Dust control method in calcium carbonate manufacturing process

Abstract

The invention provides a dust control method in a calcium carbonate manufacturing process, which comprises the following steps: preprocessing a data set of dust monitoring data and process parameters, fusing data with different sources, different formats and different timestamps, and normalizing the data into a time sequence analysis form; under preset production efficiency, cost control and quality stability constraint conditions, optimizing and searching equipment decision variable parameters, a product formula and a transportation mode to obtain an equilibrium solution of dust control and production management, and forming an improved process scheme; and the optimized process scheme is issued to a control system of each production device, the temperature, pressure and flow parameters of each process are adjusted in real time, the dust concentration change is dynamically monitored, and closed-loop optimization control of dust is formed.

Description

Technical Field

[0001] The invention relates to the field of information technology, and in particular to a dust control method in a calcium carbonate manufacturing process. Background Art

[0002] In the calcium carbonate manufacturing process, the generation of dust is closely related to the process flow. In order to achieve effective dust control, it is necessary to monitor and collect data on key parameters such as the chemical composition, particle size and pH value of dust in real time. Among them, how to collect data on multiple dust attributes is a major problem. However, the massive amount of dust monitoring data has brought huge challenges to data analysis and process optimization. How to mine the intrinsic connection between dust and manufacturing process from huge data and formulate accurate process optimization plans based on this has become a technical problem that needs to be solved urgently. The correlation analysis between dust data and manufacturing process involves multiple complex links, including data preprocessing, feature extraction, association rule mining, etc. Among them, data preprocessing requires cleaning, integration and conversion of the original dust monitoring data to improve data quality; feature extraction requires selecting the most representative and discriminative feature subsets from high-dimensional data to reduce the complexity of analysis; association rule mining requires the use of algorithms such as frequent pattern mining and sequence pattern mining to discover the implicit relationship between dust attributes and process parameters. These links are interdependent and iteratively optimized, which places extremely high demands on the accuracy, efficiency and robustness of the algorithm. In addition, the calcium carbonate manufacturing process involves multiple steps such as raw material processing, calcination, digestion, and drying, with a wide variety of process parameters, such as temperature, pressure, flow rate, speed, etc. The correlation between dust data and process parameters is diverse, dynamic, and uncertain, which further increases the difficulty of correlation analysis. At the same time, process optimization requires comprehensive decision-making in multiple dimensions such as equipment parameters, product formula, transportation methods, ventilation management, etc. The optimization process affects the entire system. How to meet the dust control goals while taking into account production efficiency, cost control, quality stability and other constraints is a multi-objective optimization problem that is difficult to solve. Summary of the invention

[0003] The present invention provides a dust control method in a calcium carbonate manufacturing process, which mainly comprises:

[0004] In the calcium carbonate manufacturing process, an infrared spectrometer is installed to monitor the chemical composition data of dust, a laser particle size analyzer is used to monitor the particle size data of dust, and a pH meter is used to monitor the pH value data of dust, so as to obtain dust monitoring data in the calcium carbonate manufacturing process. At the same time, the temperature, pressure, flow rate, and speed process parameters of raw material processing and calcination, digestion, and drying processes are obtained to construct a data set of dust monitoring data and process parameters;

[0005] Preprocess the dust monitoring data and process parameter data sets, fuse data from different sources, formats, and timestamps, and normalize the data into a time series analysis format;

[0006] Extracting a subset of dust monitoring data and process parameter features from the pre-processed dust monitoring data and process parameter data sets, combining dust attributes including dust chemical composition, particle size, and pH value attributes with process parameters including temperature, pressure, flow rate, and rotation speed, and screening out feature combinations;

[0007] Based on the feature combination, the frequent association patterns between dust properties and process parameters are obtained. The frequent association rules are obtained by using the preset association rule mining algorithm and the preset algorithm setting parameters. Based on the analysis results of dust properties and process parameters in the time series, the dust and process association rule library is obtained.

[0008] According to the dust and process association library obtained by mining, combined with the preset calcium carbonate production process flow chart, the causes of dust generation in each process are determined, the dust control points of key processes such as raw material processing and calcination, digestion and drying are determined, and process improvement measures for equipment parameters, product formula and transportation methods are formulated;

[0009] Under the preset production efficiency, cost control, and quality stability constraints, optimize the equipment decision variable parameters, product formula, and transportation method to obtain a balanced solution for dust control and production management and form an improved process plan;

[0010] The optimized process plan is sent to the control system of each production equipment, the temperature, pressure and flow parameters of each process are adjusted in real time, and the changes in dust concentration are dynamically monitored to form a closed-loop optimization control of dust.

[0011] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0012] The invention discloses a dust control method in a calcium carbonate manufacturing process. The method installs sensors on a production line to monitor dust properties and process parameters in real time, and constructs a data set of dust monitoring and process parameters. After preprocessing and feature extraction of the data, an association rule base of dust and process is established by using association rule mining and time series analysis. Based on the rule base, the invention determines the cause of dust generation in each process, determines the dust control point of the key process, and formulates process improvement measures. Under the constraints of production efficiency, cost and quality, the process parameters are optimized to form an improvement plan. The optimized plan is sent to the production equipment, the process parameters are adjusted in real time, and the dust concentration change is dynamically monitored to form a closed-loop control. By analyzing the relationship between the dust concentration change curve and the process parameters, the invention can adjust the parameters according to feedback to achieve precise control of dust. The method realizes the intelligent management of dust emissions in the calcium carbonate production process, and improves production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a flow chart of a dust control method in a calcium carbonate manufacturing process.

[0014] Figure 2 It is a schematic diagram of a dust control method in a calcium carbonate manufacturing process of the present invention.

[0015] Figure 3 It is another schematic diagram of a dust control method in a calcium carbonate manufacturing process of the present invention. DETAILED DESCRIPTION

[0016] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0017] like Figure 1-3 In this embodiment, a dust control method in a calcium carbonate manufacturing process may specifically include:

[0018] S101. During the calcium carbonate manufacturing process, an infrared spectrometer is installed to monitor the chemical composition data of dust, a laser particle size analyzer is used to monitor the particle size data of dust, and a pH meter is used to monitor the pH value data of dust, so as to obtain dust monitoring data during the calcium carbonate manufacturing process. At the same time, the temperature, pressure, flow rate, and speed process parameters of the raw material processing and calcination, digestion, and drying processes are obtained to construct a data set of dust monitoring data and process parameters.

[0019] The dust chemical composition data monitored by the infrared spectrometer and the dust particle size data monitored by the laser particle size analyzer are obtained; a dust feature matrix is ​​constructed based on the dust chemical composition data and the dust particle size data, wherein the dust feature matrix includes the chemical composition percentage of each dust sample and the multidimensional feature vector of the particle size distribution; a principal component analysis is performed on the dust feature matrix, and the principal component whose cumulative contribution rate reaches a preset threshold is selected as the dust key feature vector; a nonlinear mapping model between process parameters and dust key feature vectors is established, and the nonlinear mapping model is constructed using a random forest regression tool, wherein the process parameters are used as input variables and the dust key feature vector is used as the target variable; the importance score of each process parameter to the dust characteristics is calculated based on the nonlinear mapping model, and the key process parameters that most significantly affect the dust characteristics are determined; a gradient boosting decision tree tool is constructed, wherein the key process parameters and the dust key feature vector are used as input features, and the pH value is used as the prediction target, wherein the pH value is obtained by monitoring the pH meter; the decision tree tool is used to analyze the multidimensional correlation data set containing process parameters, dust characteristics and pH values, and a relationship model between process parameters and dust characteristics is established, wherein the process parameters are used as feature nodes, and the dust characteristics and pH values ​​are used as leaf nodes.

[0020] Specifically, a dust feature matrix was constructed based on the dust chemical composition data obtained by infrared spectrometer monitoring and the dust particle size data obtained by laser particle size analyzer monitoring, and the chemical composition percentage and particle size distribution of each dust sample were numerically processed to form a multidimensional feature vector. The feature matrix was reduced in dimension using principal component analysis, and the principal component with a cumulative contribution rate of 95% was selected as the dust key feature vector. A nonlinear mapping model was established using the random forest regression tool to determine the relationship between process parameters such as raw material processing temperature, calcination pressure, digestion flow rate, and drying speed and the dust key feature vector. During the model training process, the process parameters were used as input variables and the dust key feature vector was used as the target variable. The best fitting result was obtained through multiple iterative optimizations. Using the obtained nonlinear mapping model, the importance score of each process parameter on the dust characteristics was calculated to determine the key process parameters that most significantly affect the dust characteristics. Combined with the dust pH value data obtained by pH meter monitoring and the aforementioned nonlinear mapping model between process parameters and dust characteristics, a gradient boosting decision tree tool was constructed. In the gradient boosting decision tree, process parameters and dust key feature vectors were used as input features, and pH value was used as the prediction target. The model was trained and optimized by cross-validation method, and finally a prediction model that can accurately predict the pH value of dust under different process parameter combinations was obtained. Based on the nonlinear mapping model of process parameters and dust characteristics and the pH value prediction model, a multidimensional correlation data set containing process parameters, dust characteristics and pH value was constructed. The decision tree tool was used to analyze the data set and establish a relationship model between process parameters and dust characteristics. In the decision tree, process parameters were used as feature nodes, dust characteristics and pH values ​​were used as leaf nodes, and feature selection and tree structure construction were performed using the information gain criterion. By analyzing the decision tree model, the key process parameter combinations that affect dust characteristics were identified, providing a judgment basis for the optimization of process parameters in the calcium carbonate manufacturing process. In the calcium carbonate manufacturing process, the dust chemical composition data obtained by the infrared spectrometer include the percentages of components such as CaCO3, MgCO3, and SiO2, and the dust particle size data obtained by the laser particle size analyzer include particle size distribution parameters such as D10, D50, and D90. These data are combined into a multidimensional feature vector to construct a dust feature matrix. Apply principal component analysis to the feature matrix, and select principal components with a cumulative contribution rate of 95%, such as the first three principal components PC1, PC2, and PC3, as the key characteristic vectors of dust. Random forest regression is used to establish a nonlinear mapping model for process parameters such as raw material processing temperature, such as 800℃-1000℃, calcination pressure, such as 0.1-0.5MPa, digestion flow rate, such as 50-100L / min, and drying speed, such as 10-30rpm. In model training, 500 decision trees are used, the maximum depth of each tree is set to 10, and the minimum number of leaf node samples is set to 5. The model parameters are optimized through cross-validation to obtain the best fitting results.Using the trained random forest model, the feature importance score of each process parameter is calculated, such as temperature 0.4, pressure 0.3, flow rate 0.2, and speed 0.1, and temperature and pressure are determined to be the key process parameters that most significantly affect dust characteristics. Combined with the dust pH value data obtained by the pH meter, such as pH7.5-9.5, the gradient boosting decision tree model XGBoost is constructed. In XGBoost, the process parameters and dust key feature vectors PC1, PC2, and PC3 are used as input features, and the pH value is used as the prediction target. 100 base learners are used, the learning rate is set to 0.1, and the maximum tree depth is 6. The model is trained and optimized through 5-fold cross validation, and finally a pH value prediction model with a root mean square error RMSE of 0.15 is obtained. Based on the above model, a multidimensional correlation data set containing process parameters, dust characteristics, and pH values ​​is constructed. The decision tree algorithm C4.5 is used to analyze the data set and establish a relationship model between process parameters and dust characteristics. In the decision tree, process parameters are used as feature nodes, dust characteristics and pH values ​​are used as leaf nodes, and feature selection and tree structure construction are performed using the information gain ratio criterion. The minimum number of leaf node samples is set to 10, and the maximum tree depth is set to 8. By analyzing the decision tree model, the key process parameter combinations that affect dust characteristics are identified. For example, ideal dust characteristics and pH values ​​can be obtained when the temperature is in the range of 950℃-980℃ and the pressure is in the range of 0.3-0.4MPa.

[0021] S102. Preprocess the dust monitoring data and process parameter data sets, fuse data from different sources, different formats, and different time stamps, and normalize the data into a time series analysis format.

[0022] The dust monitoring data and process parameter data were obtained, and the data were preliminarily denoised using an adaptive moving median filter to obtain a preliminarily denoised data set. Based on the preliminarily denoised data set, the Z-score method was combined to identify outliers, and it was determined whether the Z-score value of the data point exceeded the preset threshold range. If it exceeded, it was marked as an outlier. For the outliers, the local linear regression method was used to correct them to obtain a corrected data set. The corrected data set was formatted uniformly using a data conversion tool, and the data of the infrared spectrometer, laser particle size analyzer, and pH meter were converted into a standardized JSON format to obtain a standardized data set. The Ka lman filtering algorithm was used to perform multi-source data fusion on the standardized data set, and a unified data stream was generated based on the measurement errors and time delays of different data sources. The timestamp alignment tool was used to synchronize the unified data stream, interpolate all data at a uniform time interval, and fill in the missing time point data using the cubic spline interpolation method to obtain a time-aligned normalized time series data set.

[0023] Specifically, an adaptive moving median filter is used to perform preliminary noise removal on dust monitoring data and process parameter data. The sliding window size is dynamically adjusted by calculating the local data variance. The median is calculated for the data in each window, and the original data points are replaced to obtain a preliminary denoised data set. The mirror filling method is used for edge data points to maintain data integrity. According to the statistical characteristics of dust monitoring data and process parameter data, the Z-score method and local anomaly factor algorithm are combined to identify outliers, calculate the Z-score value of the data, set the threshold to ±3, and calculate the local anomaly factor of each data point at the same time. Data that exceeds the threshold range or has a significantly higher local anomaly factor than the surrounding points are marked as outliers, and the local linear regression method is used to correct the outliers. For data from different sources and formats, the data format is unified through data conversion tools, and the data from infrared spectrometers, laser particle size analyzers, and pH meters are converted into standardized JSON format. Field mapping, unit conversion, and data type unification are performed. At the same time, the process parameter data is format converted to ensure that all data have consistent structures and field names, and the original timestamp information is retained in the process. The Ka lman filter algorithm is used to fuse multi-source data, taking into account the measurement errors and time delays of different data sources to generate a unified data stream. Then, the timestamp alignment tool is used to synchronize the fused data of different timestamps, interpolate all data at a unified time interval, use the cubic spline interpolation method to fill the missing time point data, and use piecewise linear interpolation for intervals with drastic changes. Finally, a time-aligned normalized time series data set is obtained, laying the foundation for subsequent analysis. In the calcium carbonate manufacturing process, an adaptive moving median filter is used to perform preliminary noise removal on dust monitoring data and process parameter data. In specific implementation, the local data variance within 10 minutes is calculated. If the variance is greater than 0.5, the sliding window is set to 3 data points; if the variance is less than 0.5, it is set to 7 data points. The median is calculated for the data in each window, and the original data points are replaced to obtain a preliminary denoised data set. For edge data points, the mirror filling method is used to mirror the first 3 or last 3 data points to the edge to maintain data integrity. Subsequently, the Z-score method and the local anomaly factor algorithm are combined to identify outliers. Calculate the Z-score value of the data and set the threshold to

[0024] ±3, and calculate the local anomaly factor of each data point at the same time, select 10 nearest neighbor points, and if the local anomaly factor is greater than 1.5, mark the point as an outlier. The local linear regression method is used to correct the outliers, and the 5 normal data points before and after are used to fit the straight line, and the outliers are replaced by the fitted values. For data from different sources and different formats, the data format is unified through data conversion tools. The wavelength range 4000-400cm^-1 in the infrared spectrometer data is converted to JSON format, the D10, D50, and D90 values ​​of the laser particle size analyzer are converted to micrometer units, and the data of the pH meter retains two decimal places. At the same time, the process parameter data is formatted, the temperature is unified to degrees Celsius, and the pressure is unified to MPa, ensuring that all data have consistent structures and field names. In this process, the original timestamp information is retained, accurate to milliseconds. Subsequently, the multi-source data is fused using the Kalman filter algorithm, and the measurement noise covariance matrix R is set to a diagonal matrix with diagonal elements of 0.1, 0.2, and 0.05, respectively, representing the measurement errors of different data sources. The process noise covariance matrix Q is set to 0.01×I, where I is the unit matrix. A unified data stream is generated through iterative updates. Finally, the fused data is synchronized using the timestamp alignment tool, and all data are interpolated at 1 second intervals. For most of the data, the cubic spline interpolation method is used, and the smoothing parameter is set to 0.5; for intervals with drastic changes, such as temperature changes of more than 10°C within 5 seconds, piecewise linear interpolation is used. Finally, a time-aligned normalized time series data set is obtained, laying the foundation for subsequent analysis.

[0025] S103. Extracting a subset of dust monitoring data and process parameter features from the pre-processed dust monitoring data and process parameter data sets, combining dust properties including dust chemical composition, particle size, and pH value properties with process parameters including temperature, pressure, flow rate, and rotation speed, and screening out feature combinations.

[0026] According to the pre-processed dust monitoring data and process parameter data set, the data is standardized to obtain standardized data; the kernel principal component analysis method is used to extract the feature subset of the standardized data, and the principal components with a cumulative contribution rate of a preset percentage are obtained as the feature subset by calculating the eigenvalues ​​and eigenvectors of the kernel matrix of the standardized data; for the process parameter data, the maximum information coefficient is used to calculate the nonlinear correlation between temperature, pressure, flow rate, and speed. If the maximum information coefficient is greater than the threshold, the parameter with the largest amount of information is selected as the representative to obtain the process parameter feature subset; combined with the dust chemical composition, particle size, and pH value attributes, the relationship between the attribute and the process parameter feature subset is evaluated, and the attribute with the highest mutual information value is selected and combined with the process parameter feature subset to form a preliminary feature combination; the random forest feature importance evaluation tool is used to screen the preliminary feature combination, and the feature importance ranking is obtained by calculating the contribution of each feature to the model prediction accuracy; combined with the recursive feature elimination method, the least important features are gradually removed, and the model performance is evaluated after each removal until the performance begins to decline significantly or reaches the preset feature number threshold, and the optimized feature combination is obtained.

[0027] Specifically, based on the pre-processed dust monitoring data and process parameter data set, the data is first standardized to scale all features to a mean of 0 and a variance of 1. Subsequently, the kernel principal component analysis method is used to extract feature subsets. By calculating the eigenvalues ​​and eigenvectors of the kernel matrix, the principal components with a cumulative contribution rate of 95% are selected as feature subsets. For the process parameter data, the maximum information coefficient is used to calculate the nonlinear correlation between temperature, pressure, flow rate, and speed. If the maximum information coefficient is greater than 0.8, the parameter with the largest amount of information is selected as the representative to obtain the process parameter feature subset. Combined with the dust chemical composition, particle size, and pH value attributes, the minimum redundancy maximum correlation algorithm is used to evaluate the relationship between these attributes and the process parameter feature subset. The algorithm maximizes the correlation between the feature and the target variable while minimizing the redundant information between the features, selects the top five attributes with the highest mutual information value, and combines them with the process parameter feature subset to form a preliminary feature combination. The random forest feature importance evaluation tool is used to screen the preliminary feature combination, and the feature importance ranking is obtained by calculating the contribution of each feature to the model prediction accuracy. Combined with the recursive feature elimination algorithm, the least important features are gradually removed, and the model performance is evaluated after each removal until the performance begins to decline significantly or reaches the preset feature number threshold. Finally, the optimized feature combination is obtained as the input data for subsequent analysis. In the calcium carbonate manufacturing process, the pre-treated dust monitoring data and the process parameter data set are subjected to feature extraction and selection. First, all data are standardized, and the eigenvalues ​​are scaled to a range of mean 0 and variance 1 to ensure that data of different scales are comparable. For example, the temperature data is standardized from 800-1000℃ to a range of -1.5 to 1.5. Subsequently, the kernel principal component analysis method is used to extract feature subsets, and the Gaussian kernel function is selected. The kernel parameter σ is set to 1.0. By calculating the eigenvalues ​​and eigenvectors of the kernel matrix, the principal components with a cumulative contribution rate of 95% are selected as feature subsets, and 10-15 principal components are usually obtained. For the process parameter data, the maximum information coefficient is used to calculate the nonlinear correlation between temperature, pressure, flow rate, and speed, and the α parameter is set to 0.6 and the number of sampling times is 500. If the maximum information coefficient is greater than 0.8, such as the maximum information coefficient of temperature and pressure is 0.85, the temperature with the largest amount of information is selected as the representative. Combined with the dust chemical composition, such as CaCO3 content, particle size, such as D50 value, pH value attributes, the minimum redundancy maximum correlation algorithm is used to evaluate the relationship between these attributes and the process parameter feature subset. Set the redundancy parameter β to 0.5, and select the top five attributes with the highest mutual information value by maximizing the correlation between the feature and the target variable while minimizing the redundant information between the features. For example, select CaCO3 content, D50 value, pH value, temperature and pressure as key attributes.The random forest feature importance assessment tool was used, with the number of trees set to 500 and the maximum depth set to 10, to calculate the contribution of each feature to the model prediction accuracy. Combined with the recursive feature elimination algorithm, the 10% features with the lowest importance were removed each time to evaluate the performance changes of the model. The iteration was stopped when the performance dropped by more than 1% or the number of features was reduced to 50% of the original, and the optimized feature combination was finally obtained, which usually contained 15-20 features and served as the input data for subsequent analysis.

[0028] S104, obtaining frequent association patterns between dust attributes and process parameters based on feature combinations, obtaining frequent association rules through a preset association rule mining algorithm and preset algorithm setting parameters, and obtaining a dust and process association rule library based on analysis results of dust attributes and process parameters in a time series.

[0029] The continuous data is discretized with equal frequency according to the preset number of intervals to obtain the discretized feature data; the FP-growth algorithm is used to mine frequent item sets on the discretized feature data, and frequent item sets are recursively generated by constructing FP trees and conditional FP trees; the association rule generator is used to construct association rules based on the frequent item sets; if the association rule meets the preset threshold, it is determined to be a strong association rule; the time series data set is obtained, and the frequency domain characteristics of the time column data in the time series data set are extracted using wavelet transform to obtain approximate coefficients and detail coefficients of different scales; the strongly associated rules are clustered, and the rule similarity matrix is ​​constructed by calculating the semantic similarity between the rule items. In combination with the frequency domain characteristics, the hierarchical clustering algorithm is used to merge the rules with high similarity to form a dust and process association rule library.

[0030] Specifically, according to the dust properties and process parameters in the feature combination, the continuous data is first discretized with equal frequency, and the numerical features are divided into 10 intervals. Then, the FP-growth algorithm is used to mine frequent itemsets, and the minimum support threshold is set to 0.05. By constructing FP trees and conditional FP trees, frequent itemsets are recursively generated, the item set support is calculated, and the frequent itemsets that meet the minimum support requirements are screened out. Using the obtained frequent itemsets, the association rule generator is used to construct association rules, and the minimum confidence threshold is set to 0.7. The confidence, lift, Kulczynski index and full confidence index of the rules are calculated. Strong association rules that meet the minimum confidence requirements, lift greater than 1, Kulczynski index greater than 0.6 and full confidence index greater than 0.7 are screened out. For time series data, wavelet transform is used to extract the frequency domain features of the time series, Daubechies wavelet is selected, the decomposition layer number is 3, and the approximate coefficients and detail coefficients of different scales are obtained. Then, the continuous time series data is converted into discrete transaction sets using adaptive window technology. The initial window size is set to 30 minutes. The window size is dynamically adjusted by calculating the variance of the data in the window, and the sliding step is set to 1 / 6 of the window size. The data in each window is processed by equal frequency discretization and frequent item set mining methods, and the features extracted by wavelet transform are combined to obtain association rules with timestamps and frequency domain features. The obtained association rules are integrated by the rule clustering method based on semantic similarity, and the Word2Vec tool is used to calculate the semantic similarity between rule items and construct a rule similarity matrix. Combined with the time series features, the hierarchical clustering algorithm is used, and the clustering distance threshold is set to 0.2 to merge the rules with high similarity. Finally, a dust and process association rule library is formed, which contains information such as rule content, support, confidence, lift, Kulczynski index, full confidence index, time range and frequency domain features. In the calcium carbonate manufacturing process, association rule mining is performed on dust properties and process parameters. First, continuous data such as temperature 800-1000℃ and pressure 0.1-0.5MPa are processed with equal frequency discretization and divided into 10 intervals. The FP-growth algorithm is used for frequent item set mining, and the minimum support threshold is set to 0.05. For example, by constructing the FP tree, it is found that the support of {high temperature, high pressure, large particles} is 0.06, which meets the threshold requirement. Subsequently, the association rule generator is used to construct rules, such as "high temperature, high pressure → large particles", with a calculated confidence of 0.75, a lift of 1.2, a Kulczynski index of 0.7, and a full confidence index of 0.8. For time series data, the Daubechies4 wavelet is used for 3-layer decomposition to obtain the approximate coefficient and detail coefficient. For example, the dust concentration data is subjected to wavelet transform to obtain the energy distribution of different frequency components.Adaptive window technology was used, and the initial window size was set to 30 minutes. By calculating the variance of the data in the window, the window was reduced to 15 minutes when the variance was >0.5 during the high volatility period, and expanded to 45 minutes when the variance was <0.1 during the low volatility period. The sliding step size was set to 1 / 6 of the window size. The FP-growth algorithm and association rule generator were applied to the data in each window, combined with wavelet features to obtain time-related rules. Finally, the Word2Vec tool was used to calculate the semantic similarity of the rule items, such as the similarity of "high temperature" and "temperature increase" was 0.85. The rule similarity matrix was constructed, and the hierarchical clustering algorithm was applied, with the cluster distance threshold set to 0.2. Similar rules were merged, such as "high temperature → large particles" and "temperature increase → particle size increase" were merged into one rule. Finally, an association rule library containing rule content, various indicators, time range and frequency domain characteristics was formed to provide a basis for process optimization.

[0031] The dust and process parameter data are divided into time windows according to the time series, and the analysis results of dust properties and process parameters in the time series are obtained according to the preset time series model. The association rules that meet the preset practical application value are screened out by combining the mined frequent association rules. The screened rules are organized into a database format to establish a dust and process association rule library.

[0032] The dust and process parameter data are segmented into time series by the adaptive sliding window method, and the window size is dynamically adjusted according to the data variance in the window; if the variance exceeds the preset threshold, the window is reduced, otherwise the window is expanded. The dust attribute and process parameter data in the time window are analyzed, and the data non-stationarity is eliminated by the seasonal autoregressive integrated moving average model, and the multi-scale characteristics of the time series are extracted by discrete wavelet transform. According to the results of time series analysis and the frequent association rules mined in advance, the support threshold, confidence threshold, Kulczynski index threshold and full confidence index threshold are set to determine whether the rules continue to meet all threshold conditions. The rules that meet the conditions are weighted to obtain the energy distribution and determine the weight coefficient. The weighted rules are converted into a structured data format, and a relational database table structure containing a rule table and a time window table is established; the rule table contains the rule ID, antecedent, consequent, support, confidence, lift, Kulczynski index, full confidence index, and weight coefficient fields; the time window table contains the window ID, rule ID, start time, end time, and window feature fields.

[0033] Specifically, the adaptive sliding window method is used to segment the dust and process parameter data into time series. The initial window size is set to 60 minutes. The window size is dynamically adjusted by calculating the variance of the data in the window. When the variance exceeds the preset threshold, the window is reduced, otherwise the window is expanded. The sliding step is set to 1 / 6 of the window size to generate a continuous time window sequence, each of which contains the time series data of dust properties and process parameters. The seasonal autoregressive integrated moving average model and wavelet transform are used to analyze the dust properties and process parameter data in each time window. The non-stationarity of the data is eliminated by differential operation, and the order and seasonal cycle of the model are determined by using the autocorrelation function and partial autocorrelation function. At the same time, the discrete wavelet transform is applied to extract the multi-scale features of the time series. The Daubechies wavelet is selected, the decomposition layer number is 3, and the energy distribution of different frequency components is obtained. Combining the time series analysis results with the pre-mined frequent association rules, the rules are screened by setting the support threshold to 0.1, the confidence threshold to 0.7, the Kulczynski index threshold to 0.6, and the full confidence index threshold to 0.7. It is determined whether the rule continuously meets all threshold conditions in three consecutive time windows, and the lifting degree of the rule is calculated at the same time. If the lifting degree is greater than 1.5, the rule is determined to have practical application value. The rules are weighted by using time series characteristics, and the weight coefficient is determined according to the energy distribution obtained by wavelet transform, highlighting the rules with significant time series characteristics. The selected association rules with practical application value are converted into structured data format, and the relational database table structure is designed, including the main table "rule table" and the slave table "time window table". The rule table contains fields such as rule ID, which is the primary key, antecedent, consequent, support, confidence, lifting, Kulczynski index, full confidence index, and weight coefficient. The time window table contains fields such as window ID, which is the primary key, rule ID, which is the foreign key, start time, end time, and window characteristics. Indexes are established on the rule ID of the rule table and the rule ID of the time window table to optimize the query performance and form a dust and process association rule library. In the calcium carbonate manufacturing process, time series analysis and association rule mining are performed on dust and process parameter data. First, the adaptive sliding window method is used to segment the time series, and the initial window size is set to 60 minutes. By calculating the variance of the data in the window, when the variance exceeds 0.5, the window is reduced to 30 minutes; when the variance is less than 0.1, it is expanded to 90 minutes. The sliding step is set to 1 / 6 of the window size. The seasonal autoregressive integrated moving average model and wavelet transform are applied to the data in each window. The difference operation is used to eliminate non-stationarity, and the model order is determined by the autocorrelation function and partial autocorrelation function, such as SARIMA (1,1,1) (0,1,1) 12. At the same time, the Daubechies4 wavelet is used to perform a 3-layer discrete wavelet transform to obtain the approximate coefficients and detail coefficients, and calculate the energy distribution of different frequency components.For example, the energy proportion of low-frequency components is 60%, that of medium-frequency components is 30%, and that of high-frequency components is 10%. Combining the results of time series analysis and the pre-mined frequent association rules, the screening conditions are set as support threshold 0.1, confidence threshold 0.7, Kulczynski index threshold 0.6, and full confidence index threshold 0.7. Determine whether the rule continues to meet all threshold conditions in three consecutive time windows, and calculate the degree of improvement at the same time. For example, the indicators of the rule "temperature rise → dust particle size increase" in three consecutive windows are support, including 0.12, 0.13, 0.11, confidence, including 0.75, 0.78, 0.72, Kulczynski index, including 0.65, 0.68, 0.63, full confidence index, including 0.73, 0.75, 0.71, and improvement 1.8, which is determined to be a rule with practical application value. The energy distribution obtained by wavelet transform is used to determine the weight coefficient, such as the weight of the rule is 0.6×0.7+0.3×0.2+0.1×0.1=0.49. Finally, the selected rules are stored in the relational database, and the "rule table" and "time window table" are designed. The rule table contains fields such as rule ID, which is the primary key, such as R001, antecedent, which includes temperature increase, consequent, which includes dust particle size increase, support 0.12, confidence 0.75, lift 1.8, Kulczynski index 0.65, full confidence index 0.73, and weight coefficient 0.49. The fields of the time window table include window ID, which is the primary key, such as W001, rule ID, which is the foreign key, R001, start time 2024-03-1510:00:00, end time 2024-03-1511:00:00, and window features, which can store wavelet coefficients in JSON format. Establish an index on the rule ID to optimize query performance and form a complete rule library for the association between dust and process.

[0034] S105. Based on the dust and process association library obtained through mining and combined with the preset calcium carbonate production process flow chart, determine the cause of dust generation in each process, determine the dust control points of key processes such as raw material handling and calcination, digestion, and drying, and formulate process improvement measures for equipment parameters, product formula, and transportation methods.

[0035] The preset calcium carbonate production process flow chart is converted into a directed graph, in which nodes represent processes, edges represent the associations between processes, and edge weights are set according to association rules; the shortest paths between nodes in the directed graph are calculated to obtain the causes and influencing factors of dust generation in each process; based on the causes and influencing factors of dust generation, the four key processes of raw material processing, calcination, digestion, and drying are evaluated using a fuzzy comprehensive evaluation method combined with principal component analysis, and evaluation indicators of dust concentration, particle size distribution, and chemical composition are set; the main features of the evaluation indicators are extracted through principal component analysis, the weights of each indicator are determined, and the comprehensive scores of each process are calculated. The node with the highest score is determined dust control points; for dust control points, a multi-objective optimization algorithm is used to optimize equipment parameters, product formulas, and transportation methods; a particle swarm optimization algorithm combined with a simulated annealing algorithm is used to solve the Pareto optimal solution set of the objective function, and the optimal parameter combination of each control point is obtained; based on the optimal parameter combination, a knowledge graph is used to build a process improvement measures library; key concepts are extracted from the optimal parameter combination through entity recognition methods, and the relationship between entities is identified using relationship extraction; the key concepts and the relationship between entities are stored in a graph database, and the graph database query language is used to perform relationship reasoning to obtain specific process improvement measures.

[0036] Specifically, according to the rules in the dust and process association library, the shortest path algorithm in graph theory is used to analyze the preset calcium carbonate production process flow chart, and the process flow chart is converted into a directed graph. The nodes represent the processes, the edges represent the associations between the processes, and the edge weights are set according to the association rules. The shortest paths between the nodes are calculated by the Floyd-Warshall algorithm, the key nodes and paths are identified, and the causes and influencing factors of dust generation in each process are obtained. The fuzzy comprehensive evaluation method is combined with principal component analysis to evaluate the four key processes of raw material processing, calcination, digestion, and drying, and evaluation indicators such as dust concentration, particle size distribution, and chemical composition are set. The main features of the data are extracted through principal component analysis, and the weights of each indicator are determined in combination with expert experience. The comprehensive score of each process is calculated, and the node with the highest score is determined as the dust control point. For the determined dust control points, the multi-objective optimization algorithm is used to optimize the equipment parameters, product formula, and transportation methods, and the objective functions such as dust control, product quality, and energy consumption are set, while considering the constraints such as the equipment operation range and raw material supply. The particle swarm optimization algorithm combined with the simulated annealing algorithm was used to solve the Pareto optimal solution set and obtain the optimal parameter combination of each control point. The Monte Carlo simulation method was used to evaluate the stability and reliability of the optimization scheme. By randomly perturbing the input parameters, a large amount of simulation data was generated, the expected value and variance of the objective function were calculated, and the optimization scheme with high stability was screened out. Based on the optimization results, the knowledge graph was used to build a process improvement measures library. Key concepts such as "equipment parameters" and "product formula" were extracted from the optimization scheme through entity recognition, and the relationship between entities, such as "influence" and "control", were identified using relationship extraction. The extracted information was stored in the graph database, and complex relationship reasoning was performed using the graph database query language to obtain specific process improvement measures and form an executable process improvement plan. In the process of optimizing the calcium carbonate production process, the process flow chart was first converted into a directed graph. The nodes represented the processes such as raw material processing, calcination, digestion and drying, and the edges represented the associations between the processes. The edge weights were set according to the association rules, such as the edge weight from raw material processing to calcination was 0.8. The Floyd-Warshall algorithm was used to calculate the shortest path between nodes, and the calcination process was identified as a key node. The cause of dust generation was mainly related to temperature and pressure. Subsequently, the four key processes were evaluated, and the evaluation indicators were set, including dust concentration (mg / m 3 ), particle size distribution (μm), chemical composition (%). The principal component analysis was used to extract features, and the characteristic values ​​were 2.5, 1.2, and 0.3. The weights were determined based on expert experience: 0.5, 0.3, and 0.2. The comprehensive score was calculated, and the calcination process scored the highest 0.85, which was determined as the main dust control point. For the calcination process, the optimization target was set to dust concentration ≤10mg / m 3, product purity ≥ 98%, energy consumption ≤ 500kWh / t. Consider the equipment temperature range 800-1000℃, pressure range 0.1-0.5MPa and other constraints. The particle swarm optimization algorithm with a swarm size of 50 and an iteration number of 100 is combined with the simulated annealing algorithm with an initial temperature of 100 and a cooling coefficient of 0.95 to obtain the optimal parameter combination of temperature 950℃ and pressure 0.3MPa. The Monte Carlo method is used for 1000 simulations to calculate the expected value of the objective function as a dust concentration of 8.5mg / m 3 , product purity 98.5%, energy consumption 480kWh / t, variance less than 5%, confirming that the scheme is highly stable. Finally, a knowledge graph is constructed to extract entities such as "calcination temperature" and "pressure", and to identify relationships such as "affecting dust concentration" and "controlling product purity". The information is stored in the Neo4j graph database, and the Cypher query language is used for reasoning. The improvement measures are to control the calcination temperature at 950±10℃ and maintain the pressure at 0.3±0.02MPa, forming an executable process improvement plan.

[0037] S106. Under the preset constraints of production efficiency, cost control, and quality stability, optimize the equipment decision variable parameters, product formula, and transportation method to obtain a balanced solution for dust control and production management and form an improved process plan.

[0038] According to the preset production efficiency, cost control, and quality stability constraints, the equipment decision variable parameters, product formula, and transportation method are obtained as optimization variables, and the dust control degree, production efficiency, cost, and product quality are used as objective functions. The ε-constraint method is used to construct a multi-objective optimization model, and the dust control degree is selected as the main objective function, and other objective functions are converted into constraints. The particle swarm optimization algorithm is combined with the simulated annealing algorithm to solve the optimization model, from which the particle swarm size, the maximum number of iterations, the initial temperature, and the cooling coefficient are determined. By continuously updating the particle position, velocity, and temperature parameters during the annealing process, the optimal solution is obtained in the global scope. The adaptive weight adjustment strategy is adopted to dynamically adjust the weights of each sub-objective in the objective function according to the results of each iteration. The moving window standard deviation of each sub-objective is calculated to determine the sub-objective with a large standard deviation. If the standard deviation is large, it is given a high weight. The dynamic balance between the objectives is achieved through the weight update formula, and the Pareto optimality of the current optimal solution is evaluated after each iteration. The solution obtained by the optimization algorithm is subjected to sensitivity analysis and Monte Carlo simulation to obtain the sensitivity coefficients of each parameter. Latin hypercube sampling is used to generate input parameter combinations, and Monte Carlo simulations are performed to evaluate the stability and reliability of the optimization results.

[0039] Specifically, according to the preset production efficiency, cost control, and quality stability constraints, a multi-objective optimization model is established, and the equipment decision variable parameters, product formula, and transportation method are used as optimization variables, and the dust control degree, production efficiency, cost, and product quality are used as objective functions. The ε-constraint method is used to construct the multi-objective optimization model, and the dust control degree is selected as the main objective function. Other objective functions are converted into constraints, and a series of single-objective optimization problems are generated by systematically changing the constraint limit ε value. The established optimization model is solved by using the particle swarm optimization algorithm combined with the simulated annealing algorithm. The particle swarm size is set to 100, the maximum number of iterations is 1000, the initial temperature is 100, and the cooling coefficient is 0.95. By continuously updating the particle position, velocity, and temperature parameters during the annealing process, the optimal solution is searched globally. The adaptive weight adjustment strategy is used to dynamically adjust the weights of each sub-objective in the objective function according to the results of each iteration. The moving window standard deviation of each sub-objective is calculated, and the window size is set to 10. Sub-objectives with larger standard deviations are given higher weights. The weight update formula is wi = σi / Σσj, where σi is the standard deviation of the ith sub-goal, wi is the standard deviation of the ith sub-goal, and Σσj is the sum of the standard deviations of all j sub-goals. In this way, a dynamic balance between goals is achieved, and the Pareto optimality of the current optimal solution is evaluated after each iteration. Sensitivity analysis and Monte Carlo simulation are performed on the solution obtained by the optimization algorithm. By changing the value of the input parameter, observing the change in the output result, and calculating the sensitivity coefficient of each parameter. Latin hypercube sampling is used to generate 10,000 sets of input parameter combinations, and Monte Carlo simulation is performed to evaluate the stability and reliability of the optimization results. According to the sensitivity analysis and Monte Carlo simulation results, the key parameters that have a greater impact on the results are screened out. A virtual production environment is constructed using digital twin technology, the optimization plan is imported, the actual production process is simulated, and the feasibility and effect of the plan are evaluated. Based on the simulation results, the optimization plan is fine-tuned to form the final improved process plan. In the process of optimizing the calcium carbonate production process, a multi-objective optimization model is first established, and dust control degree, production efficiency, cost and product quality are set as objective functions. Using the ε-constraint method, dust control is selected as the main goal, and other goals are converted into constraints. For example, production efficiency is set to ≥ 90%, cost ≤ 5000 yuan / ton, and product quality ≥ 98% purity. The optimization variables include calcination temperature 800-1000℃, pressure 0.1-0.5MPa, CaCO3 ratio 95-99%, and conveyor belt speed 0.5-2m / s. The particle swarm optimization algorithm is combined with the simulated annealing algorithm for solution. The particle swarm size is set to 100, the maximum iteration is 1000 times, the initial temperature is 100, and the cooling coefficient is 0.95. During the iteration process, the weight is adjusted dynamically, and the standard deviation of each goal within 10 iterations is calculated, such as the standard deviation of dust control is 0.05 and the production efficiency is 0.02, and the weight is updated accordingly. Pareto optimality is evaluated every 100 iterations, and non-dominated solutions are retained.After the optimization, sensitivity analysis and Monte Carlo simulation were performed. Latin hypercube sampling was used to generate 10,000 sets of parameter combinations, such as calcination temperature of 950±10℃ and pressure of 0.3±0.02MPa. The sensitivity coefficient was calculated and it was found that the temperature sensitivity was 0.8 and the pressure was 0.5, identifying the temperature as the key parameter. A virtual production line was constructed using digital twin technology, and the optimization scheme imported was calcination temperature of 955℃, pressure of 0.32MPa, CaCO3 ratio of 97.5%, and conveyor belt speed of 1.2m / s. A 100-hour production process was simulated to evaluate the feasibility of the scheme. According to the simulation results, the parameters were fine-tuned to form the final improvement scheme, including calcination temperature of 953℃, pressure of 0.31MPa, CaCO3 ratio of 97.8%, and conveyor belt speed of 1.1m / s. It is expected to reduce dust emissions by 15% while maintaining stable production efficiency and product quality.

[0040] S107. Send the optimized process plan to the control system of each production equipment, adjust the temperature, pressure and flow parameters of each process in real time, and dynamically monitor the change of dust concentration to form a closed-loop optimization control of dust.

[0041] Receive optimization instructions carrying process plan parameters, which are issued by the central controller; send the process plan parameters to the programmable logic controller of each production equipment through the industrial communication network according to the optimization instructions, and the programmable logic controller adjusts the operating status of the production equipment according to the process plan parameters; set up primary and standby channels and adopt cyclic redundancy check, which is used to ensure the accuracy of data transmission; automatically switch to the standby channel if communication interruption or data transmission error is detected; obtain the temperature, pressure, and flow parameters of each process, and adjust the temperature, pressure, and flow parameters in real time; obtain the dust concentration detection data of each key point of the production line, and the dust concentration detection data is collected by various types of dust detection equipment; collect dust concentration detection data at a preset frequency through the data acquisition module, and perform data preprocessing; construct a correlation model between dust concentration and process parameters; dynamically optimize the process parameters according to the correlation model, and the dynamic optimization is used to adjust the relevant process parameters.

[0042] Specifically, the optimized process parameters are sent to the programmable logic controller of each production equipment through the industrial communication network, and a communication link is established between the central controller and the control system of each process equipment to achieve real-time transmission and update of parameters. Redundant communication design is adopted, and the main and standby channels are set up to ensure the accuracy of data transmission through cyclic redundancy check. When communication interruption or data transmission error is detected, it automatically switches to the standby channel and records the abnormal situation. The temperature, pressure and flow parameters of each process are adjusted in real time by using the distributed control system, and the adaptive model predictive control algorithm is introduced to predict future output and optimize the control sequence by online identification of system model parameters. The upper and lower limits and change rate limits of parameter adjustment are set to prevent over-adjustment. Various types of dust detection equipment are used to continuously monitor the dust concentration of each key point of the production line, combined with light scattering and beta-ray attenuation methods to improve the comprehensiveness and accuracy of detection. The data acquisition module collects detection data at a frequency of 1Hz, performs data preprocessing, including outlier detection and moving average filtering, and transmits the processed data to the central control system in real time. Based on the decision tree algorithm, a correlation model between dust concentration and process parameters is constructed. The sliding window technology is used, the window size is set to 1 hour, and the model parameters are updated every 5 minutes to achieve online learning. The process parameters are dynamically optimized in combination with real-time monitoring data. When the predicted dust concentration exceeds the set threshold, the relevant process parameters are automatically adjusted. At the same time, a safety protection mechanism is set. When the parameter adjustment range exceeds the preset range, an alarm is triggered and the adjustment range is limited to ensure the safe and stable operation of the production process. In the dust control optimization of the calcium carbonate production line, the optimized process parameters are first sent to the PLC of each device through the OPCUA protocol. For example, parameters such as calcination temperature of 953℃ and pressure of 0.31MPa are transmitted to the kiln control system. A dual-channel redundant design is adopted, the main channel uses industrial Ethernet, and the backup channel uses 4G wireless network. Data integrity is ensured by 32-bit CRC check. When a communication interruption of more than 100ms is detected, it automatically switches to the backup channel. The parameters of each process are adjusted in real time. For example, in the kiln temperature control, an adaptive model predictive control algorithm is used, with a sampling period of 1s, a prediction time domain of 60s, and a control time domain of 10s. The transfer function parameters are identified online through the least squares method to predict future temperature changes and optimize the control sequence. The upper limit of temperature adjustment is set to ±5℃ / min to prevent over-adjustment. In terms of dust monitoring, light scattering and β-ray dust detectors are installed at key points such as raw material processing, calcination, and screening, with a measurement range of 0-1000mg / m 3, with an accuracy of ±2%. Data was collected at a frequency of 1 Hz, outliers were removed using the 3σ criterion, and a 5-point moving average filter was applied. The processed data was transmitted to the central control system, and the C4.5 decision tree algorithm was used to build a correlation model between dust concentration and process parameters. A 1-hour sliding window was used, the model was updated every 5 minutes, and the maximum depth of the tree was set to 5 layers. When the predicted dust concentration exceeded 50 mg / m 3 For example, if the dust concentration in the calcining process is predicted to reach 55mg / m 3 , the kiln temperature will be lowered by 2°C. At the same time, a safety protection mechanism is set to limit the temperature adjustment range to no more than ±10°C / h to ensure safe and stable operation of the production process.

[0043] Dynamically monitor the dust concentration and analyze it to obtain the change curve of the dust concentration after the optimized process. Compare the dust concentration in the change curve with the process parameters adjusted at the corresponding time to form feedback, analyze the feedback error, compare the feedback error with the preset error threshold, trigger the corresponding preset dust control strategy, adjust the process parameters according to the feedback to reduce or increase the dust concentration, and achieve dust control.

[0044] The dust concentration data collected by the online dust monitor is obtained. The dust concentration data is transmitted to the central control system at a preset frequency by the online dust monitor set at the key point of the production line; the dust concentration data is preprocessed, and the preprocessing includes removing outliers by using the median filtering algorithm and applying wavelet transform for noise reduction; the preprocessed dust concentration data is processed by the autoregressive integral sliding average model to obtain a graph of dust concentration changing with time; the graph of dust concentration changing with time and the parameter data of each process are compared and analyzed by the Granger causality test tool to determine the Granger causal relationship strength between each process parameter and dust concentration; according to the Granger causal relationship strength, the adaptive fuzzy control algorithm is used to calculate the adjustment amount of the process parameters, and the adaptive fuzzy control algorithm is calculated based on the dust concentration error value and the error change rate; the adjustment amount of the process parameters is sent to each process equipment through the distributed control system to realize automatic adjustment of the process parameters.

[0045] Specifically, an online dust monitor is used to monitor the dust concentration at key points of the production line in real time, and the monitoring data is transmitted to the central control system through the data acquisition module at a frequency of 10Hz. The collected data is preprocessed, including using median filtering to remove outliers and applying wavelet transform for noise reduction. The preprocessed data is processed using the autoregressive integral moving average model to generate a graph of dust concentration changing over time. According to the dust concentration change curve, the dust concentration data is compared and analyzed with the process parameter data of each process in the same time period through the Granger causality test tool. The Granger causal relationship strength between each process parameter and dust concentration is calculated to form a quantitative index of parameter influence. The parameter influence threshold is set to screen out the key process parameters that significantly affect the dust concentration. The dust concentration error threshold is set, and the dust concentration is monitored in real time using the exponentially weighted moving average control chart. The upper control limit UCL and lower control limit LCL are calculated. When the dust concentration is detected to exceed the control limit, the preset dust control strategy is triggered. According to the parameter influence index, the process parameters that need to be adjusted are determined, and the adjustment priority is set. Based on the adaptive fuzzy control algorithm, the adjustment amount of the process parameters is calculated in combination with the dust concentration error value and the error change rate. Through the online learning mechanism, the fuzzy rule base is dynamically updated according to the historical adjustment effect. Consider the adjustment constraints of the process parameters, such as the temperature change rate limit, the pressure adjustment range, etc., to ensure that the parameters are adjusted within a safe range. The adjustment instructions are sent to each process equipment through the distributed control system to realize the automatic adjustment of the process parameters and form a closed-loop feedback mechanism for dust control. In the dust control optimization process of the calcium carbonate production line, laser scattering dust monitors are first installed in key processes such as raw material processing, calcination and screening, with a measurement range of 0-1000mg / m 3 , accuracy ±1%. Data was collected at a frequency of 10Hz and transmitted to the central control system through the OPCUA protocol. A 5-point median filter was applied to the original data to remove outliers, and then the db4 wavelet was used for 3-layer decomposition and noise reduction. The ARIMA (2,1,1) model was used to process the data and generate a dust concentration change curve. Subsequently, the relationship between dust concentration and process parameters was analyzed by the Granger causality test, with the lag order set to 10 and the significance level α = 0.05. The Granger causal strengths of calcination temperature, pressure and raw material particle size were calculated to be 0.85, 0.72 and 0.63, respectively, which exceeded the preset threshold of 0.6 and were determined to be key parameters. The dust concentration EWMA control chart was set with a smoothing coefficient λ = 0.3, a control limit coefficient L = 3, and a calculated UCL = 55 mg / m 3 , LCL = 35 mg / m 3When the dust concentration exceeds the control limit for three consecutive points, the control strategy is triggered. Based on the adaptive fuzzy control algorithm, a 7×7 fuzzy rule matrix is ​​designed. The input variables are the dust concentration error e and the error change rate ec, and the output is the parameter adjustment amount. For example, when e=+10mg / m 3 , ec = +2mg / m 3 / min, the calcination temperature adjustment amount is -5℃. The historical adjustment effect is evaluated through a sliding time window with a window size of 30 minutes, and the fuzzy rules are dynamically updated. Consider process constraints, such as limiting the temperature change rate to ±10℃ / h. Finally, the optimized parameters are sent to the on-site PLC through the ModbusTCP protocol to achieve automatic adjustment of process parameters and form a closed-loop feedback mechanism for dust control.

[0046] The above is only a specific implementation method of this specification. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described system, module and unit can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. It should be understood that the protection scope of this specification is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this specification, and these modifications or replacements should be covered within the protection scope of this specification.

Claims

1. A dust control method in a calcium carbonate manufacturing process, characterized in that: The method comprises: In the calcium carbonate manufacturing process, an infrared spectrometer is installed to monitor the chemical composition data of dust, a laser particle size analyzer is used to monitor the particle size data of dust, and a pH meter is used to monitor the pH value data of dust, so as to obtain dust monitoring data in the calcium carbonate manufacturing process. At the same time, the temperature, pressure, flow rate, and speed process parameters of raw material processing and calcination, digestion, and drying processes are obtained to construct a data set of dust monitoring data and process parameters; Preprocess the dust monitoring data and process parameter data sets, fuse data from different sources, formats, and timestamps, and normalize the data into a time series analysis format; Extracting a subset of dust monitoring data and process parameter features from the pre-processed dust monitoring data and process parameter data sets, combining dust attributes including dust chemical composition, particle size, and pH value attributes with process parameters including temperature, pressure, flow rate, and rotation speed, and screening out feature combinations; Based on the feature combination, the frequent association patterns between dust properties and process parameters are obtained. The frequent association rules are obtained by using the preset association rule mining algorithm and the preset algorithm setting parameters. Based on the analysis results of dust properties and process parameters in the time series, the dust and process association rule library is obtained. According to the dust and process association library obtained by mining, combined with the preset calcium carbonate production process flow chart, the causes of dust generation in each process are determined, the dust control points of key processes such as raw material processing and calcination, digestion and drying are determined, and process improvement measures for equipment parameters, product formula and transportation methods are formulated; Under the preset production efficiency, cost control, and quality stability constraints, optimize the equipment decision variable parameters, product formula, and transportation method to obtain a balanced solution for dust control and production management and form an improved process plan; The optimized process plan is sent to the control system of each production equipment, the temperature, pressure and flow parameters of each process are adjusted in real time, and the changes in dust concentration are dynamically monitored to form a closed-loop optimization control of dust.

2. The method according to claim 1, characterized in that: In the calcium carbonate manufacturing process, the chemical composition data of the dust is monitored by installing an infrared spectrometer, the particle size data of the dust is monitored by a laser particle size analyzer, and the pH value data of the dust is monitored by a pH meter, so as to obtain the dust monitoring data in the calcium carbonate manufacturing process, and at the same time obtain the temperature, pressure, flow rate, and speed process parameters of the raw material processing and calcination, digestion, and drying processes, and construct a data set of dust monitoring data and process parameters, including: Obtain dust chemical composition data monitored by infrared spectrometer and dust particle size data monitored by laser particle size analyzer; A dust feature matrix is ​​constructed based on the dust chemical composition data and dust particle size data. The dust feature matrix includes the chemical composition percentage and multi-dimensional feature vector of particle size distribution of each dust sample. Perform principal component analysis on the dust feature matrix and select the principal component whose cumulative contribution rate reaches the preset threshold as the key feature vector of dust; A nonlinear mapping model between process parameters and dust key feature vectors was established. The nonlinear mapping model was constructed using a random forest regression tool, in which process parameters were used as input variables and dust key feature vectors were used as target variables. The importance score of each process parameter on dust characteristics is calculated based on the nonlinear mapping model to determine the key process parameters that most significantly affect dust characteristics; A gradient boosting decision tree tool was constructed, with key process parameters and dust key feature vectors as input features and pH value as the prediction target, where the pH value was obtained by monitoring with a pH meter; The decision tree tool is used to analyze a multidimensional correlation data set including process parameters, dust characteristics and pH values, and a relationship model between process parameters and dust characteristics is established, in which process parameters are used as feature nodes, and dust characteristics and pH values ​​are used as leaf nodes.

3. The method according to claim 1, characterized in that The dust monitoring data and process parameter data sets are preprocessed to fuse data from different sources, different formats, and different time stamps, and normalize the data into a time series analysis format, including: Obtain dust monitoring data and process parameter data, use an adaptive moving median filter to perform preliminary noise removal on the data, and obtain a preliminary denoised data set; Based on the data set after preliminary denoising, the Z-score method is combined to identify outliers and determine whether the Z-score value of the data point exceeds the preset threshold range. If it exceeds, it is marked as an outlier; For outliers, the local linear regression method is used to correct them and obtain the corrected data set; The corrected data set is formatted uniformly through a data conversion tool, and the data from the infrared spectrometer, laser particle size analyzer, and pH meter are converted into a standardized JSON format to obtain a standardized data set; The Kalman filter algorithm is used to fuse multi-source data of the standardized data set to generate a unified data stream based on the measurement errors and time delays of different data sources; The timestamp alignment tool is used to synchronize the unified data stream, interpolate all data at a uniform time interval, and use the cubic spline interpolation method to fill in the missing time point data to obtain a time-aligned normalized time series data set.

4. The method according to claim 1, characterized in that Extracting a subset of dust monitoring data and process parameter features from the pre-processed dust monitoring data and process parameter data sets, combining dust attributes including dust chemical composition, particle size, and pH value attributes with process parameters including temperature, pressure, flow rate, and rotation speed, and screening out feature combinations, including: Based on the pre-processed dust monitoring data and process parameter data set, the data is standardized to obtain standardized data; The kernel principal component analysis method is used to extract the characteristic subset of the standardized data. By calculating the eigenvalues ​​and eigenvectors of the kernel matrix of the standardized data, the principal components whose cumulative contribution rate reaches a preset percentage are obtained as the characteristic subset; For the process parameter data, the maximum information coefficient is used to calculate the nonlinear correlation between temperature, pressure, flow rate and speed. If the maximum information coefficient is greater than the threshold, the parameter with the largest amount of information is selected as the representative to obtain the process parameter feature subset. Combine the dust chemical composition, particle size and pH value attributes, evaluate the relationship between the attributes and the process parameter feature subset, select the attribute with the highest mutual information value, and combine it with the process parameter feature subset to form a preliminary feature combination; The random forest feature importance assessment tool was used to screen the preliminary feature combinations, and the feature importance ranking was obtained by calculating the contribution of each feature to the model prediction accuracy; Combined with the recursive feature elimination method, the least important features are gradually removed, and the model performance is evaluated after each removal until the performance begins to decline significantly or reaches the preset feature number threshold, and the optimized feature combination is obtained.

5. The method according to claim 1, characterized in that The frequent association patterns between dust attributes and process parameters are obtained based on feature combination, frequent association rules are obtained by using a preset association rule mining algorithm and preset algorithm setting parameters, and a dust and process association rule library is obtained based on the analysis results of dust attributes and process parameters in time series, including: Perform equal-frequency discretization processing on the continuous data according to the preset number of intervals to obtain discretized feature data; The FP-growth algorithm is used to mine frequent itemsets from discretized feature data, and frequent itemsets are recursively generated by constructing FP trees and conditional FP trees. Apply the association rule generator to construct association rules based on frequent item sets; If the association rule meets the preset threshold, it is determined to be a strong association rule; Obtain a time series data set, use wavelet transform to extract the frequency domain features of the time series data in the time series data set, and obtain approximate coefficients and detail coefficients of different scales; Cluster the strongly associated rules, construct a rule similarity matrix by calculating the semantic similarity between rule items, and combine the frequency domain features to merge the rules with high similarity using a hierarchical clustering algorithm to form a dust and process association rule library; It also includes: dividing the dust and process parameter data according to the time series to form a time window, obtaining the dust attributes and process parameters in the time series analysis results according to the preset time series model, combining the mined frequent association rules, screening out the association rules that meet the preset practical application value, organizing the screened rules into a database format, and establishing a dust and process association rule library.

6. The method according to claim 5, characterized in that The method is to divide the dust and process parameter data according to the time series to form a time window, obtain the dust attribute and process parameter analysis results in the time series according to the preset time series model, combine the mined frequent association rules, screen out the association rules that meet the preset practical application value, organize the screened rules into a database format, and establish a dust and process association rule library, including: Adaptive sliding window method is used to segment the dust and process parameter data into time series, and the window size is dynamically adjusted according to the data variance in the window; If the variance exceeds the preset threshold, the window is reduced, otherwise the window is expanded; The dust properties and process parameter data within the time window are analyzed, the seasonal autoregressive integrated moving average model is used to eliminate the data non-stationarity, and the discrete wavelet transform is used to extract the multi-scale characteristics of the time series; According to the time series analysis results and the pre-mined frequent association rules, the support threshold, confidence threshold, Kulczynski index threshold and full confidence index threshold are set to determine whether the rules continue to meet all threshold conditions; Weighting the satisfying condition rules, obtaining the energy distribution and determining the weight coefficient; Convert the weighted rules into a structured data format and establish a relational database table structure including a rule table and a time window table; The rule table contains fields for rule ID, antecedent, consequent, support, confidence, lift, Kulczynski index, total confidence index, and weight coefficient; The time window table contains window ID, rule ID, start time, end time, and window feature fields.

7. The method according to claim 1, characterized in that The dust and process association library obtained by mining is combined with the preset calcium carbonate production process flow chart to determine the cause of dust generation in each process, determine the dust control points of key processes such as raw material processing and calcination, digestion, and drying, and formulate process improvement measures for equipment parameters, product formulas, and transportation methods, including: Convert the preset calcium carbonate production process flow chart into a directed graph, where nodes represent processes, edges represent associations between processes, and edge weights are set according to association rules; Calculate the shortest path between nodes in the directed graph to obtain the causes and influencing factors of dust generation in each process; According to the causes and influencing factors of dust generation, the four key processes of raw material processing, calcination, digestion and drying were evaluated using fuzzy comprehensive evaluation method combined with principal component analysis, and evaluation indicators of dust concentration, particle size distribution and chemical composition were set; The main features of the evaluation indicators are extracted through principal component analysis, the weight of each indicator is determined, and the comprehensive score of each process is calculated. The node with the highest score is determined as the dust control point; For dust control points, use multi-objective optimization algorithms to optimize equipment parameters, product formulas and transportation methods; The particle swarm optimization algorithm is combined with the simulated annealing algorithm to solve the Pareto optimal solution set of the objective function and obtain the optimal parameter combination of each control point. Based on the optimal parameter combination, a knowledge graph is used to build a process improvement measures library; The key concepts are extracted from the optimal parameter combination through entity recognition method, and the relationship between entities is identified by relation extraction; The relationships between key concepts and entities are stored in a graph database, and the graph database query language is used to perform relationship reasoning to obtain specific process improvement measures.

8. The method according to claim 1, characterized in that: Under the preset production efficiency, cost control, and quality stability constraints, the equipment decision variable parameters, product formula, and transportation method are optimized to obtain a balanced solution for dust control and production management, and to form an improved process plan, including: According to the preset production efficiency, cost control and quality stability constraints, equipment decision variable parameters, product formula and transportation method are obtained as optimization variables, and dust control degree, production efficiency, cost and product quality are used as objective functions; The multi-objective optimization model was constructed using the ε-constraint method, the dust control degree was selected as the main objective function, and other objective functions were transformed into constraint conditions; The particle swarm optimization algorithm is combined with the simulated annealing algorithm to solve the optimization model, from which the particle swarm size, maximum number of iterations, initial temperature and cooling coefficient are determined; The optimal solution is obtained globally by continuously updating the particle position, velocity and temperature parameters during annealing; Adopting an adaptive weight adjustment strategy, the weights of each sub-objective in the objective function are dynamically adjusted according to the results of each iteration; Calculate the moving window standard deviation of each sub-target and determine the sub-target with the largest standard deviation; If the standard deviation is large, it is given a high weight; A dynamic balance between objectives is achieved through the weight update formula, and the Pareto optimality of the current optimal solution is evaluated after each iteration; Conduct sensitivity analysis and Monte Carlo simulation on the solution obtained by the optimization algorithm to obtain the sensitivity coefficient of each parameter; Latin hypercube sampling is used to generate input parameter combinations, and Monte Carlo simulations are performed to evaluate the stability and reliability of the optimization results.

9. The method according to claim 1, characterized in that: The optimized process plan is sent to the control system of each production equipment, the temperature, pressure and flow parameters of each process are adjusted in real time, and the dust concentration change is dynamically monitored to form a closed-loop optimization control of dust, including: Receive optimization instructions carrying process plan parameters, which are issued by the central controller; According to the optimization instructions, the process plan parameters are sent to the programmable logic controller of each production equipment through the industrial communication network, and the programmable logic controller adjusts the operating status of the production equipment according to the process plan parameters; Set up the primary and backup channels and use cyclic redundancy check, which is used to ensure the accuracy of data transmission; If communication interruption or data transmission error is detected, it will automatically switch to the backup channel; Obtain the temperature, pressure and flow parameters of each process, and adjust the temperature, pressure and flow parameters in real time; Obtain dust concentration detection data at each key point of the production line. The dust concentration detection data is collected by various types of dust detection equipment; The dust concentration detection data is collected at a preset frequency through the data acquisition module, and the data is pre-processed; Construct a correlation model between dust concentration and process parameters; Dynamic optimization of process parameters is performed based on the correlation model, and dynamic optimization is used to adjust relevant process parameters; It also includes: dynamically monitoring dust concentration and analyzing it to obtain the change curve of dust concentration after the optimized process, comparing the dust concentration in the change curve with the parameters of each process adjusted at the corresponding time to form feedback, analyzing the error of the feedback, comparing the error of the feedback with the preset error threshold, triggering the corresponding preset dust control strategy, adjusting each process parameter according to the feedback to reduce or increase the dust concentration, and realizing dust control.

10. The method according to claim 9, characterized in that The dust concentration is dynamically monitored and analyzed to obtain a change curve of the dust concentration after the process is optimized, the dust concentration in the change curve is compared with the process parameters adjusted at the corresponding time, feedback is formed, the feedback error is analyzed, the feedback error is compared with the preset error threshold, the corresponding preset dust control strategy is triggered, and the process parameters are adjusted according to the feedback to reduce or increase the dust concentration to achieve dust control, including: Obtain dust concentration data collected by online dust monitors, which are installed at key points of the production line and transmitted to the central control system at a preset frequency; Preprocessing is performed based on dust concentration data, including using a median filter algorithm to remove outliers and applying a wavelet transform to reduce noise; The pre-processed dust concentration data were processed using an autoregressive integrated moving average model to obtain a graph of dust concentration changing over time; Granger causality test tool is used to compare and analyze the dust concentration change curve with the process parameter data to determine the Granger causality strength between each process parameter and dust concentration. According to the strength of Granger causality, the adaptive fuzzy control algorithm is used to calculate the adjustment amount of process parameters. The adaptive fuzzy control algorithm is calculated based on the dust concentration error value and error change rate. The adjustment amount of process parameters is sent to each process equipment through the distributed control system to realize automatic adjustment of process parameters.

Images