Method and device for predicting stain resistance of wood-based composite materials

Abstract

The present invention discloses a method and device for predicting the stain resistance performance of wood-based composite materials, belonging to the field of forestry engineering. The present invention trains a machine learning model capable of predicting the stain resistance performance of wood-based composite materials, and finds out the structure-activity relationship between the stain resistance performance of wood-based composite materials and their structural data and production process data. The present invention can be applied to the production practice of stain-resistant wood-based composite materials. According to the existing conditions such as the structural data and production process data of wood veneers, the actual value of the stain resistance performance can be quickly predicted through the trained machine learning model, which has the advantages of being fast, efficient, and having high prediction accuracy, realizing the rapid and accurate prediction of the performance of the final product. At the same time, through the aforementioned structure-activity relationship, the key factors affecting the stain resistance performance can be clarified, guiding the optimization of raw materials and production processes, reducing the trial-and-error cost of enterprise processing, improving production efficiency, and further enhancing the processing level of stain-resistant wood-based composite materials.

Description

Technical Field

[0001] The present invention relates to the field of forestry engineering, and particularly to a method and device for predicting the stain resistance of wood-based composites. Background Art

[0002] Wood-based composites are a commonly used consumable in fields such as home furnishing and decoration, and can be seen everywhere in daily life. They are often used in the production of furniture such as cabinets and dining tables. With the gradual improvement of people's living standards, people are more pursuing the cleanliness and aesthetics of furniture, which requires that wood-based composites have strong stain resistance during production and processing.

[0003] The processing process of stain-resistant wood-based composites is complex, and related products may be updated in terms of processing technology. As a result, in actual processing, even experienced workers are difficult to accurately estimate the quality of the material after production, making it difficult for the produced wood-based composites to achieve the desired stain resistance effect. Moreover, when enterprises optimize raw materials and production processes, they can only continuously try and error, increasing the cost of enterprise processing trial and error and reducing production efficiency. Summary of the Invention

[0004] To solve the defects of the prior art, the present invention provides a method and device for predicting the stain resistance of wood-based composites, which realizes the rapid and accurate prediction of the stain resistance of wood-based composites, guides the optimization of raw materials and production processes, reduces the enterprise processing trial and error cost, and improves production efficiency.

[0005] The present invention provides the following technical solutions:

[0006] A method for predicting the stain resistance of wood-based composites, the method comprising:

[0007] Establishing a reference data set including multiple wood veneer data samples; wherein, each wood veneer data sample includes several characteristics of the wood veneer, the types of the characteristics include structural data and production process data, the wood veneer data sample is labeled with label data, and the label data includes stain resistance data;

[0008] Digitizing and normalizing the reference data set;

[0009] Dividing the reference data set into a training set and a test set, training a machine learning model based on a multi-layer perceptron constructed by using the training set, and verifying by using the test set to obtain a machine learning model capable of predicting the stain resistance of wood-based composites;

[0010] Performing dimensionality reduction fitting on the regression function of the trained machine learning model through a principal component regression algorithm to obtain the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data.

[0011] Further, the establishment of a reference data set including a plurality of wood veneer data samples includes:

[0012] Obtaining the structural data of the wood veneer collected with the wood veneer as a unit during the production process of the wood veneer;

[0013] Wherein, the structural data includes one or more of tree species name, thickness of impregnated film paper, surface color of impregnated film paper, and surface texture of impregnated film paper;

[0014] Obtaining the production process data of the wood veneer collected with the wood veneer as a unit during the production process of the wood veneer;

[0015] Wherein, the production process data includes one or more of surface coating dosage, electron beam curing irradiation dose, and electron beam curing irradiation voltage;

[0016] Obtaining the stain resistance data of the wood veneer after production;

[0017] Wherein, the stain resistance data includes one or more of water contact angle, oil contact angle, and stain resistance grade;

[0018] Taking the structural data and production process data as the characteristics of the wood veneer data sample, and taking the stain resistance data as the label data of the wood veneer data sample, to obtain a reference data set including a plurality of wood veneer data samples.

[0019] Further, the dimensionality reduction fitting of the regression function of the trained machine learning model by the principal component regression algorithm to obtain the structure-activity relationship between the stain resistance of the wood-based composite material and its structural data and production process data includes:

[0020] For the reference data set D, calculate the correlation coefficient γ and of different types of features of the same wood veneer data sample ij , and form a correlation coefficient matrix R with all the correlation coefficients γ ij ;

[0021]

[0022] and respectively represent the i-th and j-th features of the k-th wood veneer data sample in the reference data set D, i, j = 1, 2, …, n, n is the total number of types of features, and k = 1, 2, …, m, m is the total number of wood veneer data samples;

[0023] Calculate eigenvalues (λ1, λ2,..., λ n ) and eigenvectors μ i , μ2,..., μ n ;

[0024] Sort the eigenvalues (λ1, λ2,..., λ n ) in descending order, and calculate the information contribution rate b j corresponding to the j-th eigenvalue and the cumulative contribution rate α p ;

[0025]

[0026]

[0027] Calculate the principal component z j corresponding to each eigenvalue. The set composed of all principal components z j is the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data

[0028]

[0029]

[0030] Among them, μ jk represents the k-th dimensional value of the j-th eigenvector.

[0031] Furthermore, the process of dividing the reference data set into a training set and a test set, and using the training set to train the constructed machine learning model based on a multi-layer perceptron includes:

[0032] Divide the reference data set into a training set and a test set according to a ratio of 6:4;

[0033] Establish a machine learning model based on a multi-layer perceptron. The machine learning model includes an input layer, a hidden layer, and an output layer. Any neuron in the upper layer is connected to each neuron in the lower layer;

[0034] Use the training set to train and tune the parameters of the machine learning model;

[0035] Adopt the method of K-fold cross-validation to re-divide the training set and the test set, and repeatedly train and tune the parameters of the machine learning model until a machine learning model that can predict the stain resistance of wood-based composites is obtained.

[0036] Furthermore, the process of digitizing and normalizing the reference data set includes:

[0037] Digitize the data in the reference dataset through Python technology;

[0038] Normalize the format and units of the data in the reference dataset through Python technology.

[0039] Further, before dividing the reference dataset into a training set and a test set and training the constructed machine learning model based on a multi-layer perceptron using the training set, it further includes:

[0040] Perform quality assessment on the digitized and normalized reference dataset using a fitting algorithm to evaluate the integrity and consistency of the reference dataset;

[0041] Clean the data of the reference dataset after quality assessment to remove outliers; perform quality assessment on the reference dataset after data cleaning using the fitting algorithm again until its integrity and consistency meet the requirements.

[0042] Further, performing quality assessment on the digitized and normalized reference dataset using a fitting algorithm to evaluate the integrity and consistency of the reference dataset includes:

[0043] Divide the reference dataset into K sample subsets of equal size;

[0044] Traverse each sample subset in turn, and use the current sample subset of each traversal as the validation set and the remaining wood veneer data samples as the training set to perform quality assessment;

[0045] Use the statistical values of the evaluation metrics obtained from the K times of quality assessment as the final evaluation metrics, and evaluate the integrity and consistency of the reference dataset through the final evaluation metrics.

[0046] Further, cleaning the data of the reference dataset after quality assessment to remove outliers includes:

[0047] Represent the reference dataset as a two-dimensional array, and for any element at any position in the reference dataset, generate an observation window of a set size centered on the element at that position;

[0048] Calculate the median and standard deviation of all elements within the observation window;

[0049] If the absolute value of the difference between the element at any position and the median exceeds a set multiple of the standard deviation, use the median to replace the element at that position.

[0050] A device for predicting the stain resistance of a wood-based composite material, the device includes:

[0051] A reference dataset building module for building a reference dataset including multiple wood veneer data samples; wherein each of the wood veneer data samples includes several features of the wood veneer, the types of the features including structural data and production process data, the wood veneer data samples being labeled with label data, and the label data including stain resistance performance data;

[0052] A preprocessing module for digitizing and normalizing the reference dataset;

[0053] A model training module for dividing the reference dataset into a training set and a test set, training a machine learning model based on a multi-layer perceptron constructed by using the training set, and validating by using the test set to obtain a machine learning model capable of predicting the stain resistance performance of wood-based composites;

[0054] A structure-activity relationship determination module for performing dimensionality reduction fitting on the regression function of the trained machine learning model through a principal component regression algorithm to obtain the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data.

[0055] Further, the reference dataset building module includes:

[0056] A structural data acquisition unit for acquiring the structural data of the wood veneer collected by taking the wood veneer as a basic unit during the production process of the wood veneer;

[0057] Wherein the structural data includes one or more of tree species name, impregnated film paper thickness, impregnated film paper surface color, and impregnated film paper surface texture;

[0058] A production process data acquisition unit for acquiring the production process data of the wood veneer collected by taking the wood veneer as a basic unit during the production process of the wood veneer;

[0059] Wherein the production process data includes one or more of surface coating dosage, electron beam curing irradiation dose, and electron beam curing irradiation voltage;

[0060] A stain resistance performance data acquisition unit for acquiring the stain resistance performance data of the produced wood veneer;

[0061] Wherein the stain resistance performance data includes one or more of water contact angle, oil contact angle, and stain resistance grade;

[0062] A reference dataset building unit for using the structural data and production process data as the features of the wood veneer data samples and using the stain resistance performance data as the label data of the wood veneer data samples to obtain a reference dataset including multiple wood veneer data samples.

[0063] Furthermore, the structure-activity relationship determination module includes:

[0064] A correlation coefficient matrix calculation unit, which is used to calculate the correlation coefficients γ of different types of features of the same wood veneer data sample for the reference data set D and of ij , and all the correlation coefficients γ ij are composed of a correlation coefficient matrix R;

[0065]

[0066] and and respectively represent the i-th and j-th features of the k-th wood veneer data sample in the reference data set D, where i, j = 1, 2,..., n, n is the total number of feature types, and k = 1, 2,..., m, m is the total number of wood veneer data samples;

[0067] An eigenvalue decomposition unit, which is used to calculate eigenvalues (λ1, λ2,..., λ n ) and eigenvectors μ i , μ2,..., μ n ;

[0068] A contribution rate calculation unit, which is used to sort the eigenvalues (λ1, λ2,..., λ n ) in descending order, and calculate the information contribution rate b j corresponding to the j-th eigenvalue and the cumulative contribution rate α p ;

[0069]

[0070]

[0071] A principal component calculation unit, which is used to calculate the principal component z j corresponding to each eigenvalue. The set composed of all the principal components z j is the structure-activity relationship between the stain resistance of the wood-based composite material and its structural data and production process data;

[0072]

[0073]

[0074] Among them, μ jk represents the k-th dimensional value of the j-th eigenvector.

[0075] Furthermore, the model training module includes:

[0076] A data division unit for dividing the reference data set into a training set and a test set according to a ratio of 6:4.

[0077] A model establishment unit for establishing a machine learning model based on a multi-layer perceptron. The machine learning model includes an input layer, a hidden layer, and an output layer, and any neuron in the upper layer is respectively connected to each neuron in the lower layer.

[0078] A model training unit for training and tuning the machine learning model using the training set.

[0079] An iterative training unit for re-dividing the training set and the test set using the K-fold cross-validation method, and repeatedly training and tuning the machine learning model until a machine learning model capable of predicting the stain resistance of wood composite materials is obtained.

[0080] Furthermore, the preprocessing module includes:

[0081] A digitization unit for digitizing the data in the reference data set through Python technology.

[0082] A normalization unit for normalizing the format and unit of the data in the reference data set through Python technology.

[0083] Furthermore, the device further includes:

[0084] A data evaluation module for evaluating the quality of the digitized and normalized reference data set using a fitting algorithm, and evaluating the integrity and consistency of the reference data set.

[0085] A data cleaning module for cleaning the reference data set after quality evaluation to remove outliers; and re-evaluating the quality of the reference data set after data cleaning using the fitting algorithm until its integrity and consistency meet the requirements.

[0086] Furthermore, the data evaluation module includes:

[0087] A sample subset division unit for dividing the reference data set into K sample subsets of equal size.

[0088] A traversal training unit for sequentially traversing each of the sample subsets, and using the current sample subset during each traversal as a validation set, and the remaining wood veneer data samples as a training set for quality evaluation.

[0089] An evaluation unit for using the statistical value of the evaluation index obtained from the K times of quality evaluation as the final evaluation index, and evaluating the integrity and consistency of the reference data set through the final evaluation index.

[0090] Further, the data cleaning module includes:

[0091] An observation window generation unit, configured to represent the reference data set as a two-dimensional array, and generate an observation window of a set size centered on an element at any position in the reference data set;

[0092] An observation window calculation unit, configured to calculate the median and standard deviation of all elements within the observation window;

[0093] A judgment unit, configured to, if the absolute value of the difference between the element at any position and the median exceeds a set multiple of the standard deviation, use the median to replace the element at any position.

[0094] The present invention has the following beneficial effects:

[0095] The present invention trains a machine learning model capable of predicting the stain resistance performance of wood-based composites, and finds out the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data. The present invention can be applied to the production practice of stain-resistant wood-based composites. According to the existing conditions such as the structural data and production process data of wood veneers, the actual value of the stain resistance performance can be quickly predicted through the trained machine learning model, which has the advantages of being fast, efficient, and high prediction accuracy, and realizes the rapid and accurate prediction of the performance of the final product. At the same time, through the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data, the key factors affecting the stain resistance performance can be clarified, which can be used to guide the optimization of raw materials and production processes, reduce the trial-and-error cost of enterprise processing, improve production efficiency, and further improve the processing level of stain-resistant wood-based composites. Description of the Drawings

[0096] Figure 1 is a flowchart of the method for predicting the stain resistance performance of wood-based composites of the present invention;

[0097] Figure 2 is a schematic diagram of a multi-layer perceptron;

[0098] Figure 3 is a schematic diagram of the device for predicting the stain resistance performance of wood-based composites of the present invention. Detailed Embodiments

[0099] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the present invention to be protected, but only represents the selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative work belong to the scope of protection of the present invention.

[0100] An embodiment of the present invention provides a method for predicting the stain resistance of a wood composite material. As Figure 1 shown, the method includes:

[0101] S100: Establish a reference data set including multiple wood veneer data samples; wherein, each wood veneer data sample includes several kinds of characteristics of the wood veneer, the types of the characteristics include structural data and production process data, and the wood veneer data samples are labeled with label data, and the label data includes stain resistance data.

[0102] The structural data and production process data of the wood veneer are factors affecting the stain resistance of the wood composite material. The structural data includes the types and various performance parameters of the core board and the face board, etc. The production process data includes various components added in production, processes, etc. The stain resistance data is represented by data such as water contact angle, oil contact angle and stain resistance level. The present invention infers the stain resistance data of the wood veneer through the obtained structural data and production process data of the wood veneer.

[0103] In the actual production process of the wood veneer, taking the wood veneer type as the base unit, its structural data, production process data and stain resistance data are collected. The above structural data and production process data of each wood veneer are used as the characteristics of a data sample, and the stain resistance data of each wood veneer is used as its label data. About 600 data samples are collected in total to form a reference data set.

[0104] S200: Digitalize and normalize the reference data set.

[0105] The data in the collected wood veneer data samples includes data such as Chinese characters, English characters and numerical characters. For data such as Chinese characters and English characters, it is necessary to digitalize and normalize them to convert them into digital data recognizable by a computer.

[0106] S500: Divide the reference data set into a training set and a test set. Use the training set to train the constructed machine learning model based on a multi-layer perceptron, and use the test set for verification to obtain a machine learning model that can predict the stain resistance performance of wood-based composites.

[0107] The machine learning model used in the present invention is a regression model based on a multi-layer perceptron. Train the machine learning algorithm regression model using the training set and test the results of the algorithm using the test set to obtain a machine learning model that can predict the stain resistance performance of wood-based composites.

[0108] S600: Perform dimensionality reduction fitting on the regression function of the trained machine learning model through the principal component regression algorithm to obtain the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data.

[0109] The present invention has trained a machine learning model that can predict the stain resistance performance of wood-based composites and found the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data. The present invention can be applied to the production practice of stain-resistant wood-based composites. According to the existing conditions such as the structural data and production process data of wood veneers, the actual value of its stain resistance performance can be quickly predicted through the trained machine learning model, which has the advantages of being fast, efficient, and having high prediction accuracy, and can quickly and accurately predict the performance of the final product. At the same time, through the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data, the key factors affecting the stain resistance performance can be clarified, which can be used to guide the optimization of raw materials and production processes, reduce the trial-and-error cost of enterprise processing, improve production efficiency, and further improve the processing level of stain-resistant wood-based composites.

[0110] As an improvement of an embodiment of the present invention, the foregoing S100 includes:

[0111] S101: Obtain the structural data of the wood veneer collected with the wood veneer as the unit during the production process of the wood veneer.

[0112] Among them, the structural data includes one or more of the tree species name, the thickness of the impregnated film paper, the surface color of the impregnated film paper, and the surface texture of the impregnated film paper.

[0113] S102: Obtain the production process data of the wood veneer collected with the wood veneer as the unit during the production process of the wood veneer.

[0114] Among them, the production process data includes one or more of the surface coating dosage, the electron beam curing irradiation dose, and the electron beam curing irradiation voltage.

[0115] S103: Obtain the stain resistance performance data of the produced wood veneer.

[0116] Among them, the stain resistance performance data includes one or more of the water contact angle, oil contact angle, and stain resistance level.

[0117] S104: Use the structural data and production process data as the features of the wood veneer data samples, and use the stain resistance performance data as the label data of the wood veneer data samples to obtain a reference data set including multiple wood veneer data samples.

[0118] The present invention does not limit the method for digitizing and normalizing the reference data set. One example includes:

[0119] S201: Digitize the data in the reference data set through Python technology.

[0120] S202: Normalize the format and unit of the data in the reference data set through Python technology.

[0121] When digitizing and normalizing, the Chinese and English information such as tree species names and material names can be digitized through Python and stored in the CSV format. For example, for the data of Chinese characters and English characters such as tree species names, the decimal value of the ASCII code of the first letter of its Chinese pinyin or English is used as the result value. For example, for poplar, the ASCII code value of its first letter is 89.

[0122] During the process of data collection in the factory, there are too many useless data, and most of them are qualified data obtained according to production experience. Therefore, the data needs to be further processed. The method of the present invention also includes:

[0123] S300: Use the fitting algorithm to evaluate the quality of the digitized and normalized reference data set, and evaluate the integrity and consistency of the reference data set.

[0124] This step is used to evaluate the quality of the reference data set and provide a basis for the data availability, specifically as follows:

[0125] S301: Divide the reference data set into K sample subsets of equal size.

[0126] S302: Traverse each sample subset in turn, and use the current sample subset of each traversal as the validation set, and the remaining wood veneer data samples as the training set to establish a model and conduct quality evaluation.

[0127] S303: Use the statistical value of the evaluation index obtained from the K - time quality evaluation as the final evaluation index, and evaluate the integrity and consistency of the reference data set through the final evaluation index.

[0128] The present invention uses the K-fold cross-validation method to repeatedly divide the reference data set, conducts multiple evaluations repeatedly, and uses statistical values such as the variance or average value of the multiple evaluation results as the final evaluation index to verify the integrity and consistency of the reference data set.

[0129] S400: Perform data cleaning on the reference data set after quality evaluation to remove outliers; use the fitting algorithm to perform quality evaluation on the reference data set after data cleaning again until its integrity and consistency meet the requirements.

[0130] Data cleaning is used to remove noise during the acquisition process and reduce the impact of noise on the computer prediction model. The specific implementation methods include:

[0131] S401: Represent the reference data set as a two-dimensional array. For an element at any position in the reference data set, generate an observation window of a set size centered on the element at that position.

[0132] Exemplarily, the reference data set is represented by a two-dimensional array D = {x 1 , x 2 ,... x m} T where x k represents the k-th wood veneer data sample of the reference data set D, represents the j-th feature of the k-th wood veneer data sample of the reference data set D, k = 1, 2,..., m, where m is the total number of wood veneer data samples, and j = 1, 2,..., n, where n is the total number of types of features. Expand the two-dimensional array D = {x 1 , x 2 ,... x m} T and its form is as follows:

[0133]

[0134] For an element at any position generate an observation window of a set size centered on it.

[0135] S402: Calculate the median and standard deviation of all elements within the observation window.

[0136] Assume that the size of the entire observation window is 2r + 1. Arrange all elements in the observation window in ascending order, calculate the median, that is, the (r + 1)-th number after sorting, and calculate the standard deviation.

[0137] S403: If the absolute value of the difference between an element at any position and the median exceeds a set multiple of the standard deviation, use the median to replace the element at that position.

[0138] Assume that the set multiple is 3 times. If the element If the absolute value of the difference from the median exceeds three times the standard deviation, the element is replaced with the median.

[0139] As another improvement of the embodiment of the present invention, the foregoing S500 includes:

[0140] S501: Divide the reference data set after data cleaning into a training set and a test set according to a ratio of 6:4.

[0141] For the foregoing reference data set of about 600 wood veneer data samples, after division, about 360 sample data are used as the training set, and 240 sample data are used as the test set.

[0142] S502: Establish a machine learning model based on a multi-layer perceptron, which includes an input layer, a hidden layer, and an output layer. Any neuron in the upper layer is respectively connected to each neuron in the lower layer.

[0143] The multi-layer perceptron (MLP) can be multiple layers composed of artificial neurons of perceptrons. Each layer is fully connected to the next layer. The output of each artificial neuron on a certain layer becomes the input of all artificial neurons in the next layer. The MLP has at least three-layer structures, namely the input layer, the hidden layer, and the output layer. Among them, the MLP can have multiple hidden layers. The model structure of the MLP is as Figure 2 shown.

[0144] The mathematical expression of the MLP is as follows:

[0145] f(x) = G(b (2) + W (2) (s(b (1) + W (1) x)))

[0146] where b is the bias value, W is the weight, G is the activation function, and the activation function uses the Relu function, and its formula is as follows:

[0147] f(x) = max(0, x)

[0148] S503: Use the training set to train and tune the parameters of the machine learning model.

[0149] S504: Adopt the method of K-fold cross-validation to re-divide the training set and the test set, and repeatedly train and tune the parameters of the machine learning model until a machine learning model capable of predicting the stain resistance of wood-based composites is obtained.

[0150] The expected accuracy of the finally obtained machine learning model can exceed 95%. After the model is trained, the main factors affecting the stain resistance performance can be fitted according to the principle component regression (PCR) algorithm, and the regression function obtained by the model regression can be dimensionally reduced to formulate the structure-activity relationship formula between functions and materials, so as to analyze the structure-activity relationship of the stain resistance performance of wood-based composites. The specific implementation process is as follows:

[0151] S601: For the reference data set D, calculate the correlation coefficients γ and of different types of features of the same wood veneer data sample ij , and form a correlation coefficient matrix R with all the correlation coefficients γ ij .

[0152] D = {x 1 , x 2 ,... x m} T , with a total of m wood veneer data samples, and each wood veneer data sample x k includes n types of features, then D is a two-dimensional array of m * n.

[0153]

[0154] For the i-th and j-th features and of the k-th wood veneer data sample in the reference data set D, calculate and of the correlation coefficient γ ij , where i, j = 1, 2,..., n, n is the total number of feature types, and k = 1, 2,..., m, m is the total number of wood veneer data samples.

[0155]

[0156] and are respectively and after standardization.

[0157] S602: Calculate the eigenvalues (λ1, λ2,..., λ n ) and eigenvectors μ i , μ2,..., μ n through the correlation coefficient matrix R.

[0158] S603: Sort the eigenvalues (λ1, λ2,..., λ n ) in descending order, and calculate the information contribution rate b corresponding to the j-th eigenvaluej and the cumulative contribution rate α p .

[0159]

[0160]

[0161] S604: Calculate the principal component z corresponding to each eigenvalue j , and all the principal components z j The set composed of is the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data.

[0162]

[0163]

[0164] Among them, μ jk represents the value of the k-th dimension of the j-th eigenvector.

[0165] The embodiment of the present invention also provides a device for predicting the stain resistance of wood-based composites, as Figure 3 shown, the device includes:[[]]

[0166] A reference data set establishment module 1 for establishing a reference data set including multiple wood veneer data samples; among them, each wood veneer data sample includes several features of the wood veneer, the types of features include structural data and production process data, and the wood veneer data sample is labeled with label data, and the label data includes stain resistance data.

[0167] A preprocessing module 2 for digitizing and normalizing the reference data set.

[0168] A model training module 3 for dividing the reference data set into a training set and a test set, training the constructed machine learning model based on a multi-layer perceptron using the training set, and verifying using the test set to obtain a machine learning model capable of predicting the stain resistance of wood-based composites.

[0169] A structure-activity relationship determination module 4 for performing dimensionality reduction fitting on the regression function of the trained machine learning model through the principal component regression algorithm to obtain the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data.

[0170] The present invention trains a machine learning model capable of predicting the stain resistance of wood-based composites and finds out the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data. The present invention can be applied to the production practice of stain-resistant wood-based composites. According to the existing conditions such as the structural data and production process data of wood veneers, the actual value of the stain resistance can be quickly predicted through the trained machine learning model, which has the advantages of being fast, efficient, and having high prediction accuracy, realizing the rapid and accurate prediction of the performance of the final product. At the same time, through the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data, the key factors affecting the stain resistance can be clarified, which can be used to guide the optimization of raw materials and production processes, reduce the trial-and-error cost of enterprise processing, improve production efficiency, and further improve the processing level of stain-resistant wood-based composites.

[0171] As an improvement of an embodiment of the present invention, the aforementioned reference dataset establishment module includes:

[0172] A structural data acquisition unit for acquiring the structural data of a wood veneer collected with the wood veneer as a unit during the production process of the wood veneer.

[0173] Among them, the structural data includes one or more of the tree species name, the thickness of the impregnated film paper, the surface color of the impregnated film paper, and the surface texture of the impregnated film paper.

[0174] A production process data acquisition unit for acquiring the production process data of a wood veneer collected with the wood veneer as a unit during the production process of the wood veneer.

[0175] Among them, the production process data includes one or more of the surface coating dosage, the electron beam curing irradiation dose, and the electron beam curing irradiation voltage.

[0176] A stain resistance data acquisition unit for acquiring the stain resistance data of the wood veneer after production;

[0177] Among them, the stain resistance data includes one or more of the water contact angle, the oil contact angle, and the stain resistance level.

[0178] A reference dataset establishment unit for using the structural data and the production process data as the features of the wood veneer data samples and the stain resistance data as the label data of the wood veneer data samples to obtain a reference dataset including multiple wood veneer data samples.

[0179] An example of the preprocessing module of the present invention includes:

[0180] A digitization unit for digitizing the data in the reference dataset through Python technology.

[0181] A normalization unit for normalizing the format and unit of data in a reference dataset through Python technology.

[0182] The device of the present invention may further include:

[0183] A data evaluation module for evaluating the quality of the digitized and normalized reference dataset using a fitting algorithm, and evaluating the integrity and consistency of the reference dataset.

[0184] A data cleaning module for cleaning the reference dataset after quality evaluation to remove outliers; and re-evaluating the quality of the reference dataset after data cleaning using a fitting algorithm until its integrity and consistency meet the requirements.

[0185] Among them, the data evaluation module includes:

[0186] A sample subset division unit for dividing the reference dataset into K sample subsets of equal size.

[0187] A traversal training unit for sequentially traversing each sample subset, using the current sample subset of each traversal as a validation set, and the remaining wood veneer data samples as a training set for quality evaluation.

[0188] An evaluation unit for using the statistical value of the evaluation metrics obtained from K times of quality evaluation as the final evaluation metric, and evaluating the integrity and consistency of the reference dataset through the final evaluation metric.

[0189] The data cleaning module includes:

[0190] An observation window generation unit for representing the reference dataset as a two-dimensional array, and generating an observation window of a set size centered on the element at any position in the reference dataset.

[0191] An observation window calculation unit for calculating the median and standard deviation of all elements within the observation window.

[0192] A judgment unit for replacing the element at any position with the median if the absolute value of the difference between the element at any position and the median exceeds a set multiple of the standard deviation.

[0193] As another improvement of the embodiment of the present invention, the aforementioned model training module includes:

[0194] A data division unit for dividing the reference dataset into a training set and a test set according to a ratio of 6:4.

[0195] A model establishment unit for establishing a machine learning model based on a multi-layer perceptron. The machine learning model includes an input layer, a hidden layer, and an output layer, and any neuron in the upper layer is respectively connected to each neuron in the lower layer.

[0196] A model training unit for training and tuning parameters of a machine learning model using a training set.

[0197] An iterative training unit for re - dividing the training set and the test set by using the K - fold cross - validation method, and repeatedly training and tuning the parameters of the machine learning model until a machine learning model capable of predicting the stain resistance performance of wood - based composites is obtained.

[0198] Furthermore, the aforementioned structure - activity relationship determination module includes:

[0199] A correlation coefficient matrix calculation unit for calculating the correlation coefficient γ between different types of features of the same wood veneer data sample for the reference data set D and , and forming a correlation coefficient matrix R with all the correlation coefficients γ ij , where and respectively represent the i - th and j - th features of the k - th wood veneer data sample in the reference data set D, i, j = 1, 2, …, n, n is the total number of feature types, and k = 1, 2, …, m, m is the total number of wood veneer data samples. ij

[0200]

[0201] and

[0202]

[0203] An eigenvalue decomposition unit for calculating eigenvalues (λ1, λ2, …, λ n ) and eigenvectors μ i , μ2, …, μ n .

[0204] A contribution rate calculation unit for sorting the eigenvalues (λ1, λ2, …, λ n ) in descending order, calculating the information contribution rate b j corresponding to the j - th eigenvalue and the cumulative contribution rate α p .

[0204]

[0205]

[0206] A principal component calculation unit for calculating the principal component z j corresponding to each eigenvalue. The set composed of all the principal components z j is the structure - activity relationship between the stain resistance performance of wood - based composites and their structural data and production process data.

[0207]

[0208]

[0209] Among them, μ jk represents the k-th dimensional value of the j-th eigenvector.

[0210] For the device provided in the above embodiments, its implementation principle and the technical effects produced correspond one by one to the embodiments of the foregoing method. For the sake of brief description, for the parts not mentioned in the embodiments of the device, reference may be made to the corresponding contents in the embodiments of the foregoing method. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the modules and units described in the device can all refer to the corresponding processes in the embodiments of the foregoing method, and will not be elaborated herein.

[0211] The method for predicting the stain resistance of the wood composite material described in the above embodiments of the present invention can implement the business logic through a computer program and record it on a storage medium. The storage medium can be read and executed by a computer to achieve the effects of the solutions described in the method embodiments of this specification. Therefore, the embodiments of the present invention also provide a computer-readable storage medium for predicting the stain resistance of wood composite materials, including a memory for storing processor-executable instructions, and when the instructions are executed by the processor, the steps of the method for predicting the stain resistance of wood composite materials including the foregoing embodiments are implemented.

[0212] The storage medium may include a physical device for storing information, usually storing the information after digitization and then using a medium such as electricity, magnetism, or optics. The storage medium may include: devices for storing information using electrical energy, such as various memories, such as RAM, ROM, etc.; devices for storing information using magnetic energy, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, bubble memories, USB flash drives; devices for storing information using optical methods, such as CDs or DVDs. Of course, there are also other types of readable storage media, such as quantum memories, graphene memories, and so on.

[0213] According to the description of the method embodiments, the above-mentioned storage medium may also include other implementation manners. The implementation principle and the technical effects produced by this embodiment are the same as those of the foregoing method embodiments. For details, reference may be made to the description of the relevant method embodiments, and will not be elaborated herein one by one.

[0214] An embodiment of the present invention further provides a device for predicting the stain resistance of wood-based composites. The device can be a separate computer or can include an actual operating device that uses one or more of the methods or one or more embodiment devices described in this specification. The device for predicting the stain resistance of wood-based composites can include at least one processor and a memory storing computer-executable instructions. When the processor executes the instructions, the steps of any one or more of the above-described methods for predicting the stain resistance of wood-based composites are implemented.

[0215] According to the description of the method embodiment, the above-described device may also include other implementation manners. The implementation principle and the technical effects generated by this embodiment are the same as those of the foregoing method embodiment. For details, reference may be made to the description of the relevant method embodiment, and no further elaboration will be provided here.

[0216] Finally, it should be noted that the above embodiments are only specific implementation manners of the present invention, which are used to illustrate the technical solutions of the present invention and are not intended to limit them. The protection scope of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that any person skilled in the art within the technical scope disclosed by the present invention can still modify the technical solutions recorded in the foregoing embodiments or easily conceive of changes, or perform equivalent replacements for some of the technical features; and these modifications, changes, or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. All of them should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for predicting the stain resistance of a wood composite material, characterized in that, The method includes: Establishing a reference data set including multiple wood veneer data samples; wherein, each of the wood veneer data samples includes several characteristics of the wood veneer, the types of the characteristics include structural data and production process data, the wood veneer data samples are labeled with label data, and the label data includes stain resistance performance data; Digitizing and normalizing the reference data set; Dividing the reference data set into a training set and a test set, training a machine learning model based on a multi-layer perceptron constructed by using the training set, and validating by using the test set to obtain a machine learning model capable of predicting the stain resistance performance of wood-based composites; Performing dimensionality reduction fitting on the regression function of the trained machine learning model through a principal component regression algorithm to obtain the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data; The establishing of the reference data set including multiple wood veneer data samples includes: Obtaining the structural data of the wood veneer collected with the wood veneer as a unit during the production process of the wood veneer; Wherein, the structural data includes one or more of tree species name, impregnated film paper thickness, impregnated film paper surface color, and impregnated film paper surface texture; Obtaining the production process data of the wood veneer collected with the wood veneer as a unit during the production process of the wood veneer; Wherein, the production process data includes one or more of surface coating dosage, electron beam curing irradiation dose, and electron beam curing irradiation voltage; Obtaining the stain resistance performance data of the produced wood veneer; Wherein, the stain resistance performance data includes one or more of water contact angle, oil contact angle, and stain resistance level; Taking the structural data and production process data as the characteristics of the wood veneer data samples, and taking the stain resistance performance data as the label data of the wood veneer data samples to obtain a reference data set including multiple wood veneer data samples.

2. The method for predicting the stain resistance of the wood composite material according to claim 1, characterized in that, The performing of dimensionality reduction fitting on the regression function of the trained machine learning model through a principal component regression algorithm to obtain the structure-activity relationship between the stain resistance performance of wood-based composites and their structural data and production process data includes: For the reference data set D, calculate the features of different types of the same wood veneer data sample and correlation coefficient γ ij , and form all the correlation coefficients γ ij into a correlation coefficient matrix R; and and respectively represent the i-th and j-th features of the k-th wood veneer data sample in the reference data set D, where i, j = 1, 2, …, n, n is the total number of feature types, and k = 1, 2, …, m, m is the total number of wood veneer data samples; Calculate the eigenvalues ($\lambda_1, \lambda_2, \ldots, \lambda$ n ) and eigenvectors $\mu$ i , $\mu$ 2,… , $\mu$ n ; Sort the eigenvalues (λ1, λ2, …, λ n ) in descending order, and calculate the information contribution rate b j corresponding to the j-th eigenvalue and the cumulative contribution rate α p ; Calculate the principal component z corresponding to each eigenvalue j , all principal components z j The set composed of them is the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data; Among them, μ jk represents the k-th dimensional value of the j-th eigenvector.

3. The method for predicting the stain resistance of the wood composite material according to claim 1, characterized in that, The dividing of the reference data set into a training set and a test set, and training the machine learning model based on a multi-layer perceptron constructed by using the training set includes: Dividing the reference data set into a training set and a test set according to a ratio of 6:4; Establishing a machine learning model based on a multi-layer perceptron, the machine learning model includes an input layer, a hidden layer, and an output layer, and any neuron in the upper layer is respectively connected to each neuron in the lower layer; Training and adjusting the parameters of the machine learning model by using the training set; Re-dividing the training set and the test set by using the method of K-fold cross-validation, and repeating the training and parameter adjustment of the machine learning model until a machine learning model capable of predicting the stain resistance performance of wood-based composites is obtained.

4. The method for predicting the stain resistance of the wood composite material according to claim 1, wherein The digitizing and normalizing of the reference data set includes: Digitizing the data in the reference data set through Python technology; Normalize the format and unit of the data in the reference dataset through Python technology.

5. The method for predicting the stain resistance of the wood composite material according to any one of claims 1-4, characterized in that, Before dividing the reference dataset into a training set and a test set and using the training set to train the constructed machine learning model based on a multi-layer perceptron, it also includes: Use a fitting algorithm to evaluate the quality of the digitized and normalized reference dataset, and evaluate the integrity and consistency of the reference dataset; Perform data cleaning on the reference dataset after quality evaluation to remove outliers; use the fitting algorithm to evaluate the quality of the reference dataset again after data cleaning until its integrity and consistency meet the requirements.

6. The method for predicting the stain resistance of the wood composite material according to claim 5, characterized in that, The use of a fitting algorithm to evaluate the quality of the digitized and normalized reference dataset and evaluate the integrity and consistency of the reference dataset includes: Divide the reference dataset into K sample subsets of equal size; Traverse each of the sample subsets in turn, and use the current sample subset of each traversal as the validation set, and the remaining wood veneer data samples as the training set for quality evaluation; Use the statistical value of the evaluation index obtained from the K times of quality evaluation as the final evaluation index, and evaluate the integrity and consistency of the reference dataset through the final evaluation index.

7. The method for predicting the stain resistance of the wood composite material according to claim 5, characterized in that, The data cleaning of the reference dataset after quality evaluation to remove outliers includes: Represent the reference dataset as a two-dimensional array, and for any element at any position in the reference dataset, generate an observation window of a set size centered on the element at that position; Calculate the median and standard deviation of all elements within the observation window; If the absolute value of the difference between the element at any position and the median exceeds a set multiple of the standard deviation, use the median to replace the element at that position.

8. A device for predicting the stain resistance of a wood composite material, characterized in that, The device includes: A reference dataset establishment module for establishing a reference dataset including multiple wood veneer data samples; wherein, each wood veneer data sample includes several features of the wood veneer, the types of the features include structural data and production process data, and the wood veneer data sample is labeled with label data, and the label data includes stain resistance data; A preprocessing module for digitizing and normalizing the reference dataset; A model training module for dividing the reference dataset into a training set and a test set, using the training set to train the constructed machine learning model based on a multi-layer perceptron, and using the test set for verification to obtain a machine learning model capable of predicting the stain resistance of wood-based composites; A structure-activity relationship determination module for performing dimensionality reduction fitting on the regression function of the trained machine learning model through a principal component regression algorithm to obtain the structure-activity relationship between the stain resistance of wood-based composites and their structural data and production process data; The reference dataset establishment module includes: A structural data acquisition unit for acquiring the structural data of the wood veneer collected with the wood veneer as the basic unit during the production process of the wood veneer; Among them, the structure data includes one or more of tree species name, thickness of impregnated film paper, surface color of impregnated film paper, and surface texture of impregnated film paper; A production process data acquisition unit, configured to acquire, during the production process of the wood veneer, the production process data of the wood veneer collected with the wood veneer as a unit; Among them, the production process data includes one or more of surface coating dosage, electron beam curing irradiation dose, and electron beam curing irradiation voltage; A stain resistance performance data acquisition unit, configured to acquire the stain resistance performance data of the produced wood veneer; Among them, the stain resistance performance data includes one or more of water contact angle, oil contact angle, and stain resistance level; A reference data set establishment unit, configured to use the structure data and production process data as the features of the wood veneer data sample, and use the stain resistance performance data as the label data of the wood veneer data sample, to obtain a reference data set including multiple wood veneer data samples.

Images