Method, device and equipment for predicting removal efficiency of organic pollutants in fly ash pyrolysis

Abstract

According to the method, the device and the equipment for predicting the removal efficiency of the organic pollutants in the fly ash pyrolysis, provided by the invention, the removal efficiency of the organic pollutants in the fly ash pyrolysis process is rapidly and accurately predicted by utilizing a machine learning method, the removal level of the organic pollutants in the fly ash can be determined without an experiment, the labor cost and the time cost are greatly saved, and the working efficiency is improved. The optimal numerical value condition of the fly ash treatment process parameters can be quantitatively determined, and practical application is effectively guided; on the other hand, the organic pollutant removal efficiency can be automatically predicted in a one-key mode according to the imported fly ash parameters, pyrolysis parameters and organic pollutant characteristic data, redundant experiments are avoided, and popularization and application of a trained prediction model in engineering practice are facilitated; and thirdly, the problems that time is consumed and resources are dense when a traditional experiment method is used for processing complex interaction of multiple variables in a traditional experiment can be solved.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the technical field of environmental protection, and relates to a fly ash pyrolysis organic pollutant removal efficiency prediction method, device and equipment. BACKGROUND

[0002] Among various technologies for solid waste safe disposal, sanitary landfill and incineration are still the most widely used methods. Compared with landfill, waste incineration (MSWI, Municipal Solid Waste Incineration) reduces about 90% of the volume and about 70% of the mass. In the incineration process, two types of residues are inevitably produced: municipal solid waste incineration fly ash (MSWIFA) and bottom ash. Fly ash refers to fine particles collected from the flue gas treatment system or deposited on the inner wall of the flue. The amount of fly ash produced by municipal solid waste incineration can reach 3-5 wt% of the original waste volume. MSWIFA is usually in the form of gray or dark gray powder, rich in persistent organic pollutants (POPs) such as polychlorinated dibenzo-p-dioxins and dibenzofurans (dioxins) and polychlorinated biphenyls (PCBs), and heavy metals such as lead, cadmium and chromium. These pollutants have high toxicity, environmental persistence and bioaccumulation, and through the migration of water, soil and food chain, they have significant health risks such as carcinogenicity, teratogenicity and genetic mutation. Municipal solid waste incineration fly ash belongs to hazardous waste, and the effective removal of organic pollutants is still a key technical challenge for hazardous waste management.

[0003] To solve this problem, various treatment methods have been developed. Traditional thermal treatment includes sintering fly ash at high temperature to promote the decomposition of organic pollutants, and the residues produced thereby can be used to produce glass, ceramics or concrete aggregate. When the temperature exceeds 650℃, the toxicity equivalent of dioxins can be significantly reduced or even eliminated. In addition, pyrolysis under an inert or anaerobic atmosphere at a temperature below 600℃ provides a route with lower energy consumption and simpler operation requirements. Low-temperature pyrolysis treatment can handle a large amount of municipal solid waste incineration fly ash and is not restricted by location, and is a new technology for degrading dioxins during the current disposal of municipal solid waste incineration fly ash. Hydrothermal treatment takes advantage of the unique subcritical and supercritical properties of water at 200-374℃ and 4-22MPa to promote the dissolution and hydrolysis of organic pollutants. In addition, mechanical chemical degradation uses mechanical forces such as ball milling to induce physical and chemical transformations in fly ash, promoting the decomposition of pollutants without the need for high temperatures. Pyrolysis stands out with its high efficiency, simple operation, wide applicability and other advantages, and shows good potential for practical application, and is more in line with the goal of low-carbon emission reduction. However, the efficiency of pyrolysis is affected by multiple factors, including pyrolysis parameters (temperature, residence time, type and flow rate of carrier gas), fly ash properties (ash content, fixed carbon, volatile matter content) and pollutant properties (initial concentration, melting point, logK ow ).

[0004] Although traditional experimental methods can reveal the trend of changes in individual factors, they are often time-consuming and labor-intensive when dealing with complex interactions of multiple variables. Full-factorial design or orthogonal design may not effectively capture nonlinear interactions. Currently, the main method for treating fly ash by pyrolysis is through artificial variable control experiments to determine the removal efficiency of organic pollutants under different influencing factor conditions. This method requires a large amount of time and labor costs, and the removal efficiency of organic pollutants is difficult to provide effective guidance for fly ash pyrolysis processes in actual engineering applications. If a fast, accurate and highly interpretable method for predicting the removal efficiency of organic pollutants in fly ash pyrolysis can be developed, it will be of practical significance for further optimizing large-scale industrial fly ash pyrolysis processes and improving the removal efficiency of organic pollutants in fly ash. SUMMARY

[0005] The embodiments of the present application provide a method, device and equipment for predicting the removal efficiency of organic pollutants in fly ash pyrolysis, to solve the technical drawbacks in the prior art that the removal efficiency of organic pollutants under different influencing factor conditions still needs to be determined by artificial variable control experiments.

[0006] The technical solutions of the embodiments of the present application are as follows:

[0007] The embodiments of the present application provide a method for predicting the removal efficiency of organic pollutants in fly ash pyrolysis, which comprises:

[0008] obtain an original sample data set in the fly ash pyrolysis process; the original sample data set includes a plurality of characteristic parameters affecting the removal efficiency of organic pollutants in the fly ash pyrolysis process;

[0009] sequentially perform cleaning filtering and standardization processing on the original sample data set to obtain a prediction sample data set; and split the prediction sample data set into a training set and a test set;

[0010] based on the training set and the test set, an initial organic pollutant removal efficiency prediction model is constructed through an extreme gradient boosting machine learning algorithm;

[0011] based on the feature importance analysis in the initial organic pollutant removal efficiency prediction model, the plurality of characteristic parameters are sorted in importance to obtain an importance sorting result, and the least important feature is deleted;

[0012] a variance inflation factor value of each characteristic parameter is calculated to obtain a variance inflation factor sorting result, so as to identify the characteristic parameter most affected by multicollinearity, and the characteristic with the largest variance inflation factor is deleted;

[0013] based on a preset Pearson correlation coefficient and the plurality of characteristic parameters, a feature correlation matrix is constructed; the feature correlation matrix is used to measure the correlation between different characteristic parameters;

[0014] based on a preset hierarchical clustering algorithm, the plurality of characteristic parameters are iteratively synthesized to obtain a hierarchical clustering result;

[0015] based on the importance sorting result and the variance inflation factor value sorting result, the feature correlation result and the hierarchical clustering result are analyzed, the characteristic with high importance and small variance inflation factor value is preferentially retained, and then the target organic pollutant removal efficiency prediction model is determined through the filtered input characteristic;

[0016] based on the target organic pollutant removal efficiency prediction model, the removal efficiency of organic pollutants in the to-be-pyrolyzed process is predicted to obtain a target efficiency prediction result.

[0017] An embodiment of the present application provides a device for predicting the removal efficiency of organic pollutants in fly ash pyrolysis, which comprises:

[0018] an obtaining module, configured to obtain an original sample data set in the fly ash pyrolysis process; the original sample data set includes a plurality of characteristic parameters affecting the removal efficiency of organic pollutants in the fly ash pyrolysis process;

[0019] a processing module, configured to sequentially perform cleaning filtering and standardization processing on the original sample data set to obtain a prediction sample data set; and split the prediction sample data set into a training set and a test set;

[0020] a construction module, configured to construct an initial organic pollutant removal efficiency prediction model by an extreme gradient boosting machine learning algorithm based on the training set and the test set;

[0021] an analysis module, configured to sort importance of the plurality of feature parameters based on feature importance analysis in the initial organic pollutant removal efficiency prediction model, to obtain an importance sorting result, and to delete a feature with minimum importance;

[0022] a calculation module, configured to calculate a variance inflation factor value of each of the feature parameters, to obtain a variance inflation factor sorting result, to identify a feature parameter most affected by multicollinearity, and to delete a feature with maximum variance inflation factor;

[0023] The construction module is further configured to construct a feature correlation matrix based on a preset Pearson correlation coefficient and the plurality of feature parameters; and the feature correlation matrix is configured to measure correlation between different feature parameters.

[0024] an iteration module, configured to iteratively synthesize the plurality of feature parameters based on a preset hierarchical clustering algorithm, to obtain a hierarchical clustering result;

[0025] a determination module, configured to analyze the feature correlation result and the hierarchical clustering result based on the importance sorting result and the variance inflation factor value sorting result, to preferentially retain a feature with high importance and small variance inflation factor value, and to determine a target organic pollutant removal efficiency prediction model through the filtered input features;

[0026] a prediction module, configured to predict an organic pollutant removal efficiency in a pyrolysis process based on the target organic pollutant removal efficiency prediction model, to obtain a target efficiency prediction result.

[0027] In some embodiments, the processing module is further configured to identify and delete an outlier value and a duplicate record value in the original sample data set by using a preset box plot and a preset database constraint rule, to obtain an original sample data set after deletion; and to perform multiple imputations on the original sample data set after deletion by using a preset chain equation multiple imputation method, to obtain the prediction sample data set.

[0028] In some embodiments, the construction module is further configured to calculate a correlation coefficient value between each two feature parameters based on a Pearson correlation coefficient calculation formula, and to fill the correlation coefficient value into a corresponding position of the feature correlation matrix; to sort absolute values of the correlation coefficient values, to determine a feature parameter with an absolute value greater than 0.5 as having strong correlation, and to obtain a two-by-two combined variable with strong correlation.

[0029] In some embodiments, the iteration module is further configured to calculate the Euclidean distance between each two feature parameters in the data set; and based on a preset average distance method, combine two feature parameters with the smallest distance into a new cluster, and calculate the Euclidean distance between the new cluster and other clusters; repeat the above process to construct a tree structure; and based on the tree structure, sort the Euclidean distances to obtain the hierarchical clustering result.

[0030] In some embodiments, the plurality of feature parameters include fly ash property parameters, pyrolysis parameters and organic pollutant features; the fly ash property parameters include fixed carbon content, ash content, volatile matter content, SiO2 content, Cl content, CaO content, Fe2O3 content, Al2O3 content, Cd content, Cr content and fly ash mass; the pyrolysis parameters include temperature, time, pretreatment, gas flow and gas type; and the organic pollutant features include boiling point, melting point, concentration, Henry's constant and octanol-water partition coefficient.

[0031] In some embodiments, the fixed carbon content > temperature > concentration > pretreatment > time > volatile matter content > melting point > boiling point > octanol-water partition coefficient > Henry's constant > Cl content > Cd content > ash content > SiO2 content > CaO content > gas flow > fly ash mass > Fe2O3 content > Al2O3 content > gas type > Cr content.

[0032] The embodiment of the present application provides a fly ash pyrolysis organic pollutant removal efficiency prediction device, including: a memory, used for storing executable instructions; and a processor, used for executing the executable instructions stored in the memory to realize the fly ash pyrolysis organic pollutant removal efficiency prediction method.

[0033] The embodiment of the present application provides a computer readable storage medium, which stores executable instructions, and is used for causing a processor to execute the executable instructions to realize the fly ash pyrolysis organic pollutant removal efficiency prediction method.

[0034] The one or more technical solutions provided in the embodiment of the present application have at least the following technical effects or advantages:

[0035] The application realizes rapid and accurate prediction of the removal efficiency of organic pollutants in the fly ash pyrolysis process by using a machine learning method, determines the removal level of the organic pollutants in the fly ash without experiments, greatly saves the labor cost and time cost, and quantitatively determines the optimal numerical conditions of the fly ash disposal process parameters, effectively guiding the actual application. On the other hand, the application can automatically predict the removal efficiency of the organic pollutants according to the imported fly ash parameters, pyrolysis parameters and organic pollutant characteristic data, avoid redundant experiments, and facilitate the popularization and application of the trained prediction model in engineering practice. Thirdly, the application can overcome the problem that the traditional experimental method is time-consuming and resource-intensive when dealing with complex interactions of multiple variables. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a structural schematic diagram of a fly ash pyrolysis organic pollutant removal efficiency prediction system provided by the application;

[0037] Figure 2 is a flowchart of a fly ash pyrolysis organic pollutant removal efficiency prediction method provided by the application;

[0038] Figure 3 is an input feature correlation heat map of an organic pollutant removal efficiency prediction model provided by the application;

[0039] Figure 4 is a (a) prediction model input feature hierarchical clustering tree diagram and (b) variance inflation factor diagram provided by the application;

[0040] Figure 5 is a (a) comparison diagram of the actual value of the organic pollutant removal efficiency and the XGB model prediction value, and (b) comparison diagram of the actual value of the organic pollutant removal efficiency and the optimized XGB model prediction value provided by the application;

[0041] Figure 6 is a component structure schematic diagram of a fly ash pyrolysis organic pollutant removal efficiency prediction device provided by the application;

[0042] Figure 7 is a component structure schematic diagram of an electronic device provided by the application. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical scheme and advantages of the application clearer, the application will be further described in detail below in combination with the drawings. The described embodiments should not be regarded as limiting the application, and all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0044] In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same as or different from each other and can be combined with each other, without conflict. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of the present application belong. The terminology used in the description of the embodiments of the present application is for the purpose of describing the embodiments of the present application only and is not intended to be limiting of the present application.

[0045] The following describes an exemplary application of the fly ash pyrolysis organic pollutant removal efficiency prediction device provided by the embodiments of the present application. The fly ash pyrolysis organic pollutant removal efficiency prediction device provided by the embodiments of the present application can be implemented as a terminal or a server. In one implementation, the fly ash pyrolysis organic pollutant removal efficiency prediction device provided by the embodiments of the present application can be implemented as various types of terminals such as a notebook computer, a tablet computer, a desktop computer, a mobile device, etc. In another implementation, the fly ash pyrolysis organic pollutant removal efficiency prediction device provided by the embodiments of the present application can also be implemented as a server. The server can be a standalone physical server, a server cluster composed of multiple physical servers, or a distributed system. The server can also be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms. The terminal and the server can be directly or indirectly connected through wired or wireless communication, which is not limited in the embodiments of the present application. In the following, an exemplary application of the fly ash pyrolysis organic pollutant removal efficiency prediction device as a server will be described.

[0046] Referring to Figure 1 , Figure 1 is a structural schematic diagram of a fly ash pyrolysis organic pollutant removal efficiency prediction system 10 provided by the embodiments of the present application. The embodiments of the present application can provide a fly ash pyrolysis organic pollutant removal efficiency prediction platform, which can be implemented as a fly ash pyrolysis organic pollutant removal efficiency prediction application. The fly ash pyrolysis organic pollutant removal efficiency prediction system 10 provided by the embodiments of the present application includes a terminal 110, a network 120, and a server 130. The server 130 is a server of the fly ash pyrolysis organic pollutant removal efficiency prediction application. The server 130 can constitute a fly ash pyrolysis organic pollutant removal efficiency prediction device of the embodiments of the present application. The terminal 110 is connected to the server 130 through the network 120. The network 120 can be a wide area network or a local area network, or a combination of the two.

[0047] In some embodiments, referring to Figure 1 In predicting the removal efficiency of organic pollutants in the fly ash pyrolysis process, the terminal 110 sends an initiated organic pollutant removal efficiency prediction task to the server 130 through the network 120, and the server 130 acquires an original sample data set in the fly ash pyrolysis process in response to the organic pollutant removal efficiency prediction task sent by the terminal 110; the original sample data set is sequentially subjected to cleaning filtering and standardization processing to obtain a prediction sample data set; and the prediction sample data set is split into a training set and a test set; based on the training set and the test set, an initial organic pollutant removal efficiency prediction model is constructed through an extreme gradient boosting machine learning algorithm; based on the feature importance analysis in the initial organic pollutant removal efficiency prediction model, the importance of a plurality of feature parameters is sorted to obtain an importance sorting result, and the least important feature is deleted; the variance inflation factor value of each feature parameter is calculated to obtain a variance inflation factor sorting result, so as to identify the feature parameter most affected by multicollinearity and delete the feature with the largest variance inflation factor; based on a preset Pearson correlation coefficient and a plurality of feature parameters, a feature correlation matrix is constructed; based on a preset hierarchical clustering algorithm, a plurality of feature parameters are iteratively synthesized to obtain a hierarchical clustering result; based on the importance sorting result and the variance inflation factor value sorting result, the feature correlation result and the hierarchical clustering result are analyzed, the features with high importance and small variance inflation factor value are preferentially retained, and then the target organic pollutant removal efficiency prediction model is determined through the screened input features; based on the target organic pollutant removal efficiency prediction model, the removal efficiency of organic pollutants in the pyrolysis process is predicted to obtain a target efficiency prediction result. After obtaining the target efficiency prediction result, the server 130 sends the target efficiency prediction result to the terminal 110 through the network 120.

[0048] The embodiment of the present application provides a fly ash pyrolysis organic pollutant removal efficiency prediction method, which is shown in Figure 2 , Figure 2 is a flowchart of a fly ash pyrolysis organic pollutant removal efficiency prediction method provided by the embodiment of the present application, which will be described in combination with Figure 2 the steps shown in the figure.

[0049] In step S210, an original sample data set in the fly ash pyrolysis process is acquired; the original sample data set includes a plurality of feature parameters affecting the removal efficiency of organic pollutants in the fly ash pyrolysis process.

[0050] In some embodiments, the original sample data set includes fly ash property parameters, pyrolysis parameters and organic pollutant characteristics; the fly ash property parameters include: fixed carbon content, ash content, volatile matter content, SiO2 content, Cl content, CaO content, Fe2O3 content, Al2O3 content, Cd content, Cr content and fly ash mass; the pyrolysis parameters include: temperature, time, pretreatment, gas flow rate and gas type; the organic pollutant characteristics include: boiling point, melting point, concentration, Henry's constant and octanol-water partition coefficient.

[0051] Step S220 , cleaning, filtering and standardizing the original sample data set in sequence to obtain a predicted sample data set; and splitting the predicted sample data set into a training set and a test set.

[0052] In some embodiments, cleaning and filtering is to process erroneous, invalid or redundant data in the original sample data to improve data quality.

[0053] In some embodiments, during the data collection process, certain characteristic parameters of some samples may not be recorded, resulting in missing values. Common processing methods include deleting samples with excessive missing values; using the mean, median, or mode to fill missing values ​​for numerical and categorical data; more complex multiple imputation methods, such as Multiple Imputation by Chained Equations (MICE), can also be used to estimate and fill missing values ​​using information from other relevant features. Outliers are data points that deviate significantly from the normal range and may be caused by measurement errors or real extreme cases. Through boxplot analysis, data points below Q1-1.5IQR or above Q3+1.5IQR can be considered outliers, and outlier samples can be deleted, or outliers can be corrected or converted. Methods based on statistical distributions (such as the 3σ principle) or machine learning algorithms (such as isolation forests) can also be used to detect and process outliers. Handling erroneous values: For erroneous data that is obviously illogical or exceeds a reasonable range, such as a negative pyrolysis temperature, corrections can be made based on the actual situation. If the correct value cannot be determined, it will be treated as a missing value.

[0054] In some embodiments, after cleaning, filtering, and normalization, the data quality of the original sample dataset is improved, and the data format and scale are unified, forming a prediction sample dataset. This dataset is complete, accurate, and in a standard format, meeting the requirements of machine learning model training and prediction.

[0055] Step S230: Based on the training set and the test set, an initial organic pollutant removal efficiency prediction model is constructed by using an extreme gradient boosting machine learning algorithm.

[0056] Step S240, based on the feature importance analysis in the initial organic pollutant removal efficiency prediction model, the importance of the plurality of feature parameters is sorted to obtain an importance sorting result, and the feature with the smallest importance is deleted.

[0057] In some embodiments, the feature importance analysis is a method for evaluating the influence of each feature parameter on the prediction result of the machine learning model. In the organic pollutant removal efficiency prediction model constructed based on the XGBoost algorithm, by statistically analyzing and calculating the contribution of each feature parameter to the reduction of the loss function when the decision tree node is split during the model training process, the importance score of each feature parameter can be obtained. The higher the importance score, the greater the influence of the feature parameter on the prediction of the organic pollutant removal efficiency, and the more critical the role played in the model. Through feature importance analysis, it can be understood which feature parameters are most important for the prediction of organic pollutant removal efficiency, thereby providing a basis for subsequent feature selection, removing those features that have less influence on the prediction result, simplifying the model structure, and improving the efficiency and generalization ability of the model.

[0058] Step S250, calculating the variance inflation factor value of each feature parameter to obtain a variance inflation factor sorting result, to identify the feature parameter most affected by multicollinearity, and delete the feature with the largest variance inflation factor.

[0059] In some embodiments, the variance inflation factor (VIF) is an index for measuring the degree of multicollinearity between independent variables in a regression model. In the organic pollutant removal efficiency prediction model, when there is a strong linear correlation (i.e., multicollinearity) between multiple feature parameters, it will cause the variance of the model parameter estimation to increase, reducing the stability and prediction accuracy of the model. The VIF value evaluates the degree of multicollinearity by calculating the linear correlation between each feature parameter and other feature parameters. The larger the VIF value, the stronger the multicollinearity between the feature parameter and other feature parameters. Generally, when the VIF value is greater than 10, it indicates that there is a serious multicollinearity problem. By calculating the VIF value of each feature parameter and sorting it, the feature parameter most affected by multicollinearity can be identified, and then these features can be processed during model construction, such as deleting the feature with the largest variance inflation factor, to eliminate the adverse effects of multicollinearity on the model and improve the performance of the model.

[0060] Step S260, based on the plurality of feature parameters and a preset Pearson correlation coefficient, a feature correlation matrix is constructed; the feature correlation matrix is used to measure the correlation between different feature parameters.

[0061] In some embodiments, the feature correlation matrix is a matrix for measuring the correlation between different feature parameters. In the feature correlation matrix constructed based on the preset Pearson correlation coefficient, the elements of the matrix represent the Pearson correlation coefficients between different feature parameters. The Pearson correlation coefficient is a statistical measure of the degree of linear correlation between two variables, which takes a value ranging from -1 to 1. When the correlation coefficient is 1, it indicates that there is a complete positive linear correlation between the two variables; when the correlation coefficient is -1, it indicates that there is a complete negative linear correlation between the two variables; when the correlation coefficient is 0, it indicates that there is no linear correlation between the two variables. By constructing the feature correlation matrix, the strength and direction of the correlation between the feature parameters can be intuitively understood, providing an important reference for subsequent feature selection and model optimization. For example, if the correlation coefficient between two feature parameters is close to 1 or -1, it indicates that there is a strong linear correlation between them, and there may be information redundancy in the model, so one of the features can be considered for retention to simplify the model structure.

[0062] Step S270, based on a preset hierarchical clustering algorithm, iteratively synthesizing the plurality of feature parameters to obtain a hierarchical clustering result.

[0063] In the present application, the preset hierarchical clustering algorithm is a method for iteratively synthesizing a plurality of feature parameters according to pre-set clustering criteria and parameters. It calculates the similarity or distance between feature parameters, and gradually merges feature parameters with high similarity or close distance into clusters to form a hierarchical clustering tree. In each iteration, the algorithm selects the two closest clusters according to certain rules (such as the minimum distance method, the maximum distance method, the average distance method, etc.) to merge them until the preset termination condition (such as reaching a specified number of clusters or the distance between clusters reaching a certain threshold) is met. Through the hierarchical clustering algorithm, feature parameters with similar properties or similar effects on the removal efficiency of organic pollutants can be classified into one category, thereby helping to analyze the internal relationship between the feature parameters and providing more in-depth information for feature selection and model construction.

[0064] Step S280, based on the importance ranking result and the variance inflation factor value ranking result, analyzing the feature correlation result and the hierarchical clustering result, preferentially retaining features with high importance and small variance inflation factor values, and then determining the target organic pollutant removal efficiency prediction model through the selected input features.

[0065] In some embodiments, the target organic pollutant removal efficiency prediction model is a final model obtained after a series of feature screening and optimization on the initial organic pollutant removal efficiency prediction model. During the construction process, through comprehensive analysis of the feature importance ranking result, the variance inflation factor value ranking result, the feature correlation result and the hierarchical clustering result, the feature parameters with high importance, small variance inflation factor value and weak correlation with each other are preferentially retained. These screened input features can more effectively reflect the key information related to the organic pollutant removal efficiency in the fly ash pyrolysis process, remove redundant and unimportant features, make the model structure more concise and reasonable, and significantly improve the prediction accuracy and generalization ability. The target model can accurately predict the removal efficiency of organic pollutants according to the input characteristic parameters of the fly ash to be pyrolyzed in the actual fly ash pyrolysis process, and provide important decision basis for the optimization and control of the pyrolysis process.

[0066] Step S290, based on the target organic pollutant removal efficiency prediction model, the removal efficiency of organic pollutants in the pyrolysis process is predicted to obtain a target efficiency prediction result.

[0067] In some embodiments, the target efficiency prediction result is the predicted value of the removal efficiency of organic pollutants in the fly ash pyrolysis process output by the model after inputting the characteristic parameters of the fly ash to be pyrolyzed into the target organic pollutant removal efficiency prediction model. This prediction result reflects the expected value of the removal efficiency of organic pollutants predicted by the target model under the given fly ash properties, pyrolysis conditions and organic pollutant characteristics.

[0068] In practical application, the prediction result can help the operator to understand the removal effect of organic pollutants in the fly ash pyrolysis process in advance, and adjust the pyrolysis process parameters such as temperature, time, gas flow, etc. according to the prediction result, so as to achieve better organic pollutant removal effect, at the same time, it is also helpful to evaluate the feasibility and optimization potential of the pyrolysis process, and provide strong support for the development and application of fly ash pyrolysis treatment technology.

[0069] The present application uses machine learning method to realize the rapid and accurate prediction of the removal efficiency of organic pollutants in the fly ash pyrolysis process, which can determine the removal level of organic pollutants in fly ash without experiment, greatly saving the labor cost and time cost, and can quantitatively determine the best numerical condition of fly ash disposal process parameters, effectively guiding the practical application. On the other hand, the present application can automatically predict the removal efficiency of organic pollutants according to the imported fly ash parameters, pyrolysis parameters and organic pollutant characteristic data, avoiding redundant experiments, and facilitating the popularization and application of the trained prediction model in engineering practice. Thirdly, the present application can overcome the problem that the traditional experimental method is time-consuming and resource-intensive when dealing with complex interactions of multiple variables.

[0070] In addition, in the existing fly ash pyrolysis harmless process, the removal efficiency of organic pollutants is improved by adjusting the pyrolysis process parameters to maximize the removal efficiency. However, the optimal parameter matching required for fly ash of different properties is different, which makes the adjustment process time-consuming and labor-intensive, and the optimization of the pyrolysis process lacks guidance. The optimal removal efficiency prediction model proposed in the present application integrates the influencing factors of fly ash properties, pyrolysis parameters and organic pollutant characteristics, can provide guidance for the pyrolysis parameter matching in the removal process of specific organic pollutants in fly ash of different properties, and maximize the removal efficiency.

[0071] In some embodiments, the above step S220 can be implemented by the following steps S221 to S222:

[0072] Step S221, using a preset box plot and a preset database constraint rule, identifying and deleting outliers and duplicate record values in the original sample data set to obtain a deleted original sample data set.

[0073] Step S222, using a preset chain equation multiple imputation method, multiple imputing the deleted original sample data set to obtain the prediction sample data set.

[0074] In some embodiments, the above step S260 can be implemented by the following steps S261 to S262:

[0075] Step S261, calculating the correlation coefficient value between each two feature parameters based on the Pearson correlation coefficient calculation formula, and filling the correlation coefficient value into the corresponding position of the feature correlation matrix.

[0076] Step S262, based on the absolute value of the correlation coefficient value, sorting the absolute value greater than 0.5 to determine the feature parameters with strong correlation, and obtaining the two-by-two combined variables with strong correlation.

[0077] In some embodiments, the above step S270 can be implemented by the following steps S271 to S273:

[0078] Step S271, calculating the Euclidean distance between each two feature parameters in the data set; and based on a preset average distance method, merging the two feature parameters with the smallest distance into a new cluster, and calculating the Euclidean distance between the new cluster and other clusters.

[0079] Step S272, repeating the above process to construct a tree structure.

[0080] Step S273, based on the tree structure, sorting the Euclidean distances to obtain the hierarchical clustering result.

[0081] In some embodiments, the plurality of characteristic parameters include fly ash property parameters, pyrolysis parameters, and organic pollutant characteristics; the fly ash property parameters include fixed carbon content, ash content, volatile matter content, SiO2 content, Cl content, CaO content, Fe2O3 content, Al2O3 content, Cd content, Cr content, and fly ash mass; the pyrolysis parameters include temperature, time, pretreatment, gas flow, and gas type; and the organic pollutant characteristics include boiling point, melting point, concentration, Henry's constant, and octanol-water partition coefficient.

[0082] In some embodiments, the organic pollutant removal efficiency is defined as an output characteristic, and is represented by the following formula: In the formula, C t C0are the residual concentration and initial concentration (pg / g) at time t, respectively.

[0083] In some embodiments, the importance ranking result is fixed carbon content > temperature > concentration > pretreatment > time > volatile matter content > melting point > boiling point > octanol-water partition coefficient > Henry's constant > Cl content > Cd content > ash content > SiO2 content > CaO content > gas flow > fly ash mass > Fe2O3 content > Al2O3 content > gas type > Cr content.

[0084] In the following, an exemplary application of the embodiments of the present application in an actual application scenario will be described.

[0085] The present application provides a fly ash thermal process organic pollutant removal efficiency prediction model optimization method, which adopts systematic feature engineering and model optimization strategy to improve the prediction performance of machine learning model, and specifically includes the following steps:

[0086] An original sample data set and historical organic pollutant removal efficiency data in a fly ash pyrolysis process are obtained, and the plurality of characteristic parameters include fly ash property parameters, pyrolysis parameters, and organic pollutant characteristics, a total of 21 characteristics; the fly ash property parameters include fixed carbon content, ash content, volatile matter content, SiO2 content, Cl content, CaO content, Fe2O3 content, Al2O3 content, Cd content, Cr content, and fly ash mass; the pyrolysis parameters include temperature, time, pretreatment, gas flow, and gas type; and the organic pollutant characteristics include boiling point, melting point, concentration, Henry's constant, and octanol-water partition coefficient (logKow), and a prediction sample data set is obtained; the prediction sample data set is split into a training set and a test set, and an organic pollutant removal efficiency prediction model is established by an extreme gradient boosting (XGB) machine learning algorithm.

[0087] Feature importance analysis in machine learning models aims to quantify the contribution of input features to the model's prediction results. The basic principle is to use statistical or mathematical techniques to assess the degree of association between inputs and outputs. Model-based methods use internal information generated during model training to calculate feature importance. For example, in tree-based models, the frequency of selecting splits is used as a statistical feature of the total number of split nodes (weights) in the decision tree. The higher the frequency, the greater the importance. Information gain refers to the average or total information gain caused by feature partitioning, such as the reduction in Gini coefficient or loss function (gain). Greater gains imply greater importance. Coverage measures the number of samples affected by feature partitioning or the sum of second-order derivatives (Cover), which is suitable for handling high Gini coefficient features. Through model feature importance analysis, the importance of each feature is fixed carbon > temperature > concentration > pretreatment > time > volatile matter > melting point > boiling point > logKow > Henry's constant > Cl content > Cd content > ash > SiO2 content > CaO > gas flow > fly ash mass > Fe2O3 content > Al2O3 content > gas type > Cr content.

[0088] A feature correlation matrix is constructed based on the Pearson correlation coefficient (PCC). PCC is used to quantify the linear correlation between two variables selected for model construction. Correlation analysis helps understand whether there is an association between different variables and the strength of the association, which is measured by calculating the Pearson correlation coefficient.

[0089] Figure 3 The heatmap in FIG. 6 is constructed based on the PCC matrix, revealing the linear correlation between variables. Red indicates a negative correlation, and blue indicates a positive correlation. The saturation of the color increases with the absolute value of the correlation coefficient, and the white area corresponds to feature pairs with no significant correlation. This study found that there was a significant positive correlation between the three components of fly ash (ash, fixed carbon, and volatile matter), indicating that these components would change in a coordinated manner. The content of Cd and Cr showed strong correlation with other variables, reflecting their strong covariance with other variables, indicating that dimensionality reduction should be considered. The PCC of CaO and SiO2 presents a uniform light tone across the entire row and column, with an absolute correlation coefficient of less than 0.4, indicating independence from other variables. The correlation between boiling point and melting point is highly consistent, indicating potential collinearity.

[0090] Hierarchical clustering is an unsupervised machine learning method that builds a hierarchical structure by iteratively merging or splitting data points. It generates a tree-like clustering structure through recursive grouping based on pairwise distances. The algorithm first computes a distance matrix between all samples, then iteratively merges the closest clusters according to a specified linkage criterion, continuing the process until all samples are grouped into one cluster. The resulting hierarchy is visualized using a dendrogram, where the height of each node reflects the distance of cluster linkage, indicating the degree of dissimilarity. The essence of hierarchical clustering lies in revealing the intrinsic structure of data through distance metrics and merging strategies, providing an interpretable approach for unsupervised data analysis.

[0091] The Variance Inflation Factor (VIF) is commonly used to detect multicollinearity in multiple linear regression models. It quantifies the extent to which the variance of a regression coefficient is inflated due to linear correlation between one independent variable and others. The higher the VIF, the stronger the multicollinearity. As a core diagnostic tool in regression analysis, VIF facilitates the identification of multicollinearity by measuring linear relationships between variables, thereby enhancing the interpretability and predictive stability of the model. By calculating the VIF value for each feature, highly collinear variables can be detected and eliminated, contributing to improved model robustness.

[0092] The hierarchical clustering results are shown in FIG. 10A. Figure 4 a. Gas flow is grouped with melting point, ash content is grouped with time, Cr content is grouped with Henry’s constant, Al2O3 content is grouped with pretreatment, and Cl and mass are grouped. Figure 4 b. It is shown that ash content exhibits the highest VIF value, indicating strong multicollinearity with other variables. Cd and Cr exhibit strong correlation. However, since Cr is less important, Cr is removed while keeping Cd. The correlation between Al2O3 content and Fe2O3 content is similar, but Fe2O3 content exhibits lower multicollinearity, so Fe2O3 content is retained. Melting point and boiling point both reflect the volatility and migration ability of organic pollutants, so boiling point is removed while keeping melting point since boiling point has a higher VIF. Gas flow rate is removed since it is grouped in the same cluster as melting point.

[0093] The simplified feature set obtained through this selection process is used to retrain the XGB model while keeping the hyperparameters unchanged. A fixed random seed is applied to maintain consistency in the data set partitioning into training and testing subsets, with an 80:20 split range. This approach allows for a fair evaluation of the effectiveness of feature selection and minimizes computational costs without re-adjusting model hyperparameters. In the case of a significant decrease in model performance, hyperparameter re-optimization is performed to improve model adaptability.

[0094] After optimization, the feature dimension is reduced by 23.8%, and the prediction performance of the model on the test set is significantly improved. The RMSE is reduced to 7.6517, the R2 is increased to 0.8553, and the MAE is reduced to 3.8042 (as shown in Figure 5 The results demonstrate the effectiveness of the model optimization strategy adopted, which effectively improves the model performance while reducing the computational cost.

[0095] Figure 6 is the schematic diagram of the composition structure of the fly ash pyrolysis organic pollutant removal efficiency prediction device provided by the embodiment of the present application, as Figure 6 shown, the fly ash pyrolysis organic pollutant removal efficiency prediction device 600 comprises: an acquisition module 601, configured to acquire an original sample data set in a fly ash pyrolysis process; the original sample data set comprises a plurality of feature parameters affecting the organic pollutant removal efficiency in the fly ash pyrolysis process; a processing module 602, configured to sequentially perform cleaning filtering and standardization processing on the original sample data set to obtain a prediction sample data set; and split the prediction sample data set into a training set and a test set; a construction module 603, configured to construct an initial organic pollutant removal efficiency prediction model based on the training set and the test set through an extreme gradient boosting machine learning algorithm; an analysis module 604, configured to perform importance sorting on the plurality of feature parameters based on feature importance analysis in the initial organic pollutant removal efficiency prediction model to obtain an importance sorting result, and delete the feature with the smallest importance; a calculation module 605, configured to calculate a variance inflation factor value of each feature parameter to obtain a variance inflation factor sorting result, so as to identify the feature parameter most affected by multicollinearity, and delete the feature with the largest variance inflation factor; the construction module 603 is further configured to construct a feature correlation matrix based on a preset Pearson correlation coefficient and the plurality of feature parameters; the feature correlation matrix is used to measure the correlation between different feature parameters; an iteration module 606, configured to perform iterative synthesis on the plurality of feature parameters based on a preset hierarchical clustering algorithm to obtain a hierarchical clustering result; a determination module 607, configured to analyze the feature correlation result and the hierarchical clustering result based on the importance sorting result and the variance inflation factor value sorting result, preferentially retain the feature with high importance and small variance inflation factor value, and then determine a target organic pollutant removal efficiency prediction model through the filtered input feature; and a prediction module 608, configured to predict the organic pollutant removal efficiency in a to-be-pyrolyzed process based on the target organic pollutant removal efficiency prediction model to obtain a target efficiency prediction result.

[0096] It should be noted that the description of the device embodiment of the present application is similar to the description of the method embodiment described above, and has similar beneficial effects as the method embodiment, and therefore will not be described in detail. For technical details not disclosed in the device embodiment, please refer to the description of the method embodiment of the present application for understanding.

[0097] It should be noted that the fly ash pyrolysis organic contaminant removal efficiency prediction method described above in the embodiments of the present application can be stored in a computer readable storage medium if it is implemented in the form of a software function module and sold or used as an independent product. Based on this understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for causing a terminal to perform all or part of the methods described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a magnetic disk or an optical disk, and various program code storage media. Thus, the embodiments of the present application are not limited to any particular hardware and software combination.

[0098] Correspondingly, the embodiments of the present application provide an electronic device, Figure 7 is a schematic diagram of the composition structure of the electronic device provided by the embodiments of the present application, as Figure 7 shown, the electronic device 700 at least includes: a processor 701 and a computer readable storage medium 702 configured to store executable instructions, wherein the processor 701 generally controls the overall operation of the electronic device 700. The computer readable storage medium 702 is configured to store instructions and applications executable by the processor 701, and can also cache data to be processed by the processor 701 and each module in the electronic device 700, and can be realized by FLASH or random access memory (RAM).

[0099] The embodiments of the present application provide a storage medium storing executable instructions, wherein the executable instructions are stored, and when the executable instructions are executed by a processor, the processor will execute the method provided by the embodiments of the present application, for example, the method shown in Figure 1 .

[0100] In some embodiments, the storage medium can be a computer-readable storage medium, such as a ferroelectric memory (FRAM), a read only memory (ROM), a programmable read only memory (PROM), an erasable programmable read only memory (EPROM), an electrically erasable programmable read only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disk, or a compact disk read only memory (CD-ROM), and the like. It can also be various devices including one or any combination of the above memories.

[0101] In some embodiments, the executable instructions can be in the form of programs, software, modules, scripts, or code, written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0102] By way of example, the executable instructions can or can not correspond to a file in a file system, can be stored in a part of a file that holds other programs or data, can be stored as one or more scripts in a markup language, such as Hyper Text Markup Language (HTML), in a single file dedicated to the program in question, or in multiple files in a file system, such as a file system on a stand-alone device or on a networked device. The executable instructions may, by way of example, be deployed to execute on one electronic device, or to execute on multiple electronic devices at one location, or to execute on multiple electronic devices distributed among multiple locations and connected by a communication network.

[0103] The above description is only some embodiments of the present application, and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, and improvement within the spirit and scope of the present application shall be included in the protection scope of the present application.

[0104] It is to be understood that the terminology "one embodiment" or "an embodiment" used throughout this specification means that a particular feature, structure or characteristic described is included in at least one embodiment of the application. Therefore, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily referring to the same embodiment. Furthermore, the particular features, structures or characteristics can be combined in any suitable manner in one or more embodiments. It is to be understood that the sequence of steps in the above-described various embodiments of the present application does not mean that the execution order of the steps is prior or posterior, and the execution order of the steps should be determined according to their functions and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The sequence of the above-described embodiments of the present application is only for description, and does not represent the advantages or disadvantages of the embodiments.

[0105] It should be noted that the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element preceded by "comprises... a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus that comprises the recited element. In several embodiments of the present application, it is to be understood that the disclosed apparatus and methods can be implemented in other ways. The above-described apparatus embodiments are merely illustrative, for example, the division of the units is only a logical functional division, and actual implementation can have another division, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be omitted or not executed.

[0106] The above describes only the embodiments of the present application, but the protection scope of the present application is not limited thereto, any skilled person in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for predicting the removal efficiency of organic pollutants in fly ash pyrolysis, characterized in that: The method comprises: Acquiring an original sample data set during the fly ash pyrolysis process; the original sample data set includes multiple characteristic parameters that affect the removal efficiency of organic pollutants during the fly ash pyrolysis process; The original sample data set is cleaned, filtered and standardized in sequence to obtain a predicted sample data set; and the predicted sample data set is split into a training set and a test set; Based on the training set and the test set, constructing an initial organic pollutant removal efficiency prediction model using an extreme gradient boosting machine learning algorithm; Based on the feature importance analysis in the initial organic pollutant removal efficiency prediction model, the plurality of feature parameters are ranked by importance to obtain an importance ranking result, and the feature with the least importance is deleted; Calculating the variance inflation factor value of each of the feature parameters to obtain a variance inflation factor ranking result, so as to identify the feature parameters most affected by multicollinearity, and deleting the feature with the largest variance inflation factor; Constructing a feature correlation matrix based on a preset Pearson correlation coefficient and the multiple feature parameters; the feature correlation matrix is ​​used to measure the correlation between different feature parameters; Based on a preset hierarchical clustering algorithm, the multiple feature parameters are iteratively synthesized to obtain a hierarchical clustering result; Based on the importance ranking results and the variance inflation factor value ranking results, the feature correlation results and the hierarchical clustering results are analyzed, and features with high importance and small variance inflation factor values ​​are preferentially retained, and then a target organic pollutant removal efficiency prediction model is determined through the screened input features; Based on the target organic pollutant removal efficiency prediction model, the organic pollutant removal efficiency during the pyrolysis process is predicted to obtain a target efficiency prediction result.

2. The method according to claim 1, characterized in that The original sample data set is cleaned, filtered, and standardized in sequence to obtain a predicted sample data set, including: Using a preset box plot and preset database constraint rules, identifying and deleting outliers and duplicate record values ​​in the original sample data set to obtain a deleted original sample data set; The preset chain equation multiple interpolation method is used to perform multiple interpolations on the deleted original sample data set to obtain the predicted sample data set.

3. The method according to claim 1, characterized in that The constructing of a feature correlation matrix based on the preset Pearson correlation coefficient and the multiple feature parameters includes: Calculate the correlation coefficient between each two feature parameters according to the Pearson correlation coefficient calculation formula, and fill the correlation coefficient value into the corresponding position of the feature correlation matrix; The correlation coefficients are sorted according to their absolute values, and characteristic parameters with absolute values ​​greater than 0.5 are determined to have strong correlation, thereby obtaining pairwise combination variables with strong correlation.

4. The method according to claim 1, wherein The multiple feature parameters are iteratively synthesized based on a preset hierarchical clustering algorithm to obtain a hierarchical clustering result, including: Calculate the Euclidean distance between every two feature parameters in the data set; and based on the preset average distance method, merge the two feature parameters with the smallest distance into a new cluster, and calculate the Euclidean distance between the new cluster and other clusters; Repeat the above process to build a tree structure; Based on the dendrogram structure, the Euclidean distances are sorted to obtain the hierarchical clustering result.

5. The method according to claim 1, wherein The multiple characteristic parameters include fly ash property parameters, pyrolysis parameters and organic pollutant characteristics; The fly ash property parameters include: fixed carbon content, ash content, volatile matter content, SiO2 content, Cl content, CaO content, Fe2O3 content, Al2O3 content, Cd content, Cr content and fly ash mass; the pyrolysis parameters include: temperature, time, pretreatment, gas flow rate and gas type; the organic pollutant characteristics include: boiling point, melting point, concentration, Henry's constant and octanol-water partition coefficient.

6. The method according to claim 1, characterized in that The importance ranking result is fixed carbon content > temperature > concentration > pretreatment > time > volatile matter content > melting point > boiling point > octanol-water partition coefficient > Henry's constant > Cl content > Cd content > ash content > SiO2 content > CaO content > gas flow rate > fly ash mass > Fe2O3 content > Al2O3 content > gas type > Cr content.

7. A device for predicting the removal efficiency of organic pollutants in fly ash pyrolysis, characterized in that: The device comprises: An acquisition module is used to acquire an original sample data set during the fly ash pyrolysis process; the original sample data set includes multiple characteristic parameters that affect the removal efficiency of organic pollutants during the fly ash pyrolysis process; A processing module is used to sequentially clean, filter and standardize the original sample data set to obtain a predicted sample data set; and split the predicted sample data set into a training set and a test set; A construction module is used to construct an initial organic pollutant removal efficiency prediction model based on the training set and the test set by using an extreme gradient boosting machine learning algorithm; An analysis module, configured to perform importance ranking of the plurality of characteristic parameters based on the feature importance analysis in the initial organic pollutant removal efficiency prediction model, obtain an importance ranking result, and delete the feature with the least importance; a calculation module, configured to calculate a variance inflation factor value of each of the feature parameters, obtain a variance inflation factor ranking result, identify the feature parameters most affected by multicollinearity, and delete the feature with the largest variance inflation factor; The construction module is further configured to construct a feature correlation matrix based on a preset Pearson correlation coefficient and the plurality of feature parameters; the feature correlation matrix is ​​configured to measure the correlation between different feature parameters; An iterative module, configured to iteratively synthesize the plurality of characteristic parameters based on a preset hierarchical clustering algorithm to obtain a hierarchical clustering result; a determination module for analyzing the feature correlation results and the hierarchical clustering results based on the importance ranking results and the variance inflation factor value ranking results, preferentially retaining features with high importance and small variance inflation factor values, and then determining a target organic pollutant removal efficiency prediction model based on the screened input features; The prediction module is used to predict the organic pollutant removal efficiency during the pyrolysis process based on the target organic pollutant removal efficiency prediction model to obtain a target efficiency prediction result.

8. An electronic device, characterized in that: include: a memory for storing executable instructions; The processor is configured to implement the method for predicting the removal efficiency of organic pollutants in fly ash pyrolysis according to any one of claims 1 to 6 when executing the executable instructions stored in the memory.

9. A computer-readable storage medium storing executable instructions for causing a processor to execute the executable instructions to implement the method for predicting the removal efficiency of organic pollutants in fly ash pyrolysis according to any one of claims 1 to 6.

Images