Critical point drying control method and device based on online near-infrared monitoring

Abstract

The invention relates to the technical field of electron microscope sample pretreatment, and discloses a critical point drying control method and device based on online near-infrared monitoring, and the method comprises the steps: obtaining infrared spectrum data, and unmixing the spectrum data to obtain a feature data set of a solvent; outputting a to-be-corrected signal according to the feature data set, correcting the to-be-corrected signal with preset data, and determining a signal feature of the solvent; carrying out dynamic correction on the signal characteristics and a preset characteristic matrix, outputting concentration time sequence data of the solvent, carrying out steady state judgment by combining a preset convergence threshold value, and determining a critical state of the solvent; acquiring control strategy data according to the critical state, analyzing the control strategy data into execution response characteristics, performing multi-channel mapping and correction processing, and outputting a process control signal; and performing control gain correction on the process control signal to finish closed-loop feedback control of the final drying process. The method is based on multi-solvent spectrum sensing and concentration dynamic analysis, and self-adaptive closed-loop control of the critical point drying process is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electron microscope sample pre-treatment, and in particular to a critical point drying control method and device based on online near-infrared monitoring. BACKGROUND

[0002] Critical point drying is a key step for processing biological soft tissues, hydrogels and other fragile samples before scanning electron microscope (SEM) observation. The control accuracy directly determines the preservation quality of the three-dimensional structure of the sample. Precise judgment and adaptive control of the critical point during the drying process are the core to improve the success rate, efficiency and consistency of sample processing.

[0003] In one prior art, critical point drying control mainly uses a method based on a preset time program or fixed physical parameter threshold. Although this method introduces temperature, pressure and other sensors for process monitoring, its control logic still essentially depends on the static judgment of a single or limited physical quantity, and is not associated with the chemical composition information of the dynamic changes of the multi-solvent system during the drying process. The root cause lies in the fact that the existing method is limited by the real-time and in-situ analysis and fusion capability of multi-component solvent concentration signals, especially the inability to use a sensor network chip to realize high-throughput and parallel near-infrared spectrum data acquisition and front-end intelligent processing, so as to accurately separate and track the real-time concentration of each key solvent in a complex mixed system, resulting in a lag and inaccuracy in the judgment of the critical point.

[0004] In summary, the prior art lacks real-time and in-situ sensing capability of key chemical indicators during the drying process, and the control logic is disconnected from the physical state, lacking closed-loop feedback control capability based on real-time chemical indicator sensing. SUMMARY

[0005] The present application provides a critical point drying control method and device based on online near-infrared monitoring, aiming to solve the problem that the existing critical point drying control method is difficult to realize closed-loop feedback control of the drying system.

[0006] In a first aspect, to solve the above technical problems, the present application provides a critical point drying control method based on online near-infrared monitoring, comprising: Obtaining raw infrared spectrum data, constructing a multi-dimensional spectrum matrix from the raw infrared spectrum data and performing interference correction and spectrum unmixing processing to obtain a feature data set of each solvent; Performing modal decomposition and feature band positioning on the feature data set to output a to-be-corrected feature signal, fitting and correcting the to-be-corrected feature signal with a preset standard solvent fingerprint data to determine the signal features corresponding to each solvent; According to the signal characteristics, the instantaneous concentration is obtained by combining a preset solvent light absorption characteristic matrix, the instantaneous concentration is dynamically corrected and change trend analysis is performed, and a concentration curve of each solvent is output; Concentration time sequence data of the concentration curve is acquired, a multi-dimensional feature matrix is constructed based on the concentration time sequence data, and steady state determination is performed in combination with a preset convergence threshold, and a critical state of a drying process of each solvent is determined; An operating state vector is constructed according to the critical state, and the operating state vector is processed based on a feedback compensation adjustment mechanism of a rate deviation, and control strategy data for dynamically regulating the drying process is obtained; The control strategy data is parsed into execution response characteristics, multi-channel mapping and correction processing are performed on the execution response characteristics, and stable process control signals are output; A state time sequence diagram is generated according to the process control signals, a quality and efficiency correlation feature vector is extracted according to the state time sequence diagram and control gain correction is performed, and closed-loop feedback control of the final drying process is completed.

[0007] In a second aspect, the present application provides a critical point drying control device based on online near-infrared monitoring, characterized in that it comprises: A data acquisition module is configured to acquire original infrared spectrum data, construct a multi-dimensional spectrum matrix based on the original infrared spectrum data, and perform interference correction and spectrum unmixing processing to obtain a feature data set of each solvent; A signal feature determination module is configured to perform modal decomposition and feature frequency band positioning on the feature data set, output a to-be-corrected feature signal, perform fitting correction on the to-be-corrected feature signal and preset standard solvent fingerprint data, and determine the signal characteristics corresponding to each solvent; A concentration curve generation module is configured to obtain instantaneous concentration by combining a preset solvent light absorption characteristic matrix according to the signal characteristics, dynamically correct the instantaneous concentration, and analyze the change trend, and output a concentration curve of each solvent; A critical state determination module is configured to acquire concentration time sequence data of the concentration curve, construct a multi-dimensional feature matrix based on the concentration time sequence data, and perform steady state determination in combination with a preset convergence threshold, and determine the critical state of a drying process of each solvent; A strategy data generation module is configured to construct an operating state vector according to the critical state, and process the operating state vector based on a feedback compensation adjustment mechanism of a rate deviation, and obtain control strategy data for dynamically regulating the drying process; A process control signal generation module is configured to parse the control strategy data into execution response characteristics, perform multi-channel mapping and correction processing on the execution response characteristics, and output stable process control signals; A closed-loop feedback control module is configured to generate a state timing diagram according to the process control signal, extract a quality-efficiency correlation feature vector according to the state timing diagram, and correct a control gain to complete closed-loop feedback control of the final drying process.

[0008] Compared with the prior art, the present application has the following beneficial effects: (1) The present application realizes effective separation and extraction of characteristic signals of each solvent in a multi-solvent mixed system by performing interference correction, spectral unmixing and multi-dimensional feature modeling on the online acquired near-infrared spectral data. This method breaks through the limitation of relying only on a single physical parameter or overall spectral intensity change for judgment in the existing critical point drying process, so that the key chemical component information in the drying process can be perceived in real time and in situ, thereby providing a stable and reliable data basis for subsequent critical state determination; (2) The present application realizes instantaneous concentration inversion by modal decomposition, characteristic frequency band positioning and fingerprint data fitting correction of the solvent characteristic signal, and further outputs a continuous concentration change curve in combination with the solvent absorption characteristics. This technology establishes a quantitative correlation between spectral information and solvent concentration, so that the dynamic evolution of the multi-solvent concentration in the drying process can be accurately tracked, effectively avoiding the inaccuracy of concentration judgment caused by solvent aliasing or signal drift in the prior art; (3) The present application constructs a multi-dimensional feature matrix based on the solvent concentration time series data, and determines the steady state in combination with a convergence threshold, thereby realizing accurate identification of the critical state of the drying process. This method introduces the change rate and evolution trend of the solvent concentration into the critical point judgment logic, overcoming the hysteresis and uncertainty caused by relying on fixed time or a single threshold in traditional methods, and improving the accuracy and consistency of critical point determination; (4) Based on the determination of the critical state, the present application generates dynamic control strategy data by constructing an operating state vector and introducing a feedback compensation adjustment mechanism based on rate deviation, and performs adaptive correction on the control gain in combination with closed-loop feedback, thereby realizing real-time regulation of the drying process. This technology tightly couples the chemical state sensing result with the control execution process, so that the drying process can be dynamically adjusted according to the actual state change, thereby significantly improving the control accuracy, process stability and sample processing quality of the critical point drying. BRIEF DESCRIPTION OF DRAWINGS

[0009] Figure 1 is a critical point drying control method flowchart provided by an embodiment of the present application based on online near-infrared monitoring; Figure 2 is a critical point drying control device structure schematic diagram provided by an embodiment of the present application based on online near-infrared monitoring. DETAILED DESCRIPTION

[0010] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0011] To solve the above problems, with reference to Figure 1 The first embodiment of the present application provides a charging pile intelligent site selection method based on big data analysis, comprising the following steps: S1, obtaining original infrared spectrum data, constructing a multi-dimensional spectrum matrix according to the original infrared spectrum data and performing interference correction and spectrum unmixing processing to obtain a feature data set of each solvent; S2, modal decomposition and feature band positioning are performed on the feature data set, and a to-be-corrected feature signal is output, the to-be-corrected feature signal is fitted and corrected with a preset standard solvent fingerprint data, and the signal characteristics corresponding to each solvent are determined; S3, according to the signal characteristics, the preset solvent light absorption characteristic matrix is combined to obtain the instantaneous concentration, the instantaneous concentration is dynamically corrected and trend analysis is performed, and the concentration curve of each solvent is output; S4, obtaining the concentration time series data of the concentration curve, constructing a multi-dimensional feature matrix based on the concentration time series data and combining a preset convergence threshold to determine the critical state of the drying process of each solvent; S5, constructing an operating state vector according to the critical state, and processing the operating state vector based on a feedback compensation adjustment mechanism of rate deviation to obtain control strategy data for dynamically regulating the drying process; S6, the control strategy data is parsed into execution response characteristics, and the execution response characteristics are processed by multi-channel mapping and correction, and a stable process control signal is output; S7, generating a state time sequence diagram according to the process control signal, extracting a quality-efficiency correlation feature vector according to the state time sequence diagram and performing control gain correction, and completing closed-loop feedback control of the final drying process.

[0012] In step S1, the original infrared spectrum data is obtained, a multi-dimensional spectrum matrix is constructed according to the original infrared spectrum data, and interference correction and spectrum unmixing processing are performed to obtain a feature data set of each solvent, comprising: The original infrared spectrum data is converted into a multi-dimensional spectrum matrix, the non-stationary baseline fluctuation region in the multi-dimensional spectrum matrix is identified, and an environmental interference correction model is constructed according to the characteristics of the non-stationary baseline fluctuation region; inputting the multi-dimensional spectral matrix into the environmental interference correction model for differential correction to obtain mixed spectral data after interference elimination; performing wavelet transform decomposition and reconstruction enhancement on the mixed spectral data to obtain enhanced mixed spectral data, and performing blind source separation on the enhanced mixed spectral data by using an independent component analysis algorithm to demix independent spectral components; screening effective spectral components according to characteristic peak positions of the independent spectral components, and mapping the effective spectral components to preliminary characteristic signals of each solvent to obtain a characteristic data set of each solvent after preliminary separation.

[0013] In an implementation manner, the embodiment performs continuous or periodic spectral scanning on a mixed liquid containing multiple solvents by using an infrared spectrometer array to obtain original infrared spectral data covering a preset wave number range, such as a series of original absorption spectral data with a wave number range of 4000-400 cm -1 The original infrared spectral data are in the form of absorbance changing with wave number, and are used to reflect the absorption characteristics of different solvents in the mixed system to infrared radiation.

[0014] Further, the original infrared spectral data are arranged in time sequence, and are indexed by wave number dimension, time dimension and measurement channel dimension to construct a corresponding multi-dimensional spectral matrix, which is used to uniformly describe the time sequence evolution characteristics of spectral information in the drying process. For example, a row represents different sampling points or time, and a column represents a wave number point to form a two-dimensional or higher-dimensional matrix.

[0015] In an implementation manner, to identify the non-stationary baseline fluctuation region in the multi-dimensional spectral matrix, the spectral signals of each spectral channel are segmented and counted in time sequence, the local mean, variance and first-order differential change rate are calculated, and a preset fluctuation threshold is used to judge the fluctuation amplitude and mutation point. By clustering or connectivity analysis on the fluctuation characteristics of the channel signals, the time period and wave band with significant non-stationary baseline change are marked to form an index matrix or mask diagram of the non-stationary baseline fluctuation region, so that the spectral region affected by the environmental interference is accurately identified.

[0016] It should be noted that the fluctuation threshold is a judgment standard for judging whether the local fluctuation of the spectral signal is significant. The local mean and standard deviation of each channel signal of the multi-dimensional spectral matrix are calculated by sliding window segmentation. Then, the fluctuation range is calculated according to the historical environmental noise data collected or the stable spectral data collected in advance, and the mean change amplitude and variance range of each channel are calculated. Finally, the upper limit value of the mean change amplitude and variance range is determined as the fluctuation threshold according to the statistical distribution method.

[0017] Further, the non-stationary baseline fluctuation region of each channel in the multi-dimensional spectral matrix is identified, and the amplitude, frequency and duration are extracted. According to the extracted non-stationary baseline characteristics, a baseline model is established by selecting a polynomial fitting or an asymmetric least squares method, and the local offset and trend change parameters of each channel are calculated. Finally, the baseline model and channel parameters are combined to form an environmental interference correction model.

[0018] In an implementation, for determining the mixed spectral data, the embodiment calculates the deviation of the spectral value of each sampling point of each channel in the multi-dimensional spectral matrix from the corresponding baseline value, and uses the deviation amplitude, deviation frequency and deviation duration characteristics to determine the correction weight. Subsequently, the environmental interference correction model differentially corrects the spectral value of each sampling point according to the correction weight, that is, subtracts the weighted deviation from the spectral value and accumulates the adjustment, and continuously processes all time points and channels. In order to reduce the influence of high-frequency noise, a sliding average or low-pass filtering process can be applied to the deviation sequence during correction, so that extreme fluctuations do not excessively affect the correction result. Finally, the spectral values after differential correction of all channels and time points are integrated to form the de-interfered mixed spectral data.

[0019] In an implementation, for determining the independent spectral components, the embodiment constructs the enhanced mixed spectral data into an observation signal matrix and performs centering processing, calculates the mean value of each channel signal and deducts it from the corresponding signal to eliminate the influence of the direct current component. Then, the centered signal matrix is whitened by eigenvalue decomposition or singular value decomposition to make the channel signals orthogonal to each other and have consistent variance. Based on the whitened signal matrix, the independent component analysis algorithm is used to iteratively solve the demixing matrix, and in each iteration, the demixing result is normalized. When the change of the demixing matrix is less than a preset convergence threshold, the iteration is stopped. Specifically, in each iteration process, the norm difference between the current demixing matrix and the last iteration demixing matrix is calculated. When the norm difference is less than the preset convergence threshold, it is determined that the demixing matrix has stabilized and the iteration is terminated. Finally, by applying the demixing matrix to the enhanced mixed spectral data, a plurality of independent spectral components are demixed.

[0020] It should be noted that the preset convergence threshold is pre-set to a fixed small value according to the numerical stability requirement, and the small value is used to limit the minimum change scale of the demixing matrix update amplitude, for example, set to not more than the order of magnitude of 10 -4 or 10 -5 In another implementation, the preset convergence threshold is normalized according to the amplitude range of the enhanced mixed spectral data, so that the change is at a negligible level relative to the overall amplitude of the signal, thereby ensuring the convergence of the demixing result in the numerical sense.

[0021] In an implementation, for screening out effective spectral components, the embodiment calculates the spectral amplitude distribution and local extreme point position of each independent spectral component, extracts the characteristic peak information in the independent spectral component, including the center wavelength, peak amplitude and peak width parameters of the characteristic peak. Then, the center wavelength of the characteristic peak is compared with the preset solvent characteristic absorption wavelength range, and when the center wavelength of the characteristic peak falls within the characteristic absorption wavelength range of the corresponding solvent, the independent spectral component is marked as a candidate effective spectral component.

[0022] Further, the candidate effective spectral component is subjected to amplitude stability and signal-to-noise feature screening. Specifically, the amplitude variation rate and variance of the candidate effective spectral component at consecutive sampling times are calculated, and when both the amplitude variation rate and variance are within the preset stable interval, it is determined that the independent spectral component meets the stability requirement; at the same time, the signal-to-noise ratio parameter of the independent spectral component in the characteristic peak wavelength range is calculated, and when the signal-to-noise ratio parameter is higher than the preset signal-to-noise judgment threshold, the independent spectral component is retained as an effective spectral component, and the remaining independent spectral components that do not meet the condition are removed.

[0023] It should be noted that the preset solvent characteristic absorption wavelength range is determined based on the standard solvent fingerprint spectrum obtained in advance. Specifically, for each kind of solvent to be monitored, its standard near-infrared spectrum data is collected under independent conditions, the center wavelength corresponding to the absorption peak is extracted after smoothing the standard spectrum, and the center wavelength is taken as the reference to expand the preset wavelength tolerance interval to both sides, thereby forming the characteristic absorption wavelength range of the corresponding solvent. The preset stable interval is determined based on the reference measurement data of the independent spectral component under the condition of no external disturbance. Specifically, the amplitude data of the same independent spectral component in consecutive multiple sampling periods is statistically analyzed, the mean and standard deviation are calculated, and the mean is taken as the center to construct a stable interval according to the standard deviation range of the preset multiple. When the amplitude variation rate and variance of the independent spectral component corresponding to the running process fall within the stable interval, it is determined that it meets the stability condition. The preset signal-to-noise judgment threshold is determined by calibrating the standard solvent spectrum data. Specifically, the signal amplitude is calculated in the solvent characteristic absorption wavelength range, and the noise amplitude is estimated in the adjacent non-absorption wavelength range, and the signal-to-noise ratio is calculated based on the ratio of the two. Through statistical analysis of multiple sets of standard measurement results, the minimum signal-to-noise ratio value that can stably distinguish effective characteristic signals and noise signals is selected as the preset signal-to-noise judgment threshold.

[0024] In an implementation, for determining the feature dataset, the embodiment extracts the corresponding feature peak center wavelength and peak shape parameter according to each effective spectral component, and matches the feature peak center wavelength with the preset solvent feature absorption wavelength range. When the feature peak center wavelength of a certain effective spectral component falls within the feature absorption wavelength range corresponding to a certain solvent, the effective spectral component is established in correspondence with the solvent, and is marked as a candidate feature signal of the solvent.

[0025] Further, after the wavelength range matching is completed, for one or more candidate feature signals corresponding to the same solvent, normalization and amplitude alignment processing is performed thereon. Specifically, according to the relative intensity relationship of the corresponding absorption peaks in the standard solvent fingerprint spectrum, the amplitude of the candidate feature signal is adjusted in proportion, so that it is consistent with the standard solvent fingerprint data in the amplitude scale. Subsequently, the normalized candidate feature signal is combined in time or sampling order to construct the preliminary feature signal of the solvent. By performing the above mapping and combination process on each solvent respectively, a feature dataset containing the preliminary feature signals of multiple solvents is formed, which is output as the feature dataset after preliminary separation of each solvent.

[0026] In step S2, the feature dataset is subjected to modal decomposition and feature frequency band positioning, and a to-be-corrected feature signal is output. The to-be-corrected feature signal is fitted and corrected with the preset standard solvent fingerprint data, and the signal features corresponding to each solvent are determined, including: The feature dataset is decomposed into intrinsic modal components by using an adaptive variational modal decomposition algorithm; The intrinsic modal components are used to construct a Hilbert marginal spectrum and locate a main feature frequency band, and a to-be-corrected feature signal is generated; The residual sequence of the to-be-corrected feature signal and the preset standard solvent fingerprint data is calculated, and the least square method is used to perform fitting regression processing on the residual sequence, and a corrected feature signal is output; The independent signal waveform is reconstructed according to the corrected feature signal, and the signal features corresponding to each solvent are determined.

[0027] In an implementation, for determining the intrinsic modal components, the embodiment takes the feature dataset after preliminary separation of each solvent as an input signal to construct a to-be-decomposed signal sequence. First, the parameters of the variational modal decomposition model are initialized according to the preset number of modes, including the initial value of the center frequency of each mode and the bandwidth constraint parameter. The number of modes can be preset according to the frequency spectrum structure of the feature dataset, or adjusted in the decomposition process through an adaptive updating mechanism. Subsequently, based on the constraint optimization model of the variational modal decomposition, the input signal is represented as a superposition form of several limited bandwidth mode signals, and each mode signal and its corresponding center frequency are updated simultaneously through iteration.

[0028] Further, in each iteration process, the rest of the modal signals are fixed, and only the current modal signal is updated to concentrate around the corresponding center frequency in the frequency domain. At the same time, the center frequency of each modal signal is recalculated according to the updated modal signal to reflect the position of the main frequency of the current modal. In the iteration process, the change amount of each modal signal in the adjacent two iterations is calculated, and when the change amount is less than a preset change convergence threshold, the iteration operation is stopped. Finally, a set of converged modal signals output is the intrinsic modal component, and each intrinsic modal component corresponds to a relatively independent frequency component in the input feature data set.

[0029] It should be noted that the preset convergence threshold is set by statistical analysis of the energy difference of the modal update in the historical or prior spectral decomposition process. Specifically, the maximum energy change value corresponding to the stable stage is selected and scaled to determine the convergence threshold.

[0030] In an implementation manner, after obtaining the intrinsic modal components, the embodiment performs time-frequency characteristic analysis on each intrinsic modal component, and constructs a corresponding Hilbert marginal spectrum based on the intrinsic modal components, which is used to represent the distribution of signal energy in different frequency band ranges. By analyzing the Hilbert marginal spectrum, the system identifies the frequency band interval with high energy proportion and continuous response in time dimension, and determines it as the main characteristic frequency band. Then, the signal components in the main characteristic frequency band are extracted from each intrinsic modal component, and the extracted signal components are superimposed and reconstructed to form a characteristic signal that can reflect the solvent absorption behavior and has effectively suppressed the interference of non-characteristic frequency band. The reconstructed characteristic signal is used as the to-be-corrected characteristic signal.

[0031] In an implementation manner, the embodiment aligns the to-be-corrected characteristic signal with the preset standard solvent fingerprint data point by point to ensure the consistency of the two in the wave number dimension. Then, the system calculates the difference value between the to-be-corrected characteristic signal and the standard solvent fingerprint data at the corresponding wave number position, and arranges the difference values of each wave number position in sequence to form a residual sequence, which is used to represent the offset of the current measurement signal relative to the standard fingerprint.

[0032] Further, after obtaining the residual sequence, the embodiment performs regression fitting processing on the residual sequence to reduce the influence of systematic deviation on the signal characteristics. The regression fitting processing adopts a parameter estimation method based on the error minimization principle, determines the correction parameters for correcting the to-be-corrected characteristic signal by overall fitting the residual sequence. Subsequently, the system compensates and adjusts the to-be-corrected characteristic signal according to the correction parameters, eliminates the systematic errors caused by instrument drift, environmental changes or mixed interference, and thus obtains the corrected characteristic signal.

[0033] It should be noted that the standard solvent fingerprint data is pre-established in the system initialization stage. For each target solvent, the embodiment collects the infrared spectrum data of the single solvent under standard experimental conditions, and performs preprocessing operations such as baseline correction and noise suppression on the collected spectrum data. Subsequently, the representative spectral response characteristics of each solvent in the main absorption band are extracted from the processed spectrum, and normalized according to the wave number dimension to form the corresponding standard solvent fingerprint data.

[0034] In an implementation manner, the embodiment takes the correction characteristic signal as input, and performs segmented reconstruction processing on the correction characteristic signal according to the characteristic frequency band division results corresponding to different solvents. Specifically, the system extracts the signal components related to each solvent from the correction characteristic signal according to the main absorption band position corresponding to the solvent in the standard solvent fingerprint data, and performs superposition and smoothing processing on the extracted signal components to form an independent signal waveform representing the absorption behavior of the single solvent. The independent signal waveform is used to reflect the stable spectral response characteristics of the solvent under the current working condition, and serves as the basis for determining the signal characteristics corresponding to each solvent.

[0035] In step S3, the instantaneous concentration is obtained by inversing the signal characteristics combined with the preset solvent light absorption characteristic matrix, the instantaneous concentration is dynamically corrected and change trend analyzed, and the concentration curve of each solvent is output, including: The instantaneous concentration of each solvent is obtained by inversing the signal characteristics combined with the preset solvent light absorption characteristic matrix; Based on the preset window length and recursion strategy, a real-time dynamic response window is created, and the instantaneous concentration is input into the real-time dynamic response window for smoothing and filtering correction processing to obtain corrected component content data; The increase / decrease direction and rate of each solvent component are determined according to the difference operation result of the component content data, and the component change trend description vector of each solvent is generated based on the increase / decrease direction and the rate; The concentration curve of the dynamic change of each solvent is output by performing time domain mapping and interpolation processing on the component content data according to the component change trend description vector.

[0036] In an implementation, for determining the instantaneous concentration of each solvent, the embodiment combines a preset solvent absorbance characteristic matrix with the signal features to perform concentration inversion processing. Specifically, the solvent absorbance characteristic matrix is used to describe the unit concentration response relationship of different solvents in each characteristic frequency band, which reflects the corresponding mapping between signal amplitude change and solvent concentration. The system matches the response intensity of the signal features in the corresponding characteristic frequency band with the solvent absorbance characteristic matrix, and under the premise of assuming that the absorption contribution of each solvent has a linear superposition relationship, solves the concentration estimation value of each solvent at the current sampling time, thereby obtaining the instantaneous concentration of each solvent.

[0037] It should be noted that the solvent absorbance characteristic matrix is established in advance in the system initialization stage. For each target solvent, the embodiment collects infrared spectrum data under different known concentration levels under standard experimental conditions, and extracts the corresponding signal response intensity in the characteristic frequency band consistent with step S2. Subsequently, based on the collected multiple sets of concentration-response data, the unit concentration response relationship of each solvent in each characteristic frequency band is statistically determined, and the response relationship is organized according to the solvent category and the characteristic frequency band dimension to form the solvent absorbance characteristic matrix.

[0038] In an implementation, for dynamic correction of the instantaneous concentration, the embodiment introduces a real-time dynamic response window to process the instantaneous concentration sequence obtained by inversion. By presetting a window length (for example, data covering the last 20 sampling times), the window is continuously updated in a sliding manner. The instantaneous concentration data in the window is first subjected to a smoothing filter processing, for example, using a sliding average algorithm to suppress random noise. Subsequently, outliers in the window are detected and corrected, for example, the concentration value at a certain time jumps sharply due to instantaneous interference, which is corrected to a value consistent with the trend of the previous and subsequent time. After this processing, a set of more stable and reliable corrected component content data is output.

[0039] For example, assuming that during continuous monitoring, the instantaneous concentration value of methanol fluctuates greatly in a short time, such as 0.019, 0.021, 0.018 mol / L, etc., by taking the average of the three time points through a sliding average window, the smoothed component content concentration data is 0.0193 mol / L.

[0040] It should be noted that the window length can be dynamically adjusted according to the drying process. In an implementation, the system uses a shorter window (such as 10 points) at the beginning of drying (fast concentration change) to improve response speed. When approaching the critical point (very slight concentration change), it automatically switches to a longer window (such as 30 points) to enhance the smoothing effect and accurately capture weak trends. This adaptive mechanism ensures that the corrected data is neither distorted nor effectively filtered out of interference throughout the drying process.

[0041] In one implementation, this embodiment performs differential calculations on the smoothed component content data to quantitatively analyze the real-time changes in the concentration of each solvent, considering the direction and rate of increase or decrease of each solvent component. Specifically, the system calculates the concentration difference of each solvent at two consecutive sampling times. The direction of increase or decrease in solvent concentration is determined by the sign of the difference: a negative difference indicates a decrease in concentration, and a positive difference indicates an increase in concentration. Simultaneously, the concentration difference is divided by the sampling time interval to obtain the instantaneous rate of change of the solvent at the current moment, which quantifies the speed of concentration change.

[0042] For example, the difference operation is positioned as follows:

[0043] in, At the current sampling time, This refers to the previous sampling time. Indicates the first The instantaneous change in the concentration of a solvent. Indicates at time The measured number The concentration of the solvent.

[0044] The instantaneous rate of change is defined as follows:

[0045] in, The time interval between adjacent sampling points. Indicates the first The instantaneous rate of change of a solvent is obtained by absolute value calculation, which is the magnitude of the concentration change rate and is used to quantify the severity of the change; the specific direction of the concentration change, such as increase or decrease, is directly characterized by the sign of the result of the difference operation.

[0046] In one implementation, for generating a component change trend description vector, this embodiment constructs a component change trend description vector based on the change direction and rate information of the current period and a recent historical period. This vector integrates multiple trend features, including but not limited to: the current instantaneous change rate, the average change rate calculated based on short-term history, and the change trend of the rate (e.g., accelerating or decelerating).

[0047] In an implementation, for generating and outputting the concentration curve, the embodiment takes time as the horizontal axis and the corrected component content data as the vertical axis, and uses the information in the component change trend description vector to guide the interpolation process between data points, performs curve fitting, thereby generating a continuous and smooth concentration change curve and outputting. The curve clearly shows the complete evolution trajectory of the concentration of each solvent from the start of drying to the current time.

[0048] It should be noted that the final output concentration curve is the result of dynamic correction and trend optimization, which not only provides instantaneous readings of concentration, but also reveals the dynamic change process of concentration. This high-fidelity process curve is the core input for subsequent intelligent determination of the critical state of drying.

[0049] In step S4, the concentration time series data of the concentration curve is obtained, a multi-dimensional feature matrix is constructed based on the concentration time series data, and a steady state determination is performed in combination with a preset convergence threshold to determine the critical state of the drying process of each solvent, including: The concentration time series data of the concentration curve is obtained, and the instantaneous decay rate and acceleration feature vector of the concentration time series data are calculated; The instantaneous decay rate and the acceleration feature vector are combined to construct a multi-dimensional feature matrix, the multi-dimensional feature matrix is input into a preset critical point judgment model, and the Euclidean distance between the current state point of the solvent and the preset steady state hyperplane is calculated; If the Euclidean distance is less than the preset convergence threshold, the root mean square deviation value of the residual sequence is calculated, and if the root mean square deviation value is continuously located in the preset steady state confidence interval, it is determined that the drying process of the current solvent has reached the critical state.

[0050] In an implementation, for determining the instantaneous decay rate and the acceleration feature vector, the embodiment obtains the concentration curve of each solvent output by step S3 in real time, and samples at fixed time intervals to form a discrete concentration time series data sequence. Based on this sequence, the instantaneous decay rate at each sampling point, i.e. the first derivative of concentration with respect to time, is calculated to represent the speed of concentration change. Further, the system calculates the change rate of the decay rate, i.e. the second derivative of concentration with respect to time, as the acceleration feature vector to represent whether the concentration change trend is accelerating, decelerating or tending to be stable.

[0051] For example, assume that the concentration time series data of methanol is recorded as one point per minute, and the data points are 0.020, 0.019, 0.018 mol / L, and the data of ethanol is 0.010, 0.011, 0.012 mol / L. When the discrete concentration time series data is differentiated to obtain the instantaneous decay rate value and the acceleration feature vector, the rate can be estimated by the difference between the concentration values of adjacent time points. Taking methanol as an example, the concentration difference between the adjacent two points shows that the rate presents a negative trend, indicating that the concentration is decreasing, and further analysis of the rate trend can obtain the acceleration feature, reflecting the speed of concentration decrease.

[0052] In an implementation, to comprehensively describe the state of the system at any time, the embodiment combines and normalizes the concentration values, instantaneous decay rates and acceleration feature vectors of each solvent at the same sampling time to jointly construct a multi-dimensional feature matrix. Each row of the matrix represents a sampling time, and each column represents a state feature variable (such as solvent A concentration, solvent A decay rate, solvent A acceleration, solvent B concentration, …). The matrix maps the dynamic changes of the drying process into a high-dimensional feature space.

[0053] In an implementation, a preset critical point judgment model is used. The model is a rule model based on distance determination. The core is to calculate the Euclidean distance from the current state point to the preset steady-state hyperplane, which is used as a quantitative indicator of state proximity. The model is a binary classification machine learning model (such as a support vector machine or a neural network) trained through supervised learning. The model uses historical drying process data as the training set, and the data samples are labeled as “non-critical state” and “critical state”. The trained model can directly perform pattern recognition on the input multi-dimensional feature vector and output a probability score representing the possibility of belonging to the critical state. In this implementation, the calculation of the Euclidean distance can be part of the model's internal feature extraction or decision-making process.

[0054] It should be noted that the steady-state hyperplane is a mathematical reference plane defined in the multi-dimensional feature space, representing the theoretical state of all solvent evaporation reaching dynamic equilibrium. The hyperplane is directly defined by an analytical method. Assume that the multi-dimensional feature vector is X = [x1, x2, …, xn], where the first m dimensions (for example, the concentration values of each solvent) are state variables, and the last n-m dimensions (for example, the instantaneous decay rates and accelerations of each solvent) are change rate variables. The steady-state hyperplane S can be defined by a set of simple linear equations: for all change rate dimensions i (i > m), x_i = 0.

[0055] It should be noted that the convergence threshold is calculated by calculating the Euclidean distance between the state points corresponding to the dry end of a plurality of experiments and the steady-state hyperplane, and taking the specific high quantile (such as the 95th quantile) of the distance data set as the threshold. This method ensures the objectivity of the threshold, so that the judgment standard is consistent with the high probability statistical interval of the historical process end point.

[0056] In an implementation, the steady-state confidence interval is preset by analyzing the signal residual characteristics of the historical drying processes in the stable stage. The specific preset method is as follows: the system collects a plurality of sets of historical experimental data, and accurately extracts the spectrum residual sequence determined to reach the stable stage (i.e. after the critical state) from the data generated by the signal correction or fitting process in steps S2 or S3, which represents the deviation between the measured value and the model predicted value. For each residual sequence, the root mean square deviation value is calculated as a quantitative indicator of the inherent fluctuation intensity in the stable state of the experiment. Then, statistical analysis is performed on the root mean square deviation values calculated for all historical experiments to determine the central tendency and dispersion. Finally, according to the statistical results, a numerical interval is set, for example, the upper limit of the interval is the average value plus twice the standard deviation, and the lower limit of the interval is the average value minus twice the standard deviation, thereby forming the steady-state confidence interval.

[0057] In step S5, the running state vector is constructed according to the critical state, and the running state vector is processed based on the feedback compensation adjustment mechanism of the rate deviation to obtain control strategy data for dynamically regulating the drying process, including: The running state vector is constructed according to the drying process of the critical state, and a proportional-integral-derivative control algorithm is used to calculate the rate deviation feedback signal of the running state vector, wherein the rate deviation feedback signal represents the difference in concentration decay rate; The rate deviation feedback signal is convoluted using a preset system response lag function to obtain a time compensation deviation sequence; The time compensation deviation sequence is mapped to a preset fuzzy control rule table to generate a dynamic adjustment gain matrix, and the dynamic adjustment gain matrix is superimposed to the running state vector to generate control strategy data for dynamically regulating the drying process.

[0058] In one implementation, for generating the rate deviation feedback signal, the present embodiment constructs a running state vector integrating key process variables (including solvent concentration, its instantaneous decay rate, system temperature and pressure) based on the critical state and real-time monitoring data. Then, the system extracts the actual solvent decay rate in the vector, compares it with the preset expected target rate of the current drying stage to obtain the instantaneous deviation, and applies a proportional-integral-derivative control algorithm to process the deviation, multiplies the current deviation by a proportional gain to obtain a proportional term, multiplies the accumulated historical deviation by an integral gain to obtain an integral term, multiplies the rate of change of the deviation by a derivative gain to obtain a derivative term, and finally sums the three terms to generate a rate deviation feedback signal that comprehensively reflects the deviation size, historical accumulation and change trend.

[0059] Exemplarily, the rate deviation feedback signal is defined as,

[0060] wherein, is the rate deviation feedback signal, is the proportional term, is the integral term, is the derivative term, all of which are obtained by proportional operation, integral operation and derivative operation on the rate deviation value at the current time.

[0061] In one implementation, the present embodiment convolves the sequence of discrete rate deviation feedback signals with a preset system response lag function. This operation essentially simulates the form of the deviation signal after experiencing the inherent lag of the system, and its output is a new sequence, i.e. a time-compensated deviation sequence.

[0062] It should be noted that the preset system response lag function is obtained by performing a standard step response test (e.g. step change in heating power and high-frequency recording of temperature sensor changes) on the drying system, collecting the time series data of the system output. Based on this data, a first-order or second-order inertial model with a pure lag element is fitted using the least squares method, and the mathematical expression of this model is preset as the system response lag function.

[0063] In one implementation, for generating the control strategy data, the present embodiment presets a fuzzy control rule table. This rule table defines the nonlinear mapping relationship between the input variables (such as the size of the time-compensated deviation and its change trend) and the output variables (such as the heating power adjustment coefficient, the inlet valve opening adjustment coefficient, etc.). The system maps the current characteristic value of the time-compensated deviation sequence to this rule table, generates a dynamic adjustment gain matrix through fuzzy reasoning and defuzzification calculation. Each gain coefficient in this matrix corresponds to the adjustment strength and direction of a control execution channel.

[0064] Further, the adjustment amount corresponding to each control variable in the gain matrix is superimposed on the corresponding state component in the operating state vector, respectively (for example, the heating power adjustment gain is superimposed on the current temperature state value), thereby generating a new vector containing target control instructions, that is, control strategy data for dynamically regulating the drying process.

[0065] It should be noted that the control strategy data is not a driving signal directly sent to the actuator, but a set of intermediate instructions containing target set values and adjustment logic. The core of this step is to convert the critical state information of the drying process and the real-time dynamics into preliminary control decisions with predictability and adaptability through rate deviation feedback and system lag compensation, laying a foundation for finally generating stable and accurate process control signals.

[0066] In step S6, the control strategy data is parsed into execution response characteristics, and the execution response characteristics are subjected to multi-channel mapping and correction processing, and stable process control signals are output, including: Obtaining a background thermal noise signal monitored in real time by a temperature sensor; Parsing the control strategy data into execution response characteristics, and constructing a multi-channel instruction mapping table according to the execution response characteristics; Calculating a theoretical thermal distribution value according to the multi-channel instruction mapping table, and if the theoretical thermal distribution value deviates from the background thermal noise signal, correcting the multi-channel instruction mapping table to generate an estimated control amount; Performing residual error evaluation on the estimated control amount to obtain an accuracy correction coefficient, and performing waveform shaping on the estimated control amount according to the accuracy correction coefficient to obtain a voltage drive sequence; Filtering and smoothing the voltage drive sequence to output stable process control signals.

[0067] In an implementation manner, for parsing the control strategy data and obtaining the environmental background signal, the embodiment receives the control strategy data from step S5, which is a set of target set values and adjustment instructions. The system first parses the data set and extracts the core execution response characteristics, which include, for example, total heat power demand, air inlet flow set point, pressure adjustment direction, etc. At the same time, the system reads the background thermal noise signal in real time from the sensor network chip installed at the key part of the device, which represents the current uncontrolled inherent thermal fluctuation or environmental thermal interference of the system.

[0068] In an implementation, for generating the estimated control quantity, the embodiment queries a preset multi-channel instruction mapping table according to the parsed execution response characteristics, and preliminarily allocates the abstract adjustment requirement to specific control channels (such as a heating rod channel, a circulating fan channel, an air inlet valve channel, etc.). The mapping table defines the correspondence from the response characteristic value to the basic driving instruction (such as a duty ratio, an opening degree) of each channel. Subsequently, the system calculates the expected theoretical thermal distribution value (such as the estimated temperature of each region of the cavity) based on the preliminarily allocated driving instruction of each channel according to the device thermodynamic model. The system compares the theoretical value with the actually monitored background thermal noise signal, and if there is a significant deviation, it indicates that the current environmental interference or device state deviates from the model assumption, and the system dynamically corrects the output of the multi-channel instruction mapping table to generate a set of estimated control quantities to pre-compensate the influence of the deviation.

[0069] It should be noted that the preset multi-channel instruction mapping table is preset by combining device characteristic calibration and process data optimization. First, the basic static characteristic relationship of "driving instruction-output response" of each execution mechanism channel is established by independent testing to determine the initial mapping. Subsequently, based on the learning and fitting of the optimized control combination of each channel in the historical successful process data, the initial mapping relationship is cooperatively corrected and fine-tuned to finally form a preset lookup table or function that maps the control strategy vector (such as the total heat demand) to the specific and cooperative driving instruction vector of each channel.

[0070] In an implementation, for residual error evaluation and waveform shaping to generate a driving sequence, the embodiment sends the above-mentioned estimated control quantity to a high-fidelity device simulation model or tests it through a slight actual output to quickly obtain the residual error between the estimated output and the actual demand. Based on the residual error, an accuracy correction coefficient is calculated. Subsequently, the system uses the coefficient to finely adjust the estimated control quantity, and converts it into a voltage driving sequence or pulse width modulation sequence suitable for the response of the actuator and time sequence smoothness through waveform shaping technology (such as gradient constraint or smoothing filtering). This step ensures that the control instruction is not only statically accurate, but also dynamically smooth, avoiding impact on the execution mechanism.

[0071] Exemplarily, when the residual error of the estimated control quantity is evaluated, the actual temperature data is collected in real time by the sensor network chip integrated in the system, and compared with the model estimated data. For example, the average residual error is 0.5°C and the standard deviation is 0.12°C in a 10-second window, and the precision correction coefficient is calculated as 1.005 according to the formula k=1+0.5×0.5 / 50, wherein the first 0.5 is a preset dimensionless empirical coefficient, and the value is usually distributed between 0.1-0.5, if the temperature sensor has high precision (such as ±0.1°C), the lower limit 0.1-0.2 can be taken, if the environmental disturbance is large (such as ±1°C fluctuation), the upper limit 0.3-0.5 can be taken; 0.5 / 50 is the ratio of the average residual error to the reference temperature, and the unit of the reference temperature is °C. Then, the control quantity sequence is scaled by using the coefficient, and the high-frequency spikes are eliminated by three-point moving average filtering, and finally converted into a smooth voltage driving sequence of 0-10V through PWM modulation, so that the control output is stable and the solvent decay rate is effectively maintained constant.

[0072] In an implementation manner, for the output stable process control signal, the generated voltage driving sequence is finally filtered and smoothed by the embodiment, and a low-pass filter is used to filter out the noise that may be introduced by digital calculation or high-frequency correction, and a set of high-stability and high-precision process control signals are output. These signals are directly sent to the drivers of the actuators, so as to realize accurate and stable control of the drying process.

[0073] In step S7, the state time sequence diagram is generated according to the process control signal, the quality-efficiency correlation feature vector is extracted according to the state time sequence diagram, and the control gain is corrected, and the closed-loop feedback control of the final drying process is completed, including: The state time sequence diagram is generated by analyzing the process control signal by using a preset multi-dimensional state observer based on the real-time monitoring system; The moisture content deviation feature is extracted according to the state time sequence diagram and the predicted product quality feedback data, and the quality-efficiency correlation feature vector is generated based on the moisture content deviation feature; The quality-efficiency correlation feature vector is input into a preset closed-loop verification model, the mutual information entropy of the moisture content deviation feature and the instantaneous energy consumption rate is calculated, and the response sensitivity of the current control strategy is quantified according to the mutual information entropy; The control gain sequence is output by correcting the control gain matrix currently used by the system according to the response sensitivity, and the closed-loop feedback control of the final drying process is completed.

[0074] In an implementation, for generating the state time series, the embodiment real-time acquires the process control signals output by step S6, and simultaneously reads the instantaneous energy consumption rate from the monitoring unit. The two sets of signals are input into a preset multi-dimensional state observer. The observer analyzes the internal key state variables (such as the internal temperature gradient of the material, the effective mass transfer coefficient, etc.) that cannot be directly measured according to the state space model of the system, and synchronously aligns all the state variables with the control signals and the energy consumption rate according to time, and integrates to generate a comprehensive state time series. The graph takes time as the horizontal axis, and completely depicts the trajectories of the process variables and their mutual relationships over time.

[0075] It should be noted that the preset multi-dimensional state observer is constructed based on the physical and chemical mechanism model of the drying process. First, the state space equation describing the internal state of the system (such as the internal temperature and concentration gradient of the material) is established according to the heat and mass transfer principles. Subsequently, an observer design theory (such as the Luenberger observer design method) is applied to configure a state estimator and its gain matrix for the model, forming a multi-dimensional state observer.

[0076] In an implementation, for extracting the quality-effect correlation feature vector, the embodiment synchronously receives product quality feedback data from an online quality detection unit (such as a near-infrared spectrum or a humidity sensor), for example, real-time moisture content. The product quality data is compared and analyzed with the process variables at the corresponding time in the state time series, and the deviation between the measured value and the target value of the key quality indicator (such as moisture content) is calculated, that is, the moisture content deviation feature. The system analyzes the correlation pattern between the deviation and multiple process variables (such as temperature, pressure, and concentration change rate) in the state time series, extracts and constructs a quality-effect correlation feature vector that can represent the coupling relationship between “process control effect” and “product quality result”.

[0077] In an implementation, for calculating the mutual information entropy to quantify the response sensitivity, the embodiment inputs the above-mentioned quality-effect correlation feature vector into a preset closed-loop verification model. The core operation of the model is to calculate the mutual information entropy between the moisture content deviation feature and the instantaneous energy consumption rate. Mutual information entropy is an information theory measure used to quantify the statistical dependency strength between two variables. Specifically, the mutual information entropy is calculated according to the following formula,

[0078] wherein, represents the discretized sequence of the moisture content deviation feature, represents the discretized sequence of the instantaneous energy consumption rate, is the joint probability distribution thereof, and is the marginal probability distribution of its edge; through this calculation, the efficiency of the adjustment of the system energy consumption on the final product quality deviation under the current control strategy can be evaluated, and the efficiency value is quantified as the response sensitivity. High sensitivity means that the control action can be efficiently reflected in the quality improvement. Low sensitivity indicates that the control effect is buffered by the system inertia or interference.

[0079] It should be noted that the preset closed-loop verification model is preset through statistical learning based on historical data. The system collects the simultaneously recorded moisture content deviation time series data and instantaneous energy consumption rate time series data in the historical drying process as training samples. The core of the model is to preset a calculation framework, which first calculates the mutual information entropy between the two sets of time series data to quantify the statistical dependence; then, the calculated mutual information entropy value is associated with the actual "response sensitivity" level of the batch drying process determined by artificial calibration or process (for example, through regression analysis). The trained model has the mapping relationship between the mutual information entropy and the specific response sensitivity quantization value solidified inside, thereby becoming a verification model that can automatically and accurately evaluate the effectiveness of the current control strategy according to the real-time input of the two characteristic data.

[0080] Specifically, the closed-loop verification model needs to collect the simultaneously recorded moisture content deviation time series data and instantaneous energy consumption rate time series data in multiple complete drying batches as a training sample set; the response sensitivity true value corresponding to each batch needs to be pre-labeled (for example, divided into high, medium, and low levels or given a specific numerical value) after comprehensive evaluation of the final product quality (such as structural integrity, moisture content compliance) and the stability of the control process; the goal of model training is to learn the statistical mapping relationship between the mutual information entropy and the labeled response sensitivity; when the historical data are diverse and comprehensive enough, the trained model can effectively evaluate the sensitivity of the real-time process.

[0081] In one implementation, for the correction of the control gain and the completion of the closed loop, the embodiment corrects the core control parameter, the control gain matrix (such as the PID gain, the fuzzy rule table output gain) used in steps S5 and S6, according to the calculated response sensitivity. If the response sensitivity is too low, the values of the corresponding elements in the gain matrix are increased according to the predetermined rules to enhance the strength of the control action; if the sensitivity is too high or an oscillation trend appears, the gain is appropriately reduced. The system outputs the correction control gain sequence and feeds it back to the control strategy generation and control signal analysis module of the previous stage to update its internal parameters in real time, thereby forming an adaptive closed-loop control system based on the feedback of the final product quality and the process energy efficiency.

[0082] It should be noted that the predetermined rule is defined based on a comparison result of the response sensitivity and a preset reference sensitivity interval. Specifically, the system presets a reference interval of ideal response sensitivity determined by process requirements. When the calculated actual response sensitivity is lower than the lower limit of the interval, it indicates that the effect of control on quality is too sluggish, and at this time the predetermined rule is to multiply the relevant elements in the control gain matrix by a gain reinforcement coefficient greater than 1 (for example, 1.1 to 1.3). When the actual response sensitivity is higher than the upper limit of the interval, it indicates that the control action may be too aggressive and prone to oscillation, and at this time the predetermined rule is to multiply the relevant elements in the control gain matrix by a gain attenuation coefficient less than 1 (for example, 0.7 to 0.9). If the actual response sensitivity is within the reference interval, the gain matrix remains unchanged. The specific values of the gain reinforcement coefficient and the attenuation coefficient are a set of empirical parameters pre-calibrated by system simulation or historical debugging data.

[0083] Referring to Figure 2 The second embodiment of the present application provides a critical point drying control device based on online near-infrared monitoring, comprising: A data acquisition module is configured to acquire original infrared spectrum data, construct a multi-dimensional spectrum matrix from the original infrared spectrum data, and perform interference correction and spectrum unmixing processing to obtain a feature data set of each solvent; A signal feature determination module is configured to perform modal decomposition and feature band positioning on the feature data set, output a to-be-corrected feature signal, fit and correct the to-be-corrected feature signal with preset standard solvent fingerprint data, and determine the signal features corresponding to each solvent; A concentration curve generation module is configured to obtain instantaneous concentration by inversely calculating the signal features combined with a preset solvent light absorption characteristic matrix, dynamically correct the instantaneous concentration, and analyze the change trend to output a concentration curve of each solvent; A critical state determination module is configured to obtain concentration time series data of the concentration curve, construct a multi-dimensional feature matrix based on the concentration time series data, and determine the critical state of the drying process of each solvent by combining with a preset convergence threshold for steady state determination; A strategy data generation module is configured to construct a running state vector according to the critical state, and process the running state vector based on a feedback compensation adjustment mechanism of rate deviation to obtain control strategy data for dynamically regulating the drying process; A process control signal generation module is configured to parse the control strategy data into execution response features, perform multi-channel mapping and correction processing on the execution response features, and output stable process control signals; A closed-loop feedback control module is configured to generate a state time series graph according to the process control signals, extract a quality-effect correlation feature vector according to the state time series graph, and perform control gain correction to complete closed-loop feedback control of the final drying process.

[0084] It should be noted that the critical point drying control device based on online near-infrared monitoring provided by the embodiments of the present application is used to execute all process steps of the critical point drying control method based on online near-infrared monitoring of the above-mentioned embodiments, and the working principles and beneficial effects of the two are one-to-one correspondence, so it is not repeated here.

[0085] To sum up, the present application introduces online near-infrared spectroscopy technology to monitor the drying process in real time, constructs a multi-dimensional spectral matrix and carries out interference correction and spectral unmixing, and on this basis, combines adaptive mode decomposition and solvent fingerprint fitting to realize accurate extraction of solvent signal characteristics; further, through light absorption characteristic matrix inversion and dynamic trend analysis, a high-fidelity concentration curve is generated, and based on a multi-dimensional feature matrix and a convergence threshold judgment mechanism, intelligent identification of the drying critical state is realized; at the same time, combined with the PID feedback compensation of the rate deviation and the system lag compensation, a forward-looking control strategy is generated, and through multi-channel instruction mapping and environmental thermal noise correction technology, stable process control signals are output; on this basis, a multi-dimensional state observer is used to generate a process state panorama, and the response sensitivity is quantified through mutual information entropy, and then the adaptive correction of the control gain is realized. Through the above multi-step, multi-model collaborative perception, decision and execution closed loop, the present application effectively overcomes the precision deficiency and response delay problems caused by the dependence of the prior art on static models and lag control, and can realize overall improvement of critical point drying process control precision, stability and energy efficiency under complex solvent systems and dynamic process conditions.

[0086] The embodiments of the present application also provide an electronic device. The electronic device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, for example, a critical point drying control program based on online near-infrared monitoring. The processor implements the steps in each of the above-mentioned critical point drying control method embodiments based on online near-infrared monitoring when executing the computer program, for example Figure 1 The step S1 shown. Alternatively, the processor implements the functions of each module / unit in each of the above-mentioned device embodiments when executing the computer program, for example, a data acquisition module.

[0087] Illustratively, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the electronic device.

[0088] The electronic device can be a computing device such as a desktop computer, a notebook computer, a palm computer, a smart tablet, etc. The electronic device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the above components are only examples of the electronic device and do not constitute a limitation on the electronic device, and the electronic device can include more or fewer components than the above, or combine certain components, or different components, for example, the electronic device can also include an input / output device, a network access device, a bus, etc.

[0089] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor is the control center of the electronic device, and connects various parts of the electronic device through various interfaces and lines.

[0090] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function (such as a sound playing function, an image playing function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory device.

[0091] The modules / units integrated in the electronic device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. that can carry the computer program code. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0092] It should be noted that the above-described device embodiments are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. In addition, the connection relationship between the modules in the device embodiment provided by the present application indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0093] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above-described specific embodiments are only for the specific embodiments of the present application and do not limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A critical point drying control method based on online near-infrared monitoring, characterized in that, include: The raw infrared spectral data is acquired, and a multidimensional spectral matrix is ​​constructed based on the raw infrared spectral data. Interference correction and spectral demixing are then performed to obtain the characteristic datasets of each solvent. The feature dataset is subjected to mode decomposition and feature frequency band localization to output the feature signal to be corrected. The feature signal to be corrected is then fitted and corrected with the preset standard solvent fingerprint data to determine the signal features corresponding to each solvent. Based on the signal characteristics, the instantaneous concentration is obtained by inverting the preset solvent absorbance characteristic matrix. The instantaneous concentration is then dynamically corrected and its changing trend is analyzed, and the concentration curves of each solvent are output. Acquire the concentration time-series data of the concentration curve, construct a multidimensional feature matrix based on the concentration time-series data, and combine it with a preset convergence threshold to determine the steady state and the critical state of each solvent drying process. Based on the critical state, an operating state vector is constructed, and the operating state vector is processed based on the feedback compensation adjustment mechanism of the rate deviation to obtain control strategy data for dynamically regulating the drying process. The control strategy data is parsed into execution response features, and the execution response features are subjected to multi-channel mapping and correction processing to output a stable process control signal. A state timing diagram is generated based on the process control signal. A quality-efficiency correlation feature vector is extracted based on the state timing diagram, and the control gain is corrected to complete the closed-loop feedback control of the final drying process.

2. The critical point drying control method based on online near-infrared monitoring according to claim 1, characterized in that, The process involves acquiring raw infrared spectral data, constructing a multidimensional spectral matrix based on the raw infrared spectral data, and performing interference correction and spectral demixing to obtain a feature dataset for each solvent, including: The raw infrared spectral data is converted into a multidimensional spectral matrix, and non-stationary baseline fluctuation regions in the multidimensional spectral matrix are identified. An environmental interference correction model is constructed based on the characteristics of the non-stationary baseline fluctuation regions. The multidimensional spectral matrix is ​​input into the environmental interference correction model for differential correction to obtain interference-free mixed spectral data; The mixed spectral data is decomposed and reconstructed using wavelet transform to obtain enhanced mixed spectral data. The enhanced mixed spectral data is then subjected to blind source separation using an independent component analysis algorithm to demix the independent spectral components. Effective spectral components are selected based on the characteristic peak positions of the independent spectral components, and the effective spectral components are mapped to the preliminary characteristic signals of each solvent to obtain the feature dataset after preliminary separation of each solvent.

3. The critical point drying control method based on online near-infrared monitoring according to claim 2, characterized in that, The process of performing mode decomposition and feature frequency band localization on the feature dataset, outputting the feature signal to be corrected, and fitting and correcting the feature signal to be corrected with preset standard solvent fingerprint data to determine the signal features corresponding to each solvent includes: The feature dataset is decomposed into intrinsic mode components using an adaptive variational mode decomposition algorithm. The intrinsic mode components are used to construct the Hilbert marginal spectrum and locate the main characteristic frequency band, thereby generating the characteristic signal to be corrected; Calculate the residual sequence between the feature signal to be corrected and the preset standard solvent fingerprint data, and use the least squares method to perform fitting regression processing on the residual sequence to output the corrected feature signal; The independent signal waveforms are reconstructed based on the correction feature signals to determine the signal characteristics corresponding to each solvent.

4. The critical point drying control method based on online near-infrared monitoring according to claim 1, characterized in that, The instantaneous concentration is obtained by inverting the signal characteristics and combining them with a preset solvent absorbance characteristic matrix. The instantaneous concentration is then dynamically corrected and its changing trend analyzed, and the concentration curves of each solvent are output, including: Based on the signal characteristics, the instantaneous concentration of each solvent is obtained by inversion using a preset solvent absorption characteristic matrix; Based on the preset window length and recursion strategy, a real-time dynamic response window is created, and the instantaneous concentration is input into the real-time dynamic response window for smoothing and filtering correction to obtain the corrected component content data. Based on the differential calculation results of the component content data, the direction and rate of increase or decrease of each solvent component are determined, and based on the direction and rate of increase or decrease, a component change trend description vector of each solvent is generated; Based on the component change trend description vector, the component content data is time-domain mapped and interpolated to output the concentration curves of each solvent's dynamic changes.

5. The critical point drying control method based on online near-infrared monitoring according to claim 3, characterized in that, The process of acquiring the concentration time-series data of the concentration curve, constructing a multi-dimensional feature matrix based on the concentration time-series data, and performing steady-state determination in conjunction with a preset convergence threshold to determine the critical state of each solvent drying process includes: Acquire the concentration time-series data of the concentration curve, and calculate the instantaneous decay rate and acceleration feature vector of the concentration time-series data; The instantaneous decay rate and the acceleration feature vector are combined to construct a multidimensional feature matrix. The multidimensional feature matrix is ​​input into a preset critical point judgment model, and the Euclidean distance between the current state point of the solvent and the preset steady-state hyperplane is calculated. If the Euclidean distance is less than a preset convergence threshold, the root mean square deviation of the residual sequence is calculated. If the root mean square deviation remains within a preset steady-state confidence interval, the current solvent drying process is determined to have reached a critical state.

6. The critical point drying control method based on online near-infrared monitoring according to claim 5, characterized in that, The step of constructing an operating state vector based on the critical state and processing the operating state vector based on a feedback compensation adjustment mechanism for rate deviation to obtain control strategy data for dynamically regulating the drying process includes: An operating state vector is constructed based on the drying process at the critical state, and a rate deviation feedback signal of the operating state vector is calculated using a proportional-integral-derivative control algorithm. The rate deviation feedback signal characterizes the difference in concentration decay rate. The rate deviation feedback signal is convolved using a preset system response hysteresis function to obtain a time-compensated deviation sequence. The time-compensated deviation sequence is mapped to a preset fuzzy control rule table to generate a dynamic adjustment gain matrix. The dynamic adjustment gain matrix is ​​then superimposed on the operating state vector to generate control strategy data for dynamically regulating the drying process.

7. The critical point drying control method based on online near-infrared monitoring according to claim 1, characterized in that, The step of parsing the control strategy data into execution response features, performing multi-channel mapping and correction processing on the execution response features, and outputting a stable process control signal includes: Acquire the background thermal noise signal monitored in real time by the temperature sensor; The control strategy data is parsed into execution response features, and a multi-channel instruction mapping table is constructed based on the execution response features. The theoretical thermal distribution value is calculated based on the multi-channel command mapping table. If the theoretical thermal distribution value deviates from the background thermal noise signal, the multi-channel command mapping table is corrected to generate the estimated control quantity. The estimated control quantity is subjected to residual evaluation to obtain an accuracy correction coefficient, and the estimated control quantity is subjected to waveform shaping based on the accuracy correction coefficient to obtain a voltage drive sequence; The voltage drive sequence is filtered and smoothed to output a stable process control signal.

8. The critical point drying control method based on online near-infrared monitoring according to claim 7, characterized in that, The step of generating a state timing diagram based on the process control signal, extracting quality-efficiency correlation feature vectors based on the state timing diagram and performing control gain correction to complete the closed-loop feedback control of the final drying process includes: Based on the instantaneous energy consumption rate obtained by the real-time monitoring system, the state timing diagram is obtained by analyzing the process control signal using a preset multi-dimensional state observer. Based on the state time series diagram and the product quality feedback data of the predicted quantity, extract the moisture content deviation feature, and generate a quality-efficiency correlation feature vector based on the moisture content deviation feature; The quality-efficiency correlation feature vector is input into a preset closed-loop verification model to calculate the mutual information entropy between the moisture content deviation feature and the instantaneous energy consumption rate, and the response sensitivity of the current control strategy is quantified based on the mutual information entropy. Based on the control gain matrix currently used by the response sensitivity correction system, a corrected control gain sequence is output to complete the closed-loop feedback control of the final drying process.

9. A critical point drying control device based on online near-infrared monitoring, characterized in that, include: The data acquisition module is used to acquire raw infrared spectral data, construct a multidimensional spectral matrix based on the raw infrared spectral data, and perform interference correction and spectral demixing to obtain the characteristic dataset of each solvent. The signal feature determination module is used to perform mode decomposition and feature frequency band localization on the feature dataset, output the feature signal to be corrected, fit and correct the feature signal to be corrected with the preset standard solvent fingerprint data, and determine the signal features corresponding to each solvent. The concentration curve generation module is used to obtain the instantaneous concentration based on the signal characteristics and a preset solvent absorbance characteristic matrix, perform dynamic correction and trend analysis on the instantaneous concentration, and output the concentration curves of each solvent. The critical state determination module is used to acquire the concentration time series data of the concentration curve, construct a multi-dimensional feature matrix based on the concentration time series data, and perform steady state determination in combination with a preset convergence threshold to determine the critical state of each solvent drying process. The strategy data generation module is used to construct an operating state vector based on the critical state, and process the operating state vector based on the feedback compensation adjustment mechanism of the rate deviation to obtain control strategy data for dynamically regulating the drying process. The process control signal generation module is used to parse the control strategy data into execution response features, perform multi-channel mapping and correction processing on the execution response features, and output a stable process control signal. The closed-loop feedback control module is used to generate a state timing diagram based on the process control signal, extract the quality-efficiency correlation feature vector based on the state timing diagram and perform control gain correction to complete the closed-loop feedback control of the final drying process.

Images