Paper air permeability measurement method and paper air permeability measurement system

Abstract

A method for measuring the air permeability of paper and a system for measuring the air permeability of paper, which can measure the air permeability of paper in real time, continuously and effectively during the paper manufacturing process, and timely feedback the information to production, so that the production can be adjusted in time based on the feedback information. The method for measuring the air permeability of paper includes: obtaining the current data of the air permeability related features; inputting the current data as input data into the soft measurement model of the air permeability of paper, and the soft measurement model of the air permeability of paper outputs the predicted value of the air permeability of paper based on the input current data; according to the comparison result between the predicted value of the air permeability and the preset target value of the air permeability, selectively adjust the current data of the air permeability related features so that the predicted value of the air permeability approaches or reaches the target value of the air permeability.

Description

Technical Field

[0001] The present invention relates to a method for measuring the air permeability of paper and a system for measuring the air permeability of paper. Background Art

[0002] It is known that the concept of air permeability is widely used in the paper-making field to explain the inherent characteristics of products and is one of the important characteristics of paper during production and application. Air permeability provides a measure of fiber porosity and is closely related to other factors such as product thickness and coating.

[0003] Currently, as methods for measuring paper, there are two methods: online measurement and offline measurement. The commonly used offline measurement requires sampling and testing one or more batches of paper that have been completed after production, and the test results cannot be timely fed back to production. The existing online measurement device is a contact-type measurement device. It has an open negative-pressure cavity. During testing, the open side of the cavity is attached to the paper surface, and the paper surface is firmly adsorbed by the negative pressure in the cavity. Then, by detecting the air flow rate in the cavity under a constant negative pressure, the air permeability value of the paper is calculated. Due to the contact and adsorption of the device on the paper surface, two negative impacts are brought. First, the stability of the device's adsorption on the paper surface is closely related to the measurement stability. The frictional force on the contact surface will cause scratches and paper powder accumulation on the paper surface, and even cause tearing of the paper web. Therefore, it is necessary to balance the adsorption force and measurement stability, and at the same time, it is necessary to clean the accumulated paper powder in time to avoid the influence of the device's adsorption force on the paper surface, thereby affecting the measurement result. Second, in order to avoid the scratches on the paper surface caused by the frictional force from affecting the finished product quality, the online detection device for air permeability needs to be installed on the front side of the coater, but this results in a difference between its measurement result and the true result of the formed paper, and this difference needs to be evaluated by comparing with the laboratory sampling detection value. To sum up, even with an online detection device for air permeability, its measurement result is difficult to be recognized. Summary of the Invention

[0004] Technical Problems to be Solved by the Invention

[0005] In addition, the existing online instruments for measuring the air permeability of paper generally have problems such as high price, difficult installation, affecting product quality, and poor measurement stability. On the other hand, in offline measurement, as described above, usually one or more batches of paper that have been completed are sampled and tested. Therefore, it has hysteresis and can only obtain partial and limited measurement results, and cannot comprehensively and accurately evaluate the air permeability of the entire batch of paper. As a result, the fluctuations in the production process cannot be timely fed back and guided for adjustment, often causing impacts on the finished product quality and energy efficiency.

[0006] The present invention is formed to solve the above technical problems, and its object is to provide a method for measuring the air permeability of paper and a system for measuring the air permeability of paper, which can measure the air permeability of paper in real time, continuously and effectively before or during the production of paper, and timely feedback the information to production, so that the production (such as papermaking parameters) can be adjusted in time based on the feedback information, thereby improving the qualified rate.

[0007] Technical solution for solving technical problems

[0008] The first technical solution of the present invention provides a method for measuring the air permeability of paper, including:

[0009] Obtain the current data of the air permeability-related feature, where the air permeability-related feature is a papermaking parameter that is related to the air permeability of paper;

[0010] Input the current data as input data into the soft measurement model of paper air permeability. The soft measurement model of paper air permeability outputs a predicted value of the air permeability of paper based on the input current data. The soft measurement model of paper air permeability is a learning-completed model obtained by associating the historical data of the air permeability-related feature with the corresponding actual value of the air permeability and performing learning; and

[0011] According to the comparison result between the predicted value of the air permeability and a preset target value of the air permeability, selectively adjust the current data of the air permeability-related feature so that the predicted value of the air permeability approaches or reaches the target value of the air permeability.

[0012] According to the method for measuring the air permeability of paper described in the first technical solution, by using a learning-completed model (soft measurement model of paper air permeability) constructed by a machine learning algorithm, the air permeability of paper is measured (predicted) in real time and continuously before or during the production of paper. In this way, compared with the existing off-line measurement method, the lag problem existing in off-line measurement can be overcome, and the air permeability of the entire batch of paper can be comprehensively predicted during the production process of a certain batch or multiple batches of paper. It also overcomes the problem that the air permeability of the entire production batch of paper cannot be guaranteed due to sampling inspection, ensuring the qualified rate of paper. On the other hand, compared with the existing on-line measurement method, there is no need to purchase special on-line instruments for measuring the air permeability of paper, saving a large amount of papermaking costs to a great extent. Moreover, according to the method for measuring the air permeability of paper described in the first technical solution, since the air permeability (predicted value) of paper can be fed back in real time, the current data of relevant papermaking parameters can be adjusted in real time according to the difference degree between it and the target value of the air permeability, so that the air permeability of paper can reach the preset target air permeability.

[0013] Based on the method for measuring the air permeability of paper described in the first technical solution, in the method for measuring the air permeability of paper in the second technical solution, the current data of the air permeability correlation feature is determined by the following method:

[0014] During the manufacturing process of the paper, obtain a part or all of the actual data of the papermaking parameters and the corresponding actual air permeability values;

[0015] Perform data cleaning on a part or all of the actual data of the obtained papermaking parameters to obtain the cleaned actual data; and

[0016] Perform a correlation analysis on the cleaned actual data and the actual air permeability values to determine the current data of the air permeability correlation feature.

[0017] According to the method for measuring the air permeability of paper described in the second technical solution, it is possible to effectively obtain the data of the papermaking parameters directly related to the air permeability of the paper. Thus, it is possible to reasonably reduce the amount of data to the greatest extent, eliminate useless data, invalid data, and data that may cause calculation errors or affect the calculation accuracy, improve the calculation efficiency, and ensure the prediction accuracy of the soft measurement model.

[0018] Based on the method for measuring the air permeability of paper described in the first technical solution, in the method for measuring the air permeability of paper in the third technical solution, the air permeability correlation feature includes at least a part of the caliper pressure of the calendar, basis weight, steam consumption of the front dryer, vacuum pressure of the forming roll, surface starch addition amount, and the first caliper pressure of the press section.

[0019] According to the method for measuring the air permeability of paper described in the third technical solution, by selecting the papermaking parameters with the highest correlation with the air permeability of the paper, the model quality of the soft measurement model of the air permeability of the paper can be improved, and the prediction accuracy can be improved.

[0020] Based on the method for measuring the air permeability of paper described in the first technical solution, in the method for measuring the air permeability of paper in the fourth technical solution, the soft measurement model of the air permeability of the paper is a learning-complete model formed based on the random forest regression algorithm.

[0021] According to the paper air permeability measurement method described in the fourth technical solution, compared with other machine learning algorithms, in terms of the generalization ability of the model, it is most ideal to construct a model using the random forest regression algorithm. Specifically, the papermaking process belongs to a large-scale process technology, and there are a very large number of types of parameter variables involved. Therefore, when constructing a model based on machine learning algorithms, the sample dimension in its dataset often tends to be relatively high. The random forest regression algorithm is a learning algorithm that contains multiple decision tree algorithms (specifically an ensemble learning algorithm), and it has the advantage of being able to handle high-dimensional data, which is very suitable for processing the high-dimensional data of the papermaking process. That is to say, by using the random forest regression algorithm, a soft measurement model of paper air permeability with high accuracy and generalization ability can be constructed at a relatively low computational cost.

[0022] Based on the paper air permeability measurement method described in the fourth technical solution, in the paper air permeability measurement method of the fifth technical solution, a part of the historical data of the air permeability correlation features and the corresponding actual air permeability values are used as the training dataset, and the remaining part of the historical data of the air permeability correlation features and the corresponding actual air permeability values are used as the test dataset, and the soft measurement model of paper air permeability is formed by using the holdout method or the cross-validation method.

[0023] According to the paper air permeability measurement method described in the fifth technical solution, the generalization error of the model can be further reduced, and the occurrence of overfitting or underfitting can be prevented.

[0024] Based on the paper air permeability measurement method described in the first technical solution, in the paper air permeability measurement method of the sixth technical solution, the soft measurement model of paper air permeability is updated based on the current data of the air permeability correlation features and the corresponding actual air permeability values.

[0025] Based on the paper air permeability measurement method described in the sixth technical solution, in the paper air permeability measurement method of the seventh technical solution, when the difference between the air permeability predicted value and the actual air permeability value exceeds the threshold, the soft measurement model of paper air permeability is updated.

[0026] According to the paper air permeability measurement methods described in the sixth and seventh technical solutions, the model accuracy of the soft measurement model of paper air permeability can be continuously improved.

[0027] The present invention also provides a paper air permeability measurement system, including:

[0028] A data acquisition module, which acquires the current data of the air permeability correlation features, and the air permeability correlation features are papermaking parameters that are related to the air permeability of the paper;

[0029] An air permeability prediction module, which inputs the current data as input data into a soft measurement model for paper air permeability. The soft measurement model for paper air permeability outputs a predicted value of the air permeability of the paper based on the input current data. The soft measurement model for paper air permeability is a learned model obtained by associating historical data of the air permeability-related features with the corresponding actual values of the air permeability and performing learning; and

[0030] A data adjustment module, which selectively adjusts the current data of the air permeability-related features according to the comparison result between the predicted value of the air permeability and a preset target value of the air permeability, so that the predicted value of the air permeability approaches or reaches the target value of the air permeability.

[0031] According to the above paper air permeability measurement system, the same technical effects as those of the paper air permeability measurement method described in the first technical solution can be obtained.

[0032] The present invention also provides a computer system for measuring the air permeability of paper, including a memory, a processor, and a computer program stored on the memory. Wherein, the processor executes the computer program to implement the paper air permeability measurement method described in any one of the first technical solution to the seventh technical solution.

[0033] The present invention also provides a computer-readable storage medium, which stores a computer program or instruction. Wherein, when the computer program or instruction is executed by a processor, it implements the paper air permeability measurement method described in any one of the first technical solution to the seventh technical solution.

[0034] The present invention also provides a computer program product, including a computer program or instruction. Wherein, when the computer program or instruction is executed by a processor, it implements the paper air permeability measurement method described in any one of the first technical solution to the seventh technical solution.

[0035] Advantages of the Invention

[0036] According to the above paper air permeability measurement method and paper air permeability measurement system, the air permeability of paper can be measured in real time, continuously and effectively before or during the production of paper, and the information can be fed back to the production in a timely manner. Thus, the production (such as papermaking parameters) can be adjusted in a timely manner based on the feedback information, thereby improving the qualified rate. Brief Description of the Drawings

[0037] Figure 1 It is a flowchart showing the paper air permeability measurement method according to an embodiment of the present invention.

[0038] Figure 2 It shows Figure 1Flowchart of a modified example of the paper air permeability measurement method shown

[0039] Figure 3 represents based on Figure 1 and Figure 2 Data processing and data flow diagram of the paper air permeability measurement method shown

[0040] Figure 4 Functional block diagram of a paper air permeability measurement system according to an embodiment of the present invention

[0041] Figure 5 represents Figure 4 Functional block diagram of a modified example of the paper air permeability measurement system shown

[0042] Figure 6 represents Figure 4 Functional block diagram of another modified example of the paper air permeability measurement system shown

[0043] Figure 7 represents Figure 4 Functional block diagram of yet another modified example of the paper air permeability measurement system shown

[0044] Figure 8 Structural schematic diagram of a computer system for measuring the air permeability of paper according to an embodiment of the present invention

[0045] Figure 9 Chart showing the air permeability correlation of each air permeability-related characteristic obtained under the parameter settings in Table 1

[0046] Figure 10 Chart showing the comparison result between the prediction result of the air permeability soft measurement model established based on Table 1 and the actual value detected in the laboratory Detailed implementation mode

[0047] (Paper air permeability measurement method)

[0048] First, with reference to Figure 1 , the process (steps) of the paper air permeability measurement method according to an embodiment of the present invention will be described

[0049] Figure 1 The flowchart of the paper air permeability measurement method according to an embodiment of the present invention is shown. As Figure 1 shown, when the manufacturing of a certain batch of paper starts, it first enters step ST1

[0050] In step ST1, the current data of the air permeability related features is obtained. The "air permeability related features" mentioned here refer to the papermaking parameters that are related to the air permeability of the paper to be produced or being produced. Among them, the so-called "being related" means that the specific data settings will affect the meaning of the air permeability value of the paper to a certain extent. As the air permeability related features, for example, they include the grammage of the paper, the steam consumption of the front drying cylinder, the vacuum pressure of the forming roll, the surface starch addition amount, the first wire pressure in the press section, the salt addition amount, the second wire pressure in the press section, the filling addition ratio, the steam box pressure, the steam consumption of the rear drying cylinder, the sizing flow rate, the beating degree, the power of the refiner, the machine speed (e.g., the paper machine speed), the pulp ratio (e.g., the long fiber addition ratio, the retention aid addition amount), the sizing agent addition amount, the pulp-net speed ratio, the long fiber freeness, the freeness of short fiber No. 1, the freeness of short fiber No. 2, the vacuum degree of the wire section, the starch application amount, the steam consumption of the drying cylinder, the pressure in the press section, the nip pressure of the calender, etc. In this embodiment, the air permeability related features can be artificially selected by papermaking staff according to the actual situation or papermaking experience, or can be screened out by scientific calculation methods. Regarding the screening method based on scientific calculation, it will be described in detail in the following modification examples. After obtaining the current data of the above air permeability related features, step ST2 is entered.

[0051] In step ST2, the current data of the air permeability related features obtained in step ST1 is integrated to form a data set. Then, the formed data set is used as input data and input into the constructed soft sensor model for paper air permeability. This soft sensor model for paper air permeability outputs the predicted value of the air permeability of the paper corresponding to the data set based on the input data set. It should be noted that the above soft sensor model for paper air permeability is a learning completed model formed based on machine learning algorithms. More specifically, the above soft sensor model for paper air permeability is a learning completed model formed by using the historical data of the air permeability related features obtained in the previous paper manufacturing process and the historical values of the air permeability of the manufactured paper as sample data and using specific machine learning algorithms for learning (i.e., performing supervised learning).

[0052] In particular, regarding the above-mentioned specific machine learning algorithm, it is preferably to adopt the random forest regression algorithm. This is because the papermaking process belongs to a large-scale process, and there are a wide variety of parameters (process variables) involved. If one wants to use a machine learning algorithm to train such data and form a completed learning model, one has to face the problem of how to quickly, efficiently, and accurately process high-dimensional samples. Therefore, compared with using other machine learning algorithms, such as the boosting algorithm (a serial generation serialization method) which is also an ensemble learning algorithm, the random forest regression algorithm, as a parallel algorithm, can process samples in parallel (simultaneously process). There is no strong dependence between the base learners (individual learners), and it will not cause the calculation accuracy of the overall algorithm to be abnormal due to the abnormal calculation of a certain base learner resulting in the abnormal calculation accuracy of other base learners. In addition, if the random forest regression algorithm is adopted, the influence between each feature (attribute or also called feature quantity) can be detected during the model training process, and the importance of each feature for the calculation result (air permeability) can be given after the model training, providing a reliable basis for real-time adjustment of the parameter values. In addition, the random forest regression algorithm also has advantages such as strong anti-overfitting ability and the generalization error decreases as the number of base learners (individual decision trees) increases. Regarding the calculation method of the predicted value of air permeability, for example, the prediction of the random forest regression algorithm model for unknown samples can be achieved by taking the average of the predictions of the samples on all single decision trees. The formula is as follows:

[0053]

[0054] Among them, B represents the number of decision trees. According to the size and nature of the data set, usually several hundred to several thousand decision trees are adopted. By using the cross-validation method or observing the out-of-bag error, the most suitable value of B can be found.

[0055] In addition, regarding the training strategy (evaluation method), it is preferably to adopt the hold-out method or the cross-validation method (also called k-fold cross-validation). The so-called "hold-out method" means directly dividing the data set D into two mutually exclusive sets S and T, where one set S is used as the training set and the other set T is used as the test set, that is, D = S ∪ T, After training the model on S, use T to evaluate its test error as an estimate of the generalization error. On the other hand, the so-called "cross-validation method" means first dividing the data set D into k mutually exclusive subsets with similar sizes, that is, D = D1 ∪ D2 ∪... ∪ D k , (i ≠ j). Each subset D iAll try to maintain the consistency of the data distribution as much as possible, that is, obtained by stratified sampling from D. Then, each time the union of k - 1 subsets is used as the training set, and the remaining subset is used as the test set. In this way, k groups of training / test sets are obtained, so that k times of training and testing can be carried out, and finally the mean value of these k test results is returned. Of course, as a training strategy, it is not limited to the above two methods, and other methods such as bootstrapping can also be used.

[0056] The data set mentioned here contains a data set of multiple examples composed of historical data related to air permeability characteristics and the corresponding actual air permeability values. According to the academic naming rules of machine learning theory, in each example, the historical data of air permeability related characteristics is called a sample, and the corresponding actual air permeability value is called a label.

[0057] For example, during training, 70% of the examples are randomly selected from the data set as the training set to train the model, and the remaining 30% of the examples are used as the test set to verify the accuracy of the model trained with the training set data. At the same time, cross - validation is used when training the model to compare the accuracy of the model for different parameter values such as depth and the number of decision trees. Generally speaking, the greater the depth, the better the fitting effect, but at the same time, it also brings an increase in computational complexity and overfitting phenomena. In the case where the air permeability related characteristics are not easily affected by each other, appropriately reducing the depth has no impact on the results. On the other hand, regarding the number of decision trees, usually, as the number of base classifiers increases, the random forest usually converges to a lower generalization error. Based on the parameters shown in Table 1 below, an accuracy of about 98.5% can be obtained. It should be noted that the "accuracy" here refers to the R value (i.e., the correlation coefficient).

[0058]

Table 1

[0059]

[0060]

[0061] After the current data of the above - mentioned air permeability related characteristics is input into the soft - sensing model of paper air permeability as input data in step ST2 to estimate the predicted value of the paper's air permeability, step ST3 is entered.

[0062] Here, preferably, the predicted air permeability value is compared with the actual air permeability value subsequently detected in the laboratory to verify the effectiveness of the model for new data. For example, the soft air permeability measurement model established with the parameters shown in Table 1 above is applied in practice. During deployment, through OPC communication, the air permeability correlation features are transmitted to the soft sensor model in real time for prediction, and the obtained results are compared with the actual values detected in the laboratory. After a period of production data collection and verification, the actual application accuracy (accuracy of the R value) of the model reaches 92.14%, and R 2 reaches 84.9%, and the accuracy of the model is good. Figure 10 Shows the comparison results between the predicted values of the soft air permeability measurement model and the actual values detected in the laboratory.

[0063] In step ST3, the predicted air permeability value of the paper deduced above is compared with the target air permeability value of the paper (for the current batch) preset in advance. If the difference between the predicted air permeability value and the target air permeability value is less than the specified allowable threshold, the current papermaking parameters are not adjusted, and the operating state of the paper machine remains unchanged, and it can return to step ST1. If the difference between the two is greater than the specified allowable threshold, it enters step ST4.

[0064] In step ST4, the papermaking parameters, especially the current data of the above-mentioned periodicity correlation features, are adjusted so that the predicted air permeability value deduced subsequently approaches or reaches the above-mentioned target air permeability value. Here, the allowable threshold refers to the maximum allowable error between the predicted air permeability value and the target air permeability value.

[0065] According to the above embodiment, after establishing a soft air permeability measurement model using a machine learning algorithm, the current data of the air permeability correlation features related to the air permeability of the paper are used as input data and input into the soft air permeability measurement model, so as to output the predicted air permeability value of the paper in the current manufacturing process, and use this predicted air permeability value to guide the actual production in real time. Thus, compared with the existing offline detection method and the online detection method based on expensive online instruments, it can monitor the production process of the paper in real time and comprehensively, and ensure that the air permeability of the paper meets the expected standard.

[0066] In addition, in this embodiment, preferably, the soft air permeability measurement model is updated by means of manual update. In the case where the paper machine system itself changes greatly, automatic update is often difficult. For this reason, the model is rebuilt by reusing the paper machine data and manual measurement, and then put into online use. Thus, it can ensure that the model has high accuracy.

[0067] Next, refer to Figure 2 , and describe the process of a variant of the above embodiment of the soft air permeability measurement method for paper.

[0068] and Figure 1 The difference from the process of the soft measurement method of paper air permeability shown in the embodiment is that, in the method of this modification example, before step ST1, a data preprocessing step ST0 is further included.

[0069] Specifically, the paper air permeability is the comprehensive result of various parameters of the paper machine, and dozens or even hundreds of data on the paper machine are more or less related to the air permeability. Although it is not difficult to obtain the current actual data of most papermaking parameters, not all papermaking parameters have a large correlation with the paper air permeability. If the current actual data of papermaking parameters that have little or low correlation with the paper air permeability are also included in the input data, the dimension of each sample in the data set will become extremely high. While the calculation speed is greatly reduced, the calculation accuracy hardly improves or even decreases instead. In addition, although it is possible to screen out papermaking parameters with relatively high correlation through manual screening, this requires papermaking staff with high professional ability and experience. If there is a judgment error, it may lead to unsatisfactory actual air permeability of the entire batch of paper. On the other hand, generally speaking, due to the limitations of acquisition technology and acquisition environment, the data collected often contains unhealthy data such as missing values and outliers. If these types of data are not removed before inputting into the model, it will have an adverse impact on the calculation results.

[0070] Therefore, in this modification example, after collecting the actual data of multiple papermaking parameters through devices such as sensors, the actual data is preprocessed to obtain (or determine) the current data of the air permeability correlation features for predicting the paper air permeability.

[0071] Specifically, in step ST0, after obtaining a part or all of the actual data of papermaking parameters and the corresponding actual air permeability values during the paper manufacturing process, the obtained part or all of the actual data of papermaking parameters is subjected to data cleaning processing to obtain the cleaned actual data. Then, a correlation analysis is performed on the cleaned actual data and the corresponding actual air permeability values to obtain the current data of the air permeability correlation features.

[0072] In this modification example, as the mathematical method for correlation analysis, the Pearson product-moment correlation coefficient method is adopted. In the field of statistics, the Pearson product-moment correlation coefficient method is used to measure the degree of correlation (linear correlation degree) between two variables X and Y, and its value ranges between +1 and -1. The specific calculation formula is as follows:

[0073]

[0074] According to the above calculation formula, the definition of the Pearson correlation coefficient is the covariance of two variables divided by the product of their standard deviations.

[0075] By performing Pearson correlation analysis on the values of each papermaking parameter and the paper air permeability value, the correlation values between each papermaking parameter and the paper air permeability can be obtained quickly. If the value is positive, it represents a positive correlation between them; if the value is negative, it represents a negative correlation between them. The closer the absolute value is to 1, the stronger the correlation, and a value of 0 indicates no correlation. The following table can be referred to:

[0076] Relevance Negative Positive None -0.09 to 0.0 0.0 to 0.09 Weak -0.3 to -0.1 0.1 to 0.3 Medium -0.5 to -0.3 0.3 to 0.5 Strong -1.0 to -0.5 0.5 to 1.0

[0077] By calculating the correlation coefficient between each papermaking parameter and the air permeability, and defining the confidence interval based on the production process flow and data analysis. For example, papermaking parameters with an absolute value of the correlation with the air permeability greater than 0.2 can be selected as air permeability correlation features, including calendar line pressure, grammage, front dryer steam consumption, forming roll vacuum pressure, surface starch addition amount, first press section line pressure in the press section, salt addition amount, second press section line pressure in the press section, filler addition ratio, steam box pressure, rear dryer steam consumption, paper machine speed, long fiber addition ratio, retention aid addition amount, long fiber freeness, paper grammage, beating degree, pulp ratio, filler addition amount, calendar nip pressure, wire section vacuum, dryer steam consumption, sizing agent addition amount, starch addition amount, press section pressure, etc. Preferably, papermaking parameters with an absolute value of the correlation with the air permeability greater than 0.33 can be selected as air permeability correlation features, including calendar line pressure, grammage, front dryer steam consumption, forming roll vacuum pressure, surface starch addition amount, first press section line pressure in the press section.

[0078] Figure 9 A chart showing the air permeability correlation of each air permeability correlation feature obtained under the parameter settings in Table 1 is shown.

[0079] It should be noted here that the training of this model refers to a paper machine with stable operation. Therefore, some parameters that may theoretically affect the air permeability, such as pulp freeness, paper machine speed, fillers, etc., do not show prominent correlation in the model of the present invention. If this model refers to a paper machine with large state changes, the parameter selection will be different, but the inventive concept is the same. It should also be pointed out that the model training of the present invention is mainly aimed at uncoated printing and writing paper. Therefore, some parameters of other paper grades, such as surface coatings and coating pressures of coated paper, are not included. If the soft measurement of other paper grades is carried out using the inventive concept of the present invention, it also falls within the scope of the present invention.

[0080] In addition, as the parameters optimized by this model, it can be understood that the fluctuation of the paper grammage brings about the change of the paper thickness, which has a relatively large impact on the measured value of the air permeability. The line pressure of the calender, the vacuum pressure of the forming roll, and the first line pressure of the press section directly affect the pressure on the paper, thus having a relatively large impact on the air permeability. The steam consumption of the front drying cylinder affects the shrinkage speed and degree of the paper fibers during drying, so it also has a relatively large impact on the air permeability. The surface starch addition amount greatly affects the surface sealing of the paper and has a relatively large impact on the paper air permeability.

[0081] According to this modification example, by sorting, cleaning, and screening the collected data, it is possible to effectively obtain the data of the papermaking parameters directly related to the air permeability of the paper. Thus, it is possible to reasonably reduce the data volume to the greatest extent, eliminate useless data, invalid data, and data that may cause calculation errors or affect the calculation accuracy, improve the calculation efficiency, and ensure the prediction accuracy of the soft measurement model.

[0082] In addition, as a further preferred implementation of the above-mentioned embodiments and their modification examples, it is preferably after obtaining the actual value of the air permeability of the current batch of paper in the laboratory after completing the manufacture of the current batch of paper, integrating the data of the air permeability-related features for predicting the air permeability of this batch with the corresponding actual value of the air permeability to form a training data set, and updating the current soft measurement model of the paper air permeability. It should be noted that the update of the soft measurement model of the paper air permeability can be carried out manually or by means of automatic training. Specifically, if the paper machine system itself changes too much, it is necessary to re-use the paper machine data and re-construct the model by means of manual measurement, and then put it into online use.

[0083] More preferably, in order to save the calculation cost, the current soft measurement model of the paper air permeability is updated in the above manner only when the difference between the inferred air permeability prediction value and the actually detected air permeability actual value exceeds a specified threshold.

[0084] In addition, more preferably, after accumulating a certain number of paper-related features and the corresponding actual values of the air permeability, a new data set is formed, and the current soft measurement model of the paper air permeability is updated using the above new data set.

[0085] On the other hand, another more preferred approach is to set an alarm limit value and implement different levels of alarm functions according to the magnitude of the difference between the inferred air permeability prediction value and the actually detected air permeability actual value. When any error exceeds 5%, the system issues a high-limit alarm, and when the error exceeds 10%, a high-high-limit alarm is implemented to monitor the model accuracy.

[0086] Figure 3 Shows the data processing and data flow diagram based on the above-mentioned paper air permeability measurement method.

[0087] It should be noted that the above-mentioned drawings are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present invention and their variations, rather than for limiting purposes. It is easy to understand that the processes shown in the above-mentioned drawings do not indicate or limit the chronological order of these processes. Additionally, it is also easy to understand that these processes can be executed synchronously or asynchronously in multiple modules.

[0088] (Paper air permeability measurement system)

[0089] Next, based on the paper air permeability measurement method described above, with reference to Figures 4 to 7 , the functional composition of the paper air permeability measurement system according to an embodiment of the present invention and its variations will be described in detail. It should be noted that the following are the device embodiments of the present invention and their variations, which can be used to execute the above method embodiments and their variations. For the details not disclosed in the following device embodiments and variations of the present invention, please refer to the method embodiments and their variations of the present disclosure.

[0090] Figure 4 is a functional composition diagram of the paper air permeability measurement system S according to an embodiment of the present invention.

[0091] As Figure 4 shown, the paper air permeability measurement system S includes a data acquisition module S01, an air permeability prediction module S02, and a data adjustment module S03. The data acquisition module S01 is a functional module that at least has the function of acquiring the current data of the air permeability-related features, that is, a functional module that can execute the above step ST1. The air permeability prediction module S02 is a functional module that at least has the function of being able to execute the above step ST2, that is, it can input the acquired current data of the air permeability-related features as input data into the constructed (for example, constructed by the subsequent model generation module S04) soft measurement model of paper air permeability, so that the soft measurement model of paper air permeability outputs the predicted value of the air permeability of the paper based on the input current data. The data adjustment module S03 is a functional module that can at least execute the above steps ST3 and ST4, that is, compare the predicted value of the air permeability of the paper deduced with the target value of the air permeability of the paper (of the current batch) preset. If the difference between the predicted value of the air permeability and the target value of the air permeability is less than the specified allowable threshold, the current papermaking parameters are not adjusted, and the operating state of the paper machine remains unchanged. If the difference between the two is greater than the specified allowable threshold, the papermaking parameters, especially the current data of the above synchronization-related features, are adjusted so that the predicted value of the air permeability deduced subsequently approaches or reaches the above target value of the air permeability. Here, the allowable threshold refers to the maximum allowable error between the predicted value of the air permeability and the target value of the air permeability.

[0092] The paper air permeability measurement system S according to this embodiment can achieve the technical effects of the paper air permeability measurement method of the above embodiment.

[0093] Figure 5 It represents Figure 4 a functional block diagram of a modified example of the paper air permeability measurement system S shown, namely the paper air permeability measurement system S1.

[0094] As Figure 5 shown, in addition to including a data acquisition module S01, an air permeability prediction module S02, and a data adjustment module S03, the paper air permeability measurement system S1 further includes a feature determination module S00. The feature determination module S00 is a functional module that can at least execute the functions of the above steps ST0, that is, after collecting the actual data of multiple papermaking parameters through devices such as sensors, by preprocessing these actual data, the current data of the air permeability correlation features for predicting the air permeability of the paper is obtained (or determined).

[0095] In this way, the paper air permeability measurement system S1 of this modified example can simultaneously achieve the technical effects of the paper air permeability measurement methods of the above embodiment and its modified examples.

[0096] Figure 6 It represents Figure 4 a functional block diagram of another modified example of the paper air permeability measurement system S shown, namely the paper air permeability measurement system S2.

[0097] As Figure 6 shown, in addition to including a data acquisition module S01, an air permeability prediction module S02, and a data adjustment module S03, the paper air permeability measurement system S2 of this modified example further includes a model generation module S04. The model generation module S04 is a functional module that forms a soft measurement model of paper air permeability by using the historical data of air permeability correlation features obtained in the past paper manufacturing process and the historical values of the air permeability of the manufactured paper as sample data and learning using a specific machine learning algorithm.

[0098] Thus, compared with the above embodiment and its modified examples, the paper air permeability measurement system S2 of this modified example has the function of self - constructing a learned model, and the functionality of its system is more complete.

[0099] Figure 7 It represents Figure 4 a functional block diagram of yet another modified example of the paper air permeability measurement system S shown, namely the paper air permeability measurement system S3.

[0100] As Figure 7As shown, the paper air permeability measurement system S3 of this modification example includes, in addition to the data acquisition module S01, the air permeability prediction module S02, and the data adjustment module S03, a model update module S05. The model update module S05 is a functional module that can at least perform the following functions: after obtaining the actual air permeability value of the current batch of paper in the laboratory after completing the production of the current batch of paper, integrate the data of the air permeability-related features for predicting the air permeability of this batch with the corresponding actual air permeability value to form a training data set, and update the current soft sensor model of paper air permeability.

[0101] In addition, considering the calculation cost, preferably, the above model update module S05 updates the current soft sensor model of paper air permeability in the above manner only when the difference between the estimated air permeability prediction value and the actually detected actual air permeability value exceeds a specified threshold.

[0102] Furthermore, more preferably, after accumulating a certain number of paper-related features and the corresponding actual air permeability values, a new data set is formed, and the current soft sensor model of paper air permeability is updated using the above new data set.

[0103] In addition, as another modification example, any one of the above paper air permeability measurement systems S - S3 may further include an alarm module, and the alarm module realizes different levels of alarm functions according to the magnitude of the difference between the estimated air permeability prediction value and the actually detected actual air permeability value. For example, when any error exceeds 5%, the system issues a high limit alarm, and when the error exceeds 10%, a high-high limit alarm is realized to monitor the model accuracy.

[0104] Figure 8 It is a schematic structural diagram of a computer system 100 showing an embodiment of the present invention. It should be noted that Figure 8 The shown computer system is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present invention.

[0105] As Figure 8 shown, the computer system 100 includes a central processing unit (CPU) 101, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 102 or the program loaded from the storage unit 108 into the random access memory (RAM) 103. In the RAM 103, various programs and data required for the operation of the computer system 100 are also stored. The CPU 101, ROM 102, and RAM 103 are connected to each other through a bus 104. The input / output (I / O) interface 105 is also connected to the bus 104.

[0106] An I / O device is connected to the I / O interface 105. The I / O device may include an input unit 106 such as a keyboard and a mouse, an output unit 107 such as a liquid crystal display (LCD) and a speaker, a storage part 108 such as a hard disk, and a communication part 109 of a network interface card such as a modem. The communication part 109 performs communication processing via a network such as the Internet. The drive 110 may also be connected to the I / O interface 105 as needed. In addition, a removable medium 111 may be installed on the drive 110 as needed so that a computer program read from the removable medium 111 can be installed on the storage part 108 as needed.

[0107] In particular, with reference to Figures 1 to 3 The process described, which is a method for measuring the air permeability of paper in an embodiment of the present invention and its variations, can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product that includes a computer program carried on a computer-readable medium, the computer program containing program code for performing Figures 1 to 3 any one or more of the methods shown. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 109 and / or installed from the removable medium 111. When the computer program is executed by the central processing unit (CPU) 101, the above functions defined in the system of the present invention are executed.

[0108] It should be noted that the computer-readable medium shown in the present invention can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the above two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, in which the computer-readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and this computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0109] As another aspect, the present invention also provides a computer-readable storage medium, which can be included in the computer system described in the above embodiments, or can exist separately without being assembled into the computer system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed by a computer system, the computer system implements the methods as described in the above embodiments and their variants. For example, the above computer system can implement Figure 1 each of the steps shown.

[0110] According to one aspect of the present invention, a computer program product is provided, which includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in various alternative implementations of the above embodiments and their variants.

[0111] It should be understood that the present invention is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present invention is only limited by the appended claims.

Claims

1. A method for measuring the air permeability of paper, characterized in that, Including: Obtaining current data of air permeability related features, where the air permeability related features are papermaking parameters that are related to the air permeability of paper; Taking the current data as input data and inputting it into a soft measurement model for paper air permeability. The soft measurement model for paper air permeability outputs a predicted value of the air permeability of the paper based on the input current data. The soft measurement model for paper air permeability is a learning completed model obtained by associating historical data of the air permeability related features with corresponding actual air permeability values and performing learning; And According to the comparison result between the predicted air permeability value and a preset target air permeability value, selectively adjusting the current data of the air permeability related features so that the predicted air permeability value approaches or reaches the target air permeability value, The air permeability related features include at least a part of the caliper pressure of the calendar, basis weight, steam consumption of the front drying cylinder, vacuum pressure of the forming roll, surface starch addition amount, and the first press line pressure in the press section.

2. The method for measuring paper air permeability according to claim 1, wherein The current data of the air permeability related features is determined by the following method: Obtaining actual data of a part or all of the papermaking parameters and corresponding actual air permeability values during the manufacturing process of paper; Performing data cleaning processing on the obtained actual data of a part or all of the papermaking parameters to obtain the cleaned actual data; And Performing correlation analysis on the cleaned actual data and the actual air permeability values to determine the current data of the air permeability related features.

3. The method for measuring paper air permeability according to claim 1, wherein The soft measurement model for paper air permeability is a learning completed model formed based on the random forest regression algorithm.

4. The method for measuring paper air permeability according to claim 3, wherein Taking a part of the historical data of the air permeability related features and the corresponding actual air permeability values as a training data set, and taking the remaining part of the historical data of the air permeability related features and the corresponding actual air permeability values as a test data set, and forming the soft measurement model for paper air permeability by using the hold-out method or cross-validation method.

5. The method for measuring paper air permeability according to claim 1, wherein Updating the soft measurement model for paper air permeability based on the current data of the air permeability related features and the corresponding actual air permeability values.

6. The method for measuring paper air permeability according to claim 5, wherein When the difference between the predicted air permeability value and the actual air permeability value exceeds a threshold, updating the soft measurement model for paper air permeability.

7. A paper air permeability measurement system, characterized in that, Including: A data acquisition module, which acquires current data of air permeability related features, where the air permeability related features are papermaking parameters that are related to the air permeability of paper; An air permeability prediction module, which inputs the current data as input data into a soft measurement model for paper air permeability. The soft measurement model for paper air permeability outputs a predicted value of the air permeability of the paper based on the input current data. The soft measurement model for paper air permeability is a learned model obtained by associating historical data of the air permeability-related features with the corresponding actual air permeability values and performing learning. And A data adjustment module, which selectively adjusts the current data of the air permeability-related features according to the comparison result between the predicted value of the air permeability and a preset target value of the air permeability, so that the predicted value of the air permeability approaches or reaches the target value of the air permeability. The air permeability-related features include at least a part of the calender line pressure, grammage, steam consumption of the front drying cylinder, vacuum pressure of the forming roll, surface starch addition amount, and the first pressing line pressure in the press section of the paper.

8. The paper air permeability measurement system according to claim 7, wherein It further includes a feature determination module, and the feature determination module determines the air permeability-related features in the following manner: Obtain the actual data of a part or all of the papermaking parameters and the corresponding actual air permeability values during the manufacturing process of the paper; Perform data cleaning on the obtained actual data of a part or all of the papermaking parameters to obtain the cleaned actual data; And Perform a correlation analysis on the cleaned actual data and the actual air permeability values to determine the air permeability-related features.

9. The paper air permeability measurement system according to claim 7, wherein It further includes a model generation module, and the model generation module generates the soft measurement module for paper air permeability.

10. The paper air permeability measurement system according to claim 7, wherein It further includes a model update module, and the model update module updates the soft measurement model for paper air permeability based on the current data of the air permeability-related features and the corresponding actual air permeability values.

11. A computer system for measuring the air permeability of paper, including a memory, a processor, and a computer program stored on the memory, wherein The processor executes the computer program to implement the paper air permeability measurement method according to any one of claims 1 to 6.

12. A computer-readable storage medium storing a computer program or instruction, wherein The computer program or instruction, when executed by a processor, implements the paper air permeability measurement method according to any one of claims 1 to 6.

13. A computer program product including a computer program or instruction, wherein The computer program or instruction, when executed by a processor, implements the paper air permeability measurement method according to any one of claims 1 to 6.

Images