygen consumption according to the converter steelmaking production data; among which, the field measurable data in the converter steelmaking production data adopts the measured values, the non-critical parameters that cannot be measured in real time adopt the manually set empirical values, and the critical parameters that cannot be measured in real time adopt the predicted values ​​output by the parameter prediction model; The predicted value of oxygen consumption output by the parameter prediction model and the oxygen consumption calculated by the oxygen consumption mechanism calculation model are weighted to obtain the calculation result of converter steelmaking oxygen consumption.

2. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 1, characterized in that: The on-site measurable data include: molten iron element content, final molten steel element content, molten steel weight, molten iron weight, scrap steel weight, pig iron weight, steel tapping temperature, gas recovery volume, quicklime weight, light-burned dolomite usage, and magnesium ball usage.

3. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 2, characterized in that: The non-critical parameters that cannot be measured in real time include: the proportion of furnace lining erosion to molten iron, the proportion of smoke, the iron content in the slag, and the oxygen purity.

4. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 3, characterized in that: The key parameters that cannot be measured in real time include: incomplete combustion rate of carbon, slag ratio, carbon content of scrap steel and oxygen utilization rate.

5. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 1, characterized in that: The parameter prediction model is a multi-layer neural network model.

6. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 5, characterized in that: The training process of the parameter prediction model includes: Collect historical production data and historical oxygen consumption of converter steelmaking, preprocess the collected data, and use the preprocessed historical production data and historical oxygen consumption of converter steelmaking to construct a sample data set; The sample data set is used to train a parameter prediction model, wherein the input of the parameter prediction model is field measurable data, and the output is key parameters that cannot be measured in real time and oxygen consumption.

7. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 6, characterized in that: The preprocessing of the collected data includes: The box plot method was used to identify and remove outlier data points in the data.

8. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 6, characterized in that: The loss function of the parameter prediction model is the sum of mechanism loss and data loss, expressed as: Loss=Loss M +Loss D Among them, Loss is the loss function of the parameter prediction model; Loss M Mechanism loss; Loss D For data loss; Data Loss D The calculation formula is: Where N is the total number of training data; is the oxygen consumption predicted based on the i-th training data; O i is the real data of oxygen consumption collected in the i-th training data; Mechanism Loss M The calculation formula is: in, The oxygen consumption is calculated by substituting the key non-real-time measurable parameter predicted by the parameter prediction model based on the i-th training data into the oxygen consumption mechanism calculation model.

9. The method for calculating oxygen consumption of converter steelmaking driven by data and mechanism as claimed in claim 1, characterized in that: The expression of the oxygen consumption mechanism calculation model is: Wherein, M represents the calculated oxygen consumption; O2_ratio represents the oxygen utilization rate; Among them, W iron Indicates the weight of molten iron; W hc , W hsi , W hmn , W hp , W hs Respectively represent the content of C, Si, Mn, P, and S in molten iron; W scrap Indicates the weight of scrap steel; W sc , W ssi , W smn , W sp , W ss Respectively represent the content of C, Si, Mn, P, and S in scrap steel; W pig_iron Indicates the weight of pig iron; W pc , W psi , W pmn , W pp , W ps Respectively represent the content of C, Si, Mn, P, S in pig iron; W steel Indicates the weight of molten steel produced, W tc , W tsi , W tmn , W tp , W ts They respectively represent the content of C, Si, Mn, P and S elements in molten steel; per_C represents the proportion of carbon monoxide generated; per_S represents the proportion of sulfur oxidized to sulfur dioxide; Slag_ratio represents the slag ratio; Slag_TFe represents the iron content in the slag; Dust_ratio represents the proportion of smoke in the molten iron; Dust_FeO represents the proportion of FeO generated; Dust_Fe2O3 represents the proportion of Fe2O3 generated; per_lnr represents the proportion of lining erosion in the molten iron; lnr_C represents the carbon content in the lining.

10. A converter steelmaking oxygen consumption calculation system driven by data and mechanism, characterized in that: include: Data acquisition module for: Obtain converter steelmaking production data; The converter steelmaking production data is divided into on-site measurable data and parameters that cannot be measured in real time; wherein the parameters that cannot be measured in real time include key parameters that cannot be measured in real time and non-key parameters that cannot be measured in real time; Data processing module for: Inputting the on-site measurable data into a preset parameter prediction model, and outputting the predicted values ​​of key parameters that cannot be measured in real time and the predicted values ​​of oxygen consumption through the parameter prediction model; The preset oxygen consumption mechanism calculation model is used to calculate the oxygen consumption according to the converter steelmaking production data; among which, the field measurable data in the converter steelmaking production data adopts the measured values, the non-critical parameters that cannot be measured in real time adopt the manually set empirical values, and the critical parameters that cannot be measured in real time adopt the predicted values ​​output by the parameter prediction model; The result output module is used to weight the predicted value of oxygen consumption output by the parameter prediction model and the oxygen consumption calculated by the oxygen consumption mechanism calculation model to obtain the calculation result of converter steelmaking oxygen consumption.

Images