An intelligent monitoring system and method for fly ash production line based on multi-source data

Abstract

The present invention discloses an intelligent monitoring system and method for a fly ash production line based on multi-source data, which relates to the technical field of data analysis. The present invention calculates the initial component score of fly ash, determines the initial grade of fly ash, and determines characteristic records through historical monitoring records; calculates the environmental score, determines the characteristic process parameter set of the processing procedure, and establishes a classification model for the processing procedure; constructs prediction models for the finished product score and the environmental score respectively, classifies and establishes a minimization objective function according to the label values ​​of the characteristic records; judges the label values ​​based on the real-time composition and preparation goals of the fly ash in combination with the classification model, constructs corresponding objective functions respectively, evaluates and ranks the objective function values ​​of candidate processing procedures, and selects the procedure with the smallest objective function value as the optimal processing procedure, thereby improving the safety, resource utilization and operation and maintenance efficiency of fly ash treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to an intelligent monitoring system and method for a fly ash production line based on multi-source data. Background Art

[0002] With the deepening advancement of industrial informatization and intelligent manufacturing, data-driven production management concepts have been widely applied in many key areas such as environmental governance and hazardous waste treatment. This is especially true in the high-risk and complex process of fly ash treatment, which involves multiple physical and chemical reactions, multi-stage coordinated control, and a high emphasis on environmental compliance requirements.

[0003] However, fly ash has complex sources and highly variable composition. Even the fly ash produced daily by the same incinerator can vary significantly in moisture, pH, metal content, and other aspects. Different treatment methods require different types of additives, ratios, and reaction sequences. Treatment effectiveness is highly dependent on accurate identification of raw material properties and precise execution of process control.

[0004] At present, most fly ash treatment production lines still adopt a management method that combines segmented control with manual intervention, and are unable to dynamically adjust the feeding strategy according to the characteristics of the raw materials. In order to cope with the actual needs of fly ash treatment, which has large raw material fluctuations, complex processes and high control requirements, an intelligent monitoring system and method for fly ash production lines based on multi-source data is proposed. The system realizes a production line-level monitoring system that integrates multi-source perception, intelligent judgment and automatic control, thereby improving the safety, resource utilization and operation and maintenance efficiency of fly ash treatment. Summary of the Invention

[0005] The object of the present invention is to provide an intelligent monitoring system and method for a fly ash production line based on multi-source data to solve the problems raised in the prior art.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an intelligent monitoring method for a fly ash production line based on multi-source data, the method comprising:

[0007] Step S100: Calculate the initial composition score of fly ash through historical monitoring records, determine the initial grade of fly ash, and determine characteristic records;

[0008] Step S200: Calculate the environmental score, determine the characteristic process parameter set of the processing procedure, and establish a classification model for the processing procedure;

[0009] Step S300: Construct prediction models for finished product scores and environmental scores respectively, classify them according to the label values ​​of feature records and establish a minimization objective function;

[0010] Step S400: Based on the real-time composition and preparation target of fly ash, the label value is judged in combination with the classification model, and the corresponding objective functions are constructed respectively. The objective function values ​​of the candidate processing steps are evaluated and ranked, and the step with the smallest objective function value is selected as the optimal processing step.

[0011] Furthermore, step S100 includes:

[0012] Step S101: In the fly ash production line monitoring system, the type of finished product to be produced from the fly ash is obtained, the initial component content of the fly ash is collected, and the processing parameters of each processing step are monitored. After all processing steps are completed, the finished product is tested and a monitoring record is generated;

[0013] Step S102: Obtain historical monitoring records, summarize the historical monitoring records of the same finished product, collect the initial component content of the fly ash, and calculate the initial component score of the fly ash according to the following formula:

[0014] ;

[0015] Among them, G represents the initial composition score of fly ash, H g Expressed as the content of the initial component of item g, K g It is represented as the weight of the g-th initial component, and h is represented as the total number of initial component types;

[0016] Step S103: Summarize the fly ash initial composition scores of all historical monitoring records, preset several initial levels, determine the fly ash initial composition score interval corresponding to each initial level, and summarize the historical monitoring records of a certain initial level;

[0017] Step S104: In the historical monitoring records of a certain initial level, the test result parameters of the finished product prepared in a certain historical monitoring record are collected, and the test results are normalized and calculated to calculate the finished product test score according to the following formula:

[0018] ;

[0019] Among them, A represents the finished product inspection score, B a It is expressed as the ath normalized detection result parameter, C a It represents the weight of the a-th test result, and b represents the total number of test results;

[0020] Step S105: Preset a finished product inspection score threshold, set historical monitoring records exceeding the finished product inspection score threshold as feature records, and summarize the feature records of a certain finished product;

[0021] By weighting the raw material composition and finished product quality, an objective and quantifiable data foundation is achieved. The fly ash is divided into grades according to the initial composition score, which helps to optimize the process path of raw materials of different grades. By comparing characteristic records, it is possible to identify which processing parameter combinations are optimal for specific raw material grades, thus building a data support foundation for the intelligent system of fly ash processing, which is conducive to automated parameter adjustment and process self-optimization.

[0022] Furthermore, step S200 includes:

[0023] Step S201: Obtain a feature record set of a certain finished product, collect environmental indicator parameters of a certain processing step in a certain feature record, and calculate the environmental score according to the following formula:

[0024] ;

[0025] Among them, S represents the environmental score, D d Expressed as the value of the dth environmental indicator parameter, E d It is expressed as the weight of the dth environmental indicator parameter, and e is expressed as the total number of environmental indicator parameters;

[0026] Step S202: Preset an environmental score threshold, set feature records that exceed the environmental score threshold as abnormal records and mark them with a label value of 1, and mark feature records that do not exceed the environmental score threshold with a label value of 0;

[0027] Step S203: Obtain process parameters from the abnormal record and construct a process parameter vector P = (P1, P2, ..., Pn) T , where P1, P2, ...Pn represent the 1st, 2nd, ...nth process parameters respectively. A certain process parameter is combined with the environmental score, the Pearson correlation coefficient of the process parameter is calculated, a Pearson correlation coefficient threshold is set, and the process parameter exceeding the Pearson correlation coefficient threshold is set as a characteristic parameter to obtain a characteristic process parameter set of a certain processing step;

[0028] Step S204: Obtain all feature records corresponding to a certain processing step and establish a training data set , where Xi and Qi represent the fly ash initial composition score and characteristic process parameter value set of the i-th feature record, respectively; Li represents the label value of the i-th feature record; N represents the total number of feature records corresponding to a certain processing step; a classification model f_L(X,Q)=L corresponding to a certain processing step is established, where L∈{0,1}. X and Q in each training data set are used as input values, and L is used as the output value. The classification model corresponding to a certain processing step is obtained through training through a random forest model;

[0029] Through the environmental scoring and labeling mechanism, we can quickly identify processing records that may have environmental risks, build an anomaly database, and use Pearson correlation analysis to accurately screen out the variables with the greatest environmental impact from a variety of process parameters, providing a basis for process optimization.

[0030] The random forest model has nonlinear modeling capabilities and strong generalization capabilities. It can accurately identify under what conditions anomalies are likely to occur, assist in early intervention, and adjust key process parameters in advance through model prediction results, reduce pollutant emissions, and improve the environmental protection level of the production line.

[0031] Furthermore, step S300 includes:

[0032] Step S301: Aggregate the feature record sets of the same initial level, combine the fly ash initial composition score and the feature process parameter set of the feature record with the finished product inspection score to form a finished product inspection score data set, aggregate the finished product inspection score data sets of all feature records, establish a finished product inspection score prediction model f_A(X,Q)=A, use X and Q in each finished product inspection score data set as input values, and A as output value, train the model through a linear regression model, and obtain the finished product inspection score prediction model;

[0033] Step S302: Establish the objective function formula with a label value of 0: J1(Q)=[A'-f_A(X,Q)] 2 , where J1(Q) is the objective function value with a label value of 0, and A' represents the preset target score value;

[0034] Step S303: Obtain a feature record with a tag value of 1 for a certain processing step, combine the ash initial component score, the feature process parameter set, and the environmental score to form an environmental score data set, summarize the environmental score data sets of all feature records, and establish an environmental score prediction model f_S(X,Q)=S. Use X and Q in each environmental score data set as input values ​​and S as output value, and train the model through a linear regression model to obtain the environmental score prediction model.

[0035] Step S304: Establish the objective function formula with a label value of 1: J2(Q)=[A'-f_A(X,Q)] 2 +α×max[0,f_S(X,Q)-S'] 2 , where J2(Q) is the objective function value with a label value of 1, S' represents the preset environmental score value, and α represents the weight of the environmental score;

[0036] Instead of pursuing a single goal such as output or quality, environmental risks are quantified as part of the model, supporting the design of an optimization path with finished product scores as the target and environmental scores as the constraint;

[0037] Compared to complex black box models such as neural networks, linear regression provides higher interpretability and can clearly understand the positive and negative impact of each characteristic process parameter on the finished product score and environmental score;

[0038] It supports the production system to switch the objective function according to the real-time tag value and adapt to different operation stages. The "optimal process parameter vector" can be calculated through minimization, which is conducive to giving automatic parameter adjustment suggestions for closed-loop control systems.

[0039] Furthermore, step S400 includes:

[0040] Step S401: obtaining the type of finished product to be prepared from the real-time fly ash, collecting the real-time initial component content of the fly ash, calculating the real-time initial component score, and determining the real-time initial grade of the fly ash;

[0041] Step S402: collecting real-time process parameters of a certain processing procedure executed by fly ash, inputting the real-time initial composition score of the fly ash and the currently set characteristic process parameter value set into the classification model of the processing procedure; if the label value is 1, then in a plurality of candidate processing procedures, respectively inputting the characteristic process parameter value set into the objective function formula with the label value of 1, and calculating the objective function value of each candidate processing procedure; if the label value is 0, then in a plurality of candidate processing procedures, respectively inputting the characteristic process parameter value set into the objective function formula with the label value of 0, and calculating the objective function value of each candidate processing procedure;

[0042] Step S403: sorting all objective function values, and selecting the candidate processing step corresponding to the minimum objective function value as the optimal step;

[0043] The system no longer relies on fixed processes, but instead intelligently recommends the optimal processing procedure based on real-time raw materials and working conditions. This can address the practical problems of large fluctuations and complex composition of fly ash raw materials, enabling flexible manufacturing.

[0044] Using a classification model to determine whether environmental risks exist, environmental compliance is transformed from post-processing to pre-processing. Using the penalty term in the objective function, the probability of selecting potentially polluting processes is effectively suppressed.

[0045] Adding dual constraints on finished product scores and environmental scores to the objective function makes the optimization path more scientific and reasonable. All candidate processes can be included in the same algorithm framework for evaluation, realizing automatic process scheduling control.

[0046] Traditional process selection relies on operator experience, while this method relies on data modeling and algorithm judgment to improve consistency and response speed.

[0047] In order to better implement the above method, an intelligent monitoring system for fly ash production line based on multi-source data is proposed. The system includes a feature recording module, a classification model module, an objective function module and a real-time sorting module.

[0048] Characteristic record module: calculates the initial composition score of fly ash through historical monitoring records, determines the initial grade of fly ash, and determines characteristic records;

[0049] Classification model module: calculates environmental scores, determines the characteristic process parameter set of the processing procedure, and establishes a classification model for the processing procedure;

[0050] Objective function module: Build prediction models for finished product scores and environmental scores respectively, classify according to the label values ​​of feature records and establish the minimization objective function;

[0051] Real-time sorting module: Based on the real-time composition and preparation goals of fly ash, combined with the classification model to determine the label value, the corresponding objective function is constructed respectively, the objective function value of the candidate processing steps is evaluated and sorted, and the process with the smallest objective function value is selected as the optimal processing step.

[0052] Furthermore, the feature recording module includes an initial level determination unit and a feature recording determination unit:

[0053] Determine the initial grade unit: In the fly ash production line monitoring system, obtain the type of finished product to be prepared from fly ash, collect the initial component content of fly ash, monitor the processing parameters of each processing step, and after all processing steps are completed, test the prepared finished product to generate a monitoring record, obtain historical monitoring records, summarize the historical monitoring records of the preparation of the same finished product, collect the initial component content of fly ash, calculate the initial component score of fly ash, summarize the initial component scores of fly ash from all historical monitoring records, preset several initial grades, and determine the initial component score range of fly ash corresponding to each initial grade;

[0054] Determine the characteristic record unit: in the historical monitoring records of a certain initial level, collect the test result parameters of the finished product prepared in a certain historical monitoring record, normalize the test results, calculate the finished product inspection score, preset the finished product inspection score threshold, and set the historical monitoring records that exceed the finished product inspection score threshold as characteristic records.

[0055] Furthermore, the classification model module includes an abnormal record determination unit and a classification model establishment unit:

[0056] Determine abnormal record units: obtain a set of feature records for a certain finished product, collect environmental indicator parameters for a certain processing step in a feature record, calculate the environmental score, preset the environmental score threshold, and set feature records that exceed the environmental score threshold as abnormal records;

[0057] Establish a classification model unit: obtain process parameters in abnormal records, construct a process parameter vector, combine a certain process parameter with the environmental score, calculate the Pearson correlation coefficient of the process parameter, set the Pearson correlation coefficient threshold, set the process parameters that exceed the Pearson correlation coefficient threshold as feature parameters, obtain a feature process parameter set of a certain processing step, obtain all feature records corresponding to a certain processing step, establish a training data set, input the training data set into the random forest model, and obtain a classification model corresponding to a certain processing step.

[0058] Furthermore, the objective function module includes establishing an objective function unit with a label value of 0 and establishing an objective function unit with a label value of 1:

[0059] Establish an objective function unit with a label value of 0: summarize the feature record sets of the same initial level, combine the fly ash initial composition score and the feature process parameter set of the feature records with the finished product inspection score to form a finished product inspection score data group, summarize the finished product inspection score data group of all feature records and input it into the linear regression model, establish a finished product inspection score prediction model, and establish an objective function formula with a label value of 0;

[0060] Establish an objective function unit with a label value of 1: obtain a feature record with a label value of 1 for a certain processing step, combine the ash initial composition score and the feature process parameter set with the environmental score to form an environmental score data group, summarize the environmental score data groups of all feature records and input them into the linear regression model, establish an environmental score prediction model, and establish an objective function formula with a label value of 1.

[0061] Furthermore, the real-time sorting module includes determining the real-time initial level unit and selecting the optimal process unit:

[0062] A real-time initial grade determination unit: obtaining the type of finished product to be prepared from the real-time fly ash, collecting the real-time initial component content of the fly ash, calculating the real-time initial component score, and determining the real-time initial grade of the fly ash;

[0063] Select the optimal process unit: collect the real-time process parameters of a certain processing process of fly ash, input the real-time initial composition score of the fly ash and the current set characteristic process parameter value set into the classification model of the processing process, if the label value is 1, then in multiple candidate processing processes, respectively input the characteristic process parameter value set into the objective function formula with the label value of 1, and calculate the objective function value of each candidate processing process; if the label value is 0, then in multiple candidate processing processes, respectively input the characteristic process parameter value set into the objective function formula with the label value of 0, and calculate the objective function value of each candidate processing process, sort all the objective function values, and select the candidate processing process corresponding to the minimum objective function value as the optimal process.

[0064] Compared with existing technologies, the present invention has the following advantages: it establishes a closed-loop process from raw material analysis → finished product testing → environmental assessment → process selection, and has multiple intelligent functions such as adaptive modeling, anomaly identification, parameter tuning, and process scheduling, thus realizing the transformation of traditional fly ash treatment production lines into intelligent, green, and data-based ones;

[0065] Traditional optimization methods often only focus on a single objective such as output or quality. This invention proposes that when the label is 0, only the finished product score objective function is optimized, and when the label is 1, a multi-objective function of finished product score + environmental constraints is optimized. This method has the ability to flexibly respond to different working conditions and automatically switch optimization logic.

[0066] By marking anomalies through environmental scoring and thresholds, and using classification models to predict whether the current parameter combination may be abnormal, we can achieve pre-judgment, real-time warning, and in-process intervention, thereby improving safety and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a flow chart of an intelligent monitoring method for a fly ash production line based on multi-source data according to the present invention;

[0068] Figure 2 This is a structural schematic diagram of an intelligent monitoring system for fly ash production lines based on multi-source data according to the present invention. DETAILED DESCRIPTION

[0069] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0070] See also Figure 1 and Figure 2 The present invention provides a technical solution: an intelligent monitoring method for a fly ash production line based on multi-source data, the method comprising:

[0071] Step S100: Calculate the initial composition score of fly ash through historical monitoring records, determine the initial grade of fly ash, and determine characteristic records;

[0072] Wherein, step S100 includes:

[0073] Step S101: In the fly ash production line monitoring system, the type of finished product to be produced from the fly ash is obtained, the initial component content of the fly ash is collected, and the processing parameters of each processing step are monitored. After all processing steps are completed, the finished product is tested and a monitoring record is generated;

[0074] Step S102: Obtain historical monitoring records, summarize the historical monitoring records of the same finished product, collect the initial component content of the fly ash, and calculate the initial component score of the fly ash according to the following formula:

[0075] ;

[0076] Among them, G represents the initial composition score of fly ash, H g Expressed as the content of the initial component of item g, K g It is represented as the weight of the g-th initial component, and h is represented as the total number of initial component types;

[0077] Step S103: Summarize the fly ash initial composition scores of all historical monitoring records, preset several initial levels, determine the fly ash initial composition score interval corresponding to each initial level, and summarize the historical monitoring records of a certain initial level;

[0078] Step S104: In the historical monitoring records of a certain initial level, the test result parameters of the finished product prepared in a certain historical monitoring record are collected, and the test results are normalized and calculated to calculate the finished product test score according to the following formula:

[0079] ;

[0080] Among them, A represents the finished product inspection score, B a It is expressed as the ath normalized detection result parameter, C a It represents the weight of the a-th test result, and b represents the total number of test results;

[0081] Step S105: Preset a finished product inspection score threshold, set historical monitoring records exceeding the finished product inspection score threshold as feature records, and summarize the feature records of a certain finished product;

[0082] For example, in the first monitoring record, the finished product type is green building bricks, the initial composition of fly ash is 4.2 Cl, 1.3 Pb, 0.9 Zn, and 0.01 Hg; the temperature in process 1 is 120°C and the time is 30 min; the additive ratio in process 2 is 8%, and the pH value is controlled at 9.5; the heating temperature in process 3 is 200°C and the wind speed is 5 m / s; the compressive strength is 0.88, the water absorption rate is 0.72, and the tail gas heavy metal residue is 0.65. Assume that the weight of compressive strength is 0.4, the weight of water absorption rate is 0.3, and the weight of tail gas heavy metal residue is 0.3;

[0083] Assuming the weight of Cl is 0.3, the weight of Pb is 0.3, the weight of Zn is 0.2, and the weight of Hg is 0.2, the initial composition score of fly ash is calculated to be 1.832;

[0084] The calculated finished product inspection score is 0.763, and the preset finished product score threshold is 0.75, so the first monitoring record is a feature record.

[0085] Step S200: Calculate the environmental score, determine the characteristic process parameter set of the processing procedure, and establish a classification model for the processing procedure;

[0086] Wherein, step S200 includes:

[0087] Step S201: Obtain a feature record set of a certain finished product, collect environmental indicator parameters of a certain processing step in a certain feature record, and calculate the environmental score according to the following formula:

[0088] ;

[0089] Among them, S represents the environmental score, D d Expressed as the value of the dth environmental indicator parameter, E d It is expressed as the weight of the dth environmental indicator parameter, and e is expressed as the total number of environmental indicator parameters;

[0090] Step S202: Preset an environmental score threshold, set feature records that exceed the environmental score threshold as abnormal records and mark them with a label value of 1, and mark feature records that do not exceed the environmental score threshold with a label value of 0;

[0091] Step S203: Obtain process parameters from the abnormal record and construct a process parameter vector P = (P1, P2, ..., Pn) T , where P1, P2, ...Pn represent the 1st, 2nd, ...nth process parameters respectively. A certain process parameter is combined with the environmental score, the Pearson correlation coefficient of the process parameter is calculated, a Pearson correlation coefficient threshold is set, and the process parameter exceeding the Pearson correlation coefficient threshold is set as a characteristic parameter to obtain a characteristic process parameter set of a certain processing step;

[0092] Step S204: Obtain all feature records corresponding to a certain processing step and establish a training data set , where Xi and Qi represent the fly ash initial composition score and characteristic process parameter value set of the i-th feature record, respectively; Li represents the label value of the i-th feature record; N represents the total number of feature records corresponding to a certain processing step; a classification model f_L(X,Q)=L corresponding to a certain processing step is established, where L∈{0,1}. X and Q in each training data set are used as input values, and L is used as the output value. The classification model corresponding to a certain processing step is obtained through training through a random forest model;

[0093] For example, in the first characteristic record, the discharge heavy metal concentration is 1.8 mg / L, the weight is 0.4, the tail gas ammonia nitrogen concentration is 36.0 ppm, the weight is 0.3, and the pH deviation after solidification is 0.6, the weight is 0.3. The calculated environmental score is 11.7.

[0094] The preset environmental score threshold is 10, so the first feature record is an abnormal record and the label value is 1.

[0095] Step S300: Construct prediction models for finished product scores and environmental scores respectively, classify them according to the label values ​​of feature records and establish a minimization objective function;

[0096] Wherein, step S300 includes:

[0097] Step S301: Aggregate the feature record sets of the same initial level, combine the fly ash initial composition score and the feature process parameter set of the feature record with the finished product inspection score to form a finished product inspection score data set, aggregate the finished product inspection score data sets of all feature records, establish a finished product inspection score prediction model f_A(X,Q)=A, use X and Q in each finished product inspection score data set as input values, and A as output value, train the model through a linear regression model, and obtain the finished product inspection score prediction model;

[0098] Step S302: Establish the objective function formula with a label value of 0: J1(Q)=[A'-f_A(X,Q)] 2 , where J1(Q) is the objective function value with a label value of 0, and A' represents the preset target score value;

[0099] Step S303: Obtain a feature record with a tag value of 1 for a certain processing step, combine the ash initial component score, the feature process parameter set, and the environmental score to form an environmental score data set, summarize the environmental score data sets of all feature records, and establish an environmental score prediction model f_S(X,Q)=S. Use X and Q in each environmental score data set as input values ​​and S as output value, and train the model through a linear regression model to obtain the environmental score prediction model.

[0100] Step S304: Establish the objective function formula with a label value of 1: J2(Q)=[A'-f_A(X,Q)] 2 +α×max[0,f_S(X,Q)-S'] 2 , where J2(Q) is the objective function value with a label value of 1, S' represents the preset environmental score value, and α represents the weight of the environmental score.

[0101] Step S400: Based on the real-time composition and preparation target of fly ash, the label value is judged in combination with the classification model, and the corresponding objective functions are constructed respectively. The objective function values ​​of the candidate processing steps are evaluated and ranked, and the step with the smallest objective function value is selected as the optimal processing step.

[0102] Wherein, step S400 includes:

[0103] Step S401: obtaining the type of finished product to be prepared from the real-time fly ash, collecting the real-time initial component content of the fly ash, calculating the real-time initial component score, and determining the real-time initial grade of the fly ash;

[0104] Step S402: collecting real-time process parameters of a certain processing procedure executed by fly ash, inputting the real-time initial composition score of the fly ash and the currently set characteristic process parameter value set into the classification model of the processing procedure; if the label value is 1, then in a plurality of candidate processing procedures, respectively inputting the characteristic process parameter value set into the objective function formula with the label value of 1, and calculating the objective function value of each candidate processing procedure; if the label value is 0, then in a plurality of candidate processing procedures, respectively inputting the characteristic process parameter value set into the objective function formula with the label value of 0, and calculating the objective function value of each candidate processing procedure;

[0105] Step S403: sorting all objective function values, and selecting the candidate processing step corresponding to the minimum objective function value as the optimal step;

[0106] For example, the real-time initial component content of fly ash is cl 3.8, Pb is 1.5, Zn is 1.1, and Hg is 0.02. The calculated real-time initial component score is 1.814, and the real-time initial grade of the fly ash is determined to be level 2;

[0107] Get the current set of characteristic process parameter values, input the classification model, and get a label value of 1. You need to use the objective function formula with a label value of 1.

[0108] Assume that the numerical set of characteristic process parameters of the candidate processing steps is as shown in Table 1:

[0109] Table 1

[0110]

[0111] The calculated objective function of process A is 0.1357, the objective function of process B is 0.6775, and the objective function of process C is 0.000676. Process C is selected as the optimal processing process for the current fly ash batch.

[0112] In order to better implement the above method, an intelligent monitoring system for fly ash production line based on multi-source data is proposed. The system includes a feature recording module, a classification model module, an objective function module and a real-time sorting module.

[0113] Characteristic record module: calculates the initial composition score of fly ash through historical monitoring records, determines the initial grade of fly ash, and determines characteristic records;

[0114] The feature recording module includes an initial level determination unit and a feature recording determination unit:

[0115] Determine the initial grade unit: In the fly ash production line monitoring system, obtain the type of finished product to be prepared from fly ash, collect the initial component content of fly ash, monitor the processing parameters of each processing step, and after all processing steps are completed, test the prepared finished product to generate a monitoring record, obtain historical monitoring records, summarize the historical monitoring records of the preparation of the same finished product, collect the initial component content of fly ash, calculate the initial component score of fly ash, summarize the initial component scores of fly ash from all historical monitoring records, preset several initial grades, and determine the initial component score range of fly ash corresponding to each initial grade;

[0116] Determine the characteristic record unit: in the historical monitoring records of a certain initial level, collect the test result parameters of the finished product prepared in a certain historical monitoring record, normalize the test results, calculate the finished product inspection score, preset the finished product inspection score threshold, and set the historical monitoring records that exceed the finished product inspection score threshold as characteristic records.

[0117] Classification model module: calculates environmental scores, determines the characteristic process parameter set of the processing procedure, and establishes a classification model for the processing procedure;

[0118] The classification model module includes an abnormal record determination unit and a classification model establishment unit:

[0119] Determine abnormal record units: obtain a set of feature records for a certain finished product, collect environmental indicator parameters for a certain processing step in a feature record, calculate the environmental score, preset the environmental score threshold, and set feature records that exceed the environmental score threshold as abnormal records;

[0120] Establish a classification model unit: obtain process parameters in abnormal records, construct a process parameter vector, combine a certain process parameter with the environmental score, calculate the Pearson correlation coefficient of the process parameter, set the Pearson correlation coefficient threshold, set the process parameters that exceed the Pearson correlation coefficient threshold as feature parameters, obtain a feature process parameter set of a certain processing step, obtain all feature records corresponding to a certain processing step, establish a training data set, input the training data set into the random forest model, and obtain a classification model corresponding to a certain processing step.

[0121] Objective function module: Build prediction models for finished product scores and environmental scores respectively, classify according to the label values ​​of feature records and establish the minimization objective function;

[0122] The objective function module includes an objective function unit with a label value of 0 and an objective function unit with a label value of 1:

[0123] Establish an objective function unit with a label value of 0: summarize the feature record sets of the same initial level, combine the fly ash initial composition score and the feature process parameter set of the feature records with the finished product inspection score to form a finished product inspection score data group, summarize the finished product inspection score data group of all feature records and input it into the linear regression model, establish a finished product inspection score prediction model, and establish an objective function formula with a label value of 0;

[0124] Establish an objective function unit with a label value of 1: obtain a feature record with a label value of 1 for a certain processing step, combine the ash initial composition score and the feature process parameter set with the environmental score to form an environmental score data group, summarize the environmental score data groups of all feature records and input them into the linear regression model, establish an environmental score prediction model, and establish an objective function formula with a label value of 1.

[0125] Real-time sorting module: Based on the real-time composition and preparation goals of fly ash, combined with the classification model to determine the label value, the corresponding objective function is constructed, the objective function value of the candidate processing steps is evaluated and sorted, and the process with the smallest objective function value is selected as the optimal processing step

[0126] The real-time sorting module includes determining the real-time initial grade unit and selecting the optimal process unit:

[0127] A real-time initial grade determination unit: obtaining the type of finished product to be prepared from the real-time fly ash, collecting the real-time initial component content of the fly ash, calculating the real-time initial component score, and determining the real-time initial grade of the fly ash;

[0128] Select the optimal process unit: collect the real-time process parameters of a certain processing process of fly ash, input the real-time initial composition score of the fly ash and the current set characteristic process parameter value set into the classification model of the processing process, if the label value is 1, then in multiple candidate processing processes, respectively input the characteristic process parameter value set into the objective function formula with the label value of 1, and calculate the objective function value of each candidate processing process; if the label value is 0, then in multiple candidate processing processes, respectively input the characteristic process parameter value set into the objective function formula with the label value of 0, and calculate the objective function value of each candidate processing process, sort all the objective function values, and select the candidate processing process corresponding to the minimum objective function value as the optimal process.

[0129] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. An intelligent monitoring method for fly ash production line based on multi-source data, characterized in that: Methods include: Step S100: Calculate the initial composition score of fly ash through historical monitoring records, determine the initial grade of fly ash, and determine characteristic records; Step S200: Calculate the environmental score, determine the characteristic process parameter set of the processing procedure, and establish a classification model for the processing procedure; Step S300: Construct prediction models for finished product scores and environmental scores respectively, classify them according to the label values ​​of feature records and establish a minimization objective function; Furthermore, step S300 includes the following steps: Step S301: Aggregate the feature record sets of the same initial level, combine the fly ash initial composition score and the feature process parameter set of the feature record with the finished product inspection score to form a finished product inspection score data set, aggregate the finished product inspection score data sets of all feature records, establish a finished product inspection score prediction model f_A(X,Q)=A, use X and Q in each finished product inspection score data set as input values, and A as output value, train the model through a linear regression model, and obtain the finished product inspection score prediction model; Step S302: Establish the objective function formula with a label value of 0: J1(Q)=[A'-f_A(X,Q)] 2 , where J1(Q) is the objective function value with a label value of 0, and A' represents the preset target score value; Step S303: Obtain a feature record with a tag value of 1 for a certain processing step, combine the ash initial component score, the feature process parameter set, and the environmental score to form an environmental score data set, summarize the environmental score data sets of all feature records, and establish an environmental score prediction model f_S(X,Q)=S. Use X and Q in each environmental score data set as input values ​​and S as output value, and train the model through a linear regression model to obtain the environmental score prediction model. Step S304: Establish the objective function formula with a label value of 1: J2(Q)=[A'-f_A(X,Q)] 2 +α×max[0,f_S(X,Q)-S'] 2 , where J2(Q) is the objective function value with a label value of 1, S' represents the preset environmental score value, and α represents the weight of the environmental score; Step S400: Based on the real-time composition of fly ash and the preparation target, the label value is determined in combination with the classification model, and the corresponding objective functions are constructed. The objective function values ​​of the candidate processing steps are evaluated and ranked, and the step with the smallest objective function value is selected as the optimal processing step; Furthermore, step S400 includes the following steps: Step S401: obtaining the type of finished product to be prepared from the real-time fly ash, collecting the real-time initial component content of the fly ash, calculating the real-time initial component score, and determining the real-time initial grade of the fly ash; Step S402: collecting real-time process parameters of a certain processing procedure executed by fly ash, inputting the real-time initial composition score of the fly ash and the currently set characteristic process parameter value set into the classification model of the processing procedure; if the label value is 1, then in a plurality of candidate processing procedures, respectively inputting the characteristic process parameter value set into the objective function formula with the label value of 1, and calculating the objective function value of each candidate processing procedure; if the label value is 0, then in a plurality of candidate processing procedures, respectively inputting the characteristic process parameter value set into the objective function formula with the label value of 0, and calculating the objective function value of each candidate processing procedure; Step S403: sorting all objective function values, and selecting the candidate processing step corresponding to the minimum objective function value as the optimal step.

2. The intelligent monitoring method for fly ash production line based on multi-source data according to claim 1 is characterized in that: The step S100 includes the following steps: Step S101: In the fly ash production line monitoring system, the type of finished product to be produced from the fly ash is obtained, the initial component content of the fly ash is collected, and the processing parameters of each processing step are monitored. After all processing steps are completed, the finished product is tested and a monitoring record is generated; Step S102: Obtain historical monitoring records, summarize the historical monitoring records of the same finished product, collect the initial component content of the fly ash, and calculate the initial component score of the fly ash according to the following formula: ； Among them, G represents the initial composition score of fly ash, H g Expressed as the content of the initial component of item g, K g It is represented as the weight of the g-th initial component, and h is represented as the total number of initial component types; Step S103: Summarize the fly ash initial composition scores of all historical monitoring records, preset several initial levels, determine the fly ash initial composition score interval corresponding to each initial level, and summarize the historical monitoring records of a certain initial level; Step S104: In the historical monitoring records of a certain initial level, the test result parameters of the finished product prepared in a certain historical monitoring record are collected, and the test results are normalized and calculated to calculate the finished product test score according to the following formula: ； Among them, A represents the finished product inspection score, B a It is expressed as the ath normalized detection result parameter, C a It represents the weight of the a-th test result, and b represents the total number of test results; Step S105: preset a finished product detection score threshold, set the historical monitoring records that exceed the finished product detection score threshold as feature records, and summarize the feature records of a certain finished product.

3. The intelligent monitoring method for fly ash production line based on multi-source data according to claim 2 is characterized in that: The step S200 includes the following steps: Step S201: Obtain a feature record set of a certain finished product, collect environmental indicator parameters of a certain processing step in a certain feature record, and calculate the environmental score according to the following formula: ； Among them, S represents the environmental score, D d Expressed as the value of the dth environmental indicator parameter, E d It is expressed as the weight of the dth environmental indicator parameter, and e is expressed as the total number of environmental indicator parameters; Step S202: Preset an environmental score threshold, set feature records that exceed the environmental score threshold as abnormal records and mark them with a label value of 1, and mark feature records that do not exceed the environmental score threshold with a label value of 0; Step S203: Obtain process parameters from the abnormal record and construct a process parameter vector P = (P1, P2, ..., Pn) T , where P1, P2, ...Pn represent the 1st, 2nd, ...nth process parameters respectively. A certain process parameter is combined with the environmental score, the Pearson correlation coefficient of the process parameter is calculated, a Pearson correlation coefficient threshold is set, and the process parameter exceeding the Pearson correlation coefficient threshold is set as a characteristic parameter to obtain a characteristic process parameter set of a certain processing step; Step S204: Obtain all feature records corresponding to a certain processing step and establish a training data set , where Xi and Qi represent the fly ash initial composition score and characteristic process parameter numerical value set of the i-th feature record, respectively, Li represents the label value of the i-th feature record, and N represents the total number of feature records corresponding to a certain processing step. A classification model f_L(X,Q)=L corresponding to a certain processing step is established, where L∈{0,1}. X and Q in each training data set are used as input values, and L is used as the output value. The classification model corresponding to a certain processing step is obtained through training through the random forest model.

4. An intelligent monitoring system for a fly ash production line based on multi-source data, used to implement the intelligent monitoring method for a fly ash production line based on multi-source data according to any one of claims 1 to 3, characterized in that: The system includes a feature recording module, a classification model module, an objective function module and a real-time sorting module; The characteristic recording module calculates the initial composition score of the fly ash through historical monitoring records, determines the initial grade of the fly ash, and determines the characteristic record; The classification model module calculates the environmental score, determines the characteristic process parameter set of the processing procedure, and establishes a classification model for the processing procedure; The objective function module: constructs prediction models for finished product scores and environmental scores respectively, and classifies and establishes a minimization objective function based on the label values ​​of feature records; The real-time sorting module: based on the real-time composition and preparation objectives of fly ash, combined with the classification model to determine the label value, respectively constructs the corresponding objective function, evaluates and sorts the objective function values ​​of the candidate processing steps, and selects the step with the smallest objective function value as the optimal processing step.

5. The fly ash production line intelligent monitoring system based on multi-source data according to claim 4 is characterized in that: The feature recording module includes an initial level determination unit and a feature recording determination unit: The initial grade determination unit: in the fly ash production line monitoring system, obtains the type of finished product to be prepared from the fly ash, collects the initial component content of the fly ash, monitors the processing parameters of each processing step, and after all processing steps are completed, detects the prepared finished product, generates a monitoring record, obtains historical monitoring records, summarizes the historical monitoring records of the preparation of the same type of finished product, collects the initial component content of the fly ash, calculates the initial component score of the fly ash, summarizes the initial component scores of all historical monitoring records, presets a number of initial grades, and determines the initial component score range of the fly ash corresponding to each initial grade; The characteristic record determination unit collects test result parameters of a finished product prepared from a certain historical monitoring record in a certain initial level of historical monitoring records, performs normalized calculations on the test results, calculates a finished product test score, presets a finished product test score threshold, and sets historical monitoring records that exceed the finished product test score threshold as characteristic records.

6. The fly ash production line intelligent monitoring system based on multi-source data according to claim 5 is characterized in that: The classification model module includes an abnormal record determination unit and a classification model establishment unit: The abnormal record determination unit: obtains a feature record set of a certain finished product, collects environmental indicator parameters of a certain processing step in a certain feature record, calculates an environmental score, presets an environmental score threshold, and sets feature records exceeding the environmental score threshold as abnormal records; The classification model establishment unit: obtains process parameters in the abnormal record, constructs a process parameter vector, combines a certain process parameter with the environmental score, calculates the Pearson correlation coefficient of the process parameter, sets the Pearson correlation coefficient threshold, sets the process parameters exceeding the Pearson correlation coefficient threshold as feature parameters, obtains a feature process parameter set of a certain processing step, obtains all feature records corresponding to a certain processing step, establishes a training data set, inputs the training data set into the random forest model, and obtains a classification model corresponding to a certain processing step.

7. The fly ash production line intelligent monitoring system based on multi-source data according to claim 5 is characterized in that: The objective function module includes an objective function unit for establishing a label value of 0 and an objective function unit for establishing a label value of 1: The unit for establishing an objective function with a label value of 0 comprises: summarizing a set of feature records of the same initial level, combining the fly ash initial composition score and the feature process parameter set of the feature records with the finished product inspection score to form a finished product inspection score data group, summarizing the finished product inspection score data group of all feature records and inputting the data into a linear regression model, establishing a finished product inspection score prediction model, and establishing an objective function formula with a label value of 0; The unit for establishing an objective function with a label value of 1 is as follows: a feature record with a label value of 1 for a certain processing step is obtained, the ash initial component score and the feature process parameter set and the environmental score are combined into an environmental score data group, the environmental score data groups of all feature records are summarized and input into a linear regression model, an environmental score prediction model is established, and an objective function formula with a label value of 1 is established.

8. The fly ash production line intelligent monitoring system based on multi-source data according to claim 5 is characterized in that: The real-time sorting module includes determining the real-time initial level unit and selecting the optimal process unit: The real-time initial grade determination unit: obtains the type of finished product to be prepared from the real-time fly ash, collects the real-time initial component content of the fly ash, calculates the real-time initial component score, and determines the real-time initial grade of the fly ash; The optimal process selection unit collects the real-time process parameters of a certain processing process performed by fly ash, inputs the real-time initial composition score of the fly ash and the currently set characteristic process parameter value set into the classification model of the processing process; if the label value is 1, then in multiple candidate processing processes, the characteristic process parameter value set is respectively input into the objective function formula with the label value of 1, and the objective function value of each candidate processing process is calculated; if the label value is 0, then in multiple candidate processing processes, the characteristic process parameter value set is respectively input into the objective function formula with the label value of 0, and the objective function value of each candidate processing process is calculated; all objective function values ​​are sorted, and the candidate processing process corresponding to the minimum objective function value is selected as the optimal process.

Images