Electrical online detection method and system for purity of glyceryl monostearate finished product

Abstract

The invention belongs to the technical field of purity detection, and discloses an electrical online detection method and system for the purity of a glyceryl monostearate finished product. The method comprises the following steps: acquiring first feature information of a glyceryl monostearate finished product through a first acquisition unit, preprocessing the first feature information through a second acquisition unit to obtain second feature information, and calculating the purity of the finished product through a trained glyceryl monostearate finished product purity detection model based on the second feature information. According to the scheme, the broadband dielectric spectrum sensor and the density sensor are adopted, and the glycerin monostearate finished product purity detection model is matched, so that the scheme is high in detection speed, high in automation degree and low in cost, online real-time detection of the purity of the glycerin monostearate finished product can be realized, and the detection efficiency is remarkably improved.

Description

Technical Field

[0001] This application belongs to the field of purity detection technology, specifically relating to an online electrical detection method and system for the purity of glyceryl monostearate finished products. Background Technology

[0002] Glyceryl monostearate is a nonionic surfactant whose molecular structure contains both hydrophilic and lipophilic groups, giving it multiple functions such as emulsification, stabilization, and dispersion. It is an indispensable functional additive in industrial production. However, in industrial production, glyceryl monostearate often contains impurities such as stearic acid, glycerol, distearate, and tristearate due to limitations in synthesis processes or purification techniques. Purity directly determines the core functions of glyceryl monostearate, such as emulsification and stabilization; impurities weaken its effectiveness and lead to functional failure. Therefore, it is crucial to strictly control the purity range of the finished product during production to minimize substandard purity. Existing methods for detecting the purity of glyceryl monostearate mainly include liquid chromatography (LC) and gas chromatography (GC). The core principle of LC is to quantitatively calculate purity by utilizing the difference in retention time between glyceryl monostearate and impurities in the chromatographic column, using peak area. While existing testing technologies offer high accuracy, they are costly, complex to operate, and unsuitable for batch testing, requiring sampling and laboratory testing, which fails to meet the requirements for batch quality testing of glyceryl monostearate. Chinese Patent Publication No. CN114139643A discloses a machine vision-based method and system for quality inspection of glyceryl monostearate. This technical solution constructs a quality inspection and evaluation model using machine learning, then analyzes the acquired process parameters and color characteristic information of the glyceryl monostearate products to classify them. The model then evaluates the product quality, solving the technical problems of low detection efficiency and environmental pollution caused by chemical experiments in existing technologies that rely on multiple experimental methods to detect the purity of glyceryl monostearate.

[0003] However, existing detection technologies still suffer from long detection cycles and cannot achieve online real-time detection of glyceryl monostearate. Therefore, developing a rapid online detection method has become an urgent technical problem to be solved. Summary of the Invention

[0004] To address the problems existing in the prior art, the primary objective of this application is to provide an online electrical detection method and system for the purity of glyceryl monostearate. This application acquires first characteristic information of the glyceryl monostearate product through a first acquisition unit, preprocesses the first characteristic information to obtain second characteristic information through a second acquisition unit, and then calculates the purity value of the glyceryl monostearate product based on the second characteristic information using a purity detection model. This solution solves the problems of high cost, long processing time, and inability to perform online detection in existing purity detection methods.

[0005] To achieve the above objectives, this application provides the following technical solution:

[0006] In a first aspect, this application provides an online electrical detection method for the purity of glyceryl monostearate finished product, comprising:

[0007] Obtain the first characteristic information of the finished product of glyceryl monostearate, wherein the first characteristic information is the dielectric relaxation front position f0, dielectric relaxation front intensity Δε, and density ρ of the finished product of glyceryl monostearate;

[0008] The first feature information is preprocessed to obtain the second feature information. The preprocessing step involves environmental interference compensation processing of the first feature information. The environmental interference compensation processing includes compensating for the dielectric relaxation front position f0 using a first compensation model and compensating for the dielectric relaxation front intensity Δε using a second compensation model. The first compensation model is constructed based on the coupling relationship between the dielectric relaxation front position f0 and environmental parameters established by the VFT equation and the activation volume model. The second compensation model is constructed based on the coupling relationship between the dielectric relaxation front intensity Δε and environmental parameters established by the Clausius-Mosotti equation. The environmental parameters include the temperature T and pressure P of glyceryl monostearate.

[0009] Based on the second feature information, the purity detection model for glyceryl monostearate finished product calculates the purity of the glyceryl monostearate finished product; the purity detection model for glyceryl monostearate finished product is trained based on historical second feature information of multiple calibrated purity values.

[0010] Furthermore, prior to obtaining the first characteristic information of the finished glyceryl monostearate product, the method further includes:

[0011] Obtain a standard sample of the finished product glyceryl monostearate, wherein the standard sample is a finished product sample of glyceryl monostearate with known accurate purity;

[0012] Based on the standard sample, calibrate the reference values ​​for dielectric relaxation front position f0, dielectric relaxation front intensity Δε, and density ρ.

[0013] Furthermore, the construction steps of the first compensation model include:

[0014] Based on the VFT equation, an equation relating the dielectric relaxation front position f0 to the ambient temperature T is established.

[0015] Based on the activated volume model, and on the basis of the relationship between the dielectric relaxation front position f0 and the ambient temperature T, a modified equation for the dielectric relaxation front position f0 based on the pressure P is constructed.

[0016] A standard value for the dielectric relaxation front position f0 under standard environmental parameters is set, and a first compensation model is constructed based on the standard value and the theoretical value of the correction equation under the current environmental parameters.

[0017] Furthermore, the construction steps of the second compensation model include:

[0018] Based on the Clausius-Mossotti equation, a relationship equation between dielectric relaxation front intensity Δε and temperature T and pressure P is established.

[0019] The coupling coefficients a~e in the relational equation are calibrated by standard sample experiments, and the second compensation model is obtained based on the coupling coefficients a~e.

[0020] Furthermore, the specific steps for calibrating the coupling coefficients a~e in the relational equation through standard sample experiments include:

[0021] Standard samples of glyceryl monostearate products with different purities were prepared;

[0022] Standard environmental parameters are set, including standard temperature T0 and standard pressure P0, and environmental parameter combinations are set based on the standard environmental parameters; the environmental parameter combinations consist of environmental parameters set in a gradient.

[0023] The standard sample was tested in a combination of environmental parameters, and the measured dielectric relaxation strength Δε, temperature T, and pressure P were recorded for each combination.

[0024] The coupling coefficients a~e are obtained by fitting calculations based on standard temperature T0, standard pressure P0, measured dielectric relaxation strength Δε, temperature T, and pressure P.

[0025] Furthermore, before obtaining the second feature information, the method further includes denoising and normalizing the first feature information.

[0026] Furthermore, the training steps for the purity detection model of the glyceryl monostearate finished product are as follows:

[0027] Construct a training dataset and divide the training dataset into a training set, a validation set, and a test set;

[0028] Train the SVM model using the training set, adjust the parameters using the validation set, and continue training until the purity prediction error and accuracy on the test set are satisfactory, at which point the training ends.

[0029] Furthermore, the specific steps for constructing the training dataset are as follows:

[0030] A sample library of finished glyceryl monostearate products was constructed, which was obtained by preparing finished glyceryl monostearate samples or acquiring historical sample data.

[0031] Extract the second characteristic information and purity information of the finished product of glyceryl monostearate from the sample library;

[0032] Principal component analysis is used to compress the three principal components in the second feature information into two principal components, thus obtaining the preprocessed second feature information.

[0033] The preprocessed second feature information and the purity information are combined to obtain the training dataset.

[0034] Secondly, this application provides an online electrical detection system for the purity of glyceryl monostearate finished products, comprising:

[0035] The first acquisition unit is used to acquire the first characteristic information of the finished product of glyceryl monostearate;

[0036] The second acquisition unit is used to preprocess the first feature information and acquire environmental parameters to obtain the second feature information;

[0037] The central processing unit is used to run the purity detection model of glyceryl monostearate product and calculate the purity value of glyceryl monostearate product.

[0038] Further, the first acquisition unit includes:

[0039] A broadband dielectric spectrum sensor is used to collect the dielectric relaxation front position f0 and dielectric relaxation front intensity Δε of glyceryl monostearate finished product.

[0040] Density sensor, used to collect the density ρ of the finished product of glyceryl monostearate;

[0041] The second acquisition unit includes:

[0042] A preprocessing unit is used to preprocess the first feature information to obtain the second feature information;

[0043] An environmental parameter sensor is used to acquire the environmental parameters of the finished product, glyceryl monostearate.

[0044] The beneficial effects of this application are:

[0045] This application first utilizes the dielectric relaxation phenomenon of glyceryl monostearate molecules. Because the polar groups in glyceryl monostearate molecules are hindered from reorienting in an alternating electric field by hydrophobic chains, their dielectric relaxation time is long and their frequency is low, facilitating online real-time detection. Therefore, this application utilizes this characteristic to first detect the position and intensity characteristics of the dielectric relaxation front of glyceryl monostearate using a broadband dielectric spectrum sensor. Then, an SVM model is used to calculate the purity of the finished glyceryl monostearate product based on the changes in the position and intensity of the dielectric relaxation front. Because this application further preprocesses the position and intensity characteristics of the dielectric relaxation front, the differences in these characteristics are more easily captured by the SVM model, leading to model convergence. The converged model then provides a more accurate purity prediction for the finished glyceryl monostearate product, achieving online purity detection of the finished product. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In addition, the schematic diagrams of the device structure provided in the embodiments of this application only show its functional structural relationship and do not represent the actual structural diagram of the device.

[0047] Figure 1 A flowchart of an online electrical detection method for the purity of glyceryl monostearate product provided in this application.

[0048] Figure 2 A flowchart illustrating the steps involved in constructing the first compensation model provided in this application.

[0049] Figure 3 A flowchart illustrating the steps involved in constructing the second compensation model provided in this application.

[0050] Figure 4 A flowchart outlining the specific steps for calibrating the coupling coefficients provided in this application.

[0051] Figure 5 A flowchart illustrating the steps of denoising the first feature information provided in this application.

[0052] Figure 6 The flowchart of the Kalman filter iterative steps provided in this application is shown.

[0053] Figure 7 The flowchart of the steps for normalizing the denoised first feature information provided in this application is shown.

[0054] Figure 8 A flowchart illustrating the training steps for the purity detection model of glyceryl monostearate provided in this application.

[0055] Figure 9 A flowchart illustrating the steps involved in constructing the training dataset provided in this application.

[0056] Figure 10 A schematic diagram of an online electrical detection system for the purity of glyceryl monostearate product provided in this application. Detailed Implementation

[0057] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the specific embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0058] Many specific details are set forth in the following description in order to provide a full understanding of this application. However, this application may also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0059] Glyceryl monostearate is a nonionic surfactant whose molecular structure contains both hydrophilic and lipophilic groups, possessing multiple functions such as emulsification, stabilization, and dispersion, making it an indispensable functional additive in industrial production. The purity of glyceryl monostearate directly affects its application in food, pharmaceutical, and other fields; therefore, purity detection of glyceryl monostearate is crucial. Existing technologies mainly utilize chemical analysis, liquid chromatography, and gas chromatography to detect the purity of glyceryl monostearate. However, these methods are costly, time-consuming, and cannot be performed online. Therefore, this application utilizes the differences in physical characteristics caused by variations in the purity of glyceryl monostearate. An online detection system for the purity of glyceryl monostearate has been designed, solving the problems of existing detection technologies.

[0060] This application presents an online detection system for the purity of glyceryl monostearate. The basic principle is as follows: Glyceryl monostearate molecules contain both hydrophobic long chains and hydrophilic hydroxyl groups. The hydrophilic hydroxyl groups are polar groups and will change direction in an alternating electric field, following the direction of the electric field. However, the hydrophobic long chains are non-polar and hinder the movement of the hydroxyl groups in the alternating electric field. This causes the movement of the glyceryl monostearate molecules to lag behind the frequency of the electric field changes, resulting in dielectric relaxation. The purity of glyceryl monostearate affects the position and intensity of the dielectric relaxation front. The position f0 of the dielectric relaxation front is essentially the orientation relaxation time τ of the hydrophobic long chains and hydroxyl groups in the glyceryl monostearate molecule under an alternating electric field, because the relaxation time and the position of the relaxation front satisfy the following relationship: High-purity glyceryl monostearate has a relatively regular molecular arrangement formed by van der Waals forces and hydrogen bonds. The movement of hydroxyl groups is constrained by the glyceryl monostearate molecules. In an alternating electric field, the groups can only turn within a fixed constrained space, resulting in a stable orientation relaxation rate and a fixed relaxation time τ. Therefore, its relaxation front position f0 is fixed. When the purity of glyceryl monostearate decreases, impurity molecules mixed in the system disrupt the regularity of the molecular arrangement, leading to a larger space for group movement, inconsistent constraints, and disordered relaxation rates. Consequently, the relaxation time τ is no longer fixed, and therefore, its relaxation front position f0 is also no longer fixed. Therefore, by monitoring the drift direction and amplitude of the relaxation front position f0 of glyceryl monostearate, the purity of glyceryl monostearate can be inversely deduced. The dielectric relaxation front intensity is essentially the change in dielectric constant Δε before and after relaxation. In high-purity glycerol monostearate molecules, the ester and hydroxyl groups have high polarization in low-frequency electric fields, resulting in a large static dielectric constant. In high-frequency electric fields, the polarization of the ester and hydroxyl groups is low, leading to a small limiting dielectric constant. Therefore, high-purity glycerol monostearate has a large dielectric relaxation front intensity. When the purity of glycerol monostearate decreases, impurities reduce the number of effectively polarized molecules, resulting in a decrease in dielectric relaxation front intensity. Therefore, the purity of glycerol monostearate can be estimated by detecting the change in dielectric relaxation front intensity. Furthermore, glycerol monostearate of different purities exhibits slight differences in density ρ. Using a high-precision density meter to simultaneously detect the density values ​​of glycerol monostearate can help determine its purity. In this technical solution, the dielectric relaxation front position f0, dielectric relaxation front intensity Δε, and density ρ are processed, and then the purity test results of glyceryl monostearate are calculated and output through a model.

[0061] The present application will be further described below with reference to the accompanying drawings and embodiments.

[0062] Example 1

[0063] like Figure 1 As shown, this application provides an online electrical detection method for the purity of glyceryl monostearate, comprising:

[0064] S1. Obtain the first characteristic information of the finished product of glyceryl monostearate, wherein the first characteristic information is the dielectric relaxation front position f0, dielectric relaxation front intensity Δε and density ρ of the finished product of glyceryl monostearate;

[0065] S2. Preprocess the first feature information to obtain the second feature information; the preprocessing step is to perform environmental interference compensation processing on the first feature information; the environmental interference compensation processing includes compensating for the dielectric relaxation front position f0 using a first compensation model and compensating for the dielectric relaxation front intensity Δε using a second compensation model; the first compensation model is constructed based on establishing the coupling relationship between the dielectric relaxation front position f0 and environmental parameters using the VFT equation and the activation volume model; the second compensation model is constructed based on establishing the coupling relationship between the dielectric relaxation front intensity Δε and environmental parameters using the Clausius-Mosotti equation; the environmental parameters include the temperature T and pressure P of glyceryl monostearate;

[0066] It should be noted that during the acquisition of the first feature information, since the temperature and pressure of the finished glyceryl monostearate product affect the accuracy of its dielectric relaxation front position and intensity, environmental interference compensation processing is required for the acquired first feature information. Specifically, a first compensation model is used to compensate for the dielectric relaxation front position f0, reducing the influence of temperature and pressure on f0; a second compensation model is used to compensate for the dielectric relaxation front intensity Δε, reducing the influence of temperature and pressure on Δε. Figure 2 As shown, the specific steps for constructing the first compensation model include:

[0067] S21. Establish the relationship equation between the dielectric relaxation front position f0 and the ambient temperature T based on the VFT equation;

[0068] Specifically, before constructing the aforementioned relational equation, the relational equation between relaxation time τ and temperature T is first constructed. :

[0069]

[0070] in, is the relaxation time at infinite temperature, which is a constant; D is a characteristic parameter of the VFT equation, reflecting the strength of intermolecular interactions; This is the ideal glass transition temperature for glyceryl monostearate.

[0071] Furthermore, since f0 is related to τ:

[0072]

[0073] Therefore, the equation relating f0 to temperature T can be derived. :

[0074]

[0075] in, This represents the dielectric relaxation front position at infinitely high temperatures.

[0076] S22. Based on the activated volume model, and on the basis of the relationship between the dielectric relaxation front position f0 and the ambient temperature T, a modified equation for the pressure P on the dielectric relaxation front position f0 is further constructed.

[0077] Specifically, since pressure P affects the relaxation activation energy, a correction is needed, and the correction equation is as follows: for:

[0078]

[0079] in, The relaxation activation volume is determined by the space required for the rotation of the ester group in the glyceryl monostearate molecule and can be determined experimentally; k is the Boltzmann constant.

[0080] S23. Set the standard value of the dielectric relaxation front position f0 under standard environmental parameters, and construct a first compensation model based on the standard value and the theoretical value of the correction equation under the current environmental parameters.

[0081] Specifically, the online compensation formula of the first compensation model for:

[0082]

[0083] in, This represents the theoretical f0 value under the current environmental parameters. This is the f0 value under standard conditions.

[0084] like Figure 3 As shown, the specific steps for constructing the second compensation model are as follows:

[0085] S24. Based on the Clausius-Mosotti equation, establish the relationship equation between dielectric relaxation front intensity Δε and temperature T and pressure P;

[0086] According to the Clausius-Mosotti equation, the dielectric relaxation front intensity Δε is directly related to the molecular dipole moment and the molecular number density. As temperature increases, the molecular number density decreases, leading to a decrease in the dielectric relaxation front intensity Δε; conversely, as pressure increases, the molecular number density increases, leading to an increase in the dielectric relaxation front intensity Δε. Based on this relationship, an equation relating the dielectric relaxation front intensity Δε to temperature T and pressure P can be constructed. :

[0087]

[0088] Where a is the linear influence coefficient of temperature, b is the nonlinear influence coefficient of temperature, c is the linear influence coefficient of pressure, d is the nonlinear influence coefficient of pressure, and e is the cross-influence coefficient of temperature and pressure. Standard temperature This is the standard pressure.

[0089] S25. The coupling coefficients a~e in the relational equation are calibrated through standard sample experiments, and the second compensation model is obtained based on the coupling coefficients a~e.

[0090] Specifically, such as Figure 4 As shown, the steps for obtaining the coupling coefficients a~e include:

[0091] S251. Prepare standard samples of glyceryl monostearate products of different purities;

[0092] Specifically, standard samples of glyceryl monostearate with five purity gradients can be prepared, with multiple parallel samples prepared for each purity gradient.

[0093] S252. Set standard environmental parameters T0 and P0, and set a combination of environmental parameters based on the standard environmental parameters; the combination of environmental parameters consists of environmental parameters set according to gradients;

[0094] Specifically, a temperature parameter combination can be formed by setting 7 gradient temperature parameters based on the standard temperature T0, and a pressure parameter combination can be formed by setting 5 gradient pressure parameters based on the standard pressure P0.

[0095] S253. Place the standard sample in a combination of environmental parameters for testing, and record the measured dielectric relaxation intensity Δε and environmental parameters T and P in each combination;

[0096] Specifically, temperature and pressure parameters were combined to form 35 sets of environmental parameters. Standard samples of different purities were tested in these 35 sets of environmental parameters to obtain the dielectric relaxation intensity Δε.

[0097] S254. Based on the standard environmental parameters T0, P0, the measured dielectric relaxation intensity Δε, and the environmental parameters T and P, a fitting operation is performed to obtain the coupling coefficients a~e.

[0098] Specifically, environmental parameter deviations ΔT and ΔP are obtained based on standard and actual environmental parameters. These deviations are then standardized to avoid fitting bias caused by dimensional differences. Based on these deviations, the following equation can be derived:

[0099]

[0100] in, Δε is the relative change predicted by the model; k0 is a constant term; because the stability of Δε signals varies at different temperatures and pressures, the weighted least squares method is used to fit the coefficients. The specific steps include first defining a function that minimizes the weighted sum of squared residuals, transforming the function into a system of linear equations, and then solving for the coupling coefficients a~e through matrix operations.

[0101] Furthermore, during the production of glyceryl monostearate, the equipment operation generates 50Hz electromagnetic interference. This interference signal is directly superimposed on the acquired first characteristic information. Therefore, it is necessary to remove the noise generated by electromagnetic interference to avoid affecting the detection results. Simultaneously, considering that glyceryl monostearate may experience localized turbulence during pipeline transportation or that the sensor itself may generate minor vibrations, leading to random dynamic fluctuations in the signal, a Kalman filter prediction-update method is needed to smooth these fluctuations, ensuring that the signal change trend aligns with the actual purity changes of the material. Figure 5 As shown, the specific noise reduction steps include:

[0102] S26. Filter out electromagnetic noise in the first feature information to obtain a preprocessed signal;

[0103] It should be noted that in this embodiment, electromagnetic interference signals can be removed by using a band-stop filter. By adjusting the parameters of the band-stop filter, electromagnetic waves of different frequencies can be removed. The electromagnetic waves of different frequencies mainly include electromagnetic noise at 50Hz and 60Hz.

[0104] S27. Establish the state equation and observation equation of the preprocessed signal, wherein the state equation is used to describe the dynamic changes of the preprocessed signal, and the observation equation is used to correlate the observed values ​​and state values ​​of the preprocessed signal;

[0105] It should be noted that the expression for the state equation is: ,in It is a state variable, and , The valid signal value at the current moment. Let be the rate of change of the signal; F is the state transition matrix, and , The system sampling interval; The noise is process noise, and it follows a mean of 0 and a covariance of 0. The expression for the observation equation is: ,in, The observations are for preprocessing the signal. Let be the observation matrix, and , To measure the noise, its covariance is R.

[0106] S28. Initialize the filtering parameters, and based on the state equation and observation equation, perform Kalman filtering iteration to obtain the first feature information after denoising.

[0107] The initialization of filter parameters includes: initializing state variables. T, of which The reference signal value during the system calibration phase; the initial error covariance matrix. It is consistent with the process noise covariance Q to ensure the reliability of the initial state estimation.

[0108] Specifically, such as Figure 6 As shown, the specific steps for performing the Kalman filter iteration are as follows:

[0109] S281. Obtain the best state value at the previous time step, and calculate the predicted state value and prediction error covariance value at the current time step;

[0110] Specifically, if the optimal state value at the previous moment is The predicted state value at the current moment is The prediction error covariance is .

[0111] S282. Calculate the Kalman gain value based on the predicted state value and the prediction error covariance value;

[0112] Based on the prediction error covariance Measure the noise covariance R and calculate the Kalman gain. This is used to balance the reliability of the predicted state and the accuracy of the observed values, where:

[0113]

[0114] S283. Based on the Kalman gain value, and combined with the current observation value and the predicted state value, the optimal state value and the optimal error covariance value at the current moment are obtained;

[0115] Specifically, the optimal state value , The observation residual, i.e., the difference between the predicted value and the actual observed value, is obtained by... The optimal signal value is obtained after correction. ;Optimal error covariance I is the identity matrix, which updates the error range to prepare for the next iteration.

[0116] S284. Extract the purification signal from the current optimal state value to obtain the first feature information after denoising.

[0117] Furthermore, since the first feature information obtained after denoising includes the dielectric relaxation front position f0, the dielectric relaxation front intensity Δε, and the density ρ, the values ​​of the three are very different. Direct input will lead to the large value dominating the model and the influence of the small value being ignored. Therefore, it is necessary to normalize the first feature information after denoising to obtain the second feature information.

[0118] Specifically, such as Figure 7As shown, the specific steps for normalizing the first feature information after denoising are as follows:

[0119] S29. Obtain the first feature information dataset of the finished product of glyceryl monostearate, and statistically analyze the minimum and maximum values ​​of the three features in the dataset;

[0120] It should be noted that the first feature information dataset includes the first feature information of all samples in the sample library during the model training process, and calculates the maximum and minimum values ​​of the dielectric relaxation front position f0, dielectric relaxation front intensity Δε, and density ρ of all samples in the first feature information.

[0121] S29. Perform min-max normalization calculation on the denoised first feature information to obtain the second feature information.

[0122] It should be noted that the formula for calculating the min-max normalization is as follows:

[0123]

[0124] Where x is the real-time signal value after denoising. This is the normalized value.

[0125] S3. Based on the second feature information, the purity of the finished product of glyceryl monostearate is calculated by the purity detection model.

[0126] Specifically, such as Figure 8 As shown, the training steps of the purity detection model for the glyceryl monostearate finished product include:

[0127] S31. Construct a training dataset, and divide the training dataset into a training set, a validation set, and a test set;

[0128] It's important to note that before model training, a training dataset needs to be constructed, specifically divided into a training set, a validation set, and a test set. The training set is used to train the model, the validation set is used to adjust model parameters, and the test set is used to test the model's accuracy. For example... Figure 9 As shown, the specific steps for constructing the training dataset include:

[0129] S311. Construct a sample library of finished glyceryl monostearate products, wherein the sample library is obtained by preparing finished glyceryl monostearate samples or acquiring historical sample data;

[0130] S312. Extract the second characteristic information and purity information of the finished product of glyceryl monostearate from the sample library;

[0131] S313. The three principal components in the second feature information are compressed into two principal components by principal component analysis (PCA) to obtain the preprocessed second feature information;

[0132] It should be noted that Principal Component Analysis (PCA) compresses the three core features—dielectric relaxation front position f0, dielectric relaxation front intensity Δε, and density ρ—into two principal components. The aim is to eliminate redundant information between features while retaining the main variation information of the original data, reducing the dimensionality of the model input to improve computational efficiency. The specific steps of PCA include: randomly extracting the second feature information from the sample database; calculating the covariance matrix of each feature; solving for the eigenvalues ​​and eigenvectors of the covariance matrix; sorting the obtained eigenvalues ​​in descending order; selecting the eigenvectors corresponding to the top two eigenvalues; and finally obtaining the two principal components through linear combination of the original features and their corresponding eigenvectors.

[0133] S314. Combine the preprocessed second feature information and the purity information to obtain the training dataset.

[0134] S32. Train the SVM model using the training set, and adjust the parameters using the validation set until the purity prediction error and accuracy of the test set are acceptable.

[0135] S33. End training.

[0136] Example 2

[0137] like Figure 10 As shown, this application provides an online electrical detection system for the purity of glyceryl monostearate finished product, comprising:

[0138] The first acquisition unit is used to acquire the first feature information of the finished product of glyceryl monostearate.

[0139] Specifically, the first acquisition unit includes a broadband dielectric spectrum sensor and a density sensor. The broadband dielectric spectrum sensor is used to acquire the dielectric relaxation front position f0 and dielectric relaxation front intensity Δε of the glyceryl monostearate product. The density sensor is used to acquire the density ρ of the glyceryl monostearate product. The broadband dielectric spectrum sensor is a parallel plate electrode type broadband dielectric sensor. The density sensor adopts the Coriolis force principle.

[0140] The second acquisition unit is used to preprocess the first feature information and acquire environmental parameters to obtain the second feature information.

[0141] Central processing unit; the central processing unit is used to run the purity detection model of glyceryl monostearate finished product and calculate the purity value of glyceryl monostearate finished product.

[0142] Example 3

[0143] Combination Figure 1 The method shown and Figure 10 The system shown will be used to further illustrate how the system of this application can detect the purity value of the finished product of glyceryl monostearate online.

[0144] First, the finished product of glyceryl monostearate transported in the pipeline is selected for online purity testing. The broadband dielectric spectroscopy sensor and density sensor in the first acquisition unit detect the first characteristic information of the finished product of glyceryl monostearate flowing through the pipeline at the current moment: dielectric relaxation front position f0, dielectric relaxation front intensity Δε, and density ρ. Then, the broadband dielectric spectroscopy sensor and density sensor upload the acquired first characteristic information to the system. The second acquisition unit within the system preprocesses the first characteristic information to obtain second characteristic information. The second acquisition unit then transmits the second characteristic information to the central processing unit within the system. The central processing unit is equipped with a trained glyceryl monostearate purity detection model. The central processing unit inputs the second characteristic information into the model for calculation, finally obtaining the purity value of the finished product of glyceryl monostearate flowing through the pipeline at the current moment. In actual testing, a testing cycle can be set, and the finished product of glyceryl monostearate is tested once every cycle to ensure that the final product meets the standards.

[0145] Based on the embodiments of this application, it can be concluded that the online electrical detection method and system for the purity of glyceryl monostearate product disclosed in this application can quickly and effectively monitor the purity value of glyceryl monostearate product in real time, improve the detection efficiency of glyceryl monostearate product, reduce the detection cost, and solve the problems of long detection cycle, high cost and high detection threshold of traditional chemical analysis and other methods.

[0146] It should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this application without departing from the spirit and scope of the technical solutions of this application, and all such modifications and substitutions should be covered within the scope of the claims of this application.

Claims

1. A method for on-line electrical detection of the purity of a glycerol monostearate finished product, characterized in that, The method comprises the following steps: obtaining first characteristic information of the monostearin product, wherein the first characteristic information comprises dielectric relaxation peak position f0, dielectric relaxation peak strength Δε and density ρ of the monostearin product; preprocessing the first characteristic information to obtain second characteristic information; the preprocessing step is an environmental interference compensation processing of the first characteristic information; the environmental interference compensation processing comprises compensating the dielectric relaxation peak position f0 by a first compensation model and compensating the dielectric relaxation peak strength Δε by a second compensation model; the first compensation model is constructed based on a coupling relationship between the dielectric relaxation peak position f0 and environmental parameters established by a VFT equation and an activation volume model; the second compensation model is constructed based on a coupling relationship between the dielectric relaxation peak strength Δε and environmental parameters established based on a Clausius-Mossotti equation; the environmental parameters comprise temperature T and pressure P of the monostearin product; based on the second characteristic information, a monostearin product purity detection model is used to calculate the purity of the monostearin product; the monostearin product purity detection model is trained based on historical second characteristic information of multiple calibration purity values.

2. The method according to claim 1, wherein the method is characterized by, Before the step of obtaining the first characteristic information of the monostearin product, the method further comprises the following steps: obtaining a standard sample of the monostearin product, wherein the standard sample is a monostearin product sample with a known accurate purity; based on the standard sample, calibrating reference values of the dielectric relaxation peak position f0, the dielectric relaxation peak strength Δε and the density ρ.

3. The method according to claim 1, wherein the method is characterized by, The construction step of the first compensation model comprises the following steps: establishing a relationship equation between the dielectric relaxation peak position f0 and the environmental temperature T based on the VFT equation; based on the activation volume model, constructing a correction equation of the dielectric relaxation peak position f0 with respect to the pressure P based on the relationship equation between the dielectric relaxation peak position f0 and the environmental temperature T; setting a standard value of the dielectric relaxation peak position f0 under standard environmental parameters, and constructing the first compensation model based on the standard value and a theoretical value of the dielectric relaxation peak position f0 under current environmental parameters.

4. The method according to claim 1, wherein the method is characterized by, The construction step of the second compensation model comprises the following steps: based on the Clausius-Mossotti equation, establishing a relationship equation between the dielectric relaxation peak strength Δε and the temperature T and the pressure P; calibrating coupling coefficients a~e in the relationship equation through a standard sample experiment, and obtaining the second compensation model based on the coupling coefficients a~e.

5. The method according to claim 4, wherein the method is characterized by, The specific steps of calibrating the coupling coefficients a~e in the relationship equation through the standard sample experiment comprise the following steps: preparing a plurality of standard samples of monostearin products with different purities; setting standard environmental parameters, including a standard temperature T0 and a standard pressure P0, and setting environmental parameter combinations based on the standard environmental parameters; the environmental parameter combinations are composed of environmental parameters set in a gradient; placing the standard samples in the environmental parameter combinations for testing, and recording the measured dielectric relaxation strength Δε, the temperature T and the pressure P in each combination; performing fitting operation based on the standard temperature T0, the standard pressure P0, the measured dielectric relaxation strength Δε, the temperature T and the pressure P to obtain the coupling coefficients a~e.

6. The method according to claim 1, wherein the method is characterized by, Before the second characteristic information is obtained, the method further comprises denoising and normalizing the first characteristic information.

7. The method according to claim 1, wherein the method is characterized by, The training step of the glycerol monostearate product purity detection model is: A training data set is constructed, and the training data set is divided into a training set, a validation set, and a test set; The SVM model is trained with the training set, and the parameters are adjusted through the validation set until the purity prediction error and accuracy of the test set are qualified, and the training is ended.

8. The method according to claim 7, wherein the method is characterized by, The specific steps of constructing the training data set are: A glycerol monostearate product sample library is constructed, and the sample library is obtained by preparing glycerol monostearate product samples or obtaining historical sample data; Second characteristic information and purity information of the glycerol monostearate product in the sample library are extracted; Three characteristic principal components in the second characteristic information are compressed into two characteristic principal components through principal component analysis to obtain preprocessed second characteristic information; The preprocessed second characteristic information and the purity information are combined to obtain the training data set.

9. A system for on-line electrical detection of the purity of a glycerol monostearate product, for implementing a method for on-line electrical detection of the purity of a glycerol monostearate product according to any one of claims 1 to 8, characterized in that, It comprises: A first acquisition unit for acquiring first characteristic information of a glycerol monostearate product; A second acquisition unit for preprocessing the first characteristic information and acquiring environmental parameters to obtain second characteristic information; A central processing unit for running a glycerol monostearate product purity detection model to calculate the purity value of the glycerol monostearate product.

10. The online electrical detection system for the purity of glyceryl monostearate product according to claim 9, characterized in that, The first acquisition unit comprises: A broadband dielectric spectrum sensor for collecting the dielectric relaxation peak position f0 and the dielectric relaxation peak strength Δε of the glycerol monostearate product; A density sensor for collecting the density ρ of the glycerol monostearate product; The second acquisition unit comprises: A preprocessing component for preprocessing the first characteristic information to obtain second characteristic information; An environmental parameter sensor for acquiring environmental parameters of the glycerol monostearate product.

Images