Fixed pollution source data collection calibration method, system, medium and program product

Abstract

A method, system, medium and program product for data acquisition and calibration of fixed pollution sources, in which several concentration change data are processed in segments to obtain several data segments; the concentration change trend of various pollutants is determined, and the correlation coefficient matrix is ​​calculated; the Euclidean distance value of the correlation coefficient matrix of the current data segment and each correlation coefficient matrix before sorting is calculated; when there is a Euclidean distance greater than a preset threshold, the corresponding data segment is marked as an abnormal data segment; it is determined as a target pollutant, and a material balance equation is established; based on the material balance equation, a theoretical concentration change curve of the target pollutant is generated; the deviation value is calculated, and when the deviation value exceeds the preset dynamic error limit, the sensor correction coefficient is determined according to the process parameters; the acquisition data of the target pollutant is calibrated according to the sensor correction coefficient. The present application improves the accuracy of the calibration of fixed pollution source data acquisition, thereby reducing the impact on the formulation and implementation of pollution prevention and control measures.

Description

Technical Field

[0001] The present application belongs to the field of electronic digital data processing, and in particular, relates to a method, system, medium and program product for data acquisition and calibration of fixed pollution sources. Background Art

[0002] With the continuous advancement of industrialization, industrial pollution emissions are also increasing, posing a serious threat to the ecological environment. In order to effectively regulate industrial pollution emissions, it is necessary to monitor fixed pollution sources in real time. Traditional fixed pollution source monitoring mainly relies on manual on-site data collection on a regular basis, which has problems such as low monitoring frequency and poor real-time performance, and it is difficult to meet the needs of pollution prevention and control.

[0003] To solve the above problems, a method of automatically collecting data from fixed pollution sources using online monitoring equipment has emerged. This method installs online monitoring equipment at the emission outlet of fixed pollution sources to collect parameters such as pollutant concentration and emission rate in real time, and uploads the collected data to the monitoring center, thus realizing continuous monitoring of fixed pollution sources.

[0004] However, since online monitoring equipment is in harsh industrial environments for a long time, it is easily affected by factors such as high temperature, corrosion, and electromagnetic interference, which leads to quality problems such as drift and anomalies in the collected data. Data quality problems directly affect the calculation of pollutant emissions and the judgment of excessive emissions, reduce the accuracy of data collection and calibration of fixed pollution sources, and thus affect the formulation and implementation of pollution prevention and control measures. Summary of the invention

[0005] The present application provides a method, system, medium and program product for calibrating data collection of fixed pollution sources, which are used to improve the accuracy of calibration of data collection of fixed pollution sources, thereby reducing the impact on the formulation and implementation of pollution prevention and control measures.

[0006] In a first aspect, the present application provides a fixed pollution source data collection and calibration method, which processes the concentration change data of several pollutants at the emission outlet of the fixed pollution source in segments according to a preset time period to obtain several data segments sorted by time;

[0007] Determine the concentration change trend of various pollutants in each data segment, and calculate the correlation coefficient matrix between pollutants corresponding to each data segment according to each concentration change trend. The correlation coefficient matrix represents the degree of correlation between the concentration changes of different pollutants;

[0008] Perform Euclidean distance calculation on the correlation coefficient matrix of the current data segment and each correlation coefficient matrix before the current data segment is sorted in time, and obtain several corresponding Euclidean distance values;

[0009] When there is a Euclidean distance greater than a preset threshold, the corresponding data segment is marked as an abnormal data segment;

[0010] The pollutants with sudden changes in correlation coefficients in the abnormal data segment are identified as target pollutants, and the material balance equation for the time period corresponding to the abnormal data segment is established based on the chemical reaction relationship between the target pollutants and other pollutants;

[0011] Based on the material balance equation, combined with the process parameters and production load data of the target pollutant within the time period, a theoretical concentration change curve of the target pollutant is generated;

[0012] Calculate the deviation between the theoretical concentration change curve and the concentration data actually collected of the target pollutant;

[0013] When the deviation value exceeds the preset dynamic error limit, the sensor correction coefficient is determined according to the process parameters;

[0014] The collected data of the target pollutant is calibrated according to the sensor correction coefficient.

[0015] By adopting the above technical solution, the emission data of fixed pollution sources are processed in segments according to the preset time period, and the correlation coefficient matrix between pollutants is calculated to reflect the correlation of the concentration changes of each pollutant, and the Euclidean distance is used to judge the mutation degree of the correlation coefficient matrix to identify abnormal data segments. After the abnormal data segment is found, the material balance equation is established based on the chemical reaction relationship, and the theoretical concentration change curve is generated by combining the process parameters and production load data. The sensor correction coefficient is determined by the deviation analysis between the theoretical value and the measured value, and finally the data calibration is realized. Combining data anomaly detection with chemical reaction kinetics can not only accurately identify abnormal data, but also verify the rationality of the data from the perspective of physical and chemical mechanisms, reduce the possible misjudgment caused by relying solely on statistical methods, improve the accuracy of data calibration results, and thus improve the accuracy of fixed pollution source data collection and calibration, and reduce the impact on the formulation and implementation of pollution prevention and control measures.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, calculating the correlation coefficient matrix between pollutants corresponding to each data segment according to each concentration change trend specifically includes:

[0017] The weighted sliding window method is used to preprocess each concentration change trend to obtain a number of corresponding preprocessed concentration change data;

[0018] Perform wavelet transform on several pre-processed concentration change data, select several layers of wavelet coefficients with the highest energy proportion, and obtain the characteristic coefficient sequence of each pollutant;

[0019] Calculate the cross-correlation function between the characteristic coefficient sequences of any two pollutants;

[0020] The value of the cross-correlation function at zero time delay is divided by the product of the standard deviations of the two sequences to obtain the normalized correlation coefficient;

[0021] A symmetric matrix is ​​constructed with normalized correlation coefficients as elements to obtain a correlation coefficient matrix.

[0022] By adopting the above technical solution and using the weighted sliding window method to pre-process each concentration change trend, the noise interference in the original data can be weakened and the smoothness and stability of the data can be improved. The wavelet transform is used to extract the essential characteristics of the concentration changes of each pollutant, and the original time series data is converted into characteristic coefficients with clear physical meanings, which reduces the dimension and computational complexity of the data. By calculating the cross-correlation function between the characteristic coefficient sequences of any two pollutants, the dynamic correlation characteristics between the concentration changes of different pollutants are quantitatively characterized, reducing the distortion problem when directly using the original data to calculate the correlation. The correlation coefficient matrix obtained after normalization is not only dimensionally invariant, but also its numerical value can intuitively reflect the degree of correlation between different pollutants, which is convenient for subsequent trend judgment and abnormal diagnosis.

[0023] In conjunction with some embodiments of the first aspect, in some embodiments, determining a pollutant whose correlation coefficient in an abnormal data segment undergoes a sudden change as a target pollutant specifically includes:

[0024] The data segmented cumulative sum method is used to calculate the cumulative change curve of each element in the correlation coefficient matrix;

[0025] Perform piecewise linear fitting on the cumulative change curve to obtain the slope value of each segment;

[0026] Construct mutation detection statistics based on each segment slope value, and construct background distribution of mutation detection statistics;

[0027] Calculate the significance level of the actual mutation detection statistic relative to the background distribution;

[0028] Among the correlation coefficients whose significance level exceeds the first preset threshold, the corresponding pollutant combination is extracted, and the correlation coefficient whose significance level exceeds the first preset threshold is the significant mutation correlation coefficient;

[0029] The frequency of each pollutant in the pollutant combination appearing in the significant mutation correlation coefficient is counted;

[0030] Construct a mutation contribution index of pollutants based on frequency;

[0031] The pollutants whose mutation contribution index exceeds the second preset threshold are determined as target pollutants.

[0032] By adopting the above technical solution and analyzing the correlation coefficient matrix using the data segmented cumulative sum method, the dynamic change characteristics of the correlation strength over time can be revealed. The slope value obtained by piecewise linear fitting intuitively reflects the rate of change of the correlation coefficient at different stages. Based on the cumulative slope, the mutation detection statistic is constructed. With the help of the background distribution, it is possible to quantitatively judge whether the change in the correlation coefficient at each moment is significant, and the strength of the mutation is directly reflected by the level of significance. By extracting the pollutant combinations that frequently appear in the significant mutation correlation coefficient and constructing the mutation contribution index using the frequency information, it is possible to reasonably and quantitatively evaluate the influence weight of each pollutant on the association relationship mutation, and then accurately screen out the target pollutants that cause the anomaly according to the mutation contribution threshold.

[0033] In combination with some embodiments of the first aspect, in some embodiments, after determining the pollutant whose correlation coefficient has a sudden change in the abnormal data segment as the target pollutant, the method further includes:

[0034] A multi-level reaction network topology structure is constructed based on the chemical valence distribution of the target pollutant, wherein the multi-level reaction network topology structure includes a plurality of reaction groups, and the reaction groups include a main reaction path and a side reaction path;

[0035] Calculate the rate constants of the main reaction path and the side reaction path, and establish the reaction rate matrix;

[0036] Based on the reaction rate matrix, the law of mass action is used to establish a system of differential equations for the concentration of components in each reaction group;

[0037] Solve the differential equations to obtain the dynamic change law of component concentration;

[0038] The dynamic change law is used as the constraint condition of the material balance equation.

[0039] By adopting the above technical solutions and introducing kinetic equations based on chemometrics, the limitation of simple chemometric methods that it is difficult to accurately describe the dynamic evolution of substance concentrations in complex reaction systems can be reduced. Integrating the chemical valence information of the target pollutants into the construction of the topological structure of the reaction network can form a hierarchical framework that is consistent with the intrinsic mechanism of the pollution process, thereby enhancing the applicability of the material balance equation to the actual process. The hierarchical solution of the kinetic parameters of different reaction paths can not only reduce the difficulty of parameter identification of multi-component complex systems, but also obtain rate constants with clear physical meanings and that can reflect different reaction scales, thereby improving the interpretability of the reaction kinetic model.

[0040] In combination with some embodiments of the first aspect, in some embodiments, constructing a multi-level reaction network topology structure based on the chemical valence distribution of the target pollutant specifically includes:

[0041] According to the valence state distribution of the target pollutants, all redox reaction pairs are determined;

[0042] Calculate the standard electrode potential of each redox reaction pair and establish the thermodynamic feasibility criterion based on the standard electrode potential;

[0043] Based on the thermodynamic feasibility criterion, thermodynamically feasible reaction paths are screened and the initial reaction network is constructed;

[0044] Conduct flux analysis on the initial reaction network to calculate the substance flux of the reaction path;

[0045] Determine the weight coefficient of the reaction path based on the material flow rate, and classify the reaction paths corresponding to the weight coefficient difference less than the preset difference into the same reaction level;

[0046] Determine the material transformation relationship between each reaction level, and establish a multi-level reaction network topology based on the material transformation relationship.

[0047] By adopting the above technical solutions, the thermodynamic feasibility of the reaction system is constrained based on the valence distribution of the target pollutants, and the direction of the redox reaction is quickly determined by comparing the standard electrode potential, which improves the thermodynamic rationality of the path in the reaction network. The reaction path is quantitatively evaluated by the material flow rate, which intuitively reflects the contribution of each path to the pollution process from the perspective of flow rate, and the topological structure of the reaction network can be effectively simplified after weighting. By dynamically depicting the mutual influence and constraint relationship between different reaction paths, a multi-level network framework that can map the macroscopic evolution law of the pollution process is constructed, which is easier to apply in engineering while improving the computational efficiency, and reduces the difficulty of constructing complex reaction networks. The highly generalized reaction network topology obtained can fully reflect the multi-time scale characteristics of the pollution process, and can provide an intuitive and reasonable reaction path combination scheme for the analysis of pollution mechanisms and pollution control.

[0048] In conjunction with some embodiments of the first aspect, in some embodiments, calculating the rate constants of the main reaction path and the side reaction path specifically includes:

[0049] Construct a three-dimensional state space for each reaction group, where the three dimensions are temperature, pressure, and concentration;

[0050] A number of sampling points are set in the three-dimensional state space, and the reaction rate is measured at each sampling point;

[0051] Establishing a thermodynamic correlation function between reaction rate and state parameters, where the state parameters are data measured by temperature sensors and pressure sensors;

[0052] Fitting the pre-exponential factor and activation energy of the Arrhenius equation based on the thermodynamic correlation function;

[0053] The rate constants of the main reaction pathway and the side reaction pathway were obtained by solving the Arrhenius equation.

[0054] By adopting the above technical solution and introducing sampling measurement of three-dimensional state space, the way to obtain reaction kinetic data can be expanded on the basis of the traditional control variable method, reducing the difficulty of determining kinetic parameters in complex reaction systems. By associating the reaction rate and state parameters through thermodynamic functions, the macroscopic characterization of reaction kinetic behavior is realized under the framework of the first law of thermodynamics. The use of thermodynamic functions to fit the parameters in the Arrhenius equation reduces the inconvenience caused by frequent changes in pre-exponential factors and activation energies to engineering calculations, and can also realize the correction and optimization of reaction kinetic parameters. The reaction kinetic model can be simplified by using the solved apparent rate constant to quantitatively describe the progress of the macroscopic reaction.

[0055] In conjunction with some embodiments of the first aspect, in some embodiments, the collected data of the target pollutant is calibrated according to the sensor correction coefficient, specifically including:

[0056] Perform fast Fourier transform on the collected data of target pollutants, convert the time domain signal into frequency domain spectrum, and obtain the amplitude-frequency characteristic and phase-frequency characteristic;

[0057] The sensor correction coefficient is decomposed into an amplitude correction coefficient and a phase correction coefficient, and convolution operations are performed with the amplitude-frequency characteristic and the phase-frequency characteristic respectively to obtain a corrected frequency domain spectrum;

[0058] The inverse discrete Fourier transform is used to restore the corrected frequency domain spectrum to the time domain data of preliminary calibration;

[0059] According to the frequency distribution characteristics of the time domain data after preliminary calibration, the optimal wavelet basis function and decomposition layer number are selected to perform multi-scale decomposition on the time domain data to obtain the decomposition coefficients of each scale;

[0060] Calculate the energy distribution of each scale decomposition coefficient, perform soft threshold processing on each scale coefficient based on the preset energy threshold criterion, and reconstruct the denoised time domain data;

[0061] The state space model is established using process parameters and denoised time domain data, and the state equation and observation equation of the Kalman filter are constructed;

[0062] The denoised time domain data is recursively smoothed based on the state equation and observation equation to obtain the final calibration result.

[0063] By adopting the above technical solutions, a variety of digital signal processing technologies such as frequency domain analysis, wavelet analysis and Kalman filtering are integrated when calibrating the collected data. Fourier transform is used to realize the conversion of time and frequency domains. The influence of frequency interference on measurement accuracy can be suppressed by introducing correction coefficients and convolution in the frequency domain. Wavelet analysis is used to reduce the noise of time domain data. With the multi-scale characteristics of wavelet transform, the global properties and local details of the data can be taken into account at the same time, and the basis function can be adaptively selected according to the energy distribution, which has good adaptability and robustness. The state space model is constructed by combining the denoised measurement value and the state variable to be estimated, and the system state is optimally estimated with the help of Kalman filtering, which improves the accuracy of the measurement value.

[0064] In the second aspect, an embodiment of the present application provides a fixed pollution source data acquisition and calibration system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0065] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a system, causes the system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0066] In a fourth aspect, an embodiment of the present application provides a computer program product, characterized in that when the computer program product runs on a system, the system executes the method described in any possible implementation manner in the first aspect.

[0067] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0068] 1. The present application provides a method for calibrating data collection of fixed pollution sources. By adopting the above technical solution, the emission data of fixed pollution sources are processed in segments according to a preset time period. The correlation coefficient matrix between pollutants is calculated to reflect the correlation of the concentration changes of each pollutant, and the Euclidean distance is used to judge the mutation degree of the correlation coefficient matrix to identify abnormal data segments. After the abnormal data segment is found, a material balance equation is established based on the chemical reaction relationship, and a theoretical concentration change curve is generated by combining process parameters and production load data. The sensor correction coefficient is determined by the deviation analysis between the theoretical value and the measured value, and finally the data calibration is realized. Combining data anomaly detection with chemical reaction kinetics can not only accurately identify abnormal data, but also verify the rationality of the data from the perspective of physical and chemical mechanisms, reduce the possible misjudgment caused by relying solely on statistical methods, improve the accuracy of data calibration results, and thus improve the accuracy of fixed pollution source data collection calibration, and reduce the impact on the formulation and implementation of pollution prevention and control measures.

[0069] 2. The present application provides a fixed pollution source data collection and calibration method, which uses the data segmented cumulative sum method to analyze the correlation coefficient matrix, which can reveal the dynamic change characteristics of the correlation strength over time. The slope value obtained by piecewise linear fitting intuitively reflects the rate of change of the correlation coefficient at different stages. Based on the cumulative slope, a mutation detection statistic is constructed. With the help of the background distribution, it is possible to quantitatively judge whether the change in the correlation coefficient at each moment is significant, and the strength of the mutation is directly reflected by the level of significance. Extracting pollutant combinations that frequently appear in the significant mutation correlation coefficient and constructing a mutation contribution index using frequency information can reasonably and quantitatively evaluate the weight of the influence of each pollutant on the mutation of the association relationship, and then accurately screen out the target pollutants that cause the anomaly according to the mutation contribution threshold.

[0070] 3. The present application provides a data collection and calibration method for a fixed pollution source, which introduces sampling measurement in a three-dimensional state space. It can expand the way to obtain reaction kinetic data on the basis of the traditional control variable method, and reduce the difficulty of determining kinetic parameters in complex reaction systems. By associating the reaction rate with the state parameters through thermodynamic functions, a macroscopic characterization of the reaction kinetic behavior is achieved under the framework of the first law of thermodynamics. The parameters in the Arrhenius equation are fitted using thermodynamic functions, which reduces the inconvenience caused by frequent changes in pre-exponential factors and activation energies to engineering calculations, and can also achieve the correction and optimization of reaction kinetic parameters. The reaction kinetic model can be simplified by quantitatively describing the progress of the macroscopic reaction using the apparent rate constant obtained by solving it. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a flow chart of a fixed pollution source data collection and calibration method in an embodiment of the present application.

[0072] Figure 2 It is another flow chart of a fixed pollution source data acquisition and calibration method in an embodiment of the present application.

[0073] Figure 3 It is a schematic diagram of the physical device structure of a fixed pollution source data acquisition and calibration system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0074] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more listed items.

[0075] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0076] Combine the following Figure 1 , a fixed pollution source data acquisition and calibration method in an embodiment of the present application is described:

[0077] See also Figure 1 , which is a flow chart of a fixed pollution source data acquisition and calibration method in an embodiment of the present application.

[0078] S101, segmenting the concentration change data of several pollutants at the emission outlet of the fixed pollution source according to a preset time period to obtain several data segments sorted by time;

[0079] In this step, the system first obtains the concentration change data of several pollutants at the emission outlet of the fixed pollution source. These data usually come from online monitoring equipment installed at the emission outlet, including but not limited to infrared gas analyzers, ultraviolet differential absorption spectrometers, chemiluminescence analyzers, etc.

[0080] The system can establish a communication connection with the monitoring equipment through the data acquisition unit and receive the concentration data of various pollutants in real time. Considering that the concentration changes of different pollutants have certain dynamic characteristics, the system needs to pre-process the original continuous data.

[0081] One feasible solution is to segment the data according to preset time periods. The selection of time periods needs to comprehensively consider factors such as production conditions and the operating status of pollution control facilities to ensure the consistency and representativeness of pollutant concentration changes in each data segment.

[0082] In addition, the system can also perform operations such as outlier removal and data normalization on the segmented data to improve data quality. After segmentation, the system obtains several data segments sorted in chronological order, and each data segment contains the dynamic changes in the concentration of each pollutant during that period. These data segments will serve as the basic units for subsequent analysis and mining, laying the foundation for accurately describing the laws of pollutant emissions.

[0083] S102, determining the concentration variation trend of various pollutants in each data segment, and calculating the correlation coefficient matrix between pollutants corresponding to each data segment according to each concentration variation trend;

[0084] The system determines the concentration change trend of various pollutants in each data segment, and calculates the correlation coefficient matrix between pollutants corresponding to each data segment according to each concentration change trend. The correlation coefficient matrix represents the degree of correlation between the concentration changes of different pollutants. Among them, the correlation coefficient matrix between pollutants corresponding to each data segment is calculated according to each concentration change trend, which specifically includes:

[0085] The weighted sliding window method is used to preprocess each concentration change trend to obtain a number of corresponding preprocessed concentration change data;

[0086] Perform wavelet transform on several pre-processed concentration change data, select several layers of wavelet coefficients with the highest energy proportion, and obtain the characteristic coefficient sequence of each pollutant;

[0087] Calculate the cross-correlation function between the characteristic coefficient sequences of any two pollutants;

[0088] The value of the cross-correlation function at zero time delay is divided by the product of the standard deviations of the two sequences to obtain the normalized correlation coefficient;

[0089] A symmetric matrix is ​​constructed with normalized correlation coefficients as elements to obtain a correlation coefficient matrix.

[0090] In this step, the system needs to deeply analyze the dynamic change characteristics of pollutant concentrations in each data segment. First, the system can use mathematical statistics methods to perform trend analysis on the concentration change data of each pollutant, including but not limited to moving average, exponential smoothing, trend fitting, etc. Through trend analysis, the random fluctuation components in the concentration data can be filtered out, and the trend signal that can reflect the overall change direction of the pollutant can be extracted. The trend signal can be in the form of a function of time, such as a linear function, a quadratic function, etc., or a non-parametric curve, such as a spline curve, a kernel estimation curve, etc. Based on the concentration change trend of each pollutant, the system needs to further explore the association pattern between different pollutants. An effective method is to calculate the correlation coefficient matrix between pollutants. The correlation coefficient is a statistical indicator that measures the degree of linear correlation between two variables, and its value ranges from -1 to 1. The system can use the trend signal sequence of pollutants to obtain the correlation coefficient between them through formulas such as the Pearson correlation coefficient and the Spearman rank correlation coefficient, and use this as an element to construct a symmetric matrix. Each element in the matrix quantifies the similarity of the concentration change trends of two pollutants, and the overall structure of the matrix reflects the interactive influence network between multiple pollutants. Through these quantitative indicators, the system can comprehensively characterize the dynamic evolution of pollutant components in each data segment and reveal the intrinsic connections between different components, providing data support for subsequent abnormal diagnosis and source tracing analysis.

[0091] S103, performing Euclidean distance calculation on the correlation coefficient matrix of the current data segment and each correlation coefficient matrix before the current data segment is sorted in time, to obtain a number of corresponding Euclidean distance values;

[0092] In this step, the system needs to evaluate the degree of difference between the current data segment and the previous data segments in terms of pollutant correlation. To achieve this goal, the system introduces the concept of Euclidean metric space.

[0093] Specifically, the system regards the correlation coefficient matrix corresponding to each data segment as a point in a high-dimensional space. The number of rows and columns of the matrix determines the dimension of the space, and the value of the matrix elements determines the position of the point on each coordinate axis. On this basis, the system can use the Euclidean distance formula to calculate the distance value between the current matrix and the previous matrices. Euclidean distance is the straight-line distance between two points in Euclidean space. It has a clear physical meaning and is simple to calculate. The larger the distance value, the greater the difference in the pollution state represented by the two matrices. The system calculates the Euclidean distance between the current matrix and all historical matrices to obtain a series of distance values. These distance values ​​objectively measure the degree to which the current pollution status deviates from the previous state, which is of great significance for timely detection of abnormal changes in the pollution process.

[0094] In addition, the system can also explore other matrix similarity measurement methods, such as Mahalanobis distance, cosine similarity, etc. The combination of multiple distance indicators can more comprehensively characterize the differences between data segments. The distance calculation results lay the mathematical foundation for subsequent threshold comparison and anomaly judgment.

[0095] S104, when there is a Euclidean distance greater than a preset threshold, marking the corresponding data segment as an abnormal data segment;

[0096] In this step, the system needs to determine whether the pollution state of the current data segment is abnormal based on the Euclidean distance value. The preset threshold is an important basis for identifying anomalies. It defines a hypersphere centered on the current matrix in the Euclidean metric space, and the radius of the hypersphere is equal to the threshold. If the Euclidean distance between a historical matrix and the current matrix exceeds the hypersphere, it means that the pollution states represented by the two matrices are quite different, and can be preliminarily determined to be abnormal.

[0097] To ensure the rationality of the threshold setting, the system can use statistical methods to adaptively determine the size of the threshold, such as calculating the upper limit of the threshold based on the mean and standard deviation of the distance value, or using a clustering algorithm to divide the distance value to distinguish abnormal points from normal points. For historical data segments that exceed the threshold, the system marks them as abnormal data segments and records the time range in which the abnormality occurs, providing clues for subsequent trend analysis and cause analysis. It should be noted that relying solely on the distance threshold may result in inaccurate abnormal judgments, and a comprehensive analysis must be conducted in combination with other characteristic indicators. In subsequent steps, the system will further demonstrate the significance and key influencing factors of the abnormality, and revise and confirm the preliminary abnormality judgment results.

[0098] S105, determining the pollutant whose correlation coefficient has suddenly changed in the abnormal data segment as the target pollutant, and establishing a material balance equation for the time period corresponding to the abnormal data segment according to the chemical reaction relationship between the target pollutant and other pollutants;

[0099] The system determines the pollutants with sudden changes in correlation coefficients in the abnormal data segment as target pollutants. Specifically, the cumulative change curve of each element in the correlation coefficient matrix is ​​calculated using the data segment cumulative sum method;

[0100] Perform piecewise linear fitting on the cumulative change curve to obtain the slope value of each segment;

[0101] Construct mutation detection statistics based on each segment slope value, and construct background distribution of mutation detection statistics;

[0102] Calculate the significance level of the actual mutation detection statistic relative to the background distribution;

[0103] Among the correlation coefficients whose significance level exceeds the first preset threshold, the corresponding pollutant combination is extracted, and the correlation coefficient whose significance level exceeds the first preset threshold is the significant mutation correlation coefficient;

[0104] The frequency of each pollutant in the pollutant combination appearing in the significant mutation correlation coefficient is counted;

[0105] Construct a mutation contribution index of pollutants based on frequency;

[0106] The pollutants whose mutation contribution index exceeds the second preset threshold are determined as target pollutants.

[0107] In this step, the system needs to identify the key pollutants that cause the anomaly, that is, the target pollutants, in the abnormal data segment. Since each element of the correlation coefficient matrix corresponds to the degree of correlation between the concentration trends of two pollutants, the source of the anomaly can be found by analyzing the mutation of the correlation coefficient.

[0108] In order to scientifically judge the mutation behavior of the correlation coefficient, the system introduces the concept of piecewise cumulative sum of data. The specific approach is: expand the correlation coefficient matrix along the time axis to obtain a series of two-dimensional matrix sequences, and then apply the piecewise cumulative sum operator to each matrix element to generate a cumulative change curve. This curve reflects the dynamic evolution law of the correlation coefficient in different time periods, and the mutation point of the curve corresponds to the moment when the correlation coefficient changes drastically. Using piecewise linear fitting technology, the system can efficiently locate the mutation point of the curve and extract the slope value before and after the mutation point. The change in the size of the slope indicates the mutation strength of the correlation coefficient. Based on the slope sequence to construct a mutation detection statistic and estimate its background distribution, the system can quantitatively judge the mutation significance of each correlation coefficient. The correlation coefficient that exceeds the significance threshold is determined as a significant mutation coefficient, and its corresponding pollutant combination becomes the focus of attention.

[0109] Furthermore, the system counts the frequency of each component of the pollutant combination and introduces the indicator of mutation contribution to quantitatively evaluate the importance of each pollutant. The pollutant with the largest mutation contribution is the target pollutant.

[0110] The determination of target pollutants is inseparable from the guidance of environmental chemical reaction mechanisms. The system needs to comprehensively consider factors such as the physical and chemical properties, reaction activity, and transformation pathways of pollutants to support the establishment of subsequent material balance equations. The material balance equation takes the target pollutant as the core, combines it with the stoichiometric relationship between other pollutants, and quantitatively describes the quantitative relationship between reactants and products. It is a mathematical expression of the pollution process during abnormal periods. The equation weighs the balance of payments of each material component, reveals their growth and decline laws on a spatiotemporal scale, and provides a theoretical basis for the final concentration calibration.

[0111] S106. Based on the material balance equation, combined with the process parameters and production load data of the target pollutant within the time period, generate a theoretical concentration change curve of the target pollutant;

[0112] In this step, the system needs to solve the material balance equation to obtain the theoretical change law of the target pollutant concentration. The material balance equation is based on the law of conservation of mass and takes into account the migration and transformation behavior of pollutants during the reaction process, but solving the equation still requires certain initial conditions and boundary conditions. Process parameters are important indicators that describe the state of the production process, usually including temperature, pressure, flow, pH value, etc. These parameters directly affect the generation and emission of pollutants.

[0113] The system can collect various process parameters in real time through the industrial control system and use them as known quantities in the material balance equation. In addition, the emission intensity of pollutants is also closely related to the production load. The production load data reflects the product output, raw material consumption, etc. The system can extract these data from the production management system and use them as the source term of the equation. On the basis of clarifying the parameters of the equation, the system uses a comprehensive numerical analysis method to solve the equation, mainly including differential format construction, nonlinear equation group iteration, convergence and stability analysis, etc. The solution result is the theoretical change curve of the target pollutant concentration. The system further compares the theoretical curve with the measured concentration data to evaluate the degree of deviation between the theoretical value and the actual value. The extrapolation ability of the curve can be used for concentration prediction under abnormal conditions, and the interpolation ability of the curve can be used to repair missing data.

[0114] S107, calculating the deviation value between the theoretical concentration change curve and the concentration data actually collected of the target pollutant;

[0115] In this step, the system needs to quantitatively evaluate the degree of agreement between the theoretical concentration change curve and the measured concentration data. This process is essentially to detect the pollution status of the abnormal data segment and diagnose the measurement deviation that may exist in the pollution process. First, the system compares the actual collected data of the target pollutant in the abnormal data segment with the theoretical curve value at the corresponding moment point by point, and calculates the difference between the two. The positive or negative value of the difference reflects the relative height of the measured value to the theoretical value, and the absolute value of the difference reflects the degree of deviation.

[0116] Furthermore, the system can introduce statistical indicators to measure the deviation level of the entire abnormal data segment. Commonly used indicators include mean square error, mean absolute percentage error, and coefficient of determination. The mean square error measures the average magnitude of the overall deviation, but is easily affected by individual outliers. The mean absolute percentage error eliminates the influence of the dimension and intuitively reflects the size of the relative error. The coefficient of determination judges the goodness of fit of the theoretical curve to the measured data from the perspective of regression analysis.

[0117] The system can comprehensively score multiple evaluation indicators and ultimately draw qualitative conclusions about the deviation level, such as "serious deviation", "moderate deviation", "slight deviation", etc. The core of this step is to judge the significance of the deviation and provide a quantitative basis for detecting abnormal pollutant concentrations. If the deviation value exceeds the pre-set threshold or confidence interval, it indicates that the actual emission status of the pollutant is significantly different from the theoretical expectation, and corrective measures need to be taken. On the contrary, if the deviation value is controlled within a reasonable range, it means that the measured data is generally credible and only needs fine-tuning. It is worth noting that the setting of the threshold needs to take into account the detection rate and the false alarm rate. A threshold that is too high may miss some anomalies, and a threshold that is too low may judge normal fluctuations as abnormalities.

[0118] S108. When the deviation value exceeds the preset dynamic error limit, a sensor correction coefficient is determined according to the process parameters, and the collected data of the target pollutant is calibrated according to the sensor correction coefficient.

[0119] When the deviation value exceeds the preset dynamic error limit, the system determines the sensor correction coefficient according to the process parameters, and calibrates the collected data of the target pollutant according to the sensor correction coefficient. The calibration process is as follows: perform fast Fourier transform on the collected data of the target pollutant, convert the time domain signal into the frequency domain spectrum, and obtain the amplitude-frequency characteristic and phase-frequency characteristic;

[0120] The sensor correction coefficient is decomposed into an amplitude correction coefficient and a phase correction coefficient, and convolution operations are performed with the amplitude-frequency characteristic and the phase-frequency characteristic respectively to obtain a corrected frequency domain spectrum;

[0121] The inverse discrete Fourier transform is used to restore the corrected frequency domain spectrum to the time domain data of preliminary calibration;

[0122] According to the frequency distribution characteristics of the time domain data after preliminary calibration, the optimal wavelet basis function and decomposition layer number are selected to perform multi-scale decomposition on the time domain data to obtain the decomposition coefficients of each scale;

[0123] Calculate the energy distribution of each scale decomposition coefficient, perform soft threshold processing on each scale coefficient based on the preset energy threshold criterion, and reconstruct the denoised time domain data;

[0124] The state space model is established using process parameters and denoised time domain data, and the state equation and observation equation of the Kalman filter are constructed;

[0125] The denoised time domain data is recursively smoothed based on the state equation and observation equation to obtain the final calibration result.

[0126] In this step, the system needs to correct and calibrate the abnormal data that exceeds the error limit. There may be many reasons for the distortion of measured data, such as instrument drift, environmental interference, human factors, etc. The system first diagnoses the type and degree of abnormality based on process parameters. Different contamination processes correspond to different sensor response characteristics, which are manifested in amplitude-frequency characteristics, phase-frequency characteristics, etc.

[0127] The system maps the time domain signal of the measured data to the frequency domain space through Fourier transform to obtain the amplitude and phase information of different frequency components. By analyzing the spectrum characteristics in combination with process parameters, the system can determine whether the sensor has abnormal phenomena such as decreased sensitivity and response delay. On the basis of qualitative diagnosis, the system further quantifies the correction coefficient that describes the sensor deviation. The correction coefficient is usually a complex number, the real part corresponds to the amplitude correction, and the imaginary part corresponds to the phase correction. The value of the correction coefficient is different for components of different frequencies. The system uses convolution operation to apply the correction coefficient to the frequency domain data, which is equivalent to amplifying, compressing, and translating the original signal, and finally obtains the corrected spectrum. To ensure the calibration accuracy, the system uses inverse discrete Fourier transform to reconstruct the corrected spectrum and generate time domain data after preliminary calibration.

[0128] However, since the measured data are often accompanied by high-frequency noise and other interference signals, it is difficult to completely eliminate these effects with a simple Fourier transform. Therefore, the system processes the data in a joint time-frequency domain. Wavelet transform is a multi-scale, multi-resolution signal analysis tool that can adaptively extract local features of signals in different frequency bands. The system achieves multi-scale characterization of data by optimizing the wavelet basis function and setting the number of decomposition layers, and obtains wavelet coefficients that reflect different time-frequency characteristics. The system performs adaptive threshold processing on the energy distribution of the wavelet coefficients to filter out components with lower energy, thereby achieving the effect of noise reduction and signal enhancement. After wavelet reconstruction, smooth time domain data can be obtained.

[0129] Based on the time domain data, the system further introduces the state space model to express the dynamic change process of the data as the evolution trajectory of the state variables. As a recursive state estimation algorithm, Kalman filtering describes the prediction mechanism of the system through the state equation and the update mechanism of the system through the observation equation. It can estimate and correct the state variables online and suppress the influence of noise. The system uses the optimized state space parameters to construct the Kalman filter, and iteratively filters the reconstructed time domain data to finally obtain the dynamically calibrated target pollutant concentration data.

[0130] In the above embodiment, the emission data of fixed pollution sources are processed in segments according to a preset time period, and the correlation coefficient matrix between pollutants is calculated to reflect the correlation of the concentration changes of each pollutant, and the Euclidean distance is used to determine the mutation degree of the correlation coefficient matrix to identify abnormal data segments. After the abnormal data segment is found, a material balance equation is established based on the chemical reaction relationship, and a theoretical concentration change curve is generated by combining process parameters and production load data. The sensor correction coefficient is determined by analyzing the deviation between the theoretical value and the measured value, and finally data calibration is achieved. Combining data anomaly detection with chemical reaction kinetics can not only accurately identify abnormal data, but also verify the rationality of the data from the perspective of physical and chemical mechanisms, reduce the possible misjudgment caused by relying solely on statistical methods, improve the accuracy of data calibration results, and thus improve the accuracy of fixed pollution source data collection and calibration, and reduce the impact on the formulation and implementation of pollution prevention and control measures.

[0131] In the above embodiment, after the step of determining the pollutant with a sudden change in the correlation coefficient in the abnormal data segment as the target pollutant, and before the step of establishing the material balance equation for the time period corresponding to the abnormal data segment according to the chemical reaction relationship between the target pollutant and other pollutants, in order to more accurately establish and constrain the material balance equation, the system also needs to systematically describe the transformation law of the target pollutant in the actual process by constructing a detailed chemical reaction network model, so that the material balance equation is more in line with the actual process. Figure 2 , another fixed pollution source data acquisition and calibration method in the embodiment of the present application is described:

[0132] See also Figure 2 , is another flow chart of a fixed pollution source data acquisition and calibration method in an embodiment of the present application.

[0133] S201. Construct a multi-level reaction network topology based on the chemical valence distribution of the target pollutant;

[0134] The system constructs a multi-level reaction network topology based on the chemical valence distribution of the target pollutants. The multi-level reaction network topology includes several reaction groups, and the reaction groups include main reaction paths and side reaction paths. Specifically: According to the valence distribution of the target pollutants, all redox reaction pairs are determined;

[0135] Calculate the standard electrode potential of each redox reaction pair and establish the thermodynamic feasibility criterion based on the standard electrode potential;

[0136] Based on the thermodynamic feasibility criterion, thermodynamically feasible reaction paths are screened and the initial reaction network is constructed;

[0137] Conduct flux analysis on the initial reaction network to calculate the substance flux of the reaction path;

[0138] Determine the weight coefficient of the reaction path based on the material flow rate, and classify the reaction paths corresponding to the weight coefficient difference less than the preset difference into the same reaction level;

[0139] Determine the material transformation relationship between each reaction level, and establish a multi-level reaction network topology based on the material transformation relationship.

[0140] The system first determines all possible redox reaction pairs based on the valence distribution of the target pollutants. Then, the system calculates the standard electrode potential of each redox reaction pair, and establishes a thermodynamic feasibility criterion based on the standard electrode potential to screen thermodynamically feasible reaction paths. By performing flux analysis on the initial reaction network and calculating the material flow rate of the reaction path, the system can determine the weight coefficient of the reaction path based on the material flow rate, and classify the reaction paths corresponding to weight coefficient differences less than the preset difference into the same reaction level. Finally, the system determines the material transformation relationship between each reaction level, and establishes a multi-level reaction network topology based on the material transformation relationship.

[0141] When constructing a multi-level reaction network topology, the system can use different methods to determine the redox reaction pairs, such as theoretical deduction based on the chemical structure and properties of the target pollutant, or obtaining reaction data through experimental measurement. When calculating the standard electrode potential, the system can use the Nernst equation and combine conditional parameters such as temperature and pressure for calculation. For the establishment of thermodynamic feasibility criteria, the system can comprehensively consider thermodynamic parameters such as Gibbs free energy change and equilibrium constant. When performing flux analysis, the system can use optimization algorithms such as linear programming and nonlinear programming to solve the material flux distribution. The weight coefficient can be determined by clustering algorithms, such as K-means clustering or hierarchical clustering.

[0142] In addition, the system can also consider the influence of reaction kinetic factors when constructing the multi-level reaction network topology. By introducing kinetic parameters such as reaction rate constant and reaction order, the system can more accurately describe the transformation relationship between reactants, intermediates and products, making the constructed reaction network topology closer to the actual process. At the same time, the system can use methods such as sensitivity analysis to identify key reaction steps and material components that have a greater impact on the reaction network topology, and further optimize and simplify the reaction network model. In short, by systematically constructing a multi-level reaction network topology, it can provide an important theoretical basis and constraints for the subsequent establishment of an accurate material balance equation.

[0143] S202, calculating the rate constants of the main reaction path and the side reaction path, and establishing a reaction rate matrix;

[0144] The system calculates the rate constants of the main reaction path and the side reaction path. Specifically: it constructs a three-dimensional state space for each reaction group, where the three dimensions are temperature, pressure, and concentration;

[0145] A number of sampling points are set in the three-dimensional state space, and the reaction rate is measured at each sampling point;

[0146] Establishing a thermodynamic correlation function between reaction rate and state parameters, where the state parameters are data measured by temperature sensors and pressure sensors;

[0147] Fitting the pre-exponential factor and activation energy of the Arrhenius equation based on the thermodynamic correlation function;

[0148] The rate constants of the main reaction pathway and the side reaction pathway were obtained by solving the Arrhenius equation.

[0149] Then the reaction rate matrix is ​​established.

[0150] The system first constructs a three-dimensional state space for each reaction group, where the three dimensions are temperature, pressure, and concentration. Then, the system sets several sampling points in the three-dimensional state space and measures the reaction rate at each sampling point. By establishing a thermodynamic correlation function between the reaction rate and the state parameters (temperature and pressure), the system can fit the pre-exponential factor and activation energy of the Arrhenius equation. Finally, by solving the Arrhenius equation, the system can obtain the rate constants of the main reaction path and the side reaction path, and establish a reaction rate matrix.

[0151] When calculating the rate constant, the system can use different experimental measurement methods, such as differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), etc., to obtain reaction rate data. For the establishment of thermodynamic correlation functions, the system can select appropriate function forms, such as polynomial functions, exponential functions, etc., and use optimization algorithms such as the least squares method for parameter fitting. When solving the Arrhenius equation, the system can use numerical integration methods, such as the Runge-Kutta method and the Euler method, to obtain numerical solutions to the rate constants. The establishment of the reaction rate matrix can be based on the multi-level reaction network topology, taking into account the mutual influence and coupling relationship between the main reaction path and the side reaction path.

[0152] In addition, the system can also introduce quantum chemical calculation methods, such as density functional theory (DFT) and ab initio algorithms, when calculating rate constants, to conduct a deeper study of the reaction mechanism and obtain more accurate energy and structural information. At the same time, the system can use machine learning algorithms, such as support vector machines (SVM) and artificial neural networks (ANN), to establish a nonlinear relationship between rate constants and state parameters, and improve the efficiency and accuracy of rate constant calculations. In short, by systematically calculating the rate constants of the main reaction path and the side reaction path, and establishing a reaction rate matrix, it can provide an important mathematical basis for the subsequent solution of differential equations and the acquisition of dynamic change laws.

[0153] S203, establishing a differential equation group of component concentrations in each reaction group using the law of mass action according to the reaction rate matrix;

[0154] The system first determines the reactants, intermediates, and products in each reaction group based on the multi-level reaction network topology and reaction rate matrix. Then, the system uses the law of mass action to establish differential equations for each component in each reaction group to describe the rate of change of its concentration over time. By combining the differential equations of all components together, the system can obtain a complete set of differential equations that characterizes the dynamic change process of component concentration in the entire reaction system.

[0155] When establishing a set of differential equations, the system can use different expressions of the law of mass action, such as the first-order law of mass action, the second-order law of mass action, etc., and select according to the complexity of the reaction mechanism. For complex reaction systems, the system can introduce thermodynamic parameters such as chemical equilibrium constants and adsorption equilibrium constants to describe the effect of the reaction equilibrium state on the concentration of components. At the same time, the system can consider factors such as material transport and heat conduction during the reaction process, and introduce corresponding mathematical terms such as diffusion terms and convection terms into the differential equations to improve the accuracy of the model.

[0156] In addition, the system can also use the dimensionless method to convert the differential equations into dimensionless form when establishing the differential equations, simplifying the solution process and result analysis. For highly nonlinear differential equations, the system can use multi-scale analysis methods to divide the reaction process into different time scales, establish corresponding differential equation subsystems respectively, and reduce the complexity of the model. In short, by systematically establishing the differential equations of the component concentrations in each reaction group based on the reaction rate matrix and the law of mass action, a mathematical model basis can be provided for the subsequent solution of the dynamic change law of the component concentrations.

[0157] S204, solving the differential equations to obtain the dynamic variation law of component concentrations;

[0158] The system first determines the initial conditions and boundary conditions, such as the initial concentration of the reactants, the temperature and pressure of the reactor, etc., based on the established differential equations. Then, the system selects appropriate numerical solution methods, such as the Runge-Kutta method, the Euler method, the finite difference method, etc., to discretize and iteratively solve the differential equations. Through continuous iterative calculations, the system can obtain the numerical solution of the component concentration at each time point, thereby obtaining the dynamic change curve of the component concentration over time.

[0159] When solving differential equations, the system can use an adaptive step control strategy to dynamically adjust the time step according to the rate of change of the solution, which not only ensures the calculation accuracy but also improves the solution efficiency. For rigid differential equations, the system can use implicit solution methods, such as implicit Euler method, backward difference method, etc., to improve the stability of numerical calculations. At the same time, the system can introduce parallel computing technology and use multi-core processors or distributed computing clusters to accelerate the solution process of differential equations.

[0160] In addition, the system can also identify and optimize model parameters in combination with experimental data when solving differential equations, such as using optimization algorithms such as the least squares method and genetic algorithm to minimize the error between the model prediction value and the experimental measurement value, thereby improving the model's prediction performance. For complex reaction systems, the system can use model simplification techniques, such as sensitivity analysis and principal component analysis, to identify key factors that have a greater impact on the reaction process, simplify the structure and parameters of the differential equations, and reduce the complexity of calculations. In short, by systematically solving the differential equations and obtaining the dynamic change law of component concentrations, it can provide an important quantitative basis for subsequent optimization of reaction processes and control of product quality.

[0161] S205. Use the law of dynamic change as a constraint condition for the material balance equation.

[0162] The system first mathematically expresses the dynamic change curve of component concentration, such as using polynomial functions, spline functions, etc., to obtain continuous and smooth function expressions. Then, the system uses these function expressions as constraints of the material balance equation, together with other process parameters, design variables, etc., to build a reaction process optimization model. By solving the optimization model, the system can obtain the optimal process parameter settings and operation strategies to achieve optimal control of the reaction process.

[0163] When the dynamic change law is used as a constraint, the system can use different mathematical optimization methods, such as linear programming, nonlinear programming, mixed integer programming, etc., and select according to the characteristics of the optimization objectives and constraints. For multi-objective optimization problems, the system can introduce weighted summation method, ε-constraint method and other methods to transform multiple objective functions into single-objective optimization problems for solution. At the same time, the system can consider the influence of uncertain factors, such as uncertainty of model parameters, measurement errors, etc., introduce robustness constraints in the optimization model, and improve the reliability of the optimization results.

[0164] In addition, the system can also integrate dynamic change laws with other process models, data-driven models, etc. to build a comprehensive optimization model with multi-scale and multi-physical field coupling. Through online measurement and real-time optimization, the system can realize dynamic optimization control of the reaction process, adjust process parameters in time, and deal with interference factors such as raw material fluctuations and equipment failures. In short, by using the dynamic change law of component concentration as a constraint condition for the material balance equation, the system can achieve systematic optimization of the reaction process, improve product quality and yield, and reduce energy consumption and cost, which has important engineering application value.

[0165] In the above embodiments, by introducing kinetic equations on the basis of chemometrics, the limitation that the simple chemometric method is difficult to accurately describe the dynamic evolution of substance concentrations in complex reaction systems can be reduced. Integrating the chemical valence information of the target pollutant into the construction of the topological structure of the reaction network can form a hierarchical framework that is consistent with the intrinsic mechanism of the pollution process, thereby enhancing the applicability of the material balance equation to the actual process. The hierarchical solution of the kinetic parameters of different reaction paths can not only reduce the difficulty of parameter identification of multi-component complex systems, but also obtain rate constants with clear physical meanings and that can reflect different reaction scales, thereby improving the interpretability of the reaction kinetic model.

[0166] The following describes the system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the physical device structure of a fixed pollution source data acquisition and calibration system provided in an embodiment of the present application.

[0167] It should be noted that Figure 3 The structure of the system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0168] like Figure 3As shown, the system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through a bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0169] The following components are connected to the I / O interface 305: an input section 306 including a camera, an infrared sensor, etc.; an output section 307 including a liquid crystal display (LCD) and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0170] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0171] It should be noted that the computer-readable medium shown in the embodiment of the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.

[0172] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Among them, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0173] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the system described in the above embodiment; or may exist independently without being assembled into the system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of a system, the system implements the method provided in the above embodiment.

[0174] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0175] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0176] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0177] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A fixed pollution source data acquisition and calibration method, characterized in that: include: The concentration change data of several pollutants at the emission outlet of the fixed pollution source are processed in sections according to a preset time period to obtain several data sections sorted by time; Determine the concentration change trend of various pollutants in each of the data segments, and calculate the correlation coefficient matrix between the pollutants corresponding to each of the data segments according to each of the concentration change trends, wherein the correlation coefficient matrix represents the degree of correlation between the concentration changes of different pollutants; Performing Euclidean distance calculation on the correlation coefficient matrix of the current data segment and each correlation coefficient matrix before the current data segment is sorted in time, to obtain a number of corresponding Euclidean distance values; When the Euclidean distance is greater than a preset threshold, the corresponding data segment is marked as an abnormal data segment; The pollutant with a sudden change in correlation coefficient in the abnormal data segment is determined as the target pollutant, and the material balance equation for the time period corresponding to the abnormal data segment is established according to the chemical reaction relationship between the target pollutant and other pollutants, specifically including: Calculating the cumulative change curve of each element in the correlation coefficient matrix by using the data segmented cumulative sum method; performing segmented linear fitting on the cumulative change curve to obtain the slope value of each segment; constructing mutation detection statistics based on the slope values ​​of each segment, and constructing a background distribution of the mutation detection statistics; Calculating the significance level of the actual mutation detection statistic relative to the background distribution; Extracting corresponding pollutant combinations from the correlation coefficients whose significance level exceeds the first preset threshold value, wherein the correlation coefficients whose significance level exceeds the first preset threshold value are significant mutation correlation coefficients; Counting the frequency of each pollutant in the pollutant combination appearing in the significant mutation correlation coefficient; Constructing a mutation contribution index of the pollutant based on the frequency; Determine the pollutant whose mutation contribution index exceeds a second preset threshold as the target pollutant; Based on the material balance equation, combined with the process parameters and production load data of the target pollutant in the time period, a theoretical concentration change curve of the target pollutant is generated; Calculating the deviation between the theoretical concentration change curve and the concentration data actually collected of the target pollutant; When the deviation value exceeds a preset dynamic error limit, determining a sensor correction coefficient according to the process parameter; The collected data of the target pollutant is calibrated according to the sensor correction coefficient.

2. The method according to claim 1, characterized in that: Calculating the correlation coefficient matrix between pollutants corresponding to each data segment according to each concentration change trend specifically includes: Using a weighted sliding window method to preprocess each of the concentration change trends to obtain a number of corresponding preprocessed concentration change data; Performing wavelet transformation on the pre-processed concentration change data, selecting wavelet coefficients of several layers with the highest energy proportion, and obtaining characteristic coefficient sequences of each pollutant; Calculate the cross-correlation function between the characteristic coefficient sequences of any two pollutants; Dividing the value of the cross-correlation function at zero time delay by the product of the standard deviations of the two sequences to obtain a normalized correlation coefficient; A symmetric matrix is ​​constructed with the normalized correlation coefficients as elements to obtain a correlation coefficient matrix.

3. The method according to claim 1, characterized in that After determining the pollutant with a sudden change in correlation coefficient in the abnormal data segment as a target pollutant, the method further includes: Constructing a multi-level reaction network topology structure based on the chemical valence distribution of the target pollutant, wherein the multi-level reaction network topology structure includes a plurality of reaction groups, and the reaction groups include a main reaction path and a side reaction path; Calculating the rate constants of the main reaction pathway and the side reaction pathway, and establishing a reaction rate matrix; Establishing a differential equation system for the concentration of components in each of the reaction groups using the law of mass action according to the reaction rate matrix; Solving the differential equations to obtain the dynamic variation law of the component concentrations; The dynamic change law is used as a constraint condition of the material balance equation.

4. The method according to claim 3, characterized in that The multi-level reaction network topology structure is constructed based on the chemical valence distribution of the target pollutant, specifically including: Determining all redox reaction pairs according to the valence state distribution of the target pollutant; Calculating the standard electrode potential of each of the redox reaction pairs, and establishing a thermodynamic feasibility criterion based on the standard electrode potential; Based on the thermodynamic feasibility criterion, thermodynamically feasible reaction paths are screened to construct an initial reaction network; Performing flux analysis on the initial reaction network to calculate the substance flux of the reaction path; Determine the weight coefficient of the reaction path based on the substance flow rate, and classify the reaction paths corresponding to the weight coefficient difference less than the preset difference into the same reaction level; The material transformation relationship between each reaction level is determined, and a multi-level reaction network topology is established based on the material transformation relationship.

5. The method according to claim 3, characterized in that: The calculating of the rate constants of the main reaction path and the side reaction path specifically includes: Constructing a three-dimensional state space of each of the reaction groups, wherein the three dimensions are temperature, pressure and concentration; Setting a plurality of sampling points in the three-dimensional state space, and measuring the reaction rate at each of the sampling points; Establishing a thermodynamic correlation function between the reaction rate and a state parameter, wherein the state parameter is data measured by a temperature sensor and a pressure sensor; fitting the pre-exponential factor and activation energy of the Arrhenius equation based on the thermodynamic correlation function; The rate constants of the main reaction pathway and the side reaction pathway are obtained by solving the Arrhenius equation.

6. The method according to claim 1, characterized in that The calibrating the collected data of the target pollutant according to the sensor correction coefficient specifically includes: Performing a fast Fourier transform on the collected data of the target pollutant, converting the time domain signal into a frequency domain spectrum, and obtaining an amplitude-frequency characteristic and a phase-frequency characteristic; Decomposing the sensor correction coefficient into an amplitude correction coefficient and a phase correction coefficient, and performing convolution operations with the amplitude-frequency characteristic and the phase-frequency characteristic respectively to obtain a corrected frequency domain spectrum; Using inverse discrete Fourier transform to restore the modified frequency domain spectrum to preliminarily calibrated time domain data; According to the frequency distribution characteristics of the preliminarily calibrated time domain data, selecting the optimal wavelet basis function and the number of decomposition layers to perform multi-scale decomposition on the time domain data to obtain decomposition coefficients of each scale; Calculating the energy distribution of each scale decomposition coefficient, performing soft threshold processing on each scale coefficient based on a preset energy threshold criterion, and reconstructing denoised time domain data; Establishing a state space model using the process parameters and the denoised time domain data, and constructing a state equation and an observation equation of a Kalman filter; The denoised time domain data is recursively smoothed based on the state equation and the observation equation to obtain a final calibration result.

7. A fixed pollution source data acquisition and calibration system, characterized in that: The system comprises: One or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the system to execute the method as described in any one of claims 1-6.

8. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a system, the system is caused to execute the method according to any one of claims 1 to 6.

9. A computer program product, characterized in that When the computer program product is executed on a system, the system is caused to execute the method according to any one of claims 1 to 6.

Images