Pellet quality anomaly preanalysis and optimization method and device, medium and equipment

Abstract

The invention discloses a pellet quality anomaly preanalysis and optimization method and device, a medium and equipment, and relates to the technical field of pellet production. The method comprises the following steps: respectively acquiring pellet production data of a current time node and a historical time node; according to the pellet production data of the current time node and the historical time node, historical time sequence data of different pellet process key parameters are determined; the historical time sequence data of the key parameters of the different pelletizing processes are input into corresponding index prediction models for index prediction, prediction values of different pellet quality indexes are obtained, and the prediction values of the different pellet quality indexes comprise a total iron prediction value, a silicon dioxide prediction value, a compressive strength prediction value and a drum strength prediction value; and if it is determined that the pellet quality is abnormal according to the predicted values of the different pellet quality indexes, an operation suggestion matrix is constructed based on the operation parameters so as to optimize the pellet quality. According to the method, the quality of the pellets can be predicted and optimized in advance.

Description

Technical Field

[0001] This application relates to the field of pellet production technology, and in particular to a method, apparatus, medium and equipment for pre-analysis and optimization of pellet quality anomalies. Background Technology

[0002] Pelletizing is a crucial step in steel production. With the development of low-carbon smelting technologies, the application of high-quality pellets has become key to achieving low-carbon goals. Blast furnace ironmaking relies heavily on pellets; ensuring pellet quality not only affects smooth production but also relates to resource utilization and environmental protection objectives. Compared to sintering, vertical shaft furnace pelletizing consumes less energy, emits fewer emissions, and requires less equipment space, making it suitable for small-scale production. The good permeability of pellets not only helps improve fuel efficiency, reduce coke consumption, and lower carbon emissions but also helps optimize airflow distribution, increasing both the yield and quality of molten iron.

[0003] Currently, pellet quality is typically inspected and optimized manually. However, existing methods cannot meet the needs for real-time monitoring and adjustment of pellet quality. Detection delays often result in abnormal batches being discovered only after they have entered the next process. Consequently, quality problems in the pellets cannot be detected and corrected in a timely manner, affecting production efficiency and quality stability, and causing production losses. Summary of the Invention

[0004] In view of this, this application provides a method, apparatus, medium and equipment for pre-analysis and optimization of pellet quality anomalies, which mainly enables the early prediction and optimization of pellet quality and solves the problem of lag in pellet quality detection and correction.

[0005] According to a first aspect of this application, a method for preliminary analysis and optimization of pellet quality anomalies is provided, the method comprising: Obtain pellet production data for the current time point and historical time points respectively; Based on the pellet production data at the current time point and the pellet production data at the historical time points, determine the historical time series data corresponding to different key parameters of the pellet process. The historical time series data corresponding to the key parameters of the different pelleting process are input into the corresponding index prediction models to predict the indexes and obtain the predicted values ​​of different pellet quality indicators. The predicted values ​​of the different pellet quality indicators include the predicted values ​​of total iron, silica, compressive strength and drum strength. If a pellet quality anomaly is determined based on the predicted values ​​of the different pellet quality indicators, an operation suggestion matrix is ​​constructed based on the adjustable operation parameters. Based on the operation suggestion matrix, the quality of the pellets is optimized.

[0006] According to a second aspect of this application, a pellet quality anomaly pre-analysis and optimization device is provided, the device comprising: The acquisition unit is used to acquire pellet production data for the current time node and historical time nodes, respectively. The determining unit is used to determine the historical time series data corresponding to different key parameters of the pelleting process based on the pelleting production data at the current time node and the pelleting production data at the historical time nodes. The prediction unit is used to input the historical time series data corresponding to the key parameters of the different pelleting process into the corresponding index prediction model to predict the index and obtain the predicted values ​​of different pellet quality indicators. The predicted values ​​of the different pellet quality indicators include the predicted values ​​of total iron, silica, compressive strength and drum strength. The construction unit is used to construct an operation suggestion matrix based on adjustable operation parameters if it is determined that the pellet quality is abnormal according to the predicted values ​​of the different pellet quality indicators. An optimization unit is used to optimize the quality of the pellets based on the operation suggestion matrix.

[0007] According to a third aspect of this application, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described method for pre-analysis and optimization of pellet quality anomalies.

[0008] According to a fourth aspect of this application, an electronic device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-described pellet quality anomaly pre-analysis and optimization method.

[0009] By employing the above technical solution, this application provides a method, apparatus, medium, and equipment for pre-analysis and optimization of pellet quality anomalies. Compared with existing manual operation methods, it can determine the historical time-series data corresponding to different key parameters of the pelleting process based on the pelleting production data at the current time point and the pelleting production data at historical time points. The historical time-series data corresponding to different key parameters of the pelleting process are then input into corresponding indicator prediction models for indicator prediction, obtaining predicted values ​​for different pellet quality indicators. If a pellet quality anomaly is determined based on the predicted values ​​of different pellet quality indicators, an operation suggestion matrix is ​​constructed based on adjustable operation parameters. Finally, the pellet quality is optimized based on the operation suggestion matrix. Therefore, this application, by using an indicator prediction model to predict the predicted values ​​of different pellet quality indicators, can promptly detect pellet quality problems, thereby solving the problem of detection lag in existing technologies. Simultaneously, by constructing an operation suggestion matrix, this application can optimize pellet quality, thereby promptly resolving problems in the pelleting process, ensuring the stability of pellet quality, improving pelleting production efficiency, and avoiding production losses.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0011] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart illustrating a pellet quality anomaly pre-analysis and optimization method provided in an embodiment of this application is shown. Figure 2 A flowchart illustrating the indicator prediction model training method provided in an embodiment of this application is shown. Figure 3 A schematic diagram of a pellet quality anomaly pre-analysis and optimization device provided in an embodiment of this application is shown. Detailed Implementation

[0012] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0013] The existing manual operation methods cannot meet the needs of real-time monitoring and adjustment of pellet quality. The detection is lagging, and abnormal batches are often only discovered after entering the next process. As a result, the quality problems of pellets cannot be detected and corrected in a timely manner, which affects the production efficiency and quality stability of pellets and causes production losses.

[0014] To address the aforementioned problems, embodiments of the present invention provide a method for pre-analysis and optimization of pellet quality anomalies, such as... Figure 1 As shown, the method includes: Step 10: Obtain pellet production data for the current time node and historical time nodes respectively.

[0015] The pellet production data includes multiple parameters, such as raw material data from the material yard, mixing ratio data, green pellet quality data, and vertical shaft furnace operation data. The raw material data from the material yard specifically includes the chemical composition of the raw materials, such as TFe, FeO, CaO, SiO2, MgO, Al2O3, S, and P. The green pellet quality data specifically includes the green pellet compressive strength and green pellet drop strength. The vertical shaft furnace operation data specifically includes 60 parameters, such as the vertical shaft furnace roasting temperature, roasting zone temperature, steam drum pressure, guide wall inlet temperature, total gas accumulation, drying bed temperature, fume hood temperature, air flow rate, air main pressure, combustion chamber pressure, combustion chamber temperature, cold air pressure, cold air flow rate, gas flow rate, cooling zone temperature, and guide wall outlet temperature.

[0016] This invention is primarily applicable to scenarios involving pre-analysis and optimization of pellet quality anomalies. The executing entity of this invention is a device or equipment capable of performing pre-analysis and optimization of pellet quality anomalies, such as one mounted on a server.

[0017] In the actual production scenario of the factory, a data acquisition system deployed on-site is used to monitor and collect data on the entire pellet production process in real time. This allows for the acquisition of pellet production data at different time points. Since this embodiment of the invention performs pre-analysis of pellet quality, i.e., advance prediction, it is necessary to acquire pellet production data at the current time point and pellet production data at historical time points. Based on this data, the pellet quality at the next time point is predicted. For example, if the current time point is 11:00 AM, during data acquisition, it is necessary not only to acquire the pellet production data at 11:00 AM, but also the pellet production data before 11:00 AM, i.e., to acquire the pellet production data at 10:45 AM, 10:30 AM, 10:15 AM, and 10:00 AM respectively.

[0018] Step 20: Based on the pellet production data at the current time node and the pellet production data at the historical time nodes, determine the historical time series data corresponding to different key parameters of the pellet process.

[0019] The key parameters for different pelletizing processes include those for the first, second, third, and fourth pelletizing processes. The first pelletizing process key parameter is the one associated with the predicted total iron (TFe) value among all parameters involved in the pelletizing process; the second pelletizing process key parameter is the one associated with the predicted silica (SiO2) value among all parameters involved in the pelletizing process; the third pelletizing process key parameter is the one associated with the predicted pellet compressive strength among all parameters involved in the pelletizing process; and the fourth pelletizing process key parameter is the one associated with the predicted pellet drum strength among all parameters involved in the pelletizing process. Furthermore, the historical time-series data includes key parameter data for the current time point and key parameter data for historical time points.

[0020] In this embodiment of the invention, based on the pellet production data at the current time node and the pellet production data at historical time nodes, the historical time series data corresponding to the first pellet process key parameter, the second pellet process key parameter, the third pellet process key parameter and the fourth pellet process key parameter are determined respectively.

[0021] For example, the key parameters of the first pelletizing process (the parameters related to the predicted total iron (TFe) among all parameters involved in the pelletizing process) include the TFe ratio of the raw material chemical composition, the shaft furnace roasting temperature, the roasting zone temperature, the cold blast flow rate, and the gas flow rate. Based on the TFe ratio, shaft furnace roasting temperature, roasting zone temperature, cold blast flow rate, and gas flow rate at the current and historical time points, the historical time-series data for the key parameters of the first pelletizing process can be determined. Similarly, the historical time-series data corresponding to the key parameters of the second, third, and fourth pelletizing processes can be determined separately.

[0022] Step 30: Input the historical time series data corresponding to the key parameters of the different pelleting processes into the corresponding index prediction models to predict the indexes and obtain the predicted values ​​of different pellet quality indicators.

[0023] The predicted values ​​of the different pellet quality indicators include the predicted values ​​of total iron, silica, compressive strength, and drum strength.

[0024] In this embodiment of the invention, historical time-series data corresponding to the key parameters of the first pelleting process are input into a preset total iron prediction model for prediction, to obtain the predicted total iron value of the pellet; historical time-series data corresponding to the key parameters of the second pelleting process are input into a preset silica prediction model for prediction, to obtain the predicted silica value of the pellet; historical time-series data corresponding to the key parameters of the third pelleting process are input into a preset pellet compressive strength prediction model for prediction, to obtain the predicted compressive strength value of the pellet; and historical time-series data corresponding to the key parameters of the fourth pelleting process are input into a preset pellet drum strength prediction model for prediction, to obtain the predicted drum strength value of the pellet. Thus, through the above-mentioned index prediction models, the predicted total iron value, predicted silica value, predicted pellet compressive strength value, and predicted pellet drum strength value at the current time point can be predicted respectively.

[0025] Specifically, the preset all-iron prediction model, the preset silica prediction model, the preset pellet compressive strength prediction model, and the preset pellet drum strength prediction model can be the Adaboost model, the XGboost model, or other prediction models. This embodiment of the invention does not specifically limit these models.

[0026] To further improve the prediction accuracy of pellet quality index prediction values, this embodiment of the invention will correct the predicted pellet quality index values. Based on this, the method further includes: determining the prediction error of different pellet quality indexes at the previous time node corresponding to the current time node; and compensating the predicted values ​​of different pellet quality indexes at the current time node based on the prediction error to obtain the compensated predicted values ​​of different pellet quality indexes.

[0027] Since the predicted and actual values ​​of different pellet quality indicators at the previous time point are known, the prediction error of different pellet quality indicators at the previous time point can be calculated. Then, the prediction error of the previous time point is used to compensate for the prediction value of different pellet quality indicators at the current time point. That is, the predicted value of different pellet quality indicators after compensation is equal to the sum of the prediction error of the previous time point and the prediction value of different pellet quality indicators at the current time point.

[0028] Specifically, the total iron prediction value at the current time point is compensated based on the total iron prediction error at the previous time point, resulting in a compensated total iron prediction value. Simultaneously, the silica prediction value at the current time point is compensated based on the silica prediction error at the previous time point, resulting in a compensated silica prediction value. Similarly, the compressive strength prediction value at the current time point is compensated based on the compressive strength prediction error at the previous time point, resulting in a compensated compressive strength prediction value. Likewise, the drum strength prediction value at the current time point is compensated based on the drum strength prediction error at the previous time point, resulting in a compensated drum strength prediction value.

[0029] The embodiments of the present invention employ a real-time correction mechanism, which uses the prediction error from the previous moment for real-time monitoring and correction, and uses it as a correction factor for the current time prediction value to dynamically correct the prediction results, thereby improving the accuracy and stability of indicator prediction.

[0030] Step 40: If the pellet quality is determined to be abnormal based on the predicted values ​​of the different pellet quality indicators, then an operation suggestion matrix is ​​constructed based on the adjustable operation parameters.

[0031] The adjustable operating parameters include cold air flow rate, cold air pressure, south combustion chamber pressure, and north combustion chamber pressure.

[0032] In this embodiment of the invention, if an abnormality in pellet quality is determined based on the predicted values ​​after compensation for different pellet quality indicators, an operation suggestion matrix is ​​constructed based on adjustable operating parameters. When analyzing and optimizing pellet quality, if any one of the predicted values ​​of total iron, silica, compressive strength, and drum strength exceeds the corresponding preset index range, the abnormality in pellet quality is determined. The cold air flow rate parameter, cold air pressure parameter, south combustion chamber pressure parameter, and north combustion chamber pressure parameter are used as the operating parameters, and an operation suggestion matrix is ​​constructed based on the operating range and adjustment step size of the operating parameters.

[0033] The preset indicator range can be set according to actual business needs.

[0034] Specifically, criteria for judging abnormal pellet quality can be set based on expert experience. For example, a predicted compressive strength below 2400N or a predicted drum strength below 90% can be used as a condition to trigger abnormal pellet quality. The criteria for judging abnormalities in the predicted silica and total iron values ​​are determined based on the specific on-site production conditions and the differences in raw material types.

[0035] After determining the abnormality of the pellet quality based on the predicted values ​​of total iron, silica, compressive strength, and drum strength, an operation suggestion matrix is ​​constructed based on the step size and operating range corresponding to the cold air flow rate, cold air pressure, south combustion chamber pressure, and north combustion chamber pressure parameters, respectively. This operation suggestion matrix records multiple operation suggestions, each of which includes a suggested correction value for the cold air flow rate, cold air pressure, south combustion chamber pressure, and north combustion chamber pressure parameters.

[0036] For example, if the predicted compressive strength value for the next time point output by the preset pellet compressive strength prediction model is 2383N, and since it is less than 2400N, the pellet quality is determined to be abnormal. In this case, the adjustment step size of the cold air flow rate parameter is determined to be 500m. 3 The adjustment step for the cold air pressure parameter is 0.25 kp, the adjustment step for the south combustion chamber pressure parameter is 0.5 kp, the adjustment step for the north combustion chamber pressure parameter is 0.5 kp, and the operating range for the cold air flow parameter is ±1000 m³ / s. 3 The operating range for the cold air pressure parameter is ±0.5 kp, the operating range for the south combustion chamber pressure parameter is ±1 kp, and the operating range for the north combustion chamber pressure parameter is ±1 kp. Then, based on the adjustment step size and operating range of the cold air flow rate parameter, cold air pressure parameter, south combustion chamber pressure parameter, and north combustion chamber pressure parameter, while keeping other parameters constant, multiple operating suggestions with different correction suggestions for the cold air flow rate parameter, cold air pressure parameter, south combustion chamber pressure parameter, and north combustion chamber pressure parameter are generated, thus generating an operating suggestion matrix.

[0037] Furthermore, to ensure the accuracy of the operational parameters, this embodiment of the invention designs a method for determining operational parameters from numerous parameters. Specifically, firstly, several common parameters are identified among the key parameters of the first, second, third, and fourth pelleting processes. Then, the Recursive Feature Elimination (RFE) algorithm is used to filter out the key parameters that have a significant impact on pellet quality from the multiple common parameters, and these are used as operational parameters. Specifically, sample datasets of multiple common parameters and different pellet quality indicators are collected. For each pellet quality indicator, a parameter evaluation model is constructed. This parameter evaluation model can specifically be a random forest model. For each parameter evaluation model, the parameter with the lowest weight is eliminated each time, and this process is repeated until a predetermined number of key parameters remain, which are then output. Finally, the operational parameters are determined based on the key parameters output by each parameter evaluation model; for example, the intersection of the key parameters output by each parameter evaluation model can be taken. This allows for accurate determination of the operational parameters, and by adjusting these operational parameters, pellet quality can be optimized.

[0038] Step 50: Optimize the quality of the pellets based on the operation suggestion matrix.

[0039] In this embodiment of the invention, after constructing the operation suggestion matrix, the pellet quality is optimized based on the matrix. Specifically, the optimization result of each operation suggestion in the operation suggestion matrix is ​​predicted using the indicator prediction model; based on the optimization result of each operation suggestion, a preliminary operation optimization suggestion is output to optimize the pellet quality.

[0040] Specifically, the corrected values ​​of the cold air flow rate, cold air pressure, south combustion chamber pressure, and north combustion chamber pressure parameters in each operational suggestion matrix are used as the parameter values ​​for the current time node, while other parameters remain unchanged. Then, the index prediction model is used to predict the total iron, silica, compressive strength, and drum strength prediction values ​​corresponding to each operational suggestion, thus obtaining the optimization result for each operational suggestion. If the predicted values ​​of total iron, silica, compressive strength, and drum strength all improve in the optimization result, indicating an improvement in pellet quality, then the operational suggestion can be adopted, and operational suggestions that can optimize pellet quality indicators and improve pellet quality are output. If multiple operational suggestions can optimize pellet quality indicators and improve pellet quality, they can be ranked according to the magnitude of the optimization effect, and the target operational suggestions within a preset ranking range are output based on the ranking result, such as outputting the top 20 operational suggestions.

[0041] This invention provides a method for pre-analysis and optimization of pellet quality anomalies. Based on process data (raw materials, green pellet quality, and shaft furnace operating parameters, etc.), it predicts and optimizes pellet quality, combining expert experience and data analysis to dynamically analyze pellet quality. This method is highly adaptable and interpretable. Furthermore, this invention eliminates the need for additional testing equipment; it can accurately identify abnormal pellet quality using existing production line data, resulting in lower costs, strong generalization ability in indicator prediction, and high real-time performance. Moreover, this invention optimizes pellet quality based on an indicator prediction model, providing operational suggestions that help on-site operators make quick decisions, improve production stability, and reduce costs.

[0042] Furthermore, embodiments of the present invention also provide a training method for an indicator prediction model, such as... Figure 2 As shown, it includes: Step 60: Obtain sample pellet production data, and determine the sample historical time series data corresponding to each pellet process parameter based on the sample pellet production data.

[0043] The sample pellet production data includes process parameter data for each pellet at different time points. These process parameter data include raw material data from the material yard, mixing ratio data, green pellet quality data, and vertical shaft furnace operation data. The raw material data from the material yard specifically includes the chemical composition of the raw materials, such as TFe, FeO, CaO, SiO2, MgO, Al2O3, S, and P. The green pellet quality data specifically includes the green pellet compressive strength and green pellet drop strength. The vertical shaft furnace operation data specifically includes 60 vertical shaft furnace parameters, such as roasting temperature, roasting zone temperature, steam drum pressure, guide wall inlet temperature, total gas accumulation, drying bed temperature, fume hood temperature, air flow rate, air main pressure, combustion chamber pressure, combustion chamber temperature, cold air pressure, cold air flow rate, gas flow rate, cooling zone temperature, and guide wall outlet temperature.

[0044] In this embodiment of the invention, a data acquisition system deployed on-site monitors and collects raw pellet production data and raw indicator data in real time, thereby obtaining process parameter data (sample pellet production data) and indicator data for each pellet at different time points. Then, based on the process parameter data for each pellet at different time points, historical time-series data corresponding to each pellet process parameter is generated. Specifically, the indicator data includes total iron content data, total iron content data, pellet compressive strength, and drum strength at different time points.

[0045] In this embodiment of the invention, to ensure data quality when acquiring sample pellet production data and indicator data, it is necessary to perform data cleaning on the collected raw pellet production data and raw indicator data to obtain sample data. Based on this, the method includes: acquiring raw pellet production data; performing anomaly detection and missing data imputation on the raw pellet production data to obtain the sample pellet production data, wherein, when performing missing data imputation, different missing data imputation strategies are adopted for different missing proportions of the raw pellet production data.

[0046] Specifically, after acquiring the original pellet production data and original indicator data, this embodiment of the invention employs a multi-dimensional outlier detection mechanism to ensure data quality. First, outliers are removed using the intelligent identification capabilities of the Isolation Forest algorithm, ensuring high accuracy when processing high-dimensional data. Then, box plot technology is applied for outlier identification and removal. Finally, the domain knowledge of process experts is combined to precisely locate the root cause of the outlier data from both data characteristics and production process perspectives. In addition to outlier detection in pellet production data, this embodiment of the invention can also impute missing data. When imputes missing data, different missing data impute strategies can be adopted based on different data missing percentages. For example, for a certain pellet process parameter in the original pellet production data, if the missing data percentage is below 5%, the average or mode is used for impute based on the normal or skewed distribution of the data; if the missing data percentage is between 5% and 25%, machine learning methods are used for impute; if the missing data percentage is above 25%, this parameter data is discarded. At the same time, the data of various parameters are unified in terms of time frequency in order to complete the data cleaning of the original pellet production data and original indicator data.

[0047] Step 70: Based on the historical time-series data of the samples corresponding to each pellet process parameter, perform correlation analysis on each pellet process parameter to determine the key pellet process parameters associated with the different pellet quality indicators.

[0048] Among them, the key parameters of the pelleting process associated with different pellet quality indicators include the first, second, third, and fourth key parameters of the pelleting process. The first key parameter of the pelleting process is the parameter associated with the predicted value of total iron (TFe) among all the pelleting process parameters. The second key parameter of the pelleting process is the parameter associated with the predicted value of silicon dioxide (SiO2) among all the pelleting process parameters. The third key parameter of the pelleting process is the parameter associated with the predicted value of pellet compressive strength among all the pelleting process parameters. The fourth key parameter of the pelleting process is the parameter associated with the predicted value of pellet drum strength among all the pelleting process parameters.

[0049] In this embodiment of the invention, after obtaining the historical time-series data of samples corresponding to each pelletizing process parameter, key parameters related to pellet quality are obtained through correlation analysis based on this data. Specifically, the historical time-series data of samples corresponding to each pelletizing process parameter is analyzed in depth to ensure comprehensive and accurate production process information is obtained. This data covers all stages of pelletizing production. Based on this, and combined with on-site operational principles, the pelletizing process status quality parameters that are of concern to on-site operators are extracted. These parameters are typically key parameters in the production process.

[0050] Furthermore, Pearson, Spearman, random forest, and information entropy algorithms are employed to conduct correlation analysis from both linear and nonlinear perspectives, deeply exploring key parameters in the pelletizing process that are closely related to pellet quality. The correlation analysis in this embodiment not only focuses on traditional process parameters but also introduces multiple process state parameters, such as cold air flow rate, cold air pressure, drying bed temperature, and combustion chamber pressure. These parameters have a significant impact on the final quality of the pellets because they directly participate in the control of key aspects such as the thermodynamic process, chemical reaction, and gas flow in the shaft furnace. By analyzing the relationship between the above parameters and pellet quality indicators, it is possible to identify which parameters have a strong correlation with pellet quality, thereby helping operators quickly identify potential quality fluctuations and anomalies during production.

[0051] Step 80: Determine the sample training set based on the historical time-series data of the samples corresponding to the key parameters of the pelleting process associated with the different pelleting quality indicators.

[0052] In this embodiment of the invention, the historical time-series data of samples corresponding to key parameters of the pelleting process associated with different pellet quality indicators are divided into a test set and a training set. Specifically, the historical time-series data of samples corresponding to the first key parameter of the pelleting process are divided into a first test set and a first training set; the historical time-series data of samples corresponding to the second key parameter of the pelleting process are divided into a second test set and a second training set; the historical time-series data of samples corresponding to the third key parameter of the pelleting process are divided into a third test set and a third training set; and the historical time-series data of samples corresponding to the fourth key parameter of the pelleting process are divided into a fourth test set and a fourth training set.

[0053] Step 90: Based on the sample training set, train different indicator prediction models respectively.

[0054] The different index prediction models include a preset total iron prediction model, a preset silica prediction model, a preset pellet compressive strength prediction model, and a preset pellet drum strength prediction model. Specifically, the preset total iron prediction model, the preset silica prediction model, the preset pellet compressive strength prediction model, and the preset pellet drum strength prediction model can be the Adaboost model, the XGboost model, or other prediction models; this embodiment of the invention does not specifically limit these models.

[0055] In this embodiment of the invention, a preset prediction model for total iron, a preset prediction model for silica, a preset prediction model for pellet compressive strength, and a preset prediction model for pellet drum strength are trained based on a first training set, a second training set, a third training set, and a fourth training set, respectively. After training is completed, the model can be validated on a test set, and the performance of the prediction model can be evaluated by comparing the difference between the predicted values ​​and the actual values.

[0056] Furthermore, as Figure 1 and Figure 2 The specific implementation of the method shown in this embodiment provides a pre-analysis and optimization device for pellet quality anomalies, such as... Figure 3 As shown, the device includes: an acquisition unit 101, a determination unit 102, a prediction unit 103, a construction unit 104, and an optimization unit 105.

[0057] The acquisition unit 101 can be used to acquire pellet production data at the current time node and historical time nodes, respectively.

[0058] The determining unit 102 can be used to determine the historical time series data corresponding to different key parameters of the pelleting process based on the pelleting production data at the current time node and the pelleting production data at the historical time node.

[0059] The prediction unit 103 can be used to input the historical time series data corresponding to the key parameters of the different pelleting process into the corresponding index prediction model to predict the index and obtain the predicted values ​​of different pellet quality indicators. The predicted values ​​of the different pellet quality indicators include the predicted values ​​of total iron, silica, compressive strength and drum strength.

[0060] The construction unit 104 can be used to construct an operation suggestion matrix based on adjustable operation parameters if it is determined that the pellet quality is abnormal according to the predicted values ​​of the different pellet quality indicators.

[0061] The optimization unit 105 can be used to optimize the quality of the pellets based on the operation suggestion matrix.

[0062] In some embodiments, when different pelleting process key parameters include a first pelleting process key parameter, a second pelleting process key parameter, a third pelleting process key parameter, and a fourth pelleting process key parameter, the determining unit 102 may be specifically used to determine the historical time series data corresponding to the first pelleting process key parameter, the second pelleting process key parameter, the third pelleting process key parameter, and the fourth pelleting process key parameter based on the pelleting production data at the current time node and the pelleting production data at the historical time node.

[0063] In some embodiments, the prediction unit 103 may be specifically used to input historical time-series data corresponding to the key parameters of the first pelleting process into a preset total iron prediction model for prediction, to obtain the predicted total iron value of the pellet; input historical time-series data corresponding to the key parameters of the second pelleting process into a preset silica prediction model for prediction, to obtain the predicted silica value of the pellet; input historical time-series data corresponding to the key parameters of the third pelleting process into a preset pellet compressive strength prediction model for prediction, to obtain the predicted compressive strength value of the pellet; and input historical time-series data corresponding to the key parameters of the fourth pelleting process into a preset pellet drum strength prediction model for prediction, to obtain the predicted drum strength value of the pellet.

[0064] In some embodiments, the apparatus further includes a compensation unit.

[0065] The compensation unit is used to determine the prediction error of different pellet quality indicators of the previous time node corresponding to the current time node; and to compensate the predicted values ​​of different pellet quality indicators of the current time node based on the prediction error, so as to obtain the compensated predicted values ​​of different pellet quality indicators.

[0066] The construction unit 104 can be specifically used to construct an operation suggestion matrix based on adjustable operation parameters if the predicted value after compensation based on the different pellet quality indicators is determined to be abnormal.

[0067] In some embodiments, the construction unit 104 can also be specifically used to determine that the pellet quality is abnormal if any one of the predicted values ​​of the total iron, the silica, the compressive strength, and the drum strength exceeds the corresponding preset index range; and to construct an operation suggestion matrix based on the operation range and adjustment step size of the operation parameters, using the cold air flow rate parameter, the cold air pressure parameter, the south combustion chamber pressure parameter, and the north combustion chamber pressure parameter as the operation parameters.

[0068] In some embodiments, the optimization unit 105 may be specifically used to predict the optimization result of each operation suggestion in the operation suggestion matrix using the index prediction model; and output a preliminary operation optimization suggestion based on the optimization result of each operation suggestion to optimize the quality of the pellets.

[0069] In some embodiments, the apparatus further includes a construction unit.

[0070] The construction unit can be used to acquire sample pellet production data; determine the sample historical time-series data corresponding to each pellet process parameter based on the sample pellet production data; perform correlation analysis on each pellet process parameter based on the sample historical time-series data corresponding to each pellet process parameter, and determine the key pellet process parameters associated with the different pellet quality indicators respectively; determine the sample training set based on the sample historical time-series data corresponding to the key pellet process parameters associated with the different pellet quality indicators; and train different indicator prediction models based on the sample training set respectively.

[0071] The construction unit can be specifically used to acquire raw pellet production data; perform anomaly detection and missing data imputation on the raw pellet production data to obtain the sample pellet production data, wherein, when performing missing data imputation, different missing data imputation strategies are adopted for different missing proportions of the raw pellet production data.

[0072] It should be noted that other corresponding descriptions of the functional units involved in the pellet quality anomaly pre-analysis and optimization device provided in this embodiment can be found in [reference]. Figure 1 and Figure 2 The corresponding descriptions in [the document] will not be repeated here.

[0073] Based on the above, Figure 1 and Figure 2 Accordingly, this embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the above-described method. Figure 1 and Figure 2 The method for pre-analysis and optimization of pellet quality anomalies is shown.

[0074] Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause an electronic device (such as personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0075] Based on the above, Figure 1 and Figure 2 The method shown, and Figure 3To achieve the above objectives, the present application also provides an electronic device, specifically a personal computer, tablet computer, server, or other network device, as shown in the virtual device embodiment. This device includes a storage medium and a processor; the storage medium stores a computer program; the processor executes the computer program to achieve the above-described objectives. Figure 1 and Figure 2 The method for pre-analysis and optimization of pellet quality anomalies is shown.

[0076] Optionally, the aforementioned physical devices may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0077] Those skilled in the art will understand that the physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0078] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0079] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platform, or it can be implemented by hardware.

[0080] This invention employs an index prediction model to predict the values ​​of different pellet quality indicators, enabling timely detection of pellet quality problems and thus addressing the issue of lagging detection in existing technologies. Simultaneously, by constructing an operation suggestion matrix, this invention can optimize pellet quality, thereby promptly resolving issues during pellet generation, ensuring pellet quality stability, improving pellet production efficiency, and preventing production losses.

[0081] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application. Those skilled in the art will understand that the modules in the apparatus of the embodiment can be distributed within the apparatus of the embodiment as described, or can be modified to be located in one or more apparatuses different from this embodiment. The modules of the above-described embodiment can be combined into one module, or further divided into multiple sub-modules.

[0082] The serial numbers in this application are for descriptive purposes only and do not represent the superiority or inferiority of any particular implementation scenario. The above disclosures are merely a few specific implementation scenarios of this application; however, this application is not limited thereto, and any variations conceived by those skilled in the art should fall within the protection scope of this application.

Claims

1. A method for preliminary analysis and optimization of pellet quality anomalies, characterized in that, include: Obtain pellet production data for the current time point and historical time points respectively; Based on the pellet production data at the current time point and the pellet production data at the historical time points, determine the historical time series data corresponding to different key parameters of the pellet process. The historical time series data corresponding to the key parameters of the different pelleting process are input into the corresponding index prediction models to predict the indexes and obtain the predicted values ​​of different pellet quality indicators. The predicted values ​​of the different pellet quality indicators include the predicted values ​​of total iron, silica, compressive strength and drum strength. If a pellet quality anomaly is determined based on the predicted values ​​of the different pellet quality indicators, an operation suggestion matrix is ​​constructed based on the adjustable operation parameters. Based on the operation suggestion matrix, the quality of the pellets is optimized.

2. The method according to claim 1, characterized in that, When different key parameters of the pelletizing process include key parameters of the first, second, third, and fourth pelletizing processes, the step of determining the historical time-series data corresponding to the different key parameters of the pelletizing process based on the pelletizing production data at the current time point and the pelletizing production data at the historical time points includes: Based on the pellet production data at the current time node and the pellet production data at the historical time nodes, the historical time series data corresponding to the first pellet process key parameter, the second pellet process key parameter, the third pellet process key parameter and the fourth pellet process key parameter are determined respectively. The step of inputting historical time-series data corresponding to the key parameters of the different pelleting processes into the corresponding indicator prediction models to predict the indicators and obtain the predicted values ​​of different pellet quality indicators includes: The historical time series data corresponding to the key parameters of the first pelleting process are input into the preset total iron prediction model for prediction, and the total iron prediction value of the pellet is obtained. The historical time series data corresponding to the key parameters of the second pelleting process are input into a preset silica prediction model for prediction, and the silica prediction value of the pellets is obtained. The historical time series data corresponding to the key parameters of the third pelleting process are input into the preset pellet compressive strength prediction model for prediction, and the predicted value of the compressive strength of the pellet is obtained. The historical time series data corresponding to the key parameters of the fourth pellet process are input into the preset pellet drum strength prediction model for prediction, so as to obtain the predicted value of the pellet drum strength.

3. The method according to claim 1, characterized in that, After inputting the historical time-series data corresponding to the key parameters of the different pelleting processes into the corresponding indicator prediction models to predict the indicators and obtain the predicted values ​​of different pellet quality indicators, the method further includes: Determine the prediction error of different pellet quality indicators of the previous time point corresponding to the current time point; Based on the prediction error, the predicted values ​​of different pellet quality indicators at the current time point are compensated to obtain the compensated predicted values ​​of different pellet quality indicators. If a pellet quality anomaly is determined based on the predicted values ​​of the different pellet quality indicators, then an operation suggestion matrix is ​​constructed based on adjustable operation parameters, including: If a pellet quality anomaly is determined based on the predicted values ​​after compensation for the different pellet quality indicators, an operation suggestion matrix is ​​constructed based on the adjustable operation parameters.

4. The method according to claim 1, characterized in that, If a pellet quality anomaly is determined based on the predicted values ​​of the different pellet quality indicators, then an operation suggestion matrix is ​​constructed based on adjustable operation parameters, including: If any one of the predicted values ​​of total iron, silica, compressive strength, and drum strength exceeds the corresponding preset index range, then the pellet quality is determined to be abnormal. The cold air flow rate, cold air pressure, south combustion chamber pressure, and north combustion chamber pressure are used as the operating parameters, and an operating suggestion matrix is ​​constructed based on the operating range and adjustment step size of the operating parameters.

5. The method according to claim 1, characterized in that, The optimization of the pellet quality based on the operation suggestion matrix includes: The optimization result of each operation suggestion in the operation suggestion matrix is ​​predicted using the indicator prediction model. Based on the optimization results of each operation suggestion, an optimization suggestion for the preceding operation is output to optimize the quality of the pellets.

6. The method according to any one of claims 1-5, characterized in that, Before inputting the historical time-series data corresponding to the key parameters of the different pelleting processes into the corresponding indicator prediction models to predict the indicators and obtain the predicted values ​​of different pelleting quality indicators, the method further includes: Obtain sample pellet production data; Based on the sample pellet production data, determine the sample historical time series data corresponding to each pellet process parameter; Based on the historical time-series data of the samples corresponding to each pellet process parameter, a correlation analysis is performed on each pellet process parameter to determine the key pellet process parameters associated with the different pellet quality indicators. The sample training set is determined based on the historical time series data of the samples corresponding to the key parameters of the pelleting process associated with the different pelleting quality indicators. Based on the aforementioned sample training set, different indicator prediction models are trained respectively.

7. The method according to claim 6, characterized in that, The acquisition of sample pellet production data includes: Obtain raw pellet production data; Anomaly detection and missing data imputation are performed on the original pellet production data to obtain the sample pellet production data. When imputing missing data, different missing data imputation strategies are adopted for different missing proportions of the original pellet production data.

8. A device for pre-analysis and optimization of pellet quality anomalies, characterized in that, include: The acquisition unit is used to acquire pellet production data for the current time node and historical time nodes, respectively. The determining unit is used to determine the historical time series data corresponding to different key parameters of the pelleting process based on the pelleting production data at the current time node and the pelleting production data at the historical time nodes. The prediction unit is used to input the historical time series data corresponding to the key parameters of the different pelleting process into the corresponding index prediction model to predict the index and obtain the predicted values ​​of different pellet quality indicators. The predicted values ​​of the different pellet quality indicators include the predicted values ​​of total iron, silica, compressive strength and drum strength. The construction unit is used to construct an operation suggestion matrix based on adjustable operation parameters if it is determined that the pellet quality is abnormal according to the predicted values ​​of the different pellet quality indicators. An optimization unit is used to optimize the quality of the pellets based on the operation suggestion matrix.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

10. An electronic device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

Images