Acid slag smelting control method based on data analysis

Abstract

The invention belongs to the technical field of metallurgy, and relates to an acid slag smelting control method based on data analysis. According to the method, the historical case library is built, the parameter optimization model is built based on the deep learning technology, basic data support is provided for subsequent intelligent recommendation and dynamic adjustment of process parameters, and the generalization ability and adaptability of the parameter optimization model are enhanced through accumulation and updating of the historical case library; by generating the initial solution and the constraint condition of the parameter optimization model, the operation time of the parameter optimization model is shortened, the convergence speed of the model is improved, and intelligent recommendation and dynamic intelligent adjustment of the process parameters are realized; according to the method, feedback closed-loop control of recommended values and adjustment of process parameters is achieved through an automatic control system, the ferrochrome recovery rate and the resource utilization rate are improved, the production cost is reduced, eye shifting and furnace leakage accidents are reduced, the production safety is improved, and multi-target collaborative optimization and cost control are achieved.

Description

Technical Field

[0001] The invention belongs to the technical field of metallurgy and relates to an acid slag smelting control method based on data analysis. Background Art

[0002] Acid slag smelting is an important smelting method in the metallurgical industry, mainly used to process ores or furnace materials containing acidic oxides. Ferrochrome is an important material widely used in the industrial field, and its smelting method is acid slag smelting. With the development of the global economy and industry, the demand for ferrochrome is also increasing. However, in recent years, the price of chromium ore for producing ferrochrome has continued to rise, and companies are facing increasing cost pressure. Therefore, in order to reduce production costs, a new monitoring and control method for acid slag smelting of ferrochrome is urgently needed.

[0003] In the prior art, there are also some related solutions related to ferrochrome smelting. For example, a smelting method of a high carbon ferrochrome alloy with a Chinese patent publication number of CN114921604A comprises: adding South African chromite and Turkish chromite to a mixer according to a mass ratio (1-9): 1, and then adding bentonite and dust removal ash to obtain a mixture. Add water to the mixture to granulate into a raw material mass. The raw material mass is sieved and roasted to obtain chromite pellets, and the roasting temperature is controlled to be 1200-1350°C and the roasting time is 12-15h. The chromite pellets are mixed with slag-making agents and reducing agents and put into the furnace, and the smelting parameters of the submerged arc furnace are adjusted: that is, the pole center circle diameter is 2.4-2.6m. The electrode current density is 8-8.2A / cm2, the secondary voltage is 170-173V, the power is 220kVA, the smelting temperature is 1650°C-1750°C, and the iron is tapped after smelting for 4-5h to obtain a high carbon ferrochrome alloy. The method stabilizes the source of raw materials and makes it easy to regulate the alkali ratio, magnesium-aluminum ratio, slag type, etc. during the batching process, thereby reducing energy consumption, production costs, and improving the comprehensive recovery rate of chromium.

[0004] Another Chinese patent publication number CN117431357A is a method for controlling the process operation of an electric arc furnace through a model. In order to solve the problem of judging and adjusting the furnace condition of a fully enclosed electric arc furnace for smelting high-carbon ferrochrome, a process model for judging and adjusting the furnace condition is proposed through heat transfer analysis, theoretical calculations and statistical analysis of a large amount of production data. The operating ideas and methods of employees on the job are unified, and the problem of long-term stable operation of the electric arc furnace is solved. The technical and economic indicators such as power consumption and ore consumption have reached the world's advanced level.

[0005] Although the above schemes have proposed some solutions for ferrochrome smelting, they still have certain limitations: (1) Some existing technologies originally controlled the smelting process by idle specific raw material ratios and fixed process parameters to reduce energy consumption, production costs and improve the comprehensive recovery rate of chromium. However, they ignored the fact that the raw material ratios in the ferrochrome smelting process may be inconsistent with specific requirements and the process parameter values ​​may also change, thus failing to ensure the intelligent recommendation and dynamic adjustment of process parameters under the conditions of low energy consumption, low cost and high ferrochrome recovery rate.

[0006] (2) Some existing technologies optimize furnace condition judgment and adjustment through heat transfer analysis and production data statistics. However, their intelligence level is low and the adjustment process mainly relies on manual experience. As a result, they are unable to reduce the adjustment lag caused by manual experience and are highly subjective. They are also unable to achieve real-time precise control through data-driven means.

[0007] (3) The existing technologies lack feedback closed-loop control for realizing recommended values ​​and adjustments of process parameters through automated control systems, thus failing to achieve precise control of charge ratio and chromium content in slag, failing to reduce demand for new raw materials and production costs, failing to reduce accidents of eyelet and furnace leakage, failing to improve production safety, and thus failing to achieve multi-objective collaborative optimization and cost control. Summary of the invention

[0008] In view of this, in order to solve the problems raised in the above background technology, a method for controlling acid slag smelting based on data analysis is now proposed.

[0009] The purpose of the present invention can be achieved through the following technical solutions: The present invention provides an acid slag smelting control method based on data analysis, including: S1, building a historical case library according to the various production parameters and corresponding process parameters generated by each acid slag smelting process collected during the test in the target smelting field, and building a parameter optimization model based on deep learning technology.

[0010] S2. Perform similarity analysis on the collected production parameters in the current acid slag smelting process and the corresponding production parameters in the historical case library by calculating the Euclidean distance, determine the initial solution and constraint conditions of the current acid slag smelting process in the parameter optimization model, and run it to obtain the recommended values ​​of each production parameter.

[0011] S3. Compare each production parameter with the corresponding recommended value to calculate the deviation and determine the control requirement of each production parameter. If it is a control requirement, execute S4; otherwise, execute S2.

[0012] S4. Use the automatic control system to regulate and control the corresponding process parameters of each production parameter, and adjust each production parameter to the corresponding recommended value.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The present invention retains information by constructing a historical case library for the collected parameter information during the experiment, thereby providing basic data support for the subsequent intelligent recommendation and dynamic adjustment of process parameters, and continuously enhances the generalization ability and adaptability of the parameter optimization model by building a parameter optimization model based on deep learning technology and accumulating and updating the historical case library in real time.

[0014] 2. The present invention calculates the Euclidean distance between the various production parameters in the current acid slag smelting process and the corresponding production parameters in the historical case library for similarity analysis, generates the initial solution and constraint conditions of the parameter optimization model, shortens the running time of the parameter optimization model, improves the model convergence speed, realizes the intelligent recommendation and dynamic intelligent adjustment of process parameters, and reduces the dependence on manual experience.

[0015] 3. The present invention determines the control requirements of various production parameters and adjusts the corresponding process parameters to the corresponding recommended values ​​according to the automatic control system, thereby achieving precise control of the charge ratio and the chromium content in the slag, improving the ferrochrome recovery rate and resource utilization, reducing the demand for new raw materials, reducing production costs, and reducing eyelet and furnace leakage accidents, extending the service life of the smelting furnace, improving production safety, and thus achieving multi-objective collaborative optimization and cost control. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0017] Figure 1 It is a schematic diagram for implementing the method steps of the present invention.

[0018] Figure 2 Construct a flow chart for the historical case library of the present invention.

[0019] Figure 3 A flow chart is constructed for the parameter optimization model of the present invention.

[0020] Figure 4 Schematic diagram of the hardware structure of the computer of the present invention.

[0021] Description of the drawings: 200, bus; 201, processor; 202, memory; 203, communication interface. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Embodiment 1

[0024] See also Figure 1 As shown, the present invention provides an acid slag smelting control method based on data analysis, and the specific steps are as follows: S1, constructing a historical case library according to various production parameters and corresponding process parameters generated by each acid slag smelting process collected during the test in the target smelting field, and building a parameter optimization model based on deep learning technology.

[0025] As a preferred feasible embodiment, the production parameters include polar center circle power density, charge conductivity, chromium content in slag, furnace wall thickness and silicon dioxide content. Production parameters refer to physical quantities, chemical quantities and process status information collected in real time by sensors, instruments and other equipment during the production process.

[0026] It should be further explained that the polar circle power density, charge conductivity, chromium content in slag, furnace wall thickness and silicon dioxide content are specifically collected and obtained in the following manner: according to the standard calculation formula Get the polar power density ,in They are the total power and the pole circle diameter of the submerged arc furnace respectively. The conductivity of the charge is directly measured by a conductivity measuring instrument (such as an eddy current conductivity meter). The thickness of the furnace wall is directly measured by a thickness measuring instrument (such as an ultrasonic thickness gauge). The composition analyzer based on X-ray fluorescence spectroscopy (XRF) directly detects the silica in the slag and the reactor to obtain the chromium content and silica content in the slag.

[0027] The process parameters include the diameter of the polar circle, the content of basic oxides, the thickness of the charge, the proportion of the coke particle size distribution, the current-voltage ratio, the chromium ore particle size, the slag iron temperature, the calcium oxide content in the slag, the number of times the furnace is discharged, the refining time, the voltage and the silicon dioxide content. The process parameters refer to the controllable operating variables in the production process, which are used to optimize the process performance (such as efficiency, cost, safety).

[0028] As a preferred feasible embodiment, Figure 2As shown, the specific construction process of the historical case library includes: A1, label setting: determine whether each acid slag smelting process in the target smelting field during the test process has achieved the preset expected goal. If the expected goal is achieved, its label is set as a successful case, otherwise, its label is set as a failed case, and the labels of each acid slag smelting process in the target smelting field during the test process are set accordingly. Distinguishing between successful and failed cases through labels is convenient for subsequent analysis and optimization, and successful cases can provide optimization references for subsequent parameter optimization models, and failed cases can provide constraints for subsequent parameter optimization models.

[0029] A2. Data type matching: Match the production parameters and corresponding process parameters generated by each acid slag smelting process collected in the target smelting field after setting the label with the production parameters and process parameters of each data type stored in the information base, and obtain the data types of each production parameter and corresponding process parameter generated by each acid slag smelting process collected in the target smelting field after setting the label. Unify the data format and dimension, clarify the data type of each parameter (such as safety, efficiency, and economy), and provide a basis for subsequent analysis and modeling. In addition, through data type matching, relevant data can be quickly located and called, data management efficiency can be improved, and input features can be provided for subsequent multi-objective optimization models; according to the data type (such as safety>efficiency>economy), the priority order of parameter adjustment is determined to achieve comprehensive optimization of safety, efficiency, and economy.

[0030] A3. Construction of historical case library: The production parameters and corresponding process parameters generated by each acid slag smelting process collected in the target smelting field during the test after data type matching are stored in the database after data association using timestamps and numbers, and recorded as a historical case library, which is updated in real time. Through the association of timestamps and numbers, unified management of multi-source data is achieved; the case library is updated in real time to ensure the timeliness and accuracy of the historical case library; the model is continuously optimized through feedback closed loops to improve prediction accuracy and adjustment effects.

[0031] As a preferred feasible embodiment, the preset expected targets include the CaO content in the slag, the improvement rate of chromium ore reducibility, and the probability of occurrence of eyelet and furnace leakage problems.

[0032] It needs to be further explained that the method for obtaining the expected target includes: extracting the Ca0 content in the slag, the improvement rate of chromium ore reducibility, and the probability of eye run and furnace leakage problems of each acid slag smelting process labeled as a successful case from the historical case library, and selecting the maximum Ca0 content in the slag, the highest maximum chromium ore reducibility improvement rate, and the minimum probability of eye run and furnace leakage problems in each acid slag smelting process as the expected target.

[0033] The data types include safety data, efficiency data and economic data.

[0034] It should be explained that the safety data refers to parameters that directly affect production safety (such as the risk of eye leakage and furnace wall erosion rate). The safety data in the present invention include furnace wall thickness and silica content. The efficiency data refers to parameters that directly affect production efficiency (such as metal recovery rate and slag iron separation efficiency). The efficiency data in the present invention include polar circle power density, charge conductivity, chromium content in slag, polar circle diameter, alkaline oxide content, charge thickness, coke particle size distribution ratio, current-voltage ratio, chromium ore particle size, slag iron temperature, calcium oxide content in slag, refining time and voltage. The economic data refers to parameters that directly affect production costs (such as energy consumption and raw material consumption). The economic data in the present invention includes the number of times the furnace is discharged.

[0035] As a preferred feasible embodiment, Figure 3 As shown, the parameter optimization model building process includes: B1, data set construction and division: extract the various production parameters generated by each acid slag smelting process collected in the target smelting field during the test process from the historical case library, collectively referred to as the acid slag smelting parameter data set, and divide it into the acid slag smelting parameter training set, test set and validation set. The training set accounts for most of the data set (usually 60%-80%) and is used for model learning and parameter adjustment. The model will learn the patterns and rules of the data from the training set, and continuously optimize its own parameters to predict or classify the input data as accurately as possible. The validation set generally accounts for 10%-20% of the data set and is used to evaluate and monitor the performance of the model during the model training process. Through the performance on the validation set, it can be judged whether the model is overfitting or underfitting, and then the hyperparameters of the model (such as learning rate, number of network layers, etc.) can be adjusted to improve the generalization ability of the model. The test set also accounts for 10%-20% of the data set and is used to objectively evaluate the final performance of the model after the model training is completed. The data in the test set has not been used in the model training process, so it can more realistically reflect the performance of the model in actual applications.

[0036] B2. Model architecture selection: Select the network structure and build the model. The built model will be recorded as the basic model for parameter optimization.

[0037] As a specific example, the network structure can be a multi-layer perceptron (MLP), which consists of an input layer, multiple hidden layers and an output layer, each neuron is connected to all neurons in the previous layer, and can learn complex mapping relationships through nonlinear activation functions; or a recurrent neural network (RNN). When the production parameters have time series characteristics, such as the change of furnace temperature over time, the adjustment of raw material addition amount over time, etc., RNN can better process sequence data and capture the time dependency in the data; or a convolutional neural network (CNN). If the data has a spatial structure, such as the spatial distribution of furnace temperature or components, CNN can automatically extract the spatial features of the data through the convolution layer, and is widely used in image recognition and signal processing.

[0038] B3. Define loss function and optimizer: Select the loss function and optimizer of the parameter optimization base model. The loss function is used to measure the difference between the model's predicted value and the true value. The optimizer's role is to update the model's parameters according to the gradient of the loss function to minimize the loss function.

[0039] In a specific example, the loss function may be a mean square error loss function (MSE), which is suitable for regression problems, such as predicting the specific value of a production parameter (such as the exact value of the pressure in the furnace). It calculates the average of the square of the difference between the predicted value and the true value, so that the predicted value of the model is as close to the true value as possible.

[0040] The optimizer can be Adam optimizer, which is suitable for large-scale data sets and complex models. It has a fast convergence speed, can quickly find the appropriate parameter update direction in the early stage of training, and can adaptively adjust the learning rate during training to avoid the problem of too large or too small learning rate. In the acid slag smelting parameter optimization model, Adam can help the model converge to the optimal solution faster.

[0041] B4. Model training: The acid slag smelting parameter training set is input into the parameter optimization basic model for training. During the training process, the parameters of the parameter optimization basic model are continuously adjusted to minimize the loss function until it is stable. The parameter optimization basic model corresponding to the minimized and stable loss function is recorded as the parameter optimization model.

[0042] Specifically, the training process is to input the acid slag smelting parameter training set into the parameter optimization basic model, and the model performs forward propagation to calculate the predicted value. Then, the loss function is used to calculate the loss between the predicted value and the true value. Next, the gradient of the loss function with respect to the model parameters is calculated by the back propagation algorithm. The optimizer updates the parameters of the model according to the calculated gradient so that the value of the loss function gradually decreases. Training continues until the value of the loss function is minimized and stabilized. During the training process, a maximum number of training rounds (epochs) is usually set, or when the change in the loss function within a certain number of rounds is less than a certain threshold, the training is considered to have converged and the training is stopped. At this time, the parameter optimization basic model in this state is recorded as the parameter optimization model.

[0043] The present invention retains information by constructing a historical case library for the collected parameter information during the experiment, thereby providing basic data support for the subsequent intelligent recommendation and dynamic adjustment of process parameters. It also builds a parameter optimization model based on deep learning technology, and accumulates and updates the historical case library in real time, thereby continuously enhancing the generalization ability and adaptability of the parameter optimization model.

[0044] S2. Perform similarity analysis on the collected production parameters in the current acid slag smelting process and the corresponding production parameters in the historical case library by calculating the Euclidean distance, determine the initial solution and constraint conditions of the current acid slag smelting process in the parameter optimization model, and run it to obtain the recommended values ​​of each production parameter.

[0045] As a preferred feasible embodiment, the specific method for determining the initial solution and constraint conditions of the parameter optimization model of the current acid slag smelting process is: extracting various production parameters in the current acid slag smelting process , extract the corresponding production parameters generated by each acid slag smelting process labeled as a successful case from the historical case library, and record them as the production parameters of the current acid slag smelting process corresponding to the production parameters of each successful historical acid slag smelting process Similarly, the production parameters in the current acid slag smelting process can be obtained to correspond to the production parameters of the acid slag smelting process in each failed history. ,in , is the change of each production parameter, , The serial numbers of each successful acid slag smelting in history, , It is the serial number of each failed historical acid slag smelting.

[0046] According to the calculation formula Obtain the Euclidean distance between each successful historical acid slag smelting process and the current acid slag smelting process , select the successful historical acid slag smelting process with the smallest Euclidean distance from the current acid slag smelting process, and record its various production parameters as the initial solution of the current acid slag smelting process in the parameter optimization model. The initial solution refers to the process parameter combination that reaches or exceeds the expected target (such as the Ca0 content in the slag to about 5%, the chromium ore reducibility to increase by 25%, and the probability of eyelet and furnace leakage problems to 0) in the historical production data.

[0047] It should be explained that in the parameter optimization model, the initial solution is the starting point when the algorithm begins to search for the optimal solution. Using the production parameters of successful cases as the initial solution can accelerate the convergence of the parameter optimization model and improve the optimization efficiency.

[0048] Specifically, in the acid slag smelting process, it includes many parameters such as the size of the polar circle, current, voltage, coke particle size, chromite particle size, alkaline oxide addition, slag iron temperature, and the proportion of components in the slag. Each parameter has its own range of values, and different parameters may also influence each other. Taking voltage and current as examples, they jointly determine the power input into the furnace, and at the same time affect the electrode insertion depth, the temperature distribution in the furnace, and the reduction reaction of chromite. Assuming that each parameter has 10 different possible values, there will be 5 possible combinations of only 5 parameters. There are many different parameter combinations. The more parameters there are, the more possible combinations there are, which increases exponentially, resulting in an extremely large search space. However, most of these combinations may be invalid and fail to achieve the desired smelting effect. The initial solution of the successful case represents a set of parameter combinations that have been verified to achieve good results. For example, when adjusting the slag temperature, the temperature range corresponding to the successful case provides a reference, and a better temperature value can be directly found within this range without blindly searching in the entire possible temperature range.

[0049] Similarly, the failed historical acid slag smelting process with the smallest Euclidean distance from the current acid slag smelting process can be obtained, and its various production parameters are recorded as the constraints of the current acid slag smelting process in the parameter optimization model. Constraints refer to the process parameter combinations that did not achieve the expected goals in the historical production data.

[0050] It should be explained that in the parameter optimization model, constraints are used to limit the scope of the solution and ensure that the results meet the actual production requirements. Using the production parameters of the failed cases as constraints can avoid the model recommending unreasonable process parameters.

[0051] Specifically, the failure cases record situations where problems occur in production parameters within a specific range, representing solutions that are not feasible in the actual acid slag smelting process. By using these parameters as constraints, these invalid areas can be excluded from the solution space of the parameter optimization model, narrowing the search range. Therefore, when the parameter optimization model is looking for the optimal solution, it will not waste time and computing resources within the range of production parameters that are known to cause failures, thereby improving the search efficiency and reducing the possibility of recommending unreasonable parameters. For example, in a historical acid slag smelting process, when the furnace temperature was too high and the amount of alkaline oxide added was too low, the acid slag smelting effect did not meet the standards. Then, by setting constraints in the model to limit the furnace temperature to not exceed a certain value and requiring the amount of alkaline oxide added to not be lower than a certain lower limit, similar failures can be avoided.

[0052] The present invention calculates the Euclidean distance between the various production parameters in the current acid slag smelting process and the corresponding production parameters in the historical case library for similarity analysis, generates the initial solution and constraint conditions of the parameter optimization model, shortens the running time of the parameter optimization model, improves the model convergence speed, realizes the intelligent recommendation and dynamic intelligent adjustment of process parameters, and reduces the dependence on manual experience.

[0053] S3. Compare each production parameter with the corresponding recommended value to calculate the deviation and determine the control requirement of each production parameter. If it is a control requirement, execute S4; otherwise, execute S2.

[0054] As a preferred feasible embodiment, the specific method for determining the control requirements of each production parameter is as follows: subtracting each production parameter from the corresponding production parameter recommended value to obtain a deviation value between each production parameter and the corresponding production parameter recommended value.

[0055] If the deviation value between a certain production parameter and the corresponding production parameter recommended value is not 0, the control demand of the production parameter is recorded as the control demand required, otherwise, the control demand of the production parameter is recorded as the control demand not required, thereby obtaining the control demand of each production parameter.

[0056] S4. Use the automatic control system to regulate and control the corresponding process parameters of each production parameter, and adjust each production parameter to the corresponding recommended value.

[0057] As a preferred feasible embodiment, the specific control operations of adjusting each production parameter to the corresponding recommended value include: if the deviation value between the polar circle power density and the recommended value of the polar circle power density is greater than 0, the PLC executes an adjustment instruction for the electrode lifting device to move the electrode toward the edge of the furnace, increasing the polar circle diameter, thereby achieving the purpose of reducing the polar circle power density; if the deviation value between the polar circle power density and the recommended value of the polar circle power density is less than 0, the PLC executes an adjustment instruction for the electrode lifting device to move the electrode toward the center of the furnace, reducing the polar circle diameter, thereby achieving the purpose of increasing the polar circle power density.

[0058] If the deviation between the charge conductivity and the recommended value is greater than 0, the PLC will issue an adjustment instruction to the solvent addition system to reduce the amount of alkaline oxide solvent added, so as to achieve the purpose of reducing the charge conductivity. If the deviation between the charge conductivity and the recommended value is less than 0, the PLC will issue an adjustment instruction to the solvent addition system to increase the amount of alkaline oxide solvent added, so as to achieve the purpose of improving the charge conductivity.

[0059] If the deviation between the chromium content in the slag and the recommended value of the chromium content in the slag is not 0, the cause of the deviation of the chromium content in the slag is further determined, and combined with the basic principles of safety first, efficiency second, and economy supplemented, the automatic control system is used for regulation and operation.

[0060] If the deviation between the furnace wall thickness and the recommended value is greater than 0, the DCS will issue a voltage reduction instruction to the transformer to reduce the power entering the furnace, thereby increasing the furnace diameter in disguise and reducing the erosion of the furnace wall.

[0061] If the deviation between the silica content and the recommended silica content is greater than 0, the DCS will execute an adjustment instruction to reduce the silica content on the solvent addition system to achieve the purpose of reducing the silica content. If the deviation between the silica content and the recommended silica content is less than 0, the DCS will execute an adjustment instruction to increase the silica content on the solvent addition system to achieve the purpose of increasing the silica content.

[0062] As a preferred feasible embodiment, the specific method for determining the cause of the deviation of the chromium content in the slag is: obtaining the reduction degree and the slag-iron separation efficiency in the current acidic slag smelting process, and comparing them with the preset reduction degree threshold and the slag-iron separation efficiency threshold respectively; if the reduction degree in the current acidic slag smelting process is less than the reduction degree threshold, the cause of the deviation of the chromium content in the slag is recorded as low reduction degree; if the slag-iron separation efficiency in the current acidic slag smelting process is less than the slag-iron separation efficiency threshold, the cause of the deviation of the chromium content in the slag is recorded as low slag-iron separation efficiency; if the reduction degree and the slag-iron separation efficiency in the current acidic slag smelting process are both less than the reduction degree threshold and the slag-iron separation efficiency threshold, the cause of the deviation of the chromium content in the slag is recorded as low reduction degree and low slag-iron separation efficiency.

[0063] It should be further explained that the specific method for obtaining the reduction degree and slag-iron separation efficiency in the current acid slag smelting process is: according to the calculation formula Get the reduction degree of the current acid slag smelting process ,in They are respectively the chromium content in the slag during the current acid slag smelting process and the theoretical iron content in the raw materials.

[0064] The conductivity of the separated iron in the current acid slag smelting process is detected by using a conductivity meter to obtain the conductivity of the separated iron in the current acid slag smelting process, which is then compared with the standard conductivity of the separated iron, and the result obtained is used as the slag-iron separation efficiency in the current acid slag smelting process.

[0065] The specific method for obtaining the preset reduction degree threshold and slag-iron separation efficiency threshold is: extracting the reduction degree and slag-iron separation efficiency in each acid slag smelting process labeled as a successful case from the historical case library, averaging them to obtain the average reduction degree and average slag-iron separation efficiency in the acid slag smelting process labeled as a successful case, and using them as the preset reduction degree threshold and slag-iron separation efficiency threshold.

[0066] Specifically, the average reduction degree and average slag-iron separation efficiency of successful cases represent the qualified process level that can be obtained in the actual acid slag smelting process. Setting them as thresholds can provide clear quality targets for subsequent smelting processes.

[0067] As a preferred feasible embodiment, the specific method of controlling and operating by an automatic control system in combination with the basic priority principles of safety first, efficiency second, and economy supplemented includes: if the cause of the deviation of the chromium content in the slag is the low degree of reduction, then obtaining the data type of charge thickness, coke particle size, current-voltage ratio and chromium ore particle size corresponding to the chromium content in the slag, and sorting them according to the basic priority principles of safety first, efficiency second, and economy supplemented according to the data type, to obtain the adjustment priority of the process parameters corresponding to the chromium content in the slag.

[0068] It needs to be explained that the adjustment priority of the process parameters corresponding to the chromium content in the slag is obtained by sorting the chromium content in the slag corresponding to the charge thickness, coke particle size, current-voltage ratio and chromium ore particle size in the order of safety data, efficiency data and economic data in the data type, while there is no safety data and economic data for the chromium content in the slag corresponding to the charge thickness, coke particle size, current-voltage ratio and chromium ore particle size, and therefore there is no order of priority in the adjustment control of the chromium content in the slag corresponding to the charge thickness, coke particle size, current-voltage ratio and chromium ore particle size.

[0069] Specifically, if the deviation between the chromium content in the slag and the recommended value of the chromium content in the slag is greater than 0, the DCS will execute an adjustment instruction to increase the thickness of the charge or reduce the particle size of the coke or reduce the particle size of the chromium ore on the automatic batching system, or the DCS will execute an adjustment instruction to increase the current-voltage ratio on the transformer to achieve the purpose of improving the degree of reduction.

[0070] It should be further explained that the increase in the thickness of the material layer helps to increase the solid-solid reduction stroke and strengthen the upper reduction. The reduction in the coke particle size increases the specific surface area to increase the coke activity and the reaction contact area, thereby improving the upper solid-solid reduction capacity, while ensuring that there is enough coke layer at the bottom of the furnace to meet the liquid-solid reduction requirements, and the coke is matched with large and small particle sizes. The reduction in the chromium ore particle size increases the contact area and improves the reduction conditions. The increase in the current-voltage ratio helps to increase the electrode insertion depth and the material surface temperature, improve the chemical reaction rate, and shorten the smelting cycle.

[0071] If the cause of the deviation in the chromium content in the slag is the low efficiency of slag-iron separation, the data type of the calcium oxide content in the slag, the current-voltage ratio, the number of times the furnace is discharged and the refining time corresponding to the chromium content in the slag is obtained, and the data type is sorted according to the basic principle of safety first, efficiency second and economy supplemented, so as to obtain the priority of adjusting the process parameters corresponding to the chromium content in the slag.

[0072] It needs to be explained that the adjustment priority of the process parameters corresponding to the chromium content in the slag is obtained by sorting the chromium content in the slag corresponding to the calcium oxide content in the slag, the current-voltage ratio, the number of times out of the furnace and the refining time in the order of safety data, efficiency data and economic data in the data type, while there is no safety data in the chromium content in the slag corresponding to the calcium oxide content in the slag, the current-voltage ratio, the number of times out of the furnace and the refining time, the efficiency data include the calcium oxide content in the slag, the current-voltage ratio and the refining time, and the economic data include the number of times out of the furnace, and then in the adjustment control of the chromium content in the slag corresponding to the calcium oxide content in the slag, the current-voltage ratio, the number of times out of the furnace and the refining time, the calcium oxide content, the current-voltage ratio and the refining time are adjusted first, and then the number of times out of the furnace is adjusted, and there is no order of priority in the adjustment control of the calcium oxide content, the current-voltage ratio and the refining time.

[0073] Specifically, if the deviation between the chromium content in the slag and the recommended value of the chromium content in the slag is greater than 0, the DCS executes an adjustment instruction to add calcium oxide in the slag to the solvent addition system or to increase the current-voltage ratio to the transformer, or the PLC executes an adjustment instruction to extend the refining time to the furnace control system, so as to achieve the purpose of improving the slag-iron separation efficiency. If the deviation between the chromium content in the slag and the recommended value of the chromium content in the slag is still greater than 0 after the above operations are completed, the PLC executes an adjustment instruction to reduce the number of furnace discharges to the furnace control system, so as to achieve the purpose of improving the slag-iron separation efficiency.

[0074] It should be further explained that the increase in the CaO content in the slag to about 5% reduces the viscosity of the slag, thereby facilitating the separation of slag and iron. The increase in the current-voltage ratio can adjust the slag-iron temperature, reduce the viscosity of the slag-iron, increase the fluidity of the slag-iron, and fully separate the slag-iron, while increasing the electrode insertion depth, improving the furnace bottom temperature and the fluidity of the slag-iron. The reduction in the number of times the furnace is discharged can keep the smelting equipment in continuous operation for a longer period of time, avoiding the equipment start-up and shutdown process caused by frequent discharge. The extension of the refining time gives the chemical reaction more time to work.

[0075] The present invention determines the control requirements of various production parameters and adjusts the corresponding process parameters to the corresponding recommended values ​​according to the automatic control system, thereby achieving precise control of the charge ratio and the chromium content in the slag, improving the ferrochrome recovery rate and resource utilization, reducing the demand for new raw materials, reducing production costs, and reducing accidents of eyelet and furnace leakage, extending the service life of the smelting furnace, improving production safety, and thus achieving multi-objective collaborative optimization and cost control.

[0076] Embodiment 2

[0077] like Figure 4 As shown, in the second embodiment of the present invention, the embodiment of the present invention provides the following technical solution, a computer, including a memory 202, a processor 201, and a computer program stored in the memory 202 and executable on the processor 201, when the processor 201 executes the computer program, the method for predicting and analyzing the efficacy of shunt surgery for patients with hydrocephalus based on machine learning as described above is implemented.

[0078] Specifically, the processor 201 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0079] Among them, the memory 202 may include a large-capacity memory for data or instructions. For example, but not limitation, the memory 202 may include a hard disk drive (HDD), a floppy disk drive, a solid state drive (SSD), a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 202 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 202 may be inside or outside the data processing device. In a specific embodiment, the memory 202 is a non-volatile memory. In a specific embodiment, the memory 202 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0080] The memory 202 may be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 201 .

[0081] The processor 201 reads and executes the computer program instructions stored in the memory 202 to implement the above-mentioned method for predicting and analyzing the efficacy of shunt surgery for hydrocephalus patients based on machine learning.

[0082] In some embodiments, the computer may further include a communication interface 203 and a bus 200. Figure 4 As shown, the processor 201, the memory 202, and the communication interface 203 are connected via a bus 200 and communicate with each other.

[0083] The communication interface 203 is used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present application. The communication interface 203 can also implement data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage, and image / data processing workstations.

[0084] The bus 200 includes hardware, software or both, and couples the components of the computer to each other. The bus 200 includes but is not limited to at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example and not limitation, bus 200 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, bus 200 may include one or more buses. Although embodiments of the present application describe and illustrate a particular bus, the present application contemplates any suitable bus or interconnect.

[0085] The above contents are merely examples and explanations of the concept of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A method for controlling acidic slag smelting based on data analysis, characterized in that: include: S1. Build a historical case library based on the production parameters and corresponding process parameters generated by each acid slag smelting process collected during the test in the target smelting field, and build a parameter optimization model based on deep learning technology; S2. Perform similarity analysis on the collected production parameters in the current acid slag smelting process and the corresponding production parameters in the historical case library by calculating the Euclidean distance, determine the initial solution and constraint conditions of the current acid slag smelting process in the parameter optimization model, and run to obtain the recommended values ​​of each production parameter; S3, compare each production parameter with the corresponding recommended value to calculate the deviation, and judge the control demand of each production parameter. If it is a control demand, execute S4, otherwise, execute S2; S4. Use the automatic control system to regulate and control the corresponding process parameters of each production parameter, and adjust each production parameter to the corresponding recommended value.

2. The acid slag smelting control method based on data analysis according to claim 1, characterized in that: The specific construction process of the historical case library includes: A1. Label setting: Determine whether each acid slag smelting process in the target smelting field during the test reaches the preset expected target. If it reaches the expected target, set its label as a success case. Otherwise, set its label as a failure case. Set the label of each acid slag smelting process in the target smelting field during the test accordingly. A2. Data type matching: Match each production parameter and its corresponding process parameter generated by each acid slag smelting process collected during the test of the target smelting field after the label is set with each production parameter and each process parameter belonging to each data type stored in the information database, and obtain the data type of each production parameter and its corresponding process parameter generated by each acid slag smelting process collected during the test of the target smelting field after the label is set; A3. Construction of historical case library: The production parameters generated by each acid slag smelting process collected in the target smelting field during the test after data type matching and their corresponding process parameters are stored in the database after data association using timestamps and numbers. It is recorded as a historical case library and updated in real time.

3. The acid slag smelting control method based on data analysis according to claim 2, characterized in that: The preset expected targets include the CaO content in the slag, the improvement rate of chromium ore reducibility, and the probability of occurrence of eyelet and furnace leakage problems; The data types include safety data, efficiency data and economic data.

4. The acid slag smelting control method based on data analysis according to claim 2 is characterized in that: The parameter optimization model building process includes: B1. Dataset construction and division: Extract various production parameters generated by each acid slag smelting process collected during the test in the target smelting field from the historical case library, collectively referred to as the acid slag smelting parameter dataset, and divide it into the acid slag smelting parameter training set, test set and validation set; B2. Model architecture selection: select the network structure and build the model, and record the built model as the basic model for parameter optimization; B3. Define loss function and optimizer: Select the loss function and optimizer of the parameter optimization base model; B4. Model training: The acid slag smelting parameter training set is input into the parameter optimization basic model for training. During the training process, the parameters of the parameter optimization basic model are continuously adjusted to minimize the loss function until it is stable. The parameter optimization basic model corresponding to the minimized and stable loss function is recorded as the parameter optimization model.

5. The acid slag smelting control method based on data analysis according to claim 1, characterized in that: The specific method for determining the initial solution and constraint conditions of the parameter optimization model of the current acid slag smelting process is as follows: Extract various production parameters in the current acid slag smelting process , extract the corresponding production parameters generated by each acid slag smelting process labeled as a successful case from the historical case library, and record them as the production parameters of the current acid slag smelting process corresponding to the production parameters of each successful historical acid slag smelting process Similarly, the production parameters in the current acidic slag smelting process correspond to the production parameters of each failed historical acidic slag smelting process, where , is the change of each production parameter, , The serial number of each successful acid slag smelting in history; According to the current production parameters in the acid slag smelting process and production parameters of each successful historical acid slag smelting process The Euclidean distance between each successful historical acid slag smelting process and the current acid slag smelting process is obtained by fusion calculation , select the successful historical acid slag smelting process with the smallest Euclidean distance to the current acid slag smelting process, and record its various production parameters as the initial solution of the current acid slag smelting process in the parameter optimization model; Similarly, the failed historical acid slag smelting process with the smallest Euclidean distance from the current acid slag smelting process can be obtained, and its various production parameters are recorded as the constraints of the current acid slag smelting process in the parameter optimization model.

6. The acid slag smelting control method based on data analysis according to claim 3 is characterized in that: The specific determination method of the control requirements of each production parameter is as follows: Subtract each production parameter from the corresponding production parameter recommended value to obtain the deviation value between each production parameter and the corresponding production parameter recommended value; If the deviation value between a certain production parameter and the corresponding production parameter recommended value is not 0, the control demand of the production parameter is recorded as the control demand required, otherwise, the control demand of the production parameter is recorded as the control demand not required, thereby obtaining the control demand of each production parameter.

7. The acid slag smelting control method based on data analysis according to claim 6 is characterized by: The production parameters include polar circle power density, charge conductivity, chromium content in slag, furnace wall thickness and silicon dioxide content; The process parameters include the diameter of the polar circle, the content of basic oxides, the thickness of the charge, the particle size of coke, the current-voltage ratio, the particle size of chromium ore, the content of calcium oxide in the slag, the number of times out of the furnace, the refining time, the voltage and the silicon dioxide content.

8. The acid slag smelting control method based on data analysis according to claim 7 is characterized in that: The specific control operations for adjusting each production parameter to the corresponding recommended value include: If the deviation between the polar circle power density and the recommended value of the polar circle power density is greater than 0, the PLC executes an adjustment instruction for the electrode lifting device to move the electrode to the edge of the furnace; if the deviation between the polar circle power density and the recommended value of the polar circle power density is less than 0, the PLC executes an adjustment instruction for the electrode lifting device to move the electrode to the center of the furnace; If the deviation between the charge conductivity and the recommended value of charge conductivity is greater than 0, the PLC is used to execute an adjustment instruction for reducing the amount of alkaline oxide solvent added to the solvent adding system; if the deviation between the charge conductivity and the recommended value of charge conductivity is less than 0, the PLC is used to execute an adjustment instruction for increasing the amount of alkaline oxide solvent added to the solvent adding system; If the deviation between the chromium content in the slag and the recommended value of the chromium content in the slag is not 0, the cause of the deviation of the chromium content in the slag is further determined, and the automatic control system is used to control the operation in accordance with the basic principle of safety first, efficiency second, and economy supplementary priority; If the deviation between the furnace wall thickness and the recommended value of the furnace wall thickness is greater than 0, the DCS will issue an adjustment command to reduce the voltage to the transformer; If the deviation between the silica content and the recommended silica content is greater than 0, the DCS executes an adjustment instruction to reduce the silica content on the solvent addition system; if the deviation between the silica content and the recommended silica content is less than 0, the DCS executes an adjustment instruction to increase the silica content on the solvent addition system.

9. The acid slag smelting control method based on data analysis according to claim 8, characterized in that: The specific method for determining the cause of the deviation of the chromium content in the slag is: The reduction degree and slag-iron separation efficiency in the current acidic slag smelting process are obtained, and they are compared with the preset reduction degree threshold and slag-iron separation efficiency threshold respectively; if the reduction degree in the current acidic slag smelting process is less than the reduction degree threshold, the cause of the deviation of the chromium content in the slag is recorded as low reduction degree; if the slag-iron separation efficiency in the current acidic slag smelting process is less than the slag-iron separation efficiency threshold, the cause of the deviation of the chromium content in the slag is recorded as low slag-iron separation efficiency; if the reduction degree and slag-iron separation efficiency in the current acidic slag smelting process are both less than the reduction degree threshold and the slag-iron separation efficiency threshold, the cause of the deviation of the chromium content in the slag is recorded as low reduction degree and low slag-iron separation efficiency.

10. The acid slag smelting control method based on data analysis according to claim 9, characterized in that: The specific methods of controlling and operating by automatic control systems in accordance with the basic principle of safety first, efficiency second, and economy supplementary include: If the cause of the deviation of the chromium content in the slag is the low degree of reduction, the data types of the charge thickness, the coke particle size distribution ratio, the current-voltage ratio and the chromium ore particle size corresponding to the chromium content in the slag are obtained, and the data types are sorted according to the basic principle of safety first, efficiency second and economy supplement, to obtain the priority of adjusting the process parameters corresponding to the chromium content in the slag; If the reason for the deviation of chromium content in slag is low slag-iron separation efficiency, the data type of slag-iron temperature, calcium oxide content in slag, current-voltage ratio, number of furnace discharges and refining time corresponding to the chromium content in slag is obtained, and the data type is sorted according to the basic principle of safety first, efficiency second and economy supplementary, to obtain the priority of process parameter adjustment corresponding to the chromium content in slag.

Images