Printing ink quality online monitoring method and system based on multi-sensor fusion

Abstract

The invention provides an ink quality online monitoring method and system based on multi-sensor fusion, and relates to the technical field of printers, and the method comprises the steps: obtaining the temperature, humidity, pH value, viscosity, particle thickness, flow velocity and other data collected by a plurality of sensors in the production process of water-based ink, and employing a multi-sensor data fusion algorithm of adaptive weight, calculating a sensing weight corresponding to the sensor data, and combining the sensor data to obtain fused water-based ink quality characteristic data; establishing a water-based ink quality prediction model by adopting a support vector machine algorithm, and inputting the fused water-based ink quality feature data into the trained water-based ink quality prediction model to obtain a water-based ink quality prediction result; and if the water-based ink quality prediction result does not meet the preset condition, judging that the water-based ink quality is unqualified, sending a feedback signal in a preset format to water-based ink production equipment, and adjusting water-based ink production parameters through a PID controller of the water-based ink production equipment. According to the method, the related parameters of the ink can be monitored and reflected in real time, and the possible problem of artificial false description is solved.

Description

Technical Field

[0001] The present invention relates to printing press technology, and in particular to an online ink quality monitoring method and system based on multi-sensor fusion. Background Art

[0002] Ink is an indispensable and important material in the printing industry. The stability and consistency of its quality directly affect the quality and production efficiency of printed products. Traditional ink quality testing mainly relies on offline testing methods, such as sampling testing and laboratory analysis. There are problems such as low testing efficiency, poor real-time performance and insufficient sample representativeness, which makes it difficult to meet the requirements of modern printing industry for ink quality control.

[0003] In recent years, with the development of sensor technology and information technology, online quality monitoring methods have been widely used. By installing various sensors on the production line, real-time collection of key parameters in the production process, and combining data analysis and modeling methods, online monitoring and prediction of product quality can be achieved. In the field of ink production, some researchers and companies have begun to explore online ink quality monitoring methods based on multi-sensor fusion.

[0004] The existing technology still has the following shortcomings: First, the sensor selection and layout are not optimized enough, making it difficult to fully reflect the quality characteristics of the ink; second, the data fusion algorithm is relatively simple and does not fully utilize multi-sensor information, and the accuracy of quality assessment needs to be improved; third, there is a lack of effective quality prediction and process control mechanisms, making it difficult to achieve predictive control of ink quality. Summary of the invention

[0005] The embodiments of the present invention provide an online ink quality monitoring method and system based on multi-sensor fusion, which can solve the problems in the prior art.

[0006] According to a first aspect of the embodiments of the present invention,

[0007] Provide an online ink quality monitoring method based on multi-sensor fusion, including:

[0008] Acquire real-time data collected by multiple sensors during the ink production process, and transmit the real-time data to a data processing unit via industrial Ethernet; the data processing unit uses a Kalman filter algorithm to filter the received real-time data, remove noise and abnormal values ​​in the real-time data, obtain filtered real-time data, use a multi-sensor data fusion algorithm with adaptive weights to calculate the sensor weights corresponding to the real-time data, and combine the real-time data to obtain fused ink quality characteristic data;

[0009] A support vector machine algorithm is used to establish an ink quality prediction model, and the ink quality prediction model is obtained by training a large amount of historical ink production data and corresponding ink quality test results. The fused ink quality feature data is input into the trained ink quality prediction model to obtain an ink quality prediction result;

[0010] If the ink quality prediction result meets the preset conditions, the ink quality is determined to be qualified; if the ink quality prediction result does not meet the preset conditions, the ink quality is determined to be unqualified, and a feedback signal in a preset format is sent to the ink production equipment, and the ink production process parameters are adjusted through the PID controller of the ink production equipment. The adjusted ink production process parameters are fed back to the data processing unit to perform the next ink quality prediction until the ink quality prediction result meets the preset conditions, thereby completing the closed-loop control of online monitoring and dynamic regulation of ink quality.

[0011] In an optional embodiment,

[0012] The multi-sensor data fusion algorithm with adaptive weights is used to calculate the sensor weights corresponding to the real-time data, and the fused ink quality characteristic data obtained by combining the real-time data includes:

[0013] Calculate the signal-to-noise ratio of the real-time data collected by each sensor to obtain signal-to-noise ratio data representing the quality of the data of each sensor;

[0014] Calculate the correlation coefficients between real-time data collected by different sensors;

[0015] Based on the signal-to-noise ratio data and the mutual correlation coefficient, a sensor data quality assessment matrix is ​​constructed, wherein the diagonal elements of the sensor data quality assessment matrix are the signal-to-noise ratios of each sensor data, and the non-diagonal elements are the mutual correlation coefficients between different sensor data;

[0016] Utilizing the sensor data quality assessment matrix, calculating the adaptive weight of each sensor data to obtain a normalized sensor data adaptive weight vector;

[0017] The sensor data adaptive weight vector is weightedly fused with real-time data collected by multiple sensors to obtain fused ink quality characteristic data.

[0018] In an optional embodiment,

[0019] The ink quality prediction model is established by using the support vector machine algorithm. The ink quality prediction model is obtained by training a large amount of historical ink production data and the corresponding ink quality test results. The fused ink quality feature data is input into the trained ink quality prediction model, and the ink quality prediction results obtained include:

[0020] Collecting historical production data and corresponding ink quality test results during the ink production process, preprocessing the historical production data and the ink quality test results to obtain a training sample set;

[0021] The training sample set is trained by using a support vector machine algorithm, and an ink quality prediction model is obtained by constructing an optimization objective function and solving the optimization problem;

[0022] Acquire real-time data collected by multiple sensors during the ink production process, perform filtering and feature fusion on the real-time data, and obtain fused ink quality feature data;

[0023] The fused ink quality feature data is input into the ink quality prediction model, and a real-time ink quality prediction result is obtained through model prediction.

[0024] In an optional embodiment,

[0025] Constructing the optimization objective function and solving the optimization problem include:

[0026] min(1 / 2*||w|| 2 +C*sum(ξ i ));

[0027] sty i *(w·K(x i )+b)≥1-ξ i ,ξ i ≥0;

[0028] Among them, w is the normal vector of the hyperplane, b is the bias term of the hyperplane, C is the penalty coefficient, ξ i It is a slack variable, which is used to allow some sample classification errors;

[0029] Through the Lagrange multiplier method and duality technique, the above optimization problem is transformed into the following dual problem:

[0030] max(sum(α i )-(1 / 2)*sum(α i *y i *K(x i )));

[0031] sum(α i *y i )=0,0≤α i ≤C;

[0032] Among them, α i is the Lagrange multiplier, K(x i ) is the kernel function, x irepresents the i-th sample, y i Represents the ink quality prediction result corresponding to the i-th sample.

[0033] In an optional embodiment,

[0034] Adjusting the ink production process parameters by the PID controller of the ink production equipment and feeding back the adjusted ink production process parameters to the data processing unit includes:

[0035] Establishing a mathematical model between ink viscosity, ink temperature and ink composition ratio, wherein the mathematical model represents the quantitative relationship between ink viscosity, ink temperature and ink composition ratio;

[0036] According to the mathematical model, an ink temperature control loop and an ink composition ratio control loop are designed;

[0037] The ink temperature control loop measures the ink temperature through a temperature sensor, and controls the heating / cooling device to adjust the ink temperature according to the deviation between the temperature setting value and the actual value through a temperature PID controller;

[0038] The ink composition ratio control loop measures the ink composition ratio in real time through an online measuring device, and controls the addition amount of the ink components according to the deviation between the composition ratio set value and the actual value by the composition ratio PID controller.

[0039] In an optional embodiment,

[0040] The method further comprises:

[0041] According to the mathematical model, the deviation between the current ink viscosity value and the target viscosity value is calculated, the viscosity deviation is substituted into the mathematical model, the adjustment amount of the ink temperature and the ink component ratio is calculated, and the temperature adjustment amount and the component ratio adjustment amount are transmitted to the temperature PID controller and the component ratio PID controller respectively;

[0042] The temperature PID controller and the component ratio PID controller coordinately control the heating / cooling device and the ink component adding device according to their respective adjustment amounts, and synchronously adjust the ink temperature and the ink component ratio to stabilize the ink viscosity within the target range.

[0043] In the second aspect of the embodiments of the present application,

[0044] Provide an online ink quality monitoring system based on multi-sensor fusion, including:

[0045] The first unit is used to obtain real-time data collected by multiple sensors during the ink production process, and transmit the real-time data to a data processing unit through industrial Ethernet; the data processing unit uses a Kalman filter algorithm to filter the received real-time data, remove noise and abnormal values ​​in the real-time data, obtain filtered real-time data, use a multi-sensor data fusion algorithm with adaptive weights to calculate the sensor weights corresponding to the real-time data, and combine the real-time data to obtain fused ink quality characteristic data;

[0046] The second unit is used to establish an ink quality prediction model by using a support vector machine algorithm, obtain the ink quality prediction model by training a large amount of historical ink production data and corresponding ink quality test results, and input the fused ink quality feature data into the trained ink quality prediction model to obtain the ink quality prediction result;

[0047] The third unit is used to determine that the ink quality is qualified if the ink quality prediction result meets the preset conditions; if the ink quality prediction result does not meet the preset conditions, determine that the ink quality is unqualified, send a feedback signal in a preset format to the ink production equipment, adjust the ink production process parameters through the PID controller of the ink production equipment, and feed back the adjusted ink production process parameters to the data processing unit to perform the next ink quality prediction until the ink quality prediction result meets the preset conditions, thereby completing the closed-loop control of online monitoring and dynamic regulation of ink quality.

[0048] According to a third aspect of the embodiments of the present invention,

[0049] An electronic device is provided, comprising:

[0050] processor;

[0051] a memory for storing processor-executable instructions;

[0052] The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0053] According to a fourth aspect of the embodiments of the present invention,

[0054] A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0055] This application adopts a multi-sensor data fusion algorithm with adaptive weights, which adaptively adjusts the sensor weights according to the quality and importance of different sensor data, realizes the efficient fusion of multi-source heterogeneous sensor data, makes full use of multi-sensor information, comprehensively reflects the ink quality characteristics, and improves the accuracy and reliability of ink quality assessment.

[0056] By introducing the support vector machine algorithm, an ink quality prediction model was established. By training on a large amount of historical production data and quality inspection results, the intrinsic relationship between ink quality and production process parameters was learned, and online prediction of ink quality was achieved. The real-time and predictive nature of quality monitoring was improved, providing a predictive basis for quality control.

[0057] Based on the ink quality prediction results, determine whether the ink quality is qualified. For unqualified inks, dynamically adjust the production process parameters through the PID controller, and feed back the adjustment results to the data processing unit to form a closed-loop feedback of quality monitoring and process control, realize dynamic regulation of ink quality, and ensure the stability and consistency of ink quality.

[0058] The entire method integrates multiple links such as data acquisition, data preprocessing, multi-sensor fusion, quality prediction and process control, forming a complete online monitoring and control solution for ink quality. It is systematic and practical, and is of great significance to improving the quality of ink products and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a flow chart of an online ink quality monitoring method based on multi-sensor fusion according to an embodiment of the present invention;

[0060] Figure 2 It is a structural schematic diagram of an online ink quality monitoring system based on multi-sensor fusion according to an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0063] Figure 1 FIG. 1 is a flow chart of an online ink quality monitoring method based on multi-sensor fusion according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0064] S101. Acquire real-time data collected by multiple sensors during the ink production process, and transmit the real-time data to a data processing unit via industrial Ethernet; the data processing unit uses a Kalman filter algorithm to filter the received real-time data, remove noise and abnormal values ​​in the real-time data, obtain filtered real-time data, use a multi-sensor data fusion algorithm with adaptive weights to calculate the sensor weights corresponding to the real-time data, and combine the real-time data to obtain fused ink quality feature data;

[0065] For example, during the ink production process, multiple sensors are deployed to monitor the production environment and equipment operation status in real time. These sensors include temperature sensors, humidity sensors, pH sensors, viscosity sensors, flow rate sensors, etc. Each sensor collects corresponding physical quantity data at a certain sampling frequency, converts analog signals into digital signals, and transmits digital real-time data to the on-site data acquisition unit through communication protocols such as RS-485 or ModBus.

[0066] The data acquisition unit transmits the real-time data collected by multiple sensors to the data processing unit via industrial Ethernet. Industrial Ethernet is a high-speed, reliable industrial communication network based on the IEEE 802.3 standard that can achieve efficient transmission of real-time data. The data processing unit is usually an industrial computer or embedded system equipped with a high-performance CPU, large-capacity memory and storage devices, which can quickly process and analyze the massive real-time data received.

[0067] Since the real-time data collected by the sensor may contain noise and outliers, which will affect the accuracy of subsequent data analysis, it is necessary to filter the real-time data. The data processing unit uses the Kalman filter algorithm to filter the received real-time data, remove the noise and outliers in the data, and obtain the filtered real-time data.

[0068] In an optional embodiment,

[0069] The multi-sensor data fusion algorithm with adaptive weights is used to calculate the sensor weights corresponding to the real-time data, and the fused ink quality characteristic data obtained by combining the real-time data includes:

[0070] Calculate the signal-to-noise ratio of the real-time data collected by each sensor to obtain signal-to-noise ratio data representing the quality of the data of each sensor;

[0071] Calculate the correlation coefficients between real-time data collected by different sensors;

[0072] Based on the signal-to-noise ratio data and the mutual correlation coefficient, a sensor data quality assessment matrix is ​​constructed, wherein the diagonal elements of the sensor data quality assessment matrix are the signal-to-noise ratios of each sensor data, and the non-diagonal elements are the mutual correlation coefficients between different sensor data;

[0073] Utilizing the sensor data quality assessment matrix, calculating the adaptive weight of each sensor data to obtain a normalized sensor data adaptive weight vector;

[0074] The sensor data adaptive weight vector is weightedly fused with real-time data collected by multiple sensors to obtain fused ink quality characteristic data.

[0075] For example, for the real-time data collected by each sensor, the signal-to-noise ratio (SNR) is calculated. Assuming that the real-time data collected by sensor i is x_i(t) and the corresponding noise is n_i(t), the signal-to-noise ratio can be expressed as:

[0076] SNR_i=10*log10(sum(x_i(t)^2) / sum(n_i(t)^2));

[0077] Among them, sum() represents the sum of time t. The higher the signal-to-noise ratio, the better the quality of the sensor data and the smaller the noise interference.

[0078] Calculate the mutual correlation coefficient between different sensor data. Assuming that the real-time data collected by sensor i and sensor j are x_i(t) and x_j(t) respectively, the mutual correlation coefficient between them can be expressed as:

[0079] R_ij=sum((x_i(t)-mean(x_i))*(x_j(t)-mean(x_j))) / (std(x_i)*std(x_j));

[0080] Among them, mean() means to find the mean, and std() means to find the standard deviation. The value range of the cross-correlation coefficient is [-1,1]. The larger the absolute value, the stronger the correlation between the two sensor data.

[0081] According to the signal-to-noise ratio of each sensor data and the correlation coefficient between different sensor data, the sensor data quality assessment matrix Q is constructed:

[0082] Q = [SNR_1, R_12, ..., R_1n;

[0083] R_21,SNR_2,...,R_2n;

[0084] ...,...,...,...;

[0085] R_n1,R_n2,...,SNR_n];

[0086] Where n represents the number of sensors. The diagonal elements of the matrix Q are the signal-to-noise ratios of the sensor data, and the off-diagonal elements are the correlation coefficients between different sensor data.

[0087] Using the sensor data quality assessment matrix Q, the adaptive weight w_i of each sensor data is calculated:

[0088] w=(sum(Q,2)-diag(Q)) / (n-1);

[0089] w=w / sum(w)

[0090] Where sum(Q,2) represents the sum of each row of the matrix Q, diag(Q) represents extracting the diagonal elements of the matrix Q to form a vector, n represents the number of sensors, and sum(w) represents the sum of the vector w. Through the above calculation, the normalized sensor data adaptive weight vector w is obtained, satisfying sum(w)=1.

[0091] The adaptive weight vector w of sensor data is used to perform weighted fusion on the real-time data collected by multiple sensors to obtain the ink quality characteristic data y(t):

[0092] y(t)=sum(w.*[x_1(t);x_2(t);...;x_n(t)]);

[0093] Among them, .* represents the dot product of the vector, and [x_1(t); x_2(t); ...; x_n(t)] represents the vector formed by splicing the real-time data collected by multiple sensors in columns. Through adaptive weighted fusion, the characteristic data y(t) that comprehensively reflects the ink quality is obtained, which can be used for subsequent ink quality prediction and analysis.

[0094] The above steps provide a multi-sensor data fusion algorithm using adaptive weights. By calculating the signal-to-noise ratio and mutual correlation coefficient of the sensor data, a sensor data quality assessment matrix is ​​constructed, and the adaptive weight of the sensor data is calculated based on it. Then, the adaptive weight is used to perform weighted fusion on multiple sensor data to obtain ink quality feature data. The algorithm can adaptively adjust the importance of each sensor data, improve the accuracy and reliability of data fusion, and provide more comprehensive and accurate data support for online monitoring of ink quality.

[0095] S102. Using a support vector machine algorithm to establish an ink quality prediction model, by training a large amount of historical ink production data and corresponding ink quality test results, the ink quality prediction model is obtained, and the fused ink quality feature data is input into the trained ink quality prediction model to obtain an ink quality prediction result;

[0096] In an optional embodiment,

[0097] The ink quality prediction model is established by using the support vector machine algorithm. The ink quality prediction model is obtained by training a large amount of historical ink production data and the corresponding ink quality test results. The fused ink quality feature data is input into the trained ink quality prediction model, and the ink quality prediction results obtained include:

[0098] Collecting historical production data and corresponding ink quality test results during the ink production process, preprocessing the historical production data and the ink quality test results to obtain a training sample set;

[0099] The training sample set is trained by using a support vector machine algorithm, and an ink quality prediction model is obtained by constructing an optimization objective function and solving the optimization problem;

[0100] Acquire real-time data collected by multiple sensors during the ink production process, perform filtering and feature fusion on the real-time data, and obtain fused ink quality feature data;

[0101] The fused ink quality feature data is input into the ink quality prediction model, and a real-time ink quality prediction result is obtained through model prediction.

[0102] In an optional embodiment,

[0103] Constructing the optimization objective function and solving the optimization problem include:

[0104] min(1 / 2*||w|| 2 +C*sum(ξ i ));

[0105] sty i *(w·K(x i )+b)≥1-ξ i ,ξ i ≥0;

[0106] Among them, w is the normal vector of the hyperplane, b is the bias term of the hyperplane, C is the penalty coefficient, ξ i It is a slack variable, which is used to allow some sample classification errors;

[0107] Through the Lagrange multiplier method and duality technique, the above optimization problem is transformed into the following dual problem:

[0108] max(sum(α i )-(1 / 2)*sum(α i *y i *K(x i )));

[0109] sum(α i *y i )=0,0≤α i ≤C;

[0110] Among them, α i is the Lagrange multiplier, X(x i ) is the kernel function, x i represents the i-th sample, y i Represents the ink quality prediction result corresponding to the i-th sample.

[0111] For example, in order to establish an ink quality prediction model, it is necessary to first collect a large amount of historical ink production data and corresponding ink quality test results. These data include real-time data collected by various sensors during the ink production process, such as temperature, humidity, pH value, viscosity and flow rate, as well as the quality test results of the corresponding batch of inks, such as color, density, fineness, rheology and other indicators.

[0112] The collected historical data is preprocessed, including data cleaning, data normalization and feature selection. Data cleaning refers to removing noise, outliers and missing values ​​in the data to ensure the integrity and accuracy of the data. Data normalization refers to unifying the data collected by different sensors to the same scale and range, such as normalizing the data to the interval [0,1] to eliminate the dimensional differences between different data. Feature selection refers to selecting key features closely related to ink quality from a large number of sensor data, reducing data dimensions, and improving model training efficiency and prediction accuracy.

[0113] The ink quality prediction model was constructed using the Support Vector Machine (SVM) algorithm. SVM is a machine learning algorithm based on statistical learning theory. It divides samples of different categories by finding the optimal separation hyperplane and has good generalization ability and classification performance.

[0114] The preprocessed historical ink production data and quality inspection results are divided into a training set and a test set. The training set is used to train the SVM model, and the test set is used to evaluate the prediction performance of the model. The training process of the SVM model is as follows:

[0115] SVM supports a variety of kernel functions, such as linear kernel, polynomial kernel and Gaussian kernel, etc. According to the characteristics and distribution of ink production data, an appropriate kernel function is selected to map samples from the original space to a high-dimensional feature space, so that samples of different categories can be linearly separated in the high-dimensional space.

[0116] The preprocessed historical ink production data and quality inspection results are divided into a training set and a test set. The training set is used to train the SVM model, and the test set is used to evaluate the prediction performance of the model. The training process of the SVM model is as follows:

[0117] SVM supports a variety of kernel functions, such as linear kernel, polynomial kernel and Gaussian kernel, etc. According to the characteristics and distribution of ink production data, an appropriate kernel function is selected to map samples from the original space to a high-dimensional feature space, so that samples of different categories can be linearly separated in the high-dimensional space.

[0118] The goal of SVM is to find an optimal separation hyperplane in the feature space so that samples of different categories can be correctly divided by the hyperplane and the distance between the hyperplane and the nearest sample point is maximized. The following optimization objective function is constructed:

[0119] min(1 / 2*||w|| 2 +C*sum(ξ i ));

[0120] sty i *(w·K(x i )+b)≥1-ξ i ,ξ i ≥0;

[0121] Among them, w is the normal vector of the hyperplane, b is the bias term of the hyperplane, C is the penalty coefficient, ξ i is a slack variable used to allow for a small number of sample classification errors.

[0122] Through the Lagrange multiplier method and duality technique, the above optimization problem is transformed into the following dual problem:

[0123] max(sum(α i )-(1 / 2)*sum(α i *y i *K(x i )));

[0124] sum(α i *y i )=0,0≤α i ≤C;

[0125] Among them, α i is the Lagrange multiplier, K(x i ) is the kernel function, xi represents the i-th sample, y i represents the ink quality prediction result corresponding to the i-th sample. By solving the above dual problem, the optimal α is obtained. i , and then obtain the parameters w and b of the hyperplane.

[0126] The trained SVM model is used to predict the quality of new ink production data. For a new sample x, its predicted label y can be obtained through the decision function

[0127] After filtering, normalizing and feature fusion of the multi-sensor real-time data collected online, the fused ink quality feature data is obtained. The feature data is input into the trained SVM ink quality prediction model to obtain the real-time ink quality prediction result.

[0128] S103. If the ink quality prediction result meets the preset conditions, the ink quality is determined to be qualified; if the ink quality prediction result does not meet the preset conditions, the ink quality is determined to be unqualified, and a feedback signal in a preset format is sent to the ink production equipment. The ink production process parameters are adjusted through the PID controller of the ink production equipment, and the adjusted ink production process parameters are fed back to the data processing unit to perform the next ink quality prediction until the ink quality prediction result meets the preset conditions, thereby completing the closed-loop control of online monitoring and dynamic regulation of ink quality.

[0129] In an optional embodiment,

[0130] Adjusting the ink production process parameters by the PID controller of the ink production equipment and feeding back the adjusted ink production process parameters to the data processing unit includes:

[0131] Establishing a mathematical model between ink viscosity, ink temperature and ink composition ratio, wherein the mathematical model represents the quantitative relationship between ink viscosity, ink temperature and ink composition ratio;

[0132] According to the mathematical model, an ink temperature control loop and an ink component ratio control loop are designed. The ink temperature control loop measures the ink temperature through a temperature sensor, and controls a heating / cooling device to adjust the ink temperature according to a deviation between a temperature setting value and an actual value by a temperature PID controller. The ink component ratio control loop measures the ink component ratio in real time through an online measuring device, and controls the addition amount of the ink component according to a deviation between a component ratio setting value and an actual value by a component ratio PID controller.

[0133] According to the mathematical model, the deviation between the current ink viscosity value and the target viscosity value is calculated, the viscosity deviation is substituted into the mathematical model, the adjustment amount of the ink temperature and the ink component ratio is calculated, and the temperature adjustment amount and the component ratio adjustment amount are transmitted to the temperature PID controller and the component ratio PID controller respectively;

[0134] The temperature PID controller and the component ratio PID controller coordinately control the heating / cooling device and the ink component adding device according to their respective adjustment amounts, and synchronously adjust the ink temperature and the ink component ratio to stabilize the ink viscosity within the target range.

[0135] For example,

[0136] In the ink production process, the adjustment of a single process parameter is often difficult to meet the requirements of ink quality control, and it is necessary to achieve coordinated control of multiple key parameters. Taking ink viscosity control as an example, ink viscosity is a key indicator that affects the printing performance and quality of ink, and ink viscosity is affected by multiple factors such as ink temperature and ink composition ratio. Therefore, in order to achieve precise control of ink viscosity, it is necessary to synchronize the adjustment of ink temperature and composition ratio through the coordinated action of multiple PID controllers. The specific technical implementation steps are as follows:

[0137] Through experimental research and data analysis, a mathematical model between ink viscosity, ink temperature and ingredient ratio is established. The quantitative relationship between ink viscosity and factors such as temperature and ingredient ratio is fitted by multiple linear regression or nonlinear regression method, and the following model is obtained:

[0138] η=f(T,M_1,M_2,...,M_n);

[0139] Among them, η is the ink viscosity, T is the ink temperature, and M_1, M_2,..., M_n are the ink component ratios.

[0140] According to the mathematical model of ink viscosity, temperature and composition ratio, a multi-parameter collaborative control system structure is designed. The system includes an ink temperature control loop and an ink composition ratio control loop, each of which consists of a PID controller, an actuator and a sensor.

[0141] The ink temperature control loop measures the ink temperature through a temperature sensor, and the PID controller controls the heating / cooling device to adjust the ink temperature according to the deviation between the temperature setting value and the actual value. The ink composition ratio control loop measures the ink composition ratio in real time through equipment such as an online chromatograph or a near-infrared spectrometer, and the PID controller controls the amount of colorant, solvent and other components added according to the deviation between the composition ratio setting value and the actual value, and adjusts the ink composition ratio.

[0142] Based on the multi-parameter collaborative control system, a collaborative control algorithm is designed to achieve synchronous adjustment of ink temperature and component ratio. The core of the collaborative control algorithm is to calculate the adjustment amount of temperature and component ratio according to the ink viscosity model, and coordinate the control behavior of multiple PID controllers. The specific steps are as follows:

[0143] According to the ink viscosity model, the deviation Δη between the current ink viscosity value and the target viscosity value is calculated.

[0144] Substitute the viscosity deviation Δη into the ink viscosity model to calculate the temperature and component ratio adjustment ΔT and ΔM_1, ΔM_2, ..., ΔM_n. The temperature adjustment ΔT and component ratio adjustment ΔM_1, ΔM_2, ..., ΔM_n are respectively passed to the corresponding PID controller as the controller input.

[0145] Each PID controller calculates the controller output according to the adjustment amount and drives the actuator to adjust the ink temperature and component ratio until the ink viscosity is stabilized within the target range.

[0146] Through the above steps, the synergy of multiple PID controllers is used to achieve the synchronous adjustment of ink temperature and component ratio, effectively control ink viscosity, and ensure the stability and consistency of ink quality. At the same time, by optimizing the collaborative control strategy, the accuracy, efficiency and intelligence level of ink viscosity control are improved, providing important technical support for the production of high-quality ink products.

[0147] Figure 2 FIG. 1 is a schematic diagram of the structure of an online ink quality monitoring system based on multi-sensor fusion according to an embodiment of the present invention. Figure 2 As shown, the system comprises:

[0148] The first unit is used to obtain real-time data collected by multiple sensors during the ink production process, and transmit the real-time data to a data processing unit through industrial Ethernet; the data processing unit uses a Kalman filter algorithm to filter the received real-time data, remove noise and abnormal values ​​in the real-time data, obtain filtered real-time data, use a multi-sensor data fusion algorithm with adaptive weights to calculate the sensor weights corresponding to the real-time data, and combine the real-time data to obtain fused ink quality characteristic data;

[0149] The second unit is used to establish an ink quality prediction model by using a support vector machine algorithm, obtain the ink quality prediction model by training a large amount of historical ink production data and corresponding ink quality test results, and input the fused ink quality feature data into the trained ink quality prediction model to obtain the ink quality prediction result;

[0150] The third unit is used to determine that the ink quality is qualified if the ink quality prediction result meets the preset conditions; if the ink quality prediction result does not meet the preset conditions, determine that the ink quality is unqualified, send a feedback signal in a preset format to the ink production equipment, adjust the ink production process parameters through the PID controller of the ink production equipment, and feed back the adjusted ink production process parameters to the data processing unit to perform the next ink quality prediction until the ink quality prediction result meets the preset conditions, thereby completing the closed-loop control of online monitoring and dynamic regulation of ink quality.

[0151] According to a third aspect of the embodiments of the present invention,

[0152] An electronic device is provided, comprising:

[0153] processor;

[0154] a memory for storing processor-executable instructions;

[0155] The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0156] According to a fourth aspect of the embodiments of the present invention,

[0157] A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0158] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0159] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. The ink quality online monitoring method based on multi-sensor fusion is characterized by: include: Acquire sensor data collected by multiple sensors during the production process of water-based ink, the sensor data including temperature, humidity, pH value, viscosity and flow rate data, and transmit the sensor data to a data processing unit via industrial Ethernet; the data processing unit uses a Kalman filter algorithm to filter the received sensor data, remove unqualified values ​​and abnormal values ​​in the sensor data, obtain filtered real-time data, use a multi-sensor data fusion algorithm with adaptive weights to calculate the sensor weights corresponding to the real-time data, and combine the real-time data to obtain fused water-based ink quality characteristic data; A water-based ink quality prediction model is established by using a support vector machine algorithm, and a water-based ink quality prediction model is obtained by training a large amount of historical ink production data and corresponding ink quality test results. The fused water-based ink quality feature data is input into the trained water-based ink quality prediction model to obtain a water-based ink quality prediction result; If the water-based ink quality prediction result meets the preset conditions, the water-based ink quality is determined to be qualified; if the water-based ink quality prediction result does not meet the preset conditions, the water-based ink quality is determined to be unqualified, and a feedback signal in a preset format is sent to the water-based ink production equipment, and the water-based ink production process parameters are adjusted through the PID controller of the water-based ink production equipment, and the adjusted water-based ink production process parameters are fed back to the data processing unit to perform the next water-based ink quality prediction until the water-based ink quality prediction result meets the preset conditions, thereby completing the closed-loop control of online monitoring and dynamic regulation of water-based ink quality.

2. The method according to claim 1, characterized in that The multi-sensor data fusion algorithm with adaptive weights is used to calculate the sensor weights corresponding to the real-time data, and the fused ink quality characteristic data obtained by combining the real-time data includes: Calculate the signal-to-noise ratio of the real-time data collected by each sensor to obtain signal-to-noise ratio data representing the quality of the data of each sensor; Calculate the correlation coefficients between real-time data collected by different sensors; Based on the signal-to-noise ratio data and the mutual correlation coefficient, a sensor data quality assessment matrix is ​​constructed, wherein the diagonal elements of the sensor data quality assessment matrix are the signal-to-noise ratios of each sensor data, and the non-diagonal elements are the mutual correlation coefficients between different sensor data; Utilizing the sensor data quality assessment matrix, calculating the adaptive weight of each sensor data to obtain a normalized sensor data adaptive weight vector; The sensor data adaptive weight vector is weightedly fused with real-time data collected by multiple sensors to obtain fused ink quality characteristic data.

3. The method according to claim 1, characterized in that The ink quality prediction model is established by using the support vector machine algorithm. The ink quality prediction model is obtained by training a large amount of historical ink production data and the corresponding ink quality test results. The fused ink quality feature data is input into the trained ink quality prediction model, and the ink quality prediction results obtained include: Collecting historical production data and corresponding ink quality test results during the ink production process, preprocessing the historical production data and the ink quality test results to obtain a training sample set; The training sample set is trained by using a support vector machine algorithm, and an ink quality prediction model is obtained by constructing an optimization objective function and solving the optimization problem; Acquire real-time data collected by multiple sensors during the ink production process, perform filtering and feature fusion on the real-time data, and obtain fused ink quality feature data; The fused ink quality feature data is input into the ink quality prediction model, and a real-time ink quality prediction result is obtained through model prediction.

4. The method according to claim 3, characterized in that Constructing the optimization objective function and solving the optimization problem include: min(1 / 2*||w|| 2 +C*sum(ξ i )); s.t.y i *(w·K(x i )+b)≥1-ξ i ,ξ i ≥0; Among them, w is the normal vector of the hyperplane, b is the bias term of the hyperplane, C is the penalty coefficient, ξ i It is a slack variable, which is used to allow some sample classification errors; Through the Lagrange multiplier method and duality technique, the above optimization problem is transformed into the following dual problem: max(sum(a i )-(1 / 2)*sum(a i *y i *K(x i ))); sum(a i *y i )=0.0≤α i ≤C; Among them, α i is the Lagrange multiplier, K(x i ) is the kernel function, x i represents the i-th sample, y i Represents the ink quality prediction result corresponding to the i-th sample.

5. The method according to claim 1, characterized in that Adjusting the ink production process parameters by the PID controller of the ink production equipment and feeding back the adjusted ink production process parameters to the data processing unit includes: Establishing a mathematical model between ink viscosity, ink temperature and ink composition ratio, wherein the mathematical model represents the quantitative relationship between ink viscosity, ink temperature and ink composition ratio; According to the mathematical model, an ink temperature control loop and an ink composition ratio control loop are designed; The ink temperature control loop measures the ink temperature through a temperature sensor, and controls the heating / cooling device to adjust the ink temperature according to the deviation between the temperature setting value and the actual value through a temperature PID controller; The ink composition ratio control loop measures the ink composition ratio in real time through an online measuring device, and controls the addition amount of the ink components according to the deviation between the composition ratio set value and the actual value by the composition ratio PID controller.

6. The method according to claim 1, characterized in that The method further comprises: According to the mathematical model, the deviation between the current ink viscosity value and the target viscosity value is calculated, the viscosity deviation is substituted into the mathematical model, the adjustment amount of the ink temperature and the ink component ratio is calculated, and the temperature adjustment amount and the component ratio adjustment amount are transmitted to the temperature PID controller and the component ratio PID controller respectively; The temperature PID controller and the component ratio PID controller coordinately control the heating / cooling device and the ink component adding device according to their respective adjustment amounts, and synchronously adjust the ink temperature and the ink component ratio to stabilize the ink viscosity within the target range.

7. An ink quality online monitoring system based on multi-sensor fusion, used to implement the method described in any one of claims 1 to 6, characterized in that: include: The first unit is used to obtain real-time data collected by multiple sensors during the ink production process, and transmit the real-time data to the data processing unit via industrial Ethernet; The data processing unit uses a Kalman filter algorithm to filter the received real-time data, removes noise and abnormal values ​​in the real-time data, obtains filtered real-time data, uses an adaptive weight multi-sensor data fusion algorithm to calculate the sensor weight corresponding to the real-time data, and combines the real-time data to obtain fused ink quality characteristic data; The second unit is used to establish an ink quality prediction model by using a support vector machine algorithm, obtain the ink quality prediction model by training a large amount of historical ink production data and corresponding ink quality test results, and input the fused ink quality feature data into the trained ink quality prediction model to obtain the ink quality prediction result; The third unit is used to determine that the ink quality is qualified if the ink quality prediction result meets the preset conditions; if the ink quality prediction result does not meet the preset conditions, determine that the ink quality is unqualified, send a feedback signal in a preset format to the ink production equipment, adjust the ink production process parameters through the PID controller of the ink production equipment, and feed back the adjusted ink production process parameters to the data processing unit to perform the next ink quality prediction until the ink quality prediction result meets the preset conditions, thereby completing the closed-loop control of online monitoring and dynamic regulation of ink quality.

8. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images