A method for regulating pores of activated carbon, monitoring device

Abstract

The application discloses a kind of activated carbon pore regulation method, monitoring equipment, belongs to spectral analysis technical field, method includes: based on probability prediction model, dynamic adjustment spectral scanning frequency, and carry out real-time global spectrum acquisition;Set adaptive threshold, to extract timing slope and acceleration with dynamic sliding window, generate timing change matrix, determine abnormal spectrum, screen unexpected signal and carry out analysis, realize the accurate identification of known and new unexpected signal, by feature peak spacing back-propagation molecular configuration parameter, calculate the matching degree difference of pore size and molecular configuration, carry out matching analysis, ensure that the accuracy of pore characteristics and pollutant molecule adaptability analysis;According to the matching degree difference, carry out pore gradient regulation, real-time monitoring to verify the regulation effect, realize self-evolution, continuously improve regulation efficiency and precision.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of spectral analysis, and relates to an activated carbon pore regulation method and a monitoring device. BACKGROUND

[0002] In the field of industrial flue gas treatment, activated carbon has become a core material for desulfurization and denitrification technology due to its high specific surface area, rich pore structure and surface chemical activity. Traditional activated carbon pore regulation methods are mostly based on static processes such as steam activation, chemical immersion or high-temperature treatment, which use physical or chemical means to directionally amplify the proportion of mesopores to improve the diffusion efficiency of reactants. With the refinement of environmental protection requirements, dynamic analysis technology has gradually become a research hotspot. By monitoring the changes in pore structure during the activated carbon adsorption process in real time, the activation conditions are dynamically adjusted in combination with the working condition parameters to realize self-adaptive optimization of the pore structure.

[0003] However, the prior art does not consider the regulation adaptability defects of the dynamic analysis system when unexpected signal intrusion occurs. On the one hand, there is a lack of a pre-existing model for matching the molecular size and pore structure of out-of-design working condition pollutants such as heavy metal vapors and halogenated hydrocarbons. When such special pollutants suddenly appear in the flue gas, it is not possible to generate a pore structure adjustment strategy in real time based on their physical and chemical properties, resulting in a serious mismatch between the activated carbon surface active sites and the adsorption requirements of the pollutants. On the other hand, there is no emergency pore expansion mechanism for trace pollutants. For example, when facing mercury vapor, it is not possible to quickly generate sulfur-modified nanoscale active sites through pulse activation agent injection, which easily leads to the risk of pollutant penetration due to the lack of adsorption sites. More importantly, there is a quality control vacuum period during the regulation transition period after unexpected signal intrusion. From the detection and identification of pollutants to the completion of pore structure reconstruction, there is a lack of real-time compensation measures for the adsorption performance of activated carbon. SUMMARY

[0004] In view of the deficiencies in the prior art, the purpose of the present application is to provide an activated carbon pore regulation method and a monitoring device. By identifying unexpected signals and backstepping the molecular configuration, the matching degree is calculated in combination with the pore characteristics. A gradient regulation scheme is generated and the parameters are adjusted in real time. Full-chain monitoring and closed-loop optimization solve the problems of missing matching models, insufficient emergency mechanisms and control vacuum.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] An activated carbon pore regulation method, comprising:

[0007] Based on a probability prediction model, the spectral scanning frequency is dynamically adjusted, and real-time full-domain spectral acquisition is performed;

[0008] An adaptive threshold is set to extract the timing slope and acceleration in a dynamic sliding window, generate a timing change matrix, determine abnormal spectra, screen unexpected signals and analyze them, reverse the molecular configuration parameters through the characteristic peak spacing, calculate the difference between the pore size and the matching degree of the molecular configuration, and perform matching analysis.

[0009] According to the matching degree difference, the pore gradient is regulated, and the regulation effect is verified in real time.

[0010] Specifically, the step of adjusting the spectral scanning frequency comprises:

[0011] When the scanning device is started, a probability prediction model is called to predict the probability of the occurrence of unexpected signals in each waveband in the future period;

[0012] The waveband with a probability greater than the probability threshold is determined as a potential waveband, and the rest is a non-potential waveband;

[0013] The scanning frequency of the potential waveband is set to times of the conventional frequency, and the non-potential waveband maintains the conventional scanning frequency, and real-time global spectrum acquisition is performed synchronously.

[0014] Specifically, the step of generating a timing change matrix comprises:

[0015] The historical spectrum data of the recent period is obtained, and the baseline region in the spectrum data is located;

[0016] The baseline parameters of the baseline region are calculated, including the mean and standard deviation, and the real-time fluctuation entropy of the current spectrum data is calculated, and the fluctuation ratio is calculated in combination with the historical maximum fluctuation entropy;

[0017] In combination with the baseline parameters and the fluctuation ratio, an adaptive threshold is set;

[0018] Taking the current time as the end point, an initial window is set, time decay weights are given to the data in the initial window based on variable weight linear regression, and a weighted least squares method is used for fitting to obtain the baseline slope of the current time;

[0019] The difference between the baseline slopes of two consecutive times is calculated, and the sampling period is normalized to obtain the signal change rate :

[0020] If , the window is expanded to periods; if , the window is reduced to periods; wherein is the window adjustment threshold;

[0021] Based on the adjusted sliding window, a linear equation is fitted by weighted least squares method with time decay weight to obtain a time slope, a second-order difference of the time slope at three continuous time points is calculated and normalized by square of sampling period to obtain an acceleration, and a time change matrix is generated.

[0022] Specifically, the step of screening the unexpected signal comprises:

[0023] For each wavelength point in the time change matrix, if the time slope is greater than the adaptive threshold or the acceleration is greater than zero, it is marked as a single abnormal point, otherwise it is determined as normal fluctuation and the data is stored in a historical spectrum library;

[0024] The band distribution of the single abnormal points is counted, if 5 or more wavelength points in a certain band are single abnormal points and are not in the characteristic interval of the expected pollutant standard spectrum, it is determined as an abnormal spectrum and triggers the analysis process;

[0025] The original spectrum data corresponding to the abnormal spectrum is extracted, and a background interference spectrum library collected at the same period is merged to generate a to-be-processed matrix;

[0026] The covariance matrix of the to-be-processed matrix is calculated by principal component analysis method and is subjected to eigenvalue decomposition to obtain eigenvalues and eigenvectors, the first three principal components with cumulative variance contribution rate not less than 95% are selected, and the first principal component is removed to generate a denoised spectrum;

[0027] The pre-stored expected pollutant standard spectrum library is called, the standard spectrum library is used as the independent variable, the denoised spectrum is used as the dependent variable, and a regression model is constructed by using partial least squares regression method;

[0028] Based on the difference between the denoised spectrum and the model predicted value, a residual spectrum is calculated, the residual spectrum is subjected to peak intensity normalization processing, and an unexpected characteristic spectrum is generated.

[0029] Specifically, the specific steps of screening the unexpected signal further comprise:

[0030] The unexpected characteristic spectrum is subjected to baseline correction and moving average method smoothing to generate a smoothed spectrum, a first derivative is calculated to find candidate peak positions and screening is performed to generate a preliminary peak list;

[0031] The smoothed spectrum is subjected to peak fitting by using a Gaussian-Lorentz mixed function, key parameters of each characteristic peak are extracted and sorted by peak position, and a vibration characteristic parameter set is generated;

[0032] Each parameter in the vibration characteristic parameter set is standardized to obtain a vibration characteristic vector, a dynamic spectrum fingerprint library is called, a cosine similarity between the vibration characteristic vector and a known unexpected standard vector in the library is calculated, and the maximum value is selected as a matching similarity and a corresponding matching spectrum;

[0033] If the matching similarity is not less than a matching threshold, the signal is determined as a known unexpected signal; otherwise, the signal is determined as a new unexpected signal, and the vibration feature vector is stored in the dynamic spectrum fingerprint library.

[0034] Specifically, the step of matching analysis comprises:

[0035] For a known unexpected signal, a corresponding sub-model is mobilized, the spectrum feature is converted into a molecular geometric parameter, and a molecular configuration parameter is generated;

[0036] For a new unexpected signal, a general sub-model is called and an incremental learning mechanism is triggered to generate an initial parameter set, and based on the initial parameter set, incremental learning is carried out to generate a special sub-model, and a molecular configuration parameter of the new unexpected signal is generated.

[0037] Specifically, the step of matching analysis further comprises:

[0038] A small-angle X-ray scattering device and an X-ray photoelectron spectroscopy device built in the bed are started to generate a pore feature matrix of activated carbon;

[0039] Based on the molecular diffusion theory, the size matching degree is calculated, the polarity matching degree is calculated by the absolute difference formula, the volume matching degree is calculated based on the molecular packing theory, weights are set for each matching degree, and a comprehensive matching degree is calculated;

[0040] Based on a preset target matching degree, a matching degree difference is calculated, a secondary matching threshold is set to divide a matching level, the deduction logic is dynamically adjusted based on the matching level, and an optimized pore regulation deduction logic is generated.

[0041] Specifically, the step of pore gradient regulation comprises:

[0042] Based on the matching degree difference, the core regulation target is determined, the corresponding regulation means are called from the preset physical regulation library and chemical regulation library, the regulation priority is set based on the deduction logic, the corresponding regulation parameter calculation formula is called with the matching degree difference as the input, the parameter gradient and the time node are calculated, and a gradient regulation scheme is generated;

[0043] Based on the gradient regulation scheme, the regulation is carried out, the pore distribution is scanned according to the feedback interval during the regulation process, the regulation pore volume proportion is calculated, the actual pore change amount is calculated based on the pore volume proportion in the matching analysis, and the parameters are automatically adjusted when the actual pore change amount exceeds the expected deviation;

[0044] After the regulation is completed, a multi-dimensional matching degree evaluation system is called, the regulation matching degree difference is calculated based on the real-time monitored pore characteristics, and the regulation effect is determined and corrected.

[0045] The whole chain data of this regulation is stored in the pore regulation knowledge base according to categories, the clustering analysis is carried out on the multiple regulation data of the same non-expected signal to extract the optimal parameter combination, the recommended parameters of the gradient regulation scheme are updated, the new data is transmitted into the multi-dimensional matching degree evaluation system and the weight distribution is optimized through incremental learning.

[0046] An activated carbon pore regulation monitoring device, comprising: a spectrum acquisition terminal, a sensor array, a regulation terminal and a central control terminal;

[0047] The spectrum acquisition terminal is used for real-time global spectrum acquisition of flue gas, generates a three-dimensional original spectrum data set, and receives the instruction of the central control terminal to adjust the scanning frequency;

[0048] The sensor array is used for collecting activated carbon bed layer and flue gas parameters;

[0049] The regulation terminal is used for executing physical regulation and chemical regulation, receiving the gradient regulation scheme of the central control terminal, implementing regulation according to parameter gradient and time node and feeding back the equipment running state;

[0050] The central control terminal is used for receiving spectrum data to identify non-expected pollutants and updating a dynamic spectrum fingerprint library, processing sensor array data, generating a gradient regulation scheme, monitoring and correcting deviations in real time, verifying the effect, storing data and optimizing strategies through clustering analysis and incremental learning.

[0051] Specifically, the central control terminal comprises an identification module, an analysis module and a regulation optimization module;

[0052] The identification module is used for adjusting the spectrum scanning frequency, generating a time sequence change matrix based on an adaptive threshold and a dynamic sliding window, determining abnormal points and abnormal spectra, extracting a vibration characteristic parameter set, matching the vibration characteristic parameter set with a dynamic spectrum fingerprint library to identify known non-expected signals and new non-expected signals;

[0053] The analysis module is used for backstepping pollutant molecular configuration parameters, analyzing and generating an activated carbon pore characteristic matrix, constructing a multi-dimensional matching degree evaluation system, calculating a matching degree difference value, and dynamically adjusting pore regulation deduction logic;

[0054] The regulation optimization module is used for generating a gradient regulation scheme, calculating an actual pore change amount, automatically adjusting parameters if the actual pore change amount exceeds an expected deviation, verifying the effect after regulation, performing clustering analysis and incremental learning on knowledge base data, and optimizing regulation parameter recommended values and matching degree model weights.

[0055] The beneficial effects of the present application are:

[0056] Adopt fixed wavelength ultraviolet spectrum and variable wavelength infrared spectrum alternately scanning, combine probability prediction model dynamic adjustment scanning frequency, through self-adaptive threshold and sliding window accurate capture unexpected signal, then through principal component analysis stripping interference, characteristic peak extraction get vibration parameter, realize known and new unexpected signal accurate identification, at the same time based on vibration characteristic backstepping molecular configuration parameter, combine activated carbon pore diffraction pattern calculation matching degree difference, solved the problem of unable to generate adjustment strategy based on unexpected pollution physical and chemical characteristics in real time, avoid activated site and adsorption demand mismatch;For the problem of lack of trace pollution pore emergency reaming mechanism, based on matching degree difference formulate targeted strategy, from physical and chemical control library call means, through control parameter calculation formula generate gradient scheme, in the regulation, real-time scanning pore distribution and adjusting parameters, ensure fast response specific adsorption demand, solve the penetration risk caused by adsorption site loss;For the problem of quality control in the transition period of regulation, in the regulation process, through feedback interval monitoring pore change, combine matching degree evaluation system real-time determine effect, start fine tuning or regenerate scheme, realize from detection to reconstruction whole chain quality control, avoid transition period performance fluctuation. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 It is a kind of activated carbon pore control method schematic diagram;

[0058] Figure 2 It is the flow chart of screening unexpected signal in the present application;

[0059] Figure 3 It is the flow chart of matching analysis in the present application;

[0060] Figure 4 It is a kind of activated carbon pore control monitoring equipment structure diagram. DETAILED DESCRIPTION

[0061] The technical scheme of the present application will be described in detail below by the drawings and specific examples, it should be understood that the specific features in the embodiments and examples of the present application are detailed description of the technical scheme of the present application, and not the limitation of the technical scheme of the present application, in the case of no conflict, the technical features in the embodiments and examples of the present application can be combined with each other.

[0062] Example 1

[0063] Reference Figures 1 to 3 As shown in the present embodiment, a kind of activated carbon pore control method is introduced, including the following steps:

[0064] Step S1: Adopting the alternating scanning mode of fixed-wavelength ultraviolet spectrum and variable-wavelength infrared spectrum, calling the probability prediction model to adjust the spectrum scanning frequency, collecting the real-time global spectrum of the flue gas, setting the adaptive threshold to adapt to the background fluctuation, reducing the false judgment and omission, extracting the time series slope and acceleration with the dynamic sliding window, generating the time series change matrix, accurately capturing the signal trend, determining the abnormal points and abnormal spectrum, screening the unexpected signals through the continuous abnormal point distribution, ensuring the identification accuracy, analyzing the unexpected signals, stripping the interference, strengthening the unexpected characteristics, extracting the vibration parameters and matching with the dynamic fingerprint library, and realizing the accurate identification of known and new unexpected signals.

[0065] Specifically, the specific steps of adjusting the spectrum scanning frequency include:

[0066] When starting the scanning device, for some transient unexpected signals with short duration, such as mercury vapor of pipeline leakage, it is easy to miss detection. The pre-trained probability prediction model is called to obtain the probability of the occurrence of unexpected signals in each waveband in the future period. If the probability is greater than the preset probability threshold, it is determined as a potential waveband, otherwise, it is determined as a non-potential waveband. The potential waveband is predicted to appear in the future period, and the scanning frequency of the potential waveband is increased to times of the conventional frequency, and the non-potential waveband remains the conventional scanning frequency. The resources are concentrated to capture the potential pollutant signal. Without increasing the total scanning time, the transient pollutant capture rate is improved, the missing detection caused by long scanning interval is avoided, and the real-time global spectrum collection is simultaneously performed to generate a three-dimensional original spectrum data set containing wavelength, absorbance and timestamp. Then, the data set is sliced into a two-dimensional spectrum matrix according to the timestamp, each row corresponds to a wavelength, and each column corresponds to the absorbance of a time point. The probability prediction model is constructed based on the long short-term memory network, and the historical flue gas parameters, working condition load and raw material composition are input to predict the probability of the occurrence of pollutants, the timestamp.

[0067] Specifically, the specific steps of screening the unexpected signals include:

[0068] The fixed threshold value cannot adapt to the background fluctuation of flue gas, such as the sudden increase in humidity causing baseline drift, which is easy to cause misjudgment or omission. The historical spectrum data of the recent period is obtained, the high-frequency noise is eliminated by the Savitzky-Golay smoothing algorithm, and then the baseline component in the signal is separated by wavelet filtering. In each wavelength, the region with absorbance fluctuation <5% and without expected pollutant characteristic peak in the continuous 10 cycles is screened to locate the baseline region without target characteristics and smooth signal in each wavelength, such as the water vapor stable region in the infrared region. The arithmetic mean value of all signal values in the baseline region is calculated as the baseline mean value, and the square root of the average value of the square difference between the signal value in the baseline region and the baseline mean value is calculated as the baseline standard deviation to reflect the natural fluctuation of the baseline. The real-time fluctuation entropy is calculated by combining the signal variance and kurtosis of the current spectrum data, and the fluctuation ratio is calculated by combining the historical maximum fluctuation entropy. The fluctuation value is obtained by multiplying the sum of the fluctuation ratio and the basic coefficient by the baseline standard deviation, and the adaptive threshold value is obtained by the sum of the fluctuation value and the baseline mean value, so as to dynamically enlarge or reduce the threshold value range through the fluctuation ratio to adapt to the background fluctuation. The adaptive threshold value is obtained by multiplying the sum of the fluctuation value and the baseline mean value, so as to dynamically enlarge or reduce the threshold value range through the fluctuation ratio to adapt to the background fluctuation.

[0069] The sliding window is set based on the signal change rate, and the size of the sliding window will be adaptively adjusted according to the signal change rate, including: taking the current time as the end point, and intercepting the spectrum data segment of continuous sampling periods forward, which contains sampling period spectrum signal intensity to form an initial window. Based on the variable weight linear regression, the spectrum data at different times in the initial window is given a time decay weight, and the weight is exponentially decayed with time. The weight of the current time is 1, the weight of the previous time is 0.8, and so on. It is ensured that the more recent data has a greater impact on the current trend. The weighted least squares method is used to linearly regress the data in the window to fit the equation, so as to obtain the basic slope at the time . The basic slope of continuous two times is taken to calculate the slope difference, and the slope difference is normalized by combining the sampling period to obtain the signal change rate . The absolute value of the signal change rate is used as the basis to dynamically adjust the number of periods contained in the sliding window; if , the signal is smooth and the slope changes slowly, and the window is expanded to periods to smooth random noise through more historical data; if , the signal fluctuates and the slope changes sharply, and the window is reduced to periods to reduce the dilution of historical data to the current transient change, such as signal jump when the pollutant appears; wherein is the window adjustment threshold value.

[0070] Based on the self-adaptive sliding window, the time attenuation weight is still used to weight the least square method of the spectral data in the window, to fit the linear equation and get the slope as the time slope of the current time, to reflect the real change rate of the signal in the current window, and to calculate the second-order difference of the time slope of the three continuous time points, and to get the acceleration by combining the square normalization of the sampling period, to generate the time change matrix, including wavelength, timestamp, time slope and acceleration;

[0071] For each wavelength point in the time change matrix, if the time slope is greater than the adaptive threshold or the acceleration is greater than zero, it is marked as a single abnormal point, otherwise, it is determined as normal fluctuation, the regular scanning is maintained, and the data is stored in the historical spectrum library to update the baseline parameters;

[0072] The band distribution of single abnormal points is counted, if 5 or more wavelength points in a certain band are single abnormal points and not in the characteristic interval of the expected pollutant standard spectrum, it is determined as an abnormal spectrum, triggering the analysis process;

[0073] The original spectrum data corresponding to the abnormal spectrum is extracted, and the background interference spectrum library collected at the same period is imported, including water vapor and dust standard spectra under different humidity and dust concentration. The original spectrum data and the background interference spectrum library data are combined to generate a processing matrix. The covariance matrix of the processing matrix is calculated and decomposed by principal component analysis to obtain a group of eigenvalues and corresponding eigenvectors. The eigenvalues are arranged in descending order, and the proportion of the sum of the first The cumulative variance contribution rate is obtained, and the first three principal components with a cumulative variance contribution rate not less than 95% are selected, of which the first principal component corresponds to the background interference, and the second and third principal components correspond to the pollutant signal.

[0074] The original spectrum data of the abnormal spectrum is projected into the eigenvector space of the three principal components to obtain the principal component score matrix, that is, the projection value of the original spectrum data on each principal component. The score of the first principal component is removed, that is, the contribution of the background interference is removed, and only the scores of the second and third principal components are retained, that is, the corresponding pollutant signal. Through the transpose operation of the eigenvector matrix, the retained second and third principal component scores are inversely projected into the original wavelength space to obtain a denoising spectrum containing only the pollutant signal, which effectively removes the water vapor and dust interference and highlights the characteristic peaks of the unexpected signal.

[0075] The pre-stored expected pollutant standard spectrum library is called, the standard spectrum library is taken as an independent variable, the denoising spectrum is taken as a dependent variable, a regression model is constructed by using a partial least squares regression method, a residual spectrum is calculated based on a difference between the denoising spectrum and a model predicted value, the expected pollutant signal is removed, only the unexpected signal characteristics are reserved, and peak intensity normalization processing is performed on the residual spectrum, the maximum peak is set as 1, weak characteristic peaks are highlighted, and the strengthened unexpected characteristic spectrum is obtained;

[0076] Baseline correction is performed on the unexpected characteristic spectrum, polynomial fitting is adopted, the flat regions at both ends of the original spectrum are taken as a reference, a baseline trend line is fitted, the original spectrum data is subtracted from the baseline trend line, baseline tilting caused by instrument drift or background interference is eliminated, the corrected spectrum is smoothed by using a moving average method, sawtooth fluctuations caused by high-frequency noise are removed, the smoothed spectrum is obtained, the stability of subsequent peak shape identification is ensured, the smoothed spectrum is differentiated to find extreme points with a derivative of zero, the peak top points in the spectrum corresponding to the points where the derivative changes from positive to negative are found based on the extreme points, and are taken as candidate peak positions, extreme points with an absorbance value less than 3 times a baseline standard deviation are excluded to avoid noise misjudgment as a peak top point, candidate peak positions with peak intensity significantly higher than baseline fluctuations are reserved, a preliminary peak position list is generated, a Gaussian-Lorentz mixed function is used to perform peak fitting on the smoothed spectrum, for each candidate peak position in the preliminary peak position list, a mixed function composed of a Gaussian function and a Lorentz function in a preset proportion is set as a peak shape, the preliminary peak position is taken as a starting point, peak shape parameters are optimized by an iterative algorithm, residual error between the fitted curve and the original spectrum is calculated based on a sum of squares of differences between the fitted value and the measured value, until the residual error is less than a residual error proportion, to ensure fitting accuracy, key parameters of each characteristic peak, including a peak position, a half-peak width, a peak intensity, a peak area, are extracted, and are sequentially arranged in ascending order of the peak position, a vibration characteristic parameter set is generated;

[0077] The standardization is performed on each parameter in the vibration characteristic parameter set, and a dimensionless vibration feature vector is converted. A dynamic spectrum fingerprint library is called, a cosine similarity of the vibration feature vector and a known unexpected standard vector in the dynamic spectrum fingerprint library is calculated, and a maximum value of the cosine similarity is selected as a matching similarity. Meanwhile, the matching similarity corresponding to the unexpected standard vector is taken as a matching spectrum. If the matching similarity is not less than a matching threshold, the corresponding known contaminant in the dynamic spectrum fingerprint library is directly locked as a known unexpected signal. Otherwise, a new unexpected signal is determined, and the vibration feature vector is stored in the dynamic spectrum fingerprint library. The dynamic spectrum fingerprint library is accumulated and constructed by basic data, and is used for storing and managing the spectral characteristics and associated information of unexpected signals, providing a standard comparison benchmark for contaminant identification, and realizing self-evolution through continuous learning. A library updating mechanism is arranged in the library. Once a new unexpected signal is identified, the corresponding metadata is extracted, including the vibration characteristic parameter set, the standardized feature vector, the abnormal spectrum original data, the occurrence timestamp, the working condition parameter, the unique identification of the new contaminant, the storage in the new contaminant sub-library, and the association with the corresponding processing record, such as the subsequent pore regulation strategy and removal effect, to form an associated data chain. Through an incremental learning algorithm, the historical mean value library and the standard deviation library are adjusted, the new contaminant parameters are included in the statistical range, the adaptability of subsequent standardization processing is ensured, the feature vectors of similar contaminants are subjected to cluster analysis, similar features are merged, such as the same contaminant at different concentrations, redundant data is reduced, and the comparison efficiency is improved.

[0078] Step S2: Based on the extracted vibration characteristics, a pre-stored correlation algorithm of molecular size and vibration mode is called, the molecular configuration parameters of the contaminant molecule are inversely deduced through the characteristic peak spacing, and the matching degree difference between the pore size and the molecular configuration is calculated in combination with the online diffraction spectrum of the current pore distribution of the activated carbon, so as to perform matching analysis. According to the matching degree difference, the deduction logic is dynamically adjusted to ensure the accuracy of the analysis of the adaptability of the pore characteristics and the contaminant molecule.

[0079] Specifically, the specific steps of the matching analysis include:

[0080] Based on the vibration characteristic parameter set, the unexpected molecular configuration parameter backstepping is carried out, the high-precision and high-efficiency conversion of spectral characteristics to molecular configuration is realized, for the known unexpected signal, the corresponding sub-type model is mobilized, and the sub-type model is trained based on the quantum chemical data of the same type of unexpected signal, the spectral characteristics are converted into molecular geometric parameters, including: the molecular space configuration is backstepped through the characteristic peak spacing and peak intensity ratio, the corresponding bond length is called by using the peak position in the vibration characteristic set based on the peak position and bond length correlation table in the sub-type model, the minimum circumscribed circle diameter is calculated combined with the molecular space configuration, the polarity parameter is calculated by the characteristic peak position offset, the atomic position is calculated by the VSEPR theory (valence shell electron pair repulsion theory), the three-dimensional model is generated combined with the bond angle data, the minimum spherical volume wrapping all atoms is taken as the conservative value of the minimum envelope volume, and the molecular configuration parameters in a unified format are generated;

[0081] For new unexpected signals, a general sub-model is called, and an incremental learning mechanism is triggered, and a dedicated sub-model is generated subsequently, a multi-task learning architecture is adopted, multiple spectral characteristics are processed simultaneously, different waveband weights are dynamically allocated through an attention mechanism to automatically focus on the key waveband of unknown characteristic peaks, based on the correlation between peak position and bond energy, combined with the influence of working condition parameters on molecular vibration frequency, the preliminary range of molecular dynamic diameter is backstepped, based on the correlation between peak intensity ratio and molecular symmetry, the fuzzy interval of polarity parameter is output, multiple molecular conformations are generated by Monte Carlo simulation, combined with the molecular flexibility reflected by the half-peak width, the minimum spherical volume wrapping all atoms is taken as the conservative value of the minimum envelope volume, and the initial parameter set is generated;

[0082] Based on the initial parameter set, an incremental learning is carried out to generate a dedicated sub-model, the initial parameter set, the original spectral data and the working condition parameters are packaged as an incremental learning data package, which is transmitted to a model training cluster through a federal learning interface to ensure data privacy, the training cluster adopts a knowledge transfer strategy, freezes the core feature extraction layer in the general sub-model for identifying known unexpected signals, and only fine-tunes the model parameters for the unique features of unknown unexpected signals, such as unknown peak shape in specific waveband;

[0083] Through multiple rounds of iterative optimization, the deviation between the molecular parameters predicted by the dedicated sub-model and the virtual labels generated based on the interpolation of similar structure molecules is compared, the parameter weight is continuously adjusted, and finally the dedicated sub-model for unknown unexpected signals is generated to output the optimized molecular parameter set, and the dedicated sub-model is verified, combined with the physical correlation between the molecular parameters and the smoke diffusion characteristics, whether the parameters meet the basic physical properties of gas molecules is verified, the optimized parameters are input into a pore matching degree calculation model, if the output matching degree difference is in a reasonable interval, it means that the parameters have no obvious deviation, at this time the dedicated sub-model is stored in the system model library for direct calling in subsequent identification of the same type of pollutants, and the molecular configuration parameters of new unexpected signals are generated.

[0084] The small-angle X-ray scattering device built-in in the bed is started to perform a global scan on the activated carbon bed to generate a pore diffraction spectrum. The diffraction spectrum is subjected to peak fitting according to the Porod law. By setting a secondary diameter threshold, the pore diameter is divided into three categories, including micropore, mesopore and macropore. The diffraction spectrum is decomposed into scattering signals of micropore, mesopore and macropore. The pore volume proportion and average proportion in each pore diameter interval are calculated. In combination with the bed pressure drop distribution data, the pore distribution result is corrected. If the regional pressure drop is higher than the preset proportion of the average value, it is determined that the pore is blocked, and the pore volume weight of the corresponding pore diameter is lowered;

[0085] An X-ray photoelectron spectroscopy (XPS) device is started synchronously to analyze the distribution density of the functional groups on the surface of the activated carbon, calculate the density of the polar sites, reflect the adsorption affinity of the pores to the polar molecules, and finally generate an activated carbon pore characteristic matrix including a pore size distribution curve, an average pore diameter and a density of polar sites, so as to capture the dynamic change of the pores in real time and provide a current state benchmark for the matching degree calculation;

[0086] Based on the molecular diffusion theory, the effective pore range is determined so that the molecules can enter the pores and are not easily blocked due to excessive size. The proportion of the pore volume falling within the effective pore range in the total pore volume in the activated carbon pore size distribution is calculated to obtain the effective pore proportion. The effective pore proportion is taken as a basic value, and a size matching degree is obtained by normalizing the operation through a Sigmoid function, so as to reflect the overall adaptation degree of the molecular size and the pore size.

[0087] The absolute difference formula is used to calculate the polarity difference between the unexpected signal and the pore surface. The polarity matching degree is set to be negatively correlated with the polarity difference. The polarity difference is mapped to the polarity matching degree by using an inverse proportional function for normalization processing.

[0088] Based on the ratio of the total pore volume to the number of pores, the average volume of a single pore is calculated. The volume adaptation ratio is set by using the ratio of the volume of the smallest envelope sphere to the average volume of a single pore, so as to reflect the proportional relationship between the molecular volume and the pore space. Based on the molecular packing theory, the volume ratio interval is determined so that the molecules can fully utilize the pore space and do not affect the adsorption stability due to crowding. The volume matching degree is calculated by using a piecewise function. If the volume adaptation ratio is within the volume ratio interval, the volume matching degree is 1. If the volume adaptation ratio is less than the lower limit of the volume ratio interval, the volume matching degree is linearly reduced with the decrease of the volume adaptation ratio. If the volume adaptation ratio is greater than the upper limit of the volume ratio interval, the volume matching degree is linearly reduced with the increase of the volume adaptation ratio.

[0089] The weight is set for each matching degree to construct a multi-dimensional matching degree evaluation system, and the comprehensive matching degree is calculated. Based on the preset target matching degree, the matching degree difference is calculated, and the secondary matching threshold is set to divide the matching level, including severe mismatch, moderate mismatch and basic matching. Based on the matching level, the deduction logic is dynamically adjusted to generate the optimized pore regulation deduction logic. If it is severe mismatch, the deduction logic automatically strengthens the "emergency hole expansion" weight, preferentially increases the hole diameter, if it is moderate mismatch, the deduction logic focuses on "polarity site regulation", such as activating potential polarity sites by physical field, if it is basic matching, the deduction logic maintains "fine-tuning mode", and only optimizes the pore volume distribution.

[0090] Step S3: According to the matching degree difference, a gradient regulation scheme is constructed to regulate the pore gradient. The change of the pore and the effect are monitored in real time, and the obtained data is classified into the database. The optimal parameters are extracted by cluster analysis, the model weight is optimized by incremental learning, the self-evolution is realized, the regulation efficiency and accuracy are continuously improved, and reliable support is provided for pollutant adsorption.

[0091] Specifically, the specific steps of pore gradient regulation include:

[0092] Based on the matching degree difference, a targeted strategy is formulated to determine the core regulation target. At the same time, the corresponding regulation means are called from the preset physical regulation library and chemical regulation library, and the regulation priority is set based on the deduction logic. The specific parameter gradient and time node are calculated by calling the corresponding regulation parameter calculation formula based on the matching degree difference as input, to generate a gradient regulation scheme, including regulation means combination, parameter gradient and time node, to avoid blind operation; The regulation target includes: for severe mismatch, the core target is to improve the size matching degree, for moderate mismatch, the core target is to simultaneously optimize the size matching degree and the polarity matching degree, for basic matching, the core target is to maintain the existing matching state through fine-tuning, and for new unexpected signals, a conservative coefficient is added in the original target setting, which is based on the error range of molecular parameter backstepping, to appropriately amplify the regulation target and avoid insufficient regulation due to parameter error; The regulation means include: for severe mismatch, physical regulation is mainly used to preferentially solve the size adaptation problem, for moderate mismatch, physical and chemical regulation is used cooperatively to consider size and polarity, and for basic matching, only physical fine-tuning is used to avoid excessive intervention; The regulation parameter calculation formula is constructed based on historical regulation data and pore change law;

[0093] Based on the generated gradient regulation scheme, regulation is carried out, and during the regulation process, the pore distribution is scanned according to the feedback interval, and the regulation pore volume proportion is calculated. Based on the pore volume proportion calculated in the matching analysis, difference calculation is carried out to obtain the actual pore change amount to judge whether the current regulation exceeds the expected deviation. Once the actual pore change amount exceeds the expected deviation, the parameters are automatically adjusted to correct the pore optimization according to the target through real-time feedback, which significantly improves the accuracy of the regulation;

[0094] After the regulation is completed, a multi-dimensional matching degree evaluation system is called, and based on the real-time monitored pore characteristics, the geometric matching degree, the polarity matching degree and the spatial matching degree are recalculated to obtain the regulation comprehensive matching degree and the regulation matching degree difference, so as to determine the regulation effect according to the matching degree difference;

[0095] If the regulation matching degree difference is less than the first regulation threshold, it indicates that the current regulation is up to standard, the regulation is stopped, the key parameters of this regulation are recorded as the benchmark regulation scheme, and are stored in the pore regulation knowledge base and associated with the corresponding pollutant type to provide reference for subsequent regulation of the same type of pollutant;

[0096] If the regulation matching degree difference is between the first regulation threshold and the second regulation threshold, it indicates that it is slightly unqualified, and the fine adjustment mode is started, such as low-frequency ultrasonic vibration, which optimizes the pore connectivity through gentle physical action to further reduce the pore defects;

[0097] If the regulation matching degree difference is greater than the second regulation threshold, it indicates that the regulation is insufficient, and a new regulation scheme is generated based on the current pore characteristics and executed until the matching degree is up to standard;

[0098] The database is updated and the strategy is optimized, the whole chain data of this regulation is stored in the pore regulation knowledge base according to the category, the clustering analysis is carried out on the multiple regulation data of the same type of unexpected signal, the optimal parameter combination is extracted, the recommended parameters of the gradient regulation scheme are updated, the new data is transmitted into the multi-dimensional matching degree evaluation system, the model weight distribution is optimized through incremental learning, and the accuracy of subsequent matching degree calculation is improved.

[0099] Embodiment 2

[0100] Please refer to Figure 4 Another embodiment provided by the application: a kind of activated carbon pore regulation monitoring equipment, it include: spectrum acquisition terminal, sensor array, regulation terminal and central control terminal;

[0101] The spectrum acquisition terminal is used to adopt fixed wavelength ultraviolet spectrum and variable wavelength infrared spectrum alternately scanning mode to carry out real-time global spectrum acquisition to flue gas, generates three-dimensional original spectrum data set containing wavelength, absorbance, time stamp, and is sliced into two-dimensional spectrum matrix according to time stamp, simultaneously receives the instruction of central control terminal and adjusts scanning frequency;

[0102] The sensor array is used to collect relevant parameters of the activated carbon bed and flue gas, including generating activated carbon pore diffraction patterns by a small-angle X-ray scattering device, analyzing activated carbon surface functional group distribution density by an X-ray photoelectron spectroscopy device, collecting bed pressure drop distribution data by a pressure sensor, and collecting working condition parameters such as flue gas temperature, humidity, and dust concentration;

[0103] The regulation terminal is used to perform pore regulation operations including physical regulation and chemical regulation, receive the gradient regulation scheme of the central control terminal, implement regulation according to parameter gradients and time nodes, and feed back the equipment running state in the regulation process;

[0104] The central control terminal is used to overall coordinate the operation of the spectrum collection terminal, the sensor array, and the regulation terminal, realize intelligent management of the whole process of activated carbon pore regulation, receive spectrum data from the spectrum collection terminal, identify unexpected pollutant signals by analysis and update the dynamic spectrum fingerprint library, process activated carbon pore characteristics and working condition data collected by the sensor array, back-propagate pollutant molecular configuration parameters and calculate the matching degree difference between the pores and the pollutants, generate a gradient regulation scheme based on the matching degree difference, control the regulation terminal to perform physical or chemical regulation operations, monitor the regulation process in real time and correct deviations, verify the regulation effect, classify the whole chain data into the library, optimize the system strategy through cluster analysis and incremental learning, and finally realize closed-loop management from pollutant identification, pore matching analysis to precise regulation and effect optimization, to ensure efficient adaptation of activated carbon pores to pollutants and improve pollutant removal efficiency and system stability.

[0105] Specifically, the central control terminal includes an identification module, an analysis module, and a regulation optimization module.

[0106] The identification module is used to process spectrum data output by the spectrum collection terminal, adjust the spectrum scanning frequency to improve the transient pollutant capture rate by calling a probability prediction model, set an adaptive threshold to adapt to flue gas background fluctuations, generate a time series change matrix by extracting time series slope and acceleration through a dynamic sliding window, determine abnormal points and abnormal spectra, analyze the background interference of abnormal spectra, strengthen unexpected features, extract a vibration feature parameter set, match the vibration feature parameter set with the dynamic spectrum fingerprint library, identify known and new unexpected signals, and update the dynamic spectrum fingerprint library.

[0107] The analysis module calls a type-specific sub-model or a general sub-model to back-propagate pollutant molecular configuration parameters based on the vibration feature parameter set output by the identification module, receives activated carbon pore diffraction patterns and pressure drop data output by the sensor array, analyzes and generates an activated carbon pore characteristic matrix, constructs a multi-dimensional matching degree evaluation system, calculates size matching degree, polarity matching degree, spatial matching degree, and comprehensive matching degree difference, and dynamically adjusts pore regulation deduction logic based on the matching degree difference.

[0108] The regulation optimization module formulates a targeted regulation strategy based on the matching degree difference value and the deduction logic output by the analysis module, calls regulation means from a physical regulation library and a chemical regulation library and sets priorities, calls a regulation parameter calculation formula to generate a gradient regulation scheme, controls a regulation terminal to execute the regulation scheme, receives real-time feedback of pore change data from a sensor array, calculates an actual pore change amount, and automatically adjusts parameters if the actual pore change amount exceeds an expected deviation. After the regulation ends, a multi-dimensional matching degree evaluation system is called to verify the effect, and if the effect meets the standard, the regulation parameters are recorded as a benchmark scheme and stored in a knowledge base, and if the effect does not meet the standard, a new scheme is generated until the effect meets the standard. The knowledge base data is subjected to cluster analysis and incremental learning to optimize the regulation parameter recommended value and the matching degree model weight, and the system is self-evolved.

[0109] Working principle and effect:

[0110] By alternately scanning the flue gas with fixed-wavelength ultraviolet spectrum and variable-wavelength infrared spectrum, a probability prediction model is called to adjust the scanning frequency and focus on potential unexpected signal bands. Combined with an adaptive threshold and an adaptive sliding window, a time series change matrix is formed by extracting the time series slope and acceleration, and abnormal points are marked and abnormal spectra are screened. Then, through principal component analysis to strip background interference and partial least squares regression to remove expected signals, vibration characteristic parameters are extracted and matched with a dynamic spectrum fingerprint library to realize the identification of known and new unexpected signals.

[0111] Subsequently, based on the vibration characteristic parameters, the molecular configuration parameters are backstepped, the pore characteristics of activated carbon are obtained by small-angle X-ray scattering and X-ray photoelectron spectroscopy at the same time, the size, polarity, volume matching degree and comprehensive difference value are calculated as the regulation basis, and a targeted strategy is formulated based on the matching degree difference value to call means from a preset library and generate a gradient scheme combined with means, parameter gradient and time node based on a parameter calculation formula.

[0112] During the regulation, the pore change is monitored according to the feedback interval, and the parameters are automatically adjusted when the expected deviation is exceeded. After the end, the effect is determined by a multi-dimensional evaluation system, and if the effect meets the standard, it is stored as a benchmark scheme. If the effect does not meet the standard, fine-tuning is started, and if the effect is insufficient, a new scheme is generated. The full-chain data is stored in a knowledge base, and the parameters and model weights are optimized through cluster analysis and incremental learning to form a closed-loop optimization mechanism.

[0113] Through adaptive signal analysis, real-time capture of unexpected pollutants is realized, and a precise regulation benchmark is established based on dynamic calculation of molecule-pore matching degree. Combined with gradient regulation and closed-loop updating mechanism, dynamic adaptation of pore characteristics and pollutant demand is ensured, and the regulation adaptability problem when unexpected signals invade is solved.

[0114] The above merely describes the preferred embodiments of the present application, and the protection scope of the present application is not limited to the above-described embodiments. Any technical solution falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principles of the present application shall also be considered as falling within the protection scope of the present application.

Claims

1. A method for controlling the pore size of activated carbon, characterized in that, include: Based on a probabilistic prediction model, the spectral scanning frequency is dynamically adjusted, and real-time global spectral acquisition is performed. An adaptive threshold is set, and a dynamic sliding window is used to extract the temporal slope and acceleration, generate a temporal change matrix, identify abnormal spectra, filter out unexpected signals and analyze them, infer molecular configuration parameters through the characteristic peak spacing, calculate the difference in matching degree between pore size and molecular configuration, and perform matching analysis. Based on the matching degree difference, pore gradient control is performed, and real-time monitoring is conducted to verify the control effect. The steps for filtering out unexpected signals include: For each wavelength point in the time series variation matrix, if the time series slope is greater than the adaptive threshold or the acceleration is greater than zero, it is marked as a single abnormal point; otherwise, it is judged as a normal fluctuation and the data is stored in the historical spectrum library. The band distribution of single anomalies is statistically analyzed. If five or more consecutive wavelength points in a certain band are single anomalies and are not within the characteristic range of the expected pollutant standard spectrum, then it is determined to be an abnormal spectrum and the analysis process is triggered. Extract the original spectral data corresponding to the abnormal spectra and merge them with the background interference spectral library collected at the same time to generate a matrix to be processed; The covariance matrix of the matrix to be processed is calculated by principal component analysis and eigenvalues ​​and eigenvectors are obtained. The top three principal components with a cumulative variance contribution rate of not less than 95% are selected, and the first principal component is removed to generate a denoised spectrum. The pre-stored standard spectral library of expected pollutants is called, and a regression model is constructed using partial least squares regression with the standard spectral library as the independent variable and the denoised spectrum as the dependent variable. Based on the difference between the denoised spectrum and the model prediction, the residual spectrum is calculated, and the peak intensity of the residual spectrum is normalized to generate the unexpected feature spectrum.

2. The method for controlling the pore size of activated carbon according to claim 1, characterized in that, The steps for adjusting the spectral scanning frequency include: When the scanning device is started, a probability prediction model is invoked to predict the probability of unexpected signals appearing in each band in the future period; Bands with a probability greater than a probability threshold are identified as potential bands, and the rest are non-potential bands. Set the scanning frequency of the potential band to the normal frequency. The non-potential band maintains the regular scanning frequency, and real-time global spectral acquisition is performed simultaneously.

3. The method for controlling the pore size of activated carbon according to claim 2, characterized in that, The steps for generating the time series variation matrix include: Acquire historical spectral data for the most recent period and locate the baseline region within the spectral data; Calculate the baseline parameters of the baseline region, including the mean and standard deviation, and simultaneously calculate the real-time fluctuation entropy of the current spectral data. Combined with the historical maximum fluctuation entropy, calculate the fluctuation ratio. An adaptive threshold is set by combining the baseline parameters with the fluctuation ratio. Using the current time as the endpoint, an initial window is set. Based on variable weight linear regression, time decay weights are assigned to the data within the initial window. Weighted least squares method is used for fitting to obtain the basic slope at the current time. Calculate the difference in fundamental slope between two consecutive time points, and normalize it using the sampling period to obtain the rate of change of the signal. : like The window is expanded to One cycle; if Window minimized to One cycle; among which, Adjust the threshold for the window; Based on the adjusted sliding window, the linear equation is fitted using the weighted least squares method with time decay weights to obtain the time-series slope. The second-order difference of the time-series slope at three consecutive time points is calculated and normalized by the square of the sampling period to obtain the acceleration, thus generating the time-series change matrix.

4. The method for controlling the pore size of activated carbon according to claim 3, characterized in that, The specific steps for filtering unexpected signals also include: Baseline correction and moving average smoothing are performed on the unexpected characteristic spectrum to generate a smoothed spectrum. The first derivative is calculated to find candidate peak positions and then screened to generate a preliminary peak position list. The smooth spectrum peaks were fitted using a Gaussian-Lorentz mixture function, and the key parameters of each characteristic peak were extracted and sorted by peak position to generate a set of vibrational characteristic parameters. Each parameter in the vibration feature parameter set is standardized to obtain a vibration feature vector. The dynamic spectral fingerprint library is called to calculate the cosine similarity between the vibration feature vector and the known unexpected standard vectors in the library. The maximum value is selected as the matching similarity and the corresponding matching spectrum. If the matching similarity is not less than the matching threshold, it is determined to be a known unexpected signal; otherwise, it is determined to be a new unexpected signal, and the vibration feature vector is stored in the dynamic spectral fingerprint database.

5. The method for controlling the pore size of activated carbon according to claim 4, characterized in that, The steps of matching analysis include: For known unexpected signals, the corresponding classification sub-model is activated to convert spectral features into molecular geometric parameters and generate molecular configuration parameters. For new unexpected signals, the general sub-model is invoked and the incremental learning mechanism is triggered to generate an initial parameter set. Incremental learning is then carried out based on the initial parameter set to generate a specific sub-model and to generate the molecular configuration parameters of the new unexpected signals.

6. The method for controlling the pore size of activated carbon according to claim 5, characterized in that, The steps of matching analysis also include: The small-angle X-ray scattering device and X-ray photoelectron spectroscopy device built into the bed are activated to generate the activated carbon pore feature matrix. The size matching degree is calculated based on molecular diffusion theory, the polarity matching degree is calculated using the absolute difference formula, and the volume matching degree is calculated based on molecular packing theory. Weights are set for each matching degree and the overall matching degree is calculated. The matching degree difference is calculated based on the preset target matching degree, a secondary matching threshold is set to divide the matching level, and the deduction logic is dynamically adjusted based on the matching level to generate the optimized pore control deduction logic.

7. The method for controlling the pore size of activated carbon according to claim 6, characterized in that, The steps for pore gradient control include: Based on the matching degree difference, the core control target is identified, and the corresponding control methods are retrieved from the preset physical control library and chemical control library. The control priority is set based on the deduction logic, and the corresponding control parameter calculation formula is called with the matching degree difference as input to calculate the parameter gradient and time node, thereby generating a gradient control scheme. The control is carried out based on the gradient control scheme. During the control process, the pore distribution is scanned at the feedback interval, the control pore volume ratio is calculated, the actual pore change is calculated based on the pore volume ratio in the matching analysis, and the parameters are automatically adjusted when the actual pore change exceeds the expected deviation. After the regulation is completed, a multi-dimensional matching degree evaluation system is invoked to calculate the regulation matching degree difference based on the real-time monitored pore characteristics, so as to determine the regulation effect and make corrections. The entire chain of data from this regulation is stored in the pore regulation knowledge base according to category. Cluster analysis is performed on multiple regulation data of the same type of unexpected signals to extract the optimal parameter combination, update the recommended parameters of the gradient regulation scheme, and input the new data into the multi-dimensional matching degree evaluation system and optimize the weight allocation through incremental learning.

8. An activated carbon pore control monitoring device, used to implement the activated carbon pore control method as described in any one of claims 1-7, characterized in that, include: Spectrum acquisition terminal, sensor array, control terminal, and central control terminal; The spectral acquisition terminal is used to perform real-time full-domain spectral acquisition of flue gas, generate a three-dimensional raw spectral dataset, and receive instructions from the central control terminal to adjust the scanning frequency. The sensor array is used to collect parameters of the activated carbon bed and flue gas. The control terminal is used to perform physical and chemical control, receive the gradient control scheme from the central control terminal, implement control according to parameter gradients and time nodes, and provide feedback on the equipment operating status. The central control terminal is used to receive spectral data, identify unexpected pollutants and update the dynamic spectral fingerprint database, process sensor array data, generate gradient control schemes, monitor and correct deviations in real time, verify the effect, store the data in the database, and optimize the strategy through cluster analysis and incremental learning.

9. The activated carbon pore control and monitoring device according to claim 8, characterized in that: The central control terminal includes an identification module, an analysis module, and a regulation and optimization module; The identification module is used to adjust the spectral scanning frequency, generate a time-series change matrix based on an adaptive threshold and a dynamic sliding window, determine abnormal points and abnormal spectra, extract a set of vibration feature parameters, and match the set of vibration feature parameters with a dynamic spectral fingerprint database to identify known unexpected signals and new unexpected signals. The analysis module is used to infer the molecular configuration parameters of pollutants, analyze and generate the activated carbon pore feature matrix, construct a multi-dimensional matching degree evaluation system, calculate the matching degree difference, and dynamically adjust the pore regulation deduction logic. The regulation and optimization module is used to generate a gradient regulation scheme and calculate the actual porosity change. If the deviation exceeds the expected value, the parameters are automatically adjusted. After the regulation is completed, the effect is verified, and cluster analysis and incremental learning are performed on the knowledge base data to optimize the recommended values ​​of the regulation parameters and the weights of the matching degree model.

Images