Intelligent regulation and control sintered ore quality improving method and system

Abstract

The invention belongs to the field of sinter, and particularly discloses an intelligent regulation and control sinter quality improvement method and system, and the method comprises the steps: S1, obtaining image data, production data and historical data in the sintering process of sinter, and integrating the production data and the historical data to obtain fused data; s2, predicting a quality index of the sintered ore according to the fused data through the prediction model to obtain a prediction result; s3, performing image recognition on the image data to obtain particle size distribution data; s4, dynamically adjusting a control strategy in combination with the prediction result and the particle size distribution data; according to the invention, the performance prediction model is established to predict the quality index of the sinter and obtain the particle size distribution data, and then the prediction result and the particle size distribution data are combined to dynamically adjust the control strategy and intelligently regulate and control the key parameters of the sinter quality, so that the precise control of the sinter production process is realized. And the stability of the sinter quality and the improvement of the production efficiency are ensured.

Description

Technical Field

[0001] The present invention belongs to the field of sintered ore, and specifically relates to a method and system for improving the quality of sintered ore with intelligent regulation and control. Background Art

[0002] Sintered ore is a semi-finished iron ore material formed by sintering iron ore, coking coal and other additives at high temperature; it is usually used as a raw material for iron smelting; the sintering process mainly forms a lump by heating and melting powdered iron ore particles, improving its burden performance in the blast furnace ironmaking process;

[0003] As an important raw material for blast furnace ironmaking, the quality of sintered ore directly affects the production efficiency and energy consumption of the blast furnace. However, during the traditional sintering process of sintered ore, manual monitoring is usually carried out by staff, and the sintering parameters are adjusted according to experience, with slow response speed and unsatisfactory adjustment accuracy. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for improving the quality of sintered ore with intelligent regulation and control, so as to solve the problem that in the traditional method, the sintering parameters are adjusted by staff according to experience, and the adjustment accuracy is not ideal.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method and system for improving the quality of sintered ore with intelligent regulation and control, including:

[0007] S1. Obtain image data, production data and historical data during the sintering process of sintered ore, and integrate the production data and the historical data to obtain integrated data;

[0008] S2. Predict the quality index of the sintered ore according to the integrated data through the prediction model to obtain a prediction result;

[0009] S3. Perform image recognition on the image data to obtain particle size distribution data;

[0010] S4. Dynamically adjust the control strategy by combining the prediction result and the particle size distribution data.

[0011] Preferably, in step S1, the historical data includes production data and quality indexes within a past period of time;

[0012] The image data is a visual image of the sintered ore mixture;

[0013] The production data is one of temperature, humidity, pressure, and flow rate during the sintering process.

[0014] Preferably, in step S1, the obtained real-time data and historical data are preprocessed, and the preprocessing includes data cleaning, feature extraction, and data standardization;

[0015] The data cleaning includes removing noise data and filling in missing values;

[0016] The feature extraction includes using principal component analysis to extract main features and combining with random forest for feature importance ranking to screen out features that have a greater impact on the quality index;

[0017] The data standardization includes performing standardization processing on the features with greater impact to make them have the same dimension, obtaining a standardized feature matrix.

[0018] Preferably, in step S2, the prediction model includes a DNN prediction model and an LSTM prediction model;

[0019] The DNN prediction model is used to process static features and predict the quality index at the current moment;

[0020] The LSTM prediction model is used to process time series data and predict the quality index at a future moment.

[0021] Preferably, in step S2, the following formula is used to predict the quality index at the current moment:

[0022]

[0023] Among them, is the predicted value of the quality index at the current moment, f is the DNN function, X scaled is the standardized feature matrix, and θ is the parameter of the DNN model.

[0024] Preferably, in step S2, the following formula is used to predict the quality index at a future moment:

[0025]

[0026] Among them, is the predicted value of the quality index at time step t, f is the LSTM function, X scaled，t is the standardized feature matrix at time step t, H t-1 is the hidden state of the previous time step, and θ is the parameter of the LSTM model.

[0027] Preferably, in step S3, the image data is used for image recognition to identify the particle size of the mixture, and the image recognition includes image preprocessing, feature extraction, classifier training, and particle size distribution calculation;

[0028] The image preprocessing is used to grayscale and denoise the image data to obtain processed data;

[0029] The feature extraction is used to extract the edge features and texture features of the processed data using a CNN;

[0030] The classifier training is used to train a classifier through KNN to classify different particle size mixtures and obtain classification results;

[0031] The particle size distribution calculation is used to calculate the proportion of different particle size mixtures based on the classification results to provide data support for particle size adjustment.

[0032] Preferably, step S4 outputs a control strategy through the following formula:

[0033]

[0034] where u(t) is the control strategy, e(t) is the prediction error, is the predicted value of the prediction model, yt is the target value, and K p 、K i 、K d are the proportional, integral, and derivative gains in sequence.

[0035] Preferably, step S4 also adjusts the calculation parameters through fuzzy logic control, specifically including:

[0036] Fuzzify the prediction error and the error change rate and convert them into fuzzy variables;

[0037] Define a fuzzy rule base, which is used to adjust the parameters of the proportional, integral, and derivative gains according to different prediction errors and error change rates;

[0038] Use a fuzzy inference mechanism to adjust the calculation parameters according to the fuzzy rule base;

[0039] Convert the fuzzy variables back into actual control parameters.

[0040] The present invention also provides an intelligent regulation and control system for improving the quality of sintered ore, including:

[0041] A data acquisition module, which is used to obtain image data, production data, and historical data during the sintering process of sintered ore, and integrate the production data and the historical data to obtain fused data;

[0042] A model prediction module, which is used to predict the quality index of the sintered ore according to the fused data through the prediction model to obtain a prediction result;

[0043] An image recognition module, which is used to perform image recognition on the image data to obtain particle size distribution data;

[0044] An intelligent control module for cross - validating the prediction result with the particle size distribution data and dynamically adjusting the control strategy.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] By obtaining the production data and historical data during the sintering process of sintered ore and forming a high - quality data set for subsequent analysis and modeling, then establishing a performance prediction model to predict the quality index of sintered ore, and identifying the mixture image data during the sintering process of sintered ore to obtain the particle size distribution data, and then dynamically adjusting the control strategy by combining the prediction result with the particle size distribution data, the key parameters for intelligent control of the quality of sintered ore are adjusted, achieving precise control of the production process of sintered ore and ensuring the stability of the quality of sintered ore and the improvement of production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Among them:

[0048] Figure 1 is the block diagram of the method steps of the present invention;

[0049] Figure 2 is the block diagram of the system modules of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] To make the above - mentioned objects, features, and advantages of the present invention more obvious and understandable, the following will provide a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification.

[0051] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0052] Secondly, the so - called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it an individual or selectively exclusive embodiment with other embodiments.

[0053] As shown in the Figure 1 accompanying

[0054] Example 1: This example provides a method for improving the quality of sintered ore with intelligent regulation, including:

[0055] S1. Obtain image data, production data, and historical data during the sintering process of sintered ore, and integrate the production data and historical data to obtain the fused data;

[0056] S2. Use the prediction model to predict the quality indicators of sintered ore based on the fused data to obtain the prediction results;

[0057] S3. Perform image recognition on the image data to obtain particle size distribution data;

[0058] S4. Dynamically adjust the control strategy by combining the prediction results and the particle size distribution data.

[0059] By obtaining the production data and historical data during the sintering process of sintered ore, it is used to obtain the key parameters affecting the quality of sintered ore, such as the particle size of the mixture, the burning particle size, the water content, the alkalinity, the drum strength, etc., and form a high-quality data set for subsequent analysis and modeling. Then, a performance prediction model is established to predict the quality indicators of sintered ore, and the mixture image data during the sintering process of sintered ore is recognized to obtain the particle size distribution data. Then, the control strategy is dynamically adjusted by combining the prediction results and the particle size distribution data. For example, when the drum index in the predicted quality indicators is low and does not reach the target value, or when the particle size distribution data shows that the coarse particles are low, the system will automatically adjust the control strategy to increase the proportion of coarse particles, thereby intelligently regulating the key parameters of the quality of sintered ore and achieving precise control of the sintered ore production process, ensuring the stability of the quality of sintered ore and the improvement of production efficiency;

[0060] The predicted quality indicators are preferably used to support the adjustment of water content and alkalinity, and image recognition is preferably used to support the adjustment of particle size. And cross-validation can be carried out by comparing the predicted quality indicators with the particle size distribution data, and the parameters of the prediction model or the image processing parameters can be optimized according to the verification results to improve the accuracy.

[0061] Specifically, in step S1, the historical data is the production data and quality indicators within a past period of time;

[0062] The image data is the visual image of the sintered ore mixture;

[0063] The production data is one of the temperature, humidity, pressure, and flow rate during the sintering process.

[0064] Specifically, in step S1, the obtained real-time data and historical data are preprocessed, and the preprocessing includes data cleaning, feature extraction, and data standardization;

[0065] Data cleaning includes removing noise data and filling missing values;

[0066] Feature extraction includes using principal component analysis to extract the main features, and combining with random forest to rank the feature importance, and screening out the features that have a greater impact on the quality indicators;

[0067] Data standardization includes standardizing the features that have a greater impact to make them have the same dimension, and obtaining the standardized feature matrix.

[0068] Specifically, the prediction model in step S2 includes a DNN prediction model and an LSTM prediction model;

[0069] The DNN prediction model is used to process static features, that is, feature data independent of time, such as the particle size of the mixture, the particle size of the fuel, the water content, the alkalinity, etc. The model predicts the quality indicators at the current moment by learning the relationship between these features and the sinter quality indicators;

[0070] The LSTM prediction model is used to process time series data, that is, feature data related to time, such as data that changes with time like temperature, humidity, pressure, etc. The model predicts the quality indicators at future moments by learning the long-term dependence relationships in the time series data.

[0071] Specifically, the following formula is used in step S2 to predict the quality indicators at the current moment:

[0072]

[0073] Among them, is the predicted value of the quality indicator at the current moment, f is the DNN function, X scaled is the standardized feature matrix, and θ is the parameter of the DNN model.

[0074] Specifically, the following formula is used in step S2 to predict the quality indicators at future moments:

[0075]

[0076] Among them, is the predicted value of the quality indicator at time step t, f is the LSTM function, X scaled，t is the standardized feature matrix at time step t, H t-1 is the hidden state of the previous time step, and θ is the parameter of the LSTM model.

[0077] Specifically, in step S3, image recognition is performed on the image data to identify the particle size of the mixture. The image recognition includes image preprocessing, feature extraction, classifier training, and particle size distribution calculation;

[0078] Image preprocessing is used to grayscale and denoise the image data to obtain the processed data;

[0079] Feature extraction is used to extract the edge features and texture features of the processed data using a CNN;

[0080] Classifier training is used to train a classifier through KNN to classify different granularity mixtures and obtain classification results;

[0081] Granularity distribution calculation is used to calculate the proportion of different granularity mixtures based on the classification results, providing data support for granularity adjustment.

[0082] Specifically, step S4 outputs the control strategy through the following formula:

[0083]

[0084] where u(t) is the control strategy, e(t) is the prediction error, is the predicted value of the prediction model, yt is the target value, K p 、K i 、K d are the proportional, integral, and differential gains in sequence.

[0085] Embodiment 2: This embodiment provides a method and system for improving the quality of sintered ore through intelligent regulation, including:

[0086] S1. Obtain the image data, production data, and historical data during the sintering process of sintered ore, and integrate the production data and historical data to obtain the fused data;

[0087] S2. Predict the quality indicators of sintered ore based on the fused data through a prediction model to obtain the prediction results;

[0088] S3. Perform image recognition on the image data to obtain the granularity distribution data;

[0089] S4. Dynamically adjust the control strategy by combining the prediction results and the granularity distribution data.

[0090] By obtaining the production data and historical data during the sintering process of sintered ore, it is used to obtain the key parameters affecting the quality of sintered ore, such as the granularity of the mixture, the burning granularity, the water content, the alkalinity, the drum strength, etc., and form a high-quality data set for subsequent analysis and modeling. Then, a performance prediction model is established to predict the quality indicators of sintered ore, and the image data of the mixture during the sintering process of sintered ore is recognized to obtain the granularity distribution data. Then, the control strategy is dynamically adjusted by combining the prediction results and the granularity distribution data. For example, when the drum index in the predicted quality indicators is low and does not reach the target value, or when the granularity distribution data shows that the coarse particles are low, the system will automatically adjust the control strategy to increase the proportion of coarse particles, thereby intelligently regulating the key parameters of the quality of sintered ore and achieving precise control of the sintered ore production process, ensuring the stability of the quality of sintered ore and the improvement of production efficiency;

[0091] The prediction quality index is preferentially used to support the adjustment of water content and alkalinity, the image recognition is preferentially used to support the adjustment of particle size, and cross-validation can be carried out by comparing the prediction quality index with the particle size distribution data. According to the verification results, the prediction model parameters or image processing parameters are optimized to improve the accuracy.

[0092] Specifically, in step S1, the historical data includes production data and quality indexes within a past period of time;

[0093] The image data is the visual image of the sinter mixture;

[0094] The production data is one of temperature, humidity, pressure, and flow rate during the sintering process;

[0095] Data cleaning includes removing noise data and filling missing values;

[0096] Feature extraction includes using principal component analysis to extract main features and combining with random forest to rank the feature importance, and screening out the features that have a greater impact on the quality index;

[0097] Data standardization includes performing standardization processing on the features with greater impact to make them have the same dimension, and obtaining the standardized feature matrix.

[0098] Specifically, in step S2, the prediction model includes a DNN prediction model and an LSTM prediction model;

[0099] The DNN prediction model is used to process static features, that is, feature data independent of time, such as the particle size of the mixture, the particle size of the fuel, water content, alkalinity, etc. The model predicts the quality index at the current moment by learning the relationship between these features and the sinter quality index;

[0100] The LSTM prediction model is used to process time series data, that is, feature data related to time, such as data that changes with time like temperature, humidity, pressure, etc. The model predicts the quality index at a future moment by learning the long-term dependence relationship in the time series data.

[0101] Specifically, in step S2, the following formula is used to predict the quality index at the current moment:

[0102]

[0103] Among them, is the predicted value of the quality index at the current moment, f is the DNN function, X scaled is the standardized feature matrix, and θ is the parameter of the DNN model.

[0104] Specifically, in step S2, the following formula is used to predict the quality index at a future moment:

[0105]

[0106] Among them, is the predicted value of the quality index at time step t, f is the LSTM function, and X scaled，t is the standardized feature matrix at time step t, and H t-1 is the hidden state of the previous time step, and θ is the parameter of the LSTM model.

[0107] Specifically, in step S3, image recognition is performed on the image data to identify the particle size of the mixture. The image recognition includes image preprocessing, feature extraction, classifier training, and particle size distribution calculation;

[0108] Image preprocessing is used to grayscale and denoise the image data to obtain the processed data;

[0109] Feature extraction is used to extract the edge features and texture features of the processed data using CNN;

[0110] Classifier training is used to train the classifier through KNN to classify mixtures of different particle sizes and obtain the classification results;

[0111] Particle size distribution calculation is used to calculate the proportion of mixtures of different particle sizes based on the classification results to provide data support for particle size adjustment.

[0112] Specifically, step S4 outputs the control strategy through the following formula:

[0113]

[0114]

[0115] Among them, u(t) is the control strategy, e(t) is the prediction error, is the predicted value of the prediction model, yt is the target value, and K p 、K i 、K d are the proportional, integral, and differential gains in sequence.

[0116] Specifically, step S4 also adjusts the calculation parameters through fuzzy logic control, specifically including:

[0117] Fuzzify the prediction error and the error change rate and convert them into fuzzy variables;

[0118] Define a fuzzy rule base. The fuzzy rule base is used to adjust the parameters of the proportional, integral, and differential gains according to different prediction errors and error change rates. For example, if the error is large and the error change rate is large, then increase the proportional gain; if the error is small and the error change rate is small, then decrease the differential gain;

[0119] Use a fuzzy inference mechanism to adjust the calculation parameters according to the fuzzy rule base;

[0120] Convert the fuzzy variables back to the actual control parameters;

[0121] Through fuzzy logic control, the parameters of the PID controller can be dynamically adjusted according to the actual situation, improving the control accuracy and robustness.

[0122] As can be seen from the above: The quality prediction at the current moment is provided by the DNN model to help the system adjust the immediate production parameters. The quality prediction at future moments is provided by the LSTM model to help the system adjust the production parameters in advance and avoid quality fluctuations. The water distribution amount, fuel consumption, mixture particle size and other parameters are dynamically adjusted according to the prediction error and the error change rate through the PID controller and fuzzy logic control, realizing the precise control of the sintering process and improving the quality of sinter.

[0123] As shown in the appendix Figure 2 as follows:

[0124] Embodiment 3: This embodiment is basically the same as the previous embodiment, except that the intelligent regulation sinter quality improvement system includes:

[0125] A data acquisition module, which is used to obtain the image data, production data and historical data during the sintering process of sinter, and integrate the production data and historical data to obtain the fused data;

[0126] A model prediction module, which is used to predict the quality index of sinter according to the fused data through a prediction model to obtain a prediction result;

[0127] An image recognition module, which is used to perform image recognition on the image data to obtain the particle size distribution data;

[0128] An intelligent control module, which is used to cross-verify the prediction result with the particle size distribution data and dynamically adjust the control strategy.

[0129] Importantly, it should be noted that the construction and arrangement of the present application shown in multiple different exemplary embodiments are merely illustrative. Although only a few embodiments are described in detail in this disclosure, those who refer to this disclosure should easily understand that many modifications are possible without substantially departing from the novel teachings and advantages of the subject matter described in this application (e.g., changes in the dimensions, scales, structures, shapes and proportions of various elements, as well as parameter values (such as temperature, pressure, etc.), installation arrangements, use of materials, colors, orientations, etc.). For example, an element shown as integrally formed may be composed of multiple parts or elements, the position of the element may be inverted or otherwise changed, and the nature, number or position of discrete elements may be altered or changed. Accordingly, all such modifications are intended to be included within the scope of the present invention. The order or sequence of any process or method steps may be changed or reordered according to alternative embodiments. In the claims, any "means-plus-function" clause is intended to cover the structures that perform the recited function herein, and not only structural equivalents but also equivalent structures. Other substitutions, modifications, changes and omissions may be made in the design, operating conditions and arrangement of the exemplary embodiments without departing from the scope of the present invention. Accordingly, the present invention is not limited to specific embodiments but extends to various modifications that still fall within the scope of the appended claims.

[0130] In addition, to provide a concise description of the exemplary embodiments, not all features of the actual embodiments may be described (i.e., those features that are not relevant to the currently contemplated best mode of carrying out the present invention or those features that are not relevant to the implementation of the present invention).

[0131] It should be understood that in the development of any actual implementation, as in any engineering or design project, numerous specific implementation decisions may be made. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, the development efforts will be a routine task of design, manufacturing and production without undue experimentation.

[0132] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention may be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for improving the quality of sintered ore by intelligent control, characterized in that: include: S1. Acquire image data, production data and historical data during the sintering process of sintered ore, and integrate the production data and the historical data to obtain fused data; S2. Predicting the quality index of the sintered ore according to the fused data by using the prediction model to obtain a prediction result; S3, performing image recognition on the image data to obtain particle size distribution data; S4. Dynamically adjust the control strategy based on the prediction result and the particle size distribution data.

2. The method for improving the quality of sintered ore by intelligent control according to claim 1, characterized in that: The historical data in step S1 includes production data and quality indicators within a period of time in the past; The image data is a visual image of the sintered ore mixture; The production data is one of temperature, humidity, pressure and flow rate during the sintering process.

3. The method for improving the quality of sintered ore by intelligent control according to claim 1, characterized in that: In step S1, the real-time data and historical data acquired are preprocessed, and the preprocessing includes data cleaning, feature extraction and data standardization; The data cleaning includes removing noise data and filling missing values; The feature extraction includes extracting main features using principal component analysis, and ranking feature importance in combination with random forest to screen out features that have a greater impact on the quality index; The data standardization includes standardizing the features with greater influence so that they have the same dimension and obtaining a standardized feature matrix.

4. The method for improving the quality of sintered ore by intelligent control according to claim 3, characterized in that: The prediction model in step S2 includes a DNN prediction model and an LSTM prediction model; The DNN prediction model is used to process static features and predict the quality index at the current moment; The LSTM prediction model is used to process time series data and predict quality indicators at future moments.

5. The method for improving the quality of sintered ore by intelligent control according to claim 4, characterized in that: In step S2, the following formula is used to predict the quality index at the current moment: in, is the predicted value of the quality index at the current moment, f is the DNN function, X scaled is the standardized feature matrix, and θ is the parameter of the DNN model.

6. The method for improving the quality of sintered ore by intelligent control according to claim 4, characterized in that: In step S2, the following formula is used to predict the quality index at the future time: in, is the quality index prediction value at time step t, f is the LSTM function, X scaled，t is the standardized feature matrix at time step t, H t-1 is the hidden state of the previous time step, and θ is the parameter of the LSTM model.

7. The method for improving the quality of sintered ore by intelligent control according to claim 1, characterized in that: The step S3 performs image recognition on the image data to identify the particle size of the mixture, wherein the image recognition includes image preprocessing, feature extraction, classifier training and particle size distribution calculation; The image preprocessing is used to grayscale and remove noise from the image data to obtain processed data; The feature extraction is used to extract edge features and texture features of the processed data using CNN; The classifier training is used to train the classifier through KNN to classify mixed materials with different particle sizes and obtain classification results; The particle size distribution calculation is used to calculate the proportion of mixed materials with different particle sizes according to the classification results, providing data support for particle size adjustment.

8. The method for improving the quality of sintered ore by intelligent control according to claim 7, characterized in that: The step S4 outputs the control strategy through the following formula: Among them, u ( t ) For the control strategy, e ( t ) is the prediction error, is the predicted value of the prediction model, yt is the target value, K p , K i , K d These are proportional, integral, and derivative gains, respectively.

9. The method for improving the quality of sintered ore by intelligent control according to claim 1, characterized in that: The step S4 further adjusts the calculation parameters through fuzzy logic control, specifically including: Fuzzifying the prediction error and the error change rate and converting them into fuzzy variables; Defining a fuzzy rule base, the fuzzy rule base is used to adjust the parameters of proportional, integral and differential gains according to different prediction errors and error change rates; Using a fuzzy inference mechanism, adjusting calculation parameters according to the fuzzy rule base; The fuzzy variables are converted back into actual control parameters.

10. Intelligently controlled sinter quality improvement system, characterized in that: include: A data acquisition module is used to obtain image data, production data and historical data during the sintering process of the sintered ore, and integrate the production data and the historical data to obtain fused data; A model prediction module, used to predict the quality index of the sintered ore according to the fused data through the prediction model to obtain a prediction result; An image recognition module, used for performing image recognition on the image data to obtain particle size distribution data; The intelligent control module is used to cross-validate the prediction result with the particle size distribution data and dynamically adjust the control strategy.

Images