An on-line detection method for the quality of waterborne resin

Abstract

An on-line quality detection method for aqueous resin provided by the present invention relates to the technical field of material detection. The content includes: based on the collected original data, constructing a high-dimensional chimeric tensor model to generate a high-dimensional chimeric tensor; further, through time-frequency feature enhancement technology, obtaining a tensor with enhanced time-frequency features; based on the tensor with enhanced time-frequency features, adopting a prediction method based on the high-dimensional information chimeric tensor, and combining an adaptive perturbation factor to correct the predicted value of the key quality parameters of the aqueous resin; based on the predicted value of the key quality parameters of the aqueous resin, calculating the on-line quality detection result of the aqueous resin; by comparing the on-line quality detection result of the aqueous resin with the set quality tolerance threshold, determining whether the quality of the aqueous resin meets the standard. It solves the problem that the existing aqueous resin quality detection methods have deficiencies in environmental fluctuations, capturing non-linear relationships, and adapting to complex working conditions, resulting in a decrease in the accuracy of prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of material detection, and particularly to an on-line detection method for the quality of waterborne resin. Background Art

[0002] As an environmentally friendly polymer material, waterborne resin has been widely used in industries such as coatings, adhesives, textile auxiliaries, inks, and electronic materials. Compared with traditional solvent-based resins, waterborne resin has the advantages of low volatile organic compound emissions, non-flammability, and easy processing. Therefore, under the background of increasingly strict global environmental protection regulations, its market demand shows a continuous growth trend. With the continuous expansion of waterborne resin in high-performance application fields, the requirements for its quality stability and production consistency are also getting higher and higher. The final properties of waterborne resin, such as viscosity, solid content, molecular weight distribution, pH value, etc., not only depend on the ratio of raw materials, but are also affected by production processes, environmental factors, and equipment parameters. Therefore, how to detect the quality of waterborne resin in real time during the production process and optimize the control has become a key issue affecting product performance and production efficiency.

[0003] The production process of waterborne resin usually involves chemical reactions such as emulsion polymerization, polycondensation, and addition polymerization. Different types of resins require different production processes, such as stirring in a reaction kettle, high-temperature heating, catalyst addition, etc. During these production processes, changes in process parameters such as the uniformity of raw material mixing, reaction temperature, reaction pressure, and shear rate will directly affect the microstructure of the resin, thereby affecting its rheological properties, film-forming properties, mechanical strength, and weather resistance and other key indicators. Therefore, during the production process, real-time monitoring of the key quality parameters of waterborne resin and timely adjustment of process parameters when quality deviations are found to ensure that the performance of the final product meets the requirements is the core goal of improving product competitiveness and production efficiency.

[0004] Existing methods for detecting the quality of waterborne resin have deficiencies in terms of environmental fluctuations, capturing non-linear relationships, and adapting to complex working conditions, resulting in large errors in detection results and low prediction accuracy. Therefore, there is an urgent need to provide an on-line detection method for the quality of waterborne resin to solve the above problems. Summary of the Invention

[0005] The present invention provides an on-line detection method for the quality of waterborne resin to solve the problems that most of the existing waterborne resin quality detection methods lack the ability to dynamically monitor environmental variables, resulting in large errors in detection results under large environmental fluctuations, thus affecting the stability of the production process; usually using linear regression or simple neural network models for prediction, unable to effectively capture non-linear relationships, resulting in insufficient prediction accuracy of the detection system; and mostly relying on mathematical models with fixed parameters, difficult to adapt to complex working condition changes in the production process, such as dynamic changes in quality parameters caused by factors such as raw material formula adjustment, equipment aging, and differences between production batches, thus leading to a decrease in prediction accuracy.

[0006] An on-line detection method for the quality of waterborne resin according to the present invention comprises the following steps:

[0007] S1. Collect the original data including the physical properties of the waterborne resin and the external parameters of the production environment. Based on the original data, construct a high-dimensional chimeric tensor model to generate a high-dimensional chimeric tensor; based on the high-dimensional chimeric tensor, through time-frequency feature enhancement technology, obtain the tensor after time-frequency feature enhancement;

[0008] S2. Based on the tensor after time-frequency feature enhancement, adopt a prediction method based on the high-dimensional information chimeric tensor, and combine an adaptive perturbation factor to correct the predicted value of the key quality parameters of the waterborne resin; based on the predicted value of the key quality parameters of the waterborne resin, calculate the on-line detection result of the waterborne resin quality; set a quality tolerance threshold, and judge whether the quality of the waterborne resin meets the standard by comparing the on-line detection result of the waterborne resin quality with the quality tolerance threshold.

[0009] Preferably, the S1 specifically comprises:

[0010] Monitor the physical properties of the waterborne resin and the external parameters of the production environment through sensors; the collected original data is presented as a multi-dimensional time-varying sequence, containing information in both time and feature space; construct a high-dimensional chimeric tensor model to characterize the multi-dimensional characteristics of the waterborne resin quality by fusing time series data, physical feature vectors, and environmental parameters.

[0011] Preferably, the S1 specifically comprises:

[0012] The mathematical representation of the high-dimensional chimeric tensor model is as follows:

[0013] ,

[0014] wherein, is the high-dimensional chimeric tensor; represents the time series data obtained by the th sensor; is the time step, representing the number of time points for collecting time series data; is the th physical feature vector, describing the physical properties of the aqueous resin; is the number of physical features of the aqueous resin; is the th environmental parameter; is the number of environmental parameters; is the chimeric weight; is the tensor product.

[0015] Preferably, the S1 specifically includes:

[0016] In the implementation process of the time-frequency feature enhancement technology, by performing a complex Fourier transform on the high-dimensional chimeric tensor, the frequency features in the high-dimensional chimeric tensor are extracted, and the fluctuations of the high-dimensional chimeric tensor are smoothed by a sine function.

[0017] Preferably, the S2 specifically includes:

[0018] By performing real-time integration on the deviation between the tensor after time-frequency feature enhancement and the target value, and combining with a non-linear compensation term to calculate the adaptive perturbation factor; the calculation formula of the adaptive perturbation factor is:

[0019] ,

[0020] where, is the adaptive perturbation factor; is the upper limit of time integration; is the th sensor's time series data after time-frequency feature enhancement at time ; is the expected value of the th sensor in the historical data, and the historical data is obtained from the database; is at time the th physical feature vector after time-frequency feature enhancement; is the th physical feature vector's target value; is at time the th environmental parameter after time-frequency feature enhancement; is the th environmental parameter's reference value; , and are weight factors, respectively controlling the contributions of time series deviation, physical property deviation, and environmental parameter deviation to the adaptive perturbation factor; , , are non - linear exponents, which are used to adjust the contribution intensities of the time - series deviation, physical - property deviation, and environmental - parameter deviation respectively; is the non - linear compensation term, and are regularization factors, which are used to adjust the influence intensity of the non - linear compensation term; is the tensor after time - frequency feature enhancement.

[0021] Preferably, the S2 specifically includes:

[0022] The calculation formula for the predicted value of the key quality parameter of the water - based resin is:

[0023] ,

[0024] where, is the predicted value of the key quality parameter of the water - based resin at time , including the viscosity , solid content , pH value and molecular - weight distribution ; is the regression weight; is the angular frequency; is a constant to prevent the denominator from being zero; is the predicted value of the key quality parameter of the water - based resin at time .

[0025] Preferably, the S2 specifically includes:

[0026] Based on the predicted value of the key quality parameter of the water - based resin, evaluate whether the quality of the current water - based resin meets the standard and decide whether to adjust the production parameters.

[0027] Preferably, the S2 specifically includes:

[0028] The on - line detection result of the water - based resin quality is calculated by the following formula:

[0029] ,

[0030] where, is the quality - error evaluation value of the on - line detection; is the total number of key quality parameters; is the predicted value of the th key quality parameter at time is the th is at time The adaptive weighting factor.

[0031] Preferably, the S2 specifically includes:

[0032] Set the quality tolerance threshold. When the quality error evaluation value detected online is less than or equal to the quality tolerance threshold, the quality of the waterborne resin meets the standard and there is no need to adjust the production parameters; otherwise, the quality of the waterborne resin does not meet the standard and the production parameters need to be adjusted.

[0033] The beneficial effects of the technical solution of the present invention are:

[0034] 1. Construct a high-dimensional chimeric tensor model. By fusing time series data, physical property parameters, and environmental parameters, a comprehensive characterization of the quality of the waterborne resin is formed; by introducing time-frequency feature enhancement technology, short-term fluctuations and long-term trends in the production process of the waterborne resin are effectively separated, and the dynamic changes in the production process of the waterborne resin are accurately captured, thereby improving the accuracy and stability of the quality detection of the waterborne resin.

[0035] 2. Based on the tensor enhanced by time-frequency features, combined with the adaptive perturbation factor, the output of the online quality detection system for the waterborne resin is corrected, effectively reducing the error of the online quality detection system for the waterborne resin and enhancing the adaptability of the online quality detection system for the waterborne resin to external environmental changes.

[0036] 3. Adjust the production parameters based on the quality error evaluation value detected online, so that the quality of the waterborne resin is always kept within the standard range; realize automatic real-time monitoring during the production process and provide accurate quality prediction results, so as to reduce production costs while improving production efficiency. Description of the Drawings

[0037] Figure 1 It is a flowchart of an online quality detection method for a waterborne resin according to the present invention. Detailed Embodiments

[0038] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0040] The following specifically describes the specific solution of an on-line quality detection method for waterborne resin provided by the present invention in conjunction with the accompanying drawings.

[0041] Refer to the attached Figure 1 , which shows a flow chart of an on-line quality detection method for waterborne resin provided by an embodiment of the present invention. The method includes the following steps:

[0042] S1. Collect the raw data including the physical properties of the waterborne resin and the external parameters of the production environment. Based on the raw data, construct a high-dimensional chimeric tensor model to generate a high-dimensional chimeric tensor; based on the high-dimensional chimeric tensor, through the time-frequency feature enhancement technology, obtain the tensor after time-frequency feature enhancement.

[0043] During the production process of waterborne resin, the quality stability is closely related to multiple factors, including the proportion of raw materials, temperature, pressure, humidity changes during the reaction process, and even the operating state of the equipment. Therefore, the collected raw data should not only cover the physical properties of the waterborne resin, but also include the external parameters of the production environment, so as to ensure that all factors affecting quality changes are captured comprehensively.

[0044] First of all, the physical properties of waterborne resin, such as viscosity, solid content, molecular weight distribution, pH value, etc., need to be accurately measured by real-time on-line sensors. Viscosity is a key parameter affecting the stability and fluidity of waterborne resin, and is measured by a rotational viscometer or a vibrating viscometer. The solid content represents the proportion of solid components in waterborne resin, and is obtained by the drying method or infrared analysis method. The molecular weight distribution is the molecular structure characteristic of waterborne resin, and is measured by gel permeation chromatography (GPC) technology. The obtained molecular weight distribution can further help to understand the rheological properties of waterborne resin. The pH value measured by a pH meter reflects the acidity and alkalinity of waterborne resin. The physical properties of waterborne resin, such as viscosity, solid content, molecular weight distribution, pH value, etc. are all key parameters to ensure product quality, that is, key quality parameters.

[0045] Secondly, external factors such as temperature, humidity in the production environment and pressure inside the reactor also need to be monitored by sensors. Temperature has a direct impact on the reaction rate and the chemical reaction equilibrium between reactants, so temperature data needs to be obtained in real time through a temperature sensor. Humidity changes may also affect the solid content of waterborne resin. Especially in a high-humidity environment, the moisture in waterborne resin may affect its viscosity and solid content, thus affecting the quality of the final product. Therefore, humidity data needs to be obtained in real time through a humidity sensor. Monitor the pressure change inside the reactor through a pressure sensor, which is also a factor that cannot be ignored for understanding the stability of the reaction system.

[0046] The collected raw data is presented as a multi-dimensional time-varying sequence, containing information in both the time and feature spaces. To make full use of the raw data, a high-dimensional chimeric tensor model is constructed to characterize the multi-dimensional characteristics of the waterborne resin quality by fusing time-series data, physical feature vectors, and environmental parameters.

[0047] ,

[0048] Among them, is the high-dimensional chimeric tensor, representing the fused representation of multiple feature data in the production process of the waterborne resin; represents the time-series data obtained by the th sensor, containing the time-varying characteristics in the production process of the waterborne resin; is the time step, that is, the number of time points for collecting the time-series data; is the th physical feature vector, describing the physical properties of the waterborne resin, such as viscosity, solid content, pH value, and molecular weight distribution, etc.; is the number of physical features of the waterborne resin; is the th environmental parameter, such as influencing factors like temperature and humidity in the production environment; is the number of environmental parameters; is the chimeric weight; is the tensor product, used to combine data from different dimensions so that they can jointly represent a high-dimensional feature space. The raw data is combined through the tensor product operation to form a high-dimensional chimeric tensor for comprehensive prediction of the waterborne resin quality.

[0049] The chimeric weight in the high-dimensional chimeric tensor is calculated based on the similarity between the raw data, and is used to represent the , , correlation degree between them. The calculation formula for the chimeric weight is as follows:

[0050] ,

[0051] Among them, and are respectively the , and , Euclidean distances between them, used to measure the similarity between data; is the distance attenuation factor, controlling the influence of the Euclidean distance on the chimeric weight, with a value range of , and the specific value is calibrated through experiments; is a regularization factor used to avoid over-weighting when the data distance difference is too large.

[0052] Based on the high-dimensional chimeric tensor, a time-frequency feature enhancement technique is introduced to capture the dynamic change characteristics of the quality of waterborne resin during the production process. By performing a complex Fourier transform on the high-dimensional chimeric tensor, the frequency features in the high-dimensional chimeric tensor are extracted, and the fluctuations of the high-dimensional chimeric tensor are smoothed by a sine function to enhance the periodic change characteristics of the high-dimensional chimeric tensor. Time-frequency feature enhancement is crucial for processing production data with periodic changes because, during the production process of waterborne resin, factors such as temperature and reaction time usually exhibit periodic fluctuations.

[0053] ,

[0054] where, is the tensor after time-frequency feature enhancement; is the Hadamard product, representing element-wise multiplication; The term is the complex Fourier transform term used to extract the periodic changes of the high-dimensional chimeric tensor; represents the imaginary number; is the angular frequency, calculated by Fourier transform; is the time; The term is used to smooth the high-dimensional chimeric tensor and reduce the interference of high-frequency noise; is the oscillation frequency control parameter, depending on the periodicity of the high-dimensional chimeric tensor; is the smoothing factor used to avoid a zero denominator.

[0055] S2. Based on the tensor after time-frequency feature enhancement, a prediction method based on the high-dimensional information chimeric tensor is adopted, and the predicted values of the key quality parameters of the waterborne resin are corrected by combining an adaptive perturbation factor; based on the predicted values of the key quality parameters of the waterborne resin, the online detection results of the quality of the waterborne resin are calculated; a quality tolerance threshold is set, and by comparing the online detection results of the quality of the waterborne resin with the quality tolerance threshold, it is judged whether the quality of the waterborne resin meets the standard.

[0056] To ensure that the on-line quality detection system of waterborne resin maintains high efficiency and adaptability to complex and variable working conditions during the actual production process, an adaptive disturbance factor is introduced to dynamically adjust the stability of data processing and improve the accurate prediction ability of the quality change trend of waterborne resin. During the actual production process, key quality parameters of waterborne resin such as viscosity, solid content, molecular weight distribution, and pH value are easily affected by factors such as environmental temperature and humidity fluctuations, raw material batch differences, and equipment operating status, resulting in mutations or drifts in measurement data, and may also cause short-term misjudgments or prediction distortions in the on-line quality detection system of waterborne resin, thus affecting the accuracy of overall quality control. Therefore, it is necessary to adaptively correct the output of the on-line quality detection system of waterborne resin through a dynamically adjustable adaptive disturbance factor so that it can still provide highly reliable data prediction and decision support under complex working conditions. The adaptive disturbance factor calculates the real-time integral of the deviation between the tensor enhanced by time-frequency characteristics and the target value, and combines a non-linear compensation term to ensure that the prediction of the key quality parameters of subsequent waterborne resin can effectively adapt to data changes and avoid the problem of error amplification caused by over-response to short-term fluctuations. The calculation formula of the adaptive disturbance factor is:

[0057] ,

[0058] where, is the adaptive disturbance factor; is the upper limit of time integration; is the -th sensor's time series data after time-frequency feature enhancement at time , ); is the expected value (such as historical average or smoothed estimate) of the -th sensor in historical data, and the historical data is obtained from the database; is the -th physical feature vector after time-frequency feature enhancement at time , ); is the target value (such as target viscosity or target solid content) of the -th physical feature vector; is the -th environmental parameter after time-frequency feature enhancement at time , ); is the reference value (such as target temperature or target humidity) of the -th environmental parameter; , and is the weight factor, which controls the contributions of the time series deviation, physical property deviation, and environmental parameter deviation to the adaptive perturbation factor respectively; , , are the non - linear exponents, which are used to adjust the contribution intensities of the time series deviation, physical property deviation, and environmental parameter deviation respectively; is the non - linear compensation term, and are the regularization factors, which are used to adjust the influence intensity of the non - linear compensation term.

[0059] Obtain the prediction results of the key quality parameters in the production process of water - borne resin, such as viscosity, solid content, molecular weight distribution, and pH value, etc. Since the quality of water - borne resin is affected by multiple factors, including raw material ratio, production environment, temperature change, etc., it is necessary to integrate multi - dimensional data and dynamically adjust the short - term fluctuations and long - term trends in the time series. For this purpose, a prediction method based on high - dimensional information - embedded tensor is adopted. The data from multiple data sources (time series, physical properties, environmental parameters) are fused, and the weight is corrected by combining the adaptive perturbation factor to enhance the adaptability of the key quality parameter prediction process of water - borne resin to the real - time production environment. First, time - frequency analysis is carried out in the high - dimensional tensor space to extract its periodic and non - periodic components, and then the high - frequency components are extracted through Fourier transform to identify the short - term fluctuations that may affect the quality of water - borne resin. At the same time, a smoothing term is introduced to suppress high - frequency noise. Finally, in order to make the prediction results more stable, dynamic weighting is carried out through the adaptive perturbation factor, so that high - precision prediction can still be maintained under different production states, and thus the prediction values of the key quality parameters of water - borne resin are obtained:

[0060] ,

[0061] where, is the predicted value of the key quality parameter of water - borne resin at time , including the viscosity , solid content , pH value and molecular weight distribution of water - borne resin; is the regression weight, which is used to measure the importance of the tensor after time - frequency feature enhancement in the prediction process; is the angular frequency, which represents the periodic change of the correction term; is a constant to prevent the denominator from being zero; is the predicted value of the key quality parameter of water - borne resin at time .

[0062] After obtaining the predicted values of the key quality parameters of the aqueous resin, it is necessary to calculate the online detection results of the aqueous resin quality, that is, to evaluate whether the quality of the current aqueous resin meets the standards and decide whether it is necessary to adjust the production parameters. The final result of the online detection of the aqueous resin quality is calculated by the following formula:

[0063] ,

[0064] where, is the quality error evaluation value of the online detection; is the total number of key quality parameters; is at time the predicted value of the th key quality parameter; is the ideal target value of the th key quality parameter; is the adaptive weighting factor at time , indicating the importance of the

[0065] ,

[0066] where, is the error attenuation factor, used to control the influence of different key quality parameters on the overall quality assessment; is the error compensation term, is the th error compensation factor of the key quality parameter, used to prevent the error of the key quality parameter from being too large and affecting the overall quality assessment; is the regularization factor, used to adjust the smoothness of the adaptive weighting factor.

[0067] Set a quality tolerance threshold through the expert experience method. By comparing the quality tolerance threshold with the quality error evaluation value of the online detection, it is judged whether the quality of the aqueous resin meets the standards. If , it means that the quality parameters of the current aqueous resin are within the acceptable range and there is no need to adjust the production parameters; otherwise, it means that the quality of the aqueous resin deviates from the target value and it is necessary to adjust the production parameters to optimize the quality.

[0068] In summary, an online detection method for the quality of aqueous resin is completed.

[0069] The sequence of the embodiments of the invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0070] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

[0071] The above embodiments are only used to illustrate the technical solutions of the present invention, not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An on-line detection method for the quality of aqueous resin, characterized in that, It includes the following steps: S1. Collect the original data including the physical properties of the waterborne resin and the external parameters of the production environment. Based on the original data, construct a high-dimensional chimeric tensor model to generate a high-dimensional chimeric tensor. Based on the high-dimensional chimeric tensor, through the time-frequency feature enhancement technology, obtain the tensor after time-frequency feature enhancement; S2. Based on the tensor after time-frequency feature enhancement, adopt a prediction method based on the high-dimensional information chimeric tensor, and perform real-time integration on the deviation between the tensor after time-frequency feature enhancement and the target value. Combine the non-linear compensation term to calculate the adaptive perturbation factor, and the calculation formula is: ， Among them, is the adaptive perturbation factor; is the upper limit of time integration; is the th sensor's time series data after time-frequency feature enhancement at time ; is the expected value of the th sensor in historical data, and the historical data is obtained from the database; is the th physical feature vector after time-frequency feature enhancement at time ; is the target value of the th physical feature vector; is the th environmental parameter after time-frequency feature enhancement at time ; is the reference value of the th environmental parameter; , and are weight factors; , , are non-linear exponents; is the non-linear compensation term, and are regularization factors; is the tensor after time-frequency feature enhancement; And combine the adaptive perturbation factor to correct the predicted value of the key quality parameters of the waterborne resin, and the calculation formula is: ， Among them, is the predicted value of the key quality parameters of the aqueous resin at time , including the viscosity of the aqueous resin , solid content , pH value and molecular weight distribution ; is the regression weight; is the angular frequency; is a constant to prevent the denominator from being zero; is the predicted value of the key quality parameters of the aqueous resin at time ; Based on the predicted value of the key quality parameters of the waterborne resin, calculate the on-line detection result of the waterborne resin quality. Set the quality tolerance threshold, and judge whether the quality of the waterborne resin meets the standard by comparing the on-line detection result of the waterborne resin quality with the quality tolerance threshold.

2. The online detection method for the quality of the aqueous resin according to claim 1, characterized in that The S1 specifically includes: Monitor the physical properties of the waterborne resin and the external parameters of the production environment through sensors; the collected original data is presented as a multi-dimensional time-varying sequence, which contains information in time and feature space; construct a high-dimensional chimeric tensor model to characterize the multi-dimensional characteristics of the waterborne resin quality by fusing time series data, physical feature vectors and environmental parameters.

3. The online detection method for the quality of the aqueous resin according to claim 2, wherein, The S1 specifically includes: The mathematical representation of the high-dimensional chimeric tensor model is as follows: ， Among them, is a high-dimensional chimeric tensor; represents the time series data obtained by the th sensor; is the time step, indicating the number of time points for collecting the time series data; is the th physical feature vector, describing the physical properties of the aqueous resin; is the number of physical features of the aqueous resin; is the th environmental parameter; is the number of environmental parameters; is the chimeric weight; is the tensor product.

4. The on-line detection method for the quality of the aqueous resin according to claim 3, wherein, The S1 specifically includes: In the implementation process of the time-frequency feature enhancement technology, perform a complex Fourier transform on the high-dimensional chimeric tensor, extract the frequency features in the high-dimensional chimeric tensor, and smooth the fluctuations of the high-dimensional chimeric tensor through a sine function.

5. The on-line quality detection method for the aqueous resin according to claim 1, characterized in that The S2 specifically includes: Based on the predicted value of the key quality parameters of the waterborne resin, evaluate whether the quality of the current waterborne resin meets the standard and decide whether it is necessary to adjust the production parameters.

6. The on-line detection method for the quality of the aqueous resin according to claim 1, characterized in that The S2 specifically includes: The on-line detection result of the waterborne resin quality is calculated by the following formula: ， Among them, is the quality error evaluation value of on-line detection; is the total number of key quality parameters; is at time the predicted value of the th key quality parameter; is at time the adaptive weighting factor.

7. The on-line detection method for the quality of the aqueous resin according to claim 6, wherein The S2 specifically includes: Set the quality tolerance threshold. When the on-line detected quality error evaluation value is less than or equal to the quality tolerance threshold, the quality of the waterborne resin meets the standard and there is no need to adjust the production parameters; otherwise, the quality of the waterborne resin does not meet the standard and the production parameters need to be adjusted.

Images