Digital prediction method and system for air permeability of wool knitted fabric

Abstract

The invention relates to the technical field of knitting quality analysis, in particular to a digital prediction method and system for air permeability of wool knitted fabric, and the method comprises the steps: obtaining indexes and operation parameters of a fabric material and an air permeability index; correlation influence weights are obtained according to the change correlation of the air permeability indexes along with the indexes of the fabric materials; obtaining a same-type fabric set based on the indexes of the fabric materials in combination with the association influence weights; dividing the same-type fabric set according to the operation parameters to obtain a same-parameter fabric set; according to the air permeability indexes of the fabrics in the same-parameter fabric set, the optimal air permeability index of the fabrics is obtained; and constructing a ventilation index prediction model based on the indexes of the fabric material according to the optimal ventilation index and the associated influence weight. The method aims at solving the problems that indexes of different fabric materials have different influence degrees on the air permeability index, and the air permeability index difference is large due to the fact that the same kind of wool knitted fabric is influenced by equipment factors, and the purpose of improving prediction accuracy is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of knitting quality analysis, and in particular to a digital prediction method and system for the air permeability of wool knitted fabrics. Background Art

[0002] In the textile field, the air permeability of fabrics is an important indicator to measure the fabric materials, and the air permeability of fabrics directly affects the wearing comfort and functionality. With the improvement of consumers' requirements for fabric comfort and the expansion of the outdoor sports and functional clothing markets, the demand for the air permeability of wool knitted fabrics is increasing day by day. To meet this demand, textile enterprises are actively seeking more efficient and accurate air permeability prediction methods. Traditional air permeability testing methods usually rely on physical experiments, which require a large number of operations and steps to implement, and it is difficult to predict the air permeability performance of fabrics in advance at the design stage.

[0003] With the development of computer technology and digital simulation, researchers have begun to explore using digital means to predict the air permeability of fabrics through fabric materials; in the field of wool knitted fabrics, the air permeability is affected by various material factors, including fiber type, yarn structure, fabric density, etc. In order to accurately evaluate the air permeability performance of wool knitted fabrics before production, a digital prediction model that can reflect material factors has been established. However, when constructing the digital prediction model of the air permeability of wool knitted fabrics by existing methods, due to the different degrees of influence of the indicators of different fabric materials on the air permeability of wool knitted fabrics, and the different air permeability indexes of the same type of wool knitted fabrics caused by equipment factors during the weaving process, the accuracy of the prediction results of the constructed digital prediction model of the air permeability of wool knitted fabrics decreases. Summary of the Invention

[0004] The present invention provides a digital prediction method and system for the air permeability of wool knitted fabrics to solve the existing problems.

[0005] The digital prediction method and system for the air permeability of wool knitted fabrics of the present invention adopt the following technical solutions:

[0006] An embodiment of the present invention provides a digital prediction method for the air permeability of wool knitted fabrics, and the method includes the following steps:

[0007] Obtain the indicators of the fabric materials and operation parameters of each batch of fabrics, and obtain the air permeability index of the fabrics;

[0008] According to the correlation between the air permeability index and the change of the fabric material indicators in different batches of fabrics, obtain the correlation influence weight of each fabric material indicator on the air permeability index;

[0009] Based on the differences in the indicators of the fabric materials in different batches of fabrics, combined with the associated influence weight of the fabric material indicators on the air permeability index, the fabrics are classified to obtain a set of fabrics of the same type;

[0010] According to the differences in the operation parameters, the fabrics of all batches in the set of fabrics of the same type are classified to obtain a set of fabrics with the same parameters;

[0011] According to the proportion distribution of the air permeability index of the fabrics in the set of fabrics with the same parameters, the air permeability index of each batch of fabrics in the set of fabrics with the same parameters is corrected to obtain the optimal air permeability index of the fabrics;

[0012] According to the optimal air permeability index and the associated influence weight, an air permeability index prediction model based on the fabric material indicators is constructed.

[0013] Preferably, the specific steps for obtaining the associated influence weight include:

[0014] For the m-th fabric material indicator, obtain the indicator values of the m-th fabric material indicator in all batches of fabrics, and denote the sequence composed of the indicator values as the indicator sequence of the m-th fabric material indicator; denote the sequence composed of the air permeability indices of all batches of fabrics as the air permeability index sequence;

[0015] Obtain the fluctuation degree θ of the m-th fabric material indicator according to the indicator sequence m ;

[0016] The corrected discrimination coefficient ρ m of the m-th fabric material indicator is calculated as follows:

[0017]

[0018] where ρ0 is the preset initial discrimination coefficient; M is the number of fabric material indicators, is the mean value of the fluctuation degrees of the M fabric material indicators;

[0019] Obtain the corrected discrimination coefficients of all fabric material indicators, and normalize the corrected discrimination coefficients of each fabric material indicator using the maximum-minimum normalization algorithm to obtain the final discrimination coefficient of each fabric material indicator;

[0020] Take the final discrimination coefficient of each fabric material indicator as the discrimination coefficient parameter of the grey correlation method, and use the grey correlation method to calculate the correlation degree between the indicator sequence of each fabric material indicator and the air permeability index sequence, denoted as the associated influence weight of each fabric material indicator on the air permeability index.

[0021] Preferably, the specific steps for obtaining the fluctuation degree include:

[0022] The calculation method for the degree of fluctuation of the m-th fabric material index is as follows:

[0023]

[0024] where A m is the standard deviation of all the index values in the index sequence of the m-th fabric material index, and N is the number of fabric batches;

[0025] B m,i is the absolute value of the difference between the i-th index value and the (i - 1)-th index value in the index sequence of the m-th fabric material index, and B m,j is the absolute value of the difference between the j-th index value and the (j - 1)-th index value in the index sequence of the m-th fabric material index, and γ1 is a preset first hyperparameter;

[0026] || is the absolute value function.

[0027] Preferably, the specific steps for obtaining the set of fabrics of the same type include:

[0028] 1) Use the elbow method to obtain several first initial cluster centers;

[0029] 2) Obtain the similarity of the fabric material indexes of each batch of fabrics and the fabrics corresponding to each first initial cluster center, and classify each batch of fabrics into each first initial cluster center to obtain several first intermediate clusters, where the classification condition is that the similarity of the fabric material indexes of each batch of fabrics in the first intermediate cluster and the fabric material indexes of the cluster center of the first intermediate cluster is the largest;

[0030] 3) Calculate the average value of the similarity of the fabric material indexes of each fabric in each first intermediate cluster to other fabrics, and record the fabric with the largest average value as the new first cluster center;

[0031] 4) Repeat steps 2) and 3) until no new first cluster centers are generated, and record the batch set of fabrics in the clusters corresponding to each current first cluster center as a set of fabrics of the same type, and obtain several sets of fabrics of the same type.

[0032] Preferably, the specific steps for obtaining the similarity of the fabric material indexes include:

[0033] Record the associated influence weight of the m-th fabric material index on the air permeability index as g m ;

[0034] The similarity A p,q between the fabric material indexes of the p-th batch and the q-th batch of fabrics is calculated as follows:

[0035]

[0036] where γ2 is a preset second hyperparameter, || is the absolute value function; γ3 is a preset third hyperparameter;

[0037] V m,p is the index of the m-th fabric material of the p-th batch of fabrics, and V m,q is the index of the m-th fabric material of the q-th batch of fabrics, and M is the number of indices of fabric materials.

[0038] Preferably, the specific steps for obtaining the same-parameter fabric set include:

[0039] (1) Using the elbow method to obtain several second initial cluster centers in the obtained set of fabrics of the same type;

[0040] (2) Obtaining the consistency of operation parameters between each batch of fabrics and the fabrics corresponding to each second initial cluster center, and classifying each batch of fabrics to each second initial cluster center to obtain several second intermediate clusters, where the classification condition is that the consistency of operation parameters between each batch of fabrics in the second intermediate cluster and the cluster center of the second intermediate cluster is the largest;

[0041] (3) Calculating the mean value of the consistency of operation parameters between each fabric in each second intermediate cluster and other fabrics, and recording the fabric with the largest mean value as the new second cluster center;

[0042] (4) Repeating steps (2) and (3) until no new second cluster centers are generated, and recording the batch set composed of the fabrics in the cluster corresponding to each current obtained second cluster center as a same-parameter fabric set, and obtaining several same-parameter fabric sets.

[0043] Preferably, the specific steps for obtaining the consistency of operation parameters include:

[0044] In the n-th set of fabrics of the same type, the consistency β ′ of operation parameters between the p-th ′ batch of fabrics and the q-th n,p′,q′ batch of fabrics is calculated as follows:

[0045]

[0046] where γ4 is a preset fourth hyperparameter;

[0047] K is the number of operation parameters, and H n,p′,k is the k-th operation parameter of the p-th ′ batch of fabrics in the n-th set of fabrics of the same type, and H n,q′,k is the k-th operation parameter of the q-th ′ batch of fabrics in the n-th set of fabrics of the same type;

[0048] ED() is the Euclidean distance function.

[0049] Preferably, the specific steps for obtaining the optimal air permeability index include:

[0050] If the air permeability index measurement coefficient satisfies being greater than or equal to 1 / 2 of the maximum air permeability index of all fabrics in the same-parameter fabric set, record the air permeability index of each fabric as the optimal air permeability index of each fabric;

[0051] If the air permeability index measurement coefficient satisfies being less than 1 / 2 of the maximum air permeability index of all fabrics in the same-parameter fabric set, record the fabrics in the same-parameter fabric set with an air permeability index less than 1 / 2 of the maximum air permeability index as the to-be-repaired front fabrics in the same-parameter fabric set, record the sum of the air permeability index of the to-be-repaired front fabrics in the same-parameter fabric set and 1 / 2 of the maximum air permeability index as the optimal air permeability index of each to-be-repaired front fabric, and record the air permeability index of the other fabrics in the same-parameter fabric set except the to-be-repaired front fabrics as the optimal air permeability index of each fabric.

[0052] Preferably, the specific steps for obtaining the air permeability index measurement coefficient include:

[0053] Record the difference between the maximum air permeability index and the minimum air permeability index of all fabrics in each same-parameter fabric set as the air permeability index measurement coefficient in the same-parameter fabric set.

[0054] The present invention also proposes a digital prediction system for the air permeability of wool knitted fabrics. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0055] The beneficial effects of the technical solution of the present invention are as follows: By obtaining the indexes of the fabric materials and operation parameters of each batch of fabrics, the air permeability index of the fabrics is obtained; according to the correlation between the air permeability index and the indexes of the fabric materials in different batches of fabrics, the correlation influence weight of each index of the fabric material on the air permeability index is obtained; the influence of different indexes of the fabric material on the air permeability index is highlighted through the correlation influence weight; based on the difference of the indexes of the fabric materials in different batches of fabrics, combined with the correlation influence weight of the indexes of the fabric material on the air permeability index, the fabrics are classified to obtain a set of fabrics of the same type; all fabrics are classified based on the same indexes of the fabric materials, so that the fabrics in the set of fabrics of the same type are fabrics of approximate materials, and the air permeability indexes of the fabrics of approximate materials are approximate; according to the difference of the operation parameters, the fabrics of all batches in the set of fabrics of the same type are classified to obtain a set of fabrics with the same parameters; further classification is carried out under the set of fabrics of the same type to obtain a set of fabrics with the same parameters, eliminating interference factors such as equipment factors under the same type of fabrics; according to the proportion distribution of the air permeability indexes of the fabrics in the set of fabrics with the same parameters, the air permeability index of each batch of fabrics in the set of fabrics with the same parameters is corrected to obtain the optimal air permeability index of the fabrics; according to the optimal air permeability index and the correlation influence weight, an air permeability index prediction model based on the indexes of the fabric materials is constructed. After eliminating the interference of equipment factors under the same type of fabrics, the air permeability indexes in the set of fabrics with the same parameters are optimally corrected, thereby improving the prediction accuracy of the air permeability index prediction model based on the indexes of the fabric materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings without creative efforts based on these drawings.

[0057] Figure 1 It is a flowchart of the steps of a method for digitally predicting the air permeability of a wool knitted fabric according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, describe in detail the specific implementation manners, structures, features and effects of a method and system for digitally predicting the air permeability of a wool knitted fabric according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs.

[0060] The following specifically describes the specific solutions of a method and system for digitally predicting the air permeability of a wool knitted fabric provided by the present invention in conjunction with the accompanying drawings.

[0061] Please refer to Figure 1 , which shows a flowchart of the steps of a method for digitally predicting the air permeability of a wool knitted fabric provided by an embodiment of the present invention. The method includes the following steps:

[0062] Step S001: Obtain the indicators of the fabric material and operating parameters of each batch of fabric, and obtain the air permeability index of the fabric.

[0063] It should be noted that the purpose of this embodiment is to predict the air permeability based on the indicators of the fabric material of the wool knitted fabric, and correct the air permeability in combination with the operating parameters, so as to construct an air permeability index prediction model based on the indicators of the fabric material; therefore, first, it is necessary to collect the indicators of the fabric material and the air permeability index of the fabric used to construct the air permeability index prediction model.

[0064] Specifically, production data of several batches of fabric are obtained from the production record database of the wool weaving factory. The production data includes information on various fabric materials of each batch of fabric and equipment information for producing each batch of fabric; among them, the information on various fabric materials in this embodiment includes the fiber type number, yarn structure number, fabric density, etc. of the fabric; the equipment information for producing each batch of fabric includes the knitting needle pitch, cloth tension, wool feeding speed, etc.

[0065] Furthermore, the values of the information on all fabric materials and the equipment information for producing each batch of fabric are respectively standardized to obtain multiple fabric material indicators and several operating parameters for each batch of fabric;

[0066] It should be noted that when standardizing the values of the fabric material information and equipment information, a unique and different value between 0 and 1 is assigned to the fabric material information or equipment information; as an example, for a fabric with a fabric density of 100, where the fabric density is usually between 80 and 160, the standardized result of the fabric density of this fabric is 0.25, and the larger the value, the higher the fabric density.

[0067] It should be particularly noted that due to the different requirements and uses of fabrics in different batches, the weaving processes of fabrics in different batches are different, which in turn leads to the lack of fabric material information and equipment information for some batches. In this embodiment, the indexes and operation parameters of the fabric materials with null collection results are assigned a value of 0. For example, for different wool knitted products, there are ironing processes and steaming processes during shaping. The ironing process requires air extraction to accelerate cooling, while the steaming process does not require air extraction, so the air extraction process only appears in the batches corresponding to the fabrics that need to be ironed. Therefore, the time of the air extraction process in the steaming process is assigned a value of 0.

[0068] Further, obtain the air permeability index of each batch of fabric in the production record database. The unit of the air permeability index is: one gram of water vapor per square meter per day (g / m2 / 24Hr). The air permeability index is an existing index in the weaving industry and will not be elaborated in this embodiment.

[0069] Step S002: According to the correlation between the air permeability index and the indexes of the fabric materials in different batches of fabrics, obtain the correlation influence weight of each index of the fabric material on the air permeability index.

[0070] It should be noted that the air permeability index of the fabric is mainly determined by the fabric material index that determines the style of the fabric. Therefore, the influence degrees of different fabric material indexes on the air permeability index are different. For example, the size of the fabric density affects the structure and pore size of the fabric, and its influence degree on the air permeability index is greater than that of other indexes of the fabric material. Therefore, in this embodiment, first, by analyzing the change correlation between each fabric material index and the air permeability index in different fabrics, that is, the consistency of the air permeability index with the change of each fabric material index, the correlation influence weight of each index of the fabric material on the air permeability index is obtained.

[0071] Specifically, for the index of the m-th fabric material, obtain the index values of the index of the m-th fabric material in all batches of fabrics, and record the sequence composed of the index values as the index sequence of the index of the m-th fabric material; record the sequence composed of the air permeability indexes of all batches of fabrics as the air permeability index sequence;

[0072] It should be noted that when constructing the index sequence and the air permeability index sequence in this embodiment, the sequence number order of the sequences is the order of the production batches in the production record database of the wool weaving factory, and the orders of the index sequence and the air permeability index sequence correspond one by one.

[0073] It should be noted that in this embodiment, the grey relational analysis method is used to analyze the change relationship between the indexes of each fabric material and the air permeability index in the wool knitted fabric. The grey relational analysis method calculates the correlation degree of different sequences depending on the overall trend and data shape of the data, and the noise existing in the data will affect the data trend, resulting in a decrease in the correlation degree between the index of the fabric material and the air permeability index obtained by the grey relational analysis method.

[0074] It should be further noted that when calculating the correlation degree by the traditional gray correlation degree method, a resolution coefficient needs to be set, and its default value is usually 0.5. The value of the resolution coefficient can reflect the difference and the sensitivity or insensitivity of the correlation between index sequences in the calculation process, and thus affect the accuracy of the calculation result of the correlation degree. When there are outliers in the index sequence of the fabric material index and its fluctuation is larger than that of other index sequences, its sensitivity should be appropriately reduced to avoid the reduction of the correlation degree caused by abnormal data. Therefore, in this embodiment, based on the fluctuation difference between the sequence values and other sequence values in the sequence fluctuations of each fabric material index, the fluctuation degree of each fabric material index is obtained.

[0075] Specifically, the fluctuation degree θ m of the m-th fabric material index is calculated as follows:

[0076]

[0077] where A m is the standard deviation of all index values in the index sequence of the m-th fabric material index, N is the number of fabric batches, that is, the number of sequences in the index sequence of the m-th fabric material index, where j≠i, so j∈(1,N - 1); B m,i is the absolute value of the difference between the i-th index value and the (i - 1)-th index value in the index sequence of the m-th fabric material index, B m,j is the absolute value of the difference between the j-th index value and the (j - 1)-th index value in the index sequence of the m-th fabric material index, γ1 is a preset first hyperparameter used to avoid the denominator being 0. In this embodiment, γ1 = 0.01 is taken as an example, and || is the absolute value function.

[0078] It should be noted that the standard deviation A m is used to measure the trend change of the sequence in the index sequence of the m-th fabric material index. When its change is larger, it indicates that the trend change of the index of this fabric material is more obvious, and it is more difficult to be affected by individual outliers because of its high variability. At this time, the resolution coefficient of the m-th fabric material index should be larger; |B m,i -B m,j | represents the difference in the increments of two sequence values in the index sequence. This value can reflect the change difference of different index values in the sequence. The smaller the difference, the more similar the change of the data in the sequence. Through the standard deviation and the change similarity of the index values, it reflects that the sequence has a large change amplitude and similar changes, thus hiding abnormal manifestations, that is, it can better reflect the change relationship of the index of this fabric material. Therefore, when analyzing its correlation with the air permeability index by the gray correlation degree method, the index of this fabric material can better reflect the trend of the data.

[0079] It should be noted that after obtaining the fluctuation degrees of the indexes of each fabric material, in this embodiment, the discrimination coefficient is increased for the indexes of the fabric materials whose fluctuation degrees are greater than the average value of the fluctuation degrees of the indexes of all fabric materials, and the discrimination coefficient is decreased for the indexes of the fabric materials whose fluctuation degrees are less than the average value of the fluctuation degrees of the indexes of all fabric materials, so that the calculated correlation degree depends on the change trends and abnormal manifestations of the indexes of different fabric materials.

[0080] Preferably, the specific method for adjusting the discrimination coefficient based on the fluctuation degree of the index of the fabric material to obtain the final discrimination coefficient of the index of each fabric material is as follows:

[0081]

[0082] Among them, ρ0 is the preset initial discrimination coefficient, and in this embodiment, ρ0 = 0.5 is taken as an example for description; M is the number of indexes of the fabric material, is the average value of the fluctuation degrees of the M indexes of the fabric material, θ m is the fluctuation degree of the index of the m-th fabric material; ρ m is the corrected discrimination coefficient of the index of the m-th fabric material.

[0083] Furthermore, the corrected discrimination coefficients of the indexes of all fabric materials are obtained, and the corrected discrimination coefficients of the indexes of each fabric material are normalized using the maximum-minimum normalization algorithm to obtain the final discrimination coefficient of the index of each fabric material; the maximum-minimum normalization algorithm is a well-known prior art and will not be elaborated in this embodiment.

[0084] Furthermore, the final discrimination coefficient of the index of each fabric material is used as the discrimination coefficient parameter of the grey correlation degree method, and the correlation degree between the index sequence of the index of each fabric material and the air permeability index sequence is calculated using the grey correlation degree method, denoted as the correlation influence weight of the index of each fabric material on the air permeability index. The value range of the correlation influence weight is [0, 1]. It should be noted that calculating the correlation degree between two sequences using the grey correlation degree method is a well-known technique, and the specific calculation process is not limited in this embodiment.

[0085] Step S003: Based on the difference situations of the indexes of the fabric materials in different batches of fabrics, combined with the correlation influence weight of the indexes of the fabric materials on the air permeability index, the fabrics are classified to obtain the same-type fabric set; according to the difference of the operation parameters, the fabrics of all batches in the same-type fabric set are classified to obtain the same-parameter fabric set.

[0086] It should be noted that the purpose of this embodiment is to utilize the influence relationship between the indexes of the fabric material and the air permeability index, and then construct a prediction model of the air permeability index based on the indexes of the fabric material. Therefore, the indexes of the fabric material that affect the air permeability index and the corresponding air permeability index need to be used as the original data for training the prediction model.

[0087] Furthermore, it should be noted that due to the differences in the indexes of the fabric materials of different batches, the air permeability indexes of the produced fabrics are different. For different batches of fabrics with the same indexes of the fabric material, the differences in equipment parameters or the quality of wool raw materials, as well as the influence factors such as the vibration of wool during the operation of the equipment, change the pore structure and spacing of wool fibers during the weaving process, and then lead to different air permeability indexes in different regions of some batches of fabrics. The air permeability indexes obtained through sampling detection are different in different regions, that is, there are problems of untrustworthiness of the air permeability index of fabrics with the same indexes of the fabric material and a large distribution range of the collected air permeability indexes. Then, when using the untreated original data to construct a prediction model for the influence of the air permeability index, the relationship between the indexes of the fabric material and the air permeability index will be weakened, and then the accuracy of the constructed prediction model will be reduced. Therefore, in this embodiment, first, based on the similarity of the indexes of the fabric materials of different batches of fabrics, combined with the associated influence weights of the indexes of each fabric material on the air permeability index, all batches of fabrics are divided to obtain several sets of fabrics of the same type, and each set of fabrics of the same type contains several fabrics with the same or similar fabric material information.

[0088] Specifically, the specific steps for dividing all batches of fabrics into several sets of fabrics of the same type according to the similarity of the indexes of the fabric materials of different batches of fabrics, combined with the associated influence weights of the indexes of each fabric material on the air permeability index, are as follows:

[0089] The associated influence weight of the index of the m-th fabric material on the air permeability index is denoted as g m ;

[0090] The similarity A p,q of the indexes of the fabric materials of the p-th batch and the q-th batch of fabrics is calculated as follows:

[0091]

[0092] where γ2 is a preset second hyperparameter used to avoid the denominator being zero. In this embodiment, γ2 = 0.1 is taken as an example, and || is the absolute value function; γ3 is a preset third hyperparameter. In this embodiment, γ3 = 0.5 is taken as an example, which is used to make the range of γ3 + g m be [0.5, 1.5]; V m,p is the m-th index of the fabric material of the p-th batch of fabrics, and V m,qis the index of the m-th fabric material for the q-th batch of fabrics, and M is the number of fabric material indexes.

[0093] It should be noted that |V m,p -V m,q | represents the similarity of the indexes of the same fabric material for different batches of fabrics. The smaller the value, the greater the similarity. Then the similarity A of the fabric material indexes of the p-th batch and the q-th batch of fabrics p,q also has a larger value; since the set of fabrics of the same type to be constructed in this embodiment is different batches of fabrics with the indexes of the same fabric material, and its ultimate purpose is to obtain the corresponding air permeability index under this type of fabric through the same fabric, so it is necessary to divide different batches of fabrics by the indexes of the key fabric materials as much as possible. Therefore, in this embodiment, the weight of the associated influence of the fabric material index on the air permeability index is used as the weight for calculating the similarity of different fabric material indexes, so that the calculated similarity of different fabrics only considers the key fabric material indexes, eliminates the fabric material indexes that have nothing to do with the air permeability index, and changes the value range of the associated influence weight through the third hyperparameter, so as to use the associated influence weight to have an adaptive enhancement effect on the similarity.

[0094] Further, obtain the similarity of the fabric material indexes of every two batches of fabrics.

[0095] It should be noted that after obtaining the similarity of the fabric material indexes of every two batches of fabrics, in this embodiment, the similarity is used as a measure to measure whether the fabrics belong to the same kind of fabric, and then all batches of fabrics are divided to obtain several sets of fabrics of the same type.

[0096] Preferably, the specific steps for dividing the fabrics according to the similarity of the fabric material indexes of different batches of fabrics to obtain several sets of fabrics of the same type are as follows:

[0097] 1) Use the elbow method to obtain several first initial clustering centers, where each first initial clustering center is a batch of fabrics;

[0098] 2) Obtain the similarity of the fabric material indexes of each batch of fabrics and the fabrics corresponding to each first initial clustering center, and classify each batch of fabrics into each first initial clustering center to obtain several first intermediate clusters. The classification condition is that the similarity of the fabric material indexes of each batch of fabrics in the first intermediate cluster and the clustering center of the first intermediate cluster is the largest;

[0099] 3) Calculate the mean value of the similarity of the fabric material indexes of each fabric in each first intermediate cluster to other fabrics, and record the fabric with the largest mean value as the new first clustering center;

[0100] 4) Repeat steps 2) and 3) until no new first clustering centers are generated. Denote the batch set composed of the fabrics in the cluster corresponding to each currently obtained first clustering center as a same-type fabric set, and obtain several same-type fabric sets.

[0101] It should be noted that the fabric material indexes of different batches of fabrics in the same-type fabric set are the same, indicating fabrics with the same design, treatment, requirements, etc. As the same kind of fabric, the air permeability indexes of the same kind of fabric are approximately the same. However, in the actual production process, there are also the influences of operating parameters on the air permeability index of the fabric. The purpose of this embodiment is to optimally correct the air permeability index of the same kind of fabric, reduce the distribution range of the air permeability index, and thus improve the prediction accuracy of the air permeability index prediction model. However, due to the influence of operating parameters, there are tolerances in the woven wool knitted fabrics. Therefore, in this embodiment, according to the operating parameters of different batches of fabrics in each same-type fabric set, the operating parameter consistency of every two fabrics is obtained, and the fabrics in the same-type fabric set are divided to obtain several same-parameter fabric sets.

[0102] Preferably, the specific steps for obtaining the operating parameter consistency of every two batches of fabrics according to the operating parameters of different batches of fabrics in each same-type fabric set are as follows:

[0103] In the nth same-type fabric set, for the pth ′ batch of fabric and the qth ′ batch of fabric, the calculation method of the operating parameter consistency β n,p′,q′ is as follows:

[0104]

[0105] where γ4 is a preset fourth hyperparameter used to avoid the denominator being zero. In this embodiment, γ4 = 0.1 is taken as an example; K is the number of operating parameters, H n,p′,k is the kth operating parameter of the pth ′ batch of fabric in the nth same-type fabric set, and H n,q′,k is the kth operating parameter of the qth ′ batch of fabric in the nth same-type fabric set; ED() is the Euclidean distance function.

[0106] Furthermore, the specific steps for dividing the fabrics in the same-type fabric set according to the operating parameter consistency of every two batches of fabrics in each same-type fabric set to obtain several same-parameter fabric sets are as follows:

[0107] (1) Use the elbow method to obtain several second initial clustering centers in the obtained same-type fabric set, where each second initial clustering center is a batch of fabric in the same-type fabric set;

[0108] (2) Obtain the consistency of the operation parameters between the fabrics of each batch and the fabrics corresponding to each second initial cluster center, and classify the fabrics of each batch into each second initial cluster center to obtain several second intermediate clusters. The classification condition is that the consistency of the operation parameters between each batch of fabrics in the second intermediate cluster and the cluster center of the second intermediate cluster is the largest;

[0109] (3) Calculate the mean value of the consistency of the operation parameters between each fabric in each second intermediate cluster and other fabrics, and record the fabric with the largest mean value as the new second cluster center;

[0110] (4) Repeat steps (2) and (3) until no new second cluster center is generated. Then, record the batch set of fabrics in the cluster corresponding to each currently obtained second cluster center as a same-parameter fabric set, and obtain several same-parameter fabric sets.

[0111] It should be noted that the steps of the method for obtaining the same-type fabric set and the same-parameter fabric set are the steps of the K-means clustering algorithm. The K-means clustering algorithm and the elbow method described in this embodiment are both well-known existing technologies, and will not be elaborated in this embodiment.

[0112] Step S004: According to the proportion distribution of the air permeability index of the fabrics in the same-parameter fabric set, correct the air permeability index of each batch of fabrics in the same-parameter fabric set to obtain the optimal air permeability index of the fabrics; construct an air permeability index prediction model based on the index of the fabric material according to the optimal air permeability index and the associated influence weight.

[0113] It should be noted that the same-parameter fabric set represents fabrics with the same equipment operation parameters and the same type, so their air permeability indices are approximate. However, due to tolerances in the operation, the fluctuation range of the air permeability index may be too large. Therefore, it is necessary to correct the air permeability index of the fabrics in the same-parameter fabric set with a large fluctuation range. And since the air permeability index of the fabric depends on the area of the fabric with the highest air permeability, that is, the gas that cannot pass through the area with a low air permeability index will pass through the area with a high air permeability index of the same batch of fabrics. Therefore, in this embodiment, when correcting the air permeability index, the optimal correction method is used for correction.

[0114] Preferably, record the difference between the maximum air permeability index and the minimum air permeability index of all fabrics in each same-parameter fabric set as the air permeability index measurement coefficient in the same-parameter fabric set; the specific steps for correcting the air permeability index of each fabric according to the air permeability index measurement coefficient in the same-parameter fabric set to obtain the optimal air permeability index of each fabric are as follows:

[0115] (1) If the air permeability index measurement coefficient satisfies being greater than or equal to 1 / 2 of the maximum air permeability index of all fabrics in the fabric set with the same parameters, it indicates that the air permeability index distribution of all fabrics in the fabric set with the same parameters is smaller compared to the maximum air permeability index distribution. Therefore, no correction is required, and the air permeability index of each fabric is recorded as the optimal air permeability index of each fabric.

[0116] (2) If the air permeability index measurement coefficient satisfies being less than 1 / 2 of the maximum air permeability index of all fabrics in the fabric set with the same parameters, it indicates that the air permeability index distribution of all fabrics in the fabric set with the same parameters is larger compared to the maximum air permeability index distribution. Therefore, optimal correction is required. The fabrics in the fabric set with the same parameters that are less than 1 / 2 of the maximum air permeability index are recorded as the fabrics to be corrected in the fabric set with the same parameters. The sum of the air permeability index of the fabrics to be corrected in the fabric set with the same parameters and 1 / 2 of the maximum air permeability index is recorded as the optimal air permeability index of each fabric to be corrected. The air permeability index of the other fabrics in the fabric set with the same parameters except the fabrics to be corrected is recorded as the optimal air permeability index of each fabric.

[0117] Among the indicators of all fabric materials, the indicators of the fabric materials that satisfy the weight of the associated influence of the fabric material indicators on the air permeability index being greater than the preset association threshold are recorded as the indicators of the fabric materials that can be referred to.

[0118] The indicators of the fabric materials that can be referred to and the air permeability index of each fabric form a sample. All samples are obtained to construct a sample set, and an air permeability index prediction model based on the indicators of the fabric materials is trained through the sample set.

[0119] It should be noted that the air permeability index prediction model based on the indicators of the fabric materials described in this embodiment is a neural network model. In this embodiment, a CNN model is used for training, and the loss function for training is the LOSS loss function.

[0120] Another embodiment of the present invention provides a digital prediction system for the air permeability of wool knitted fabrics. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the above method steps S001 to step S004 are implemented.

[0121] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A digital prediction method for the breathability of a wool knitted fabric, characterized in that, The method includes the following steps: Obtain the indexes of the fabric materials and operation parameters of each batch of fabrics, and obtain the air permeability index of the fabrics; According to the correlation between the air permeability index and the indexes of the fabric materials in different batches of fabrics, obtain the correlation influence weight of each index of the fabric material on the air permeability index; Based on the differences in the indexes of the fabric materials in different batches of fabrics, combine the correlation influence weight of the indexes of the fabric materials on the air permeability index to divide the fabrics, and obtain a set of fabrics of the same type; Divide the fabrics of all batches in the set of fabrics of the same type according to the differences in the operation parameters, and obtain a set of fabrics with the same parameters; According to the proportion distribution of the air permeability indexes of the fabrics in the set of fabrics with the same parameters, correct the air permeability index of each batch of fabrics in the set of fabrics with the same parameters, and obtain the optimal air permeability index of the fabrics; According to the optimal air permeability index and the correlation influence weight, construct an air permeability index prediction model based on the indexes of the fabric materials.

2. The digital prediction method for the air permeability of a wool knitted fabric according to claim 1, wherein, The specific steps for obtaining the correlation influence weight include: For the m-th index of the fabric material, obtain the index values of the m-th index of the fabric material in all batches of fabrics, and denote the sequence composed of the index values as the index sequence of the m-th index of the fabric material; denote the sequence composed of the air permeability indexes of all batches of fabrics as the air permeability index sequence; Obtain the fluctuation degree θ of the index of the m-th fabric material according to the index sequence m ; The correction discrimination coefficient ρ of the m-th fabric material index m is calculated as follows: where ρ0 is a preset initial resolution coefficient; M is the number of indices of the fabric material, which is the mean of the fluctuation degrees of the M indices of the fabric material; Obtain the corrected discrimination coefficients of all indexes of the fabric materials, and normalize the corrected discrimination coefficients of each index of the fabric material using the maximum-minimum normalization algorithm to obtain the final discrimination coefficients of each index of the fabric material; Take the final discrimination coefficient of each index of the fabric material as the discrimination coefficient parameter of the grey correlation degree method, and use the grey correlation degree method to calculate the correlation degree between the index sequence of each index of the fabric material and the air permeability index sequence, which is denoted as the correlation influence weight of each index of the fabric material on the air permeability index.

3. The digital prediction method for the air permeability of a wool knitted fabric according to claim 2, wherein The specific steps for obtaining the degree of fluctuation include: The calculation method for the degree of fluctuation of the m-th index of the fabric material is: Among them, A m is the standard deviation of all the index values in the index sequence of the m-th fabric material index, and N is the number of fabric batches; B m,i is the absolute value of the difference between the i-th index value and the (i - 1)-th index value in the index sequence of the m-th fabric material's index, B m,j is the absolute value of the difference between the j-th index value and the (j - 1)-th index value in the index sequence of the m-th fabric material's index, and γ1 is a preset first hyperparameter; || is the absolute value function.

4. The digital prediction method for the air permeability of a wool knitted fabric according to claim 1, wherein, The specific steps for obtaining the set of fabrics of the same type include: 1) Use the elbow method to obtain several first initial clustering centers; 2) Obtain the similarity of the indexes of the fabric materials between each batch of fabrics and the fabrics corresponding to each first initial clustering center, and classify each batch of fabrics into each first initial clustering center to obtain several first intermediate clusters, and the classification condition is that the similarity of the indexes of the fabric materials between each batch of fabrics in the first intermediate cluster and the clustering center of the first intermediate cluster is the largest; 3) Calculate the mean value of the similarity of the indexes of the fabric materials between each fabric in each first intermediate cluster and other fabrics, and denote the fabric with the largest mean value as the new first clustering center; 4) Repeat steps 2) and 3) until no new first clustering centers are generated, and denote the set of batches of fabrics composed of the fabrics in each cluster corresponding to the current obtained first clustering centers as a set of fabrics of the same type, and obtain several sets of fabrics of the same type.

5. The digital prediction method for the air permeability of a wool knitted fabric according to claim 4, wherein, The specific steps for obtaining the similarity of the indexes of the fabric materials include: Let the weight of the correlation influence of the index of the m-th fabric material on the air permeability index be denoted as g m ; Similarity A of the fabric material indicators of the p-th batch and the q-th batch of fabrics p,q The calculation method is as follows: where γ2 is a preset second hyperparameter, || is the absolute value function; γ3 is a preset third hyperparameter; V m,p is the index of the m-th fabric material of the p-th batch of fabrics, V m,q is the index of the m-th fabric material of the q-th batch of fabrics, and M is the number of indices of fabric materials.

6. The digital prediction method for the air permeability of a wool knitted fabric according to claim 1, characterized in that The specific steps for obtaining the set of fabrics with the same parameters include: (1) Use the elbow method to obtain several second initial cluster centers in the set of fabrics of the same type; (2) Obtain the consistency of operation parameters between each batch of fabrics and the fabrics corresponding to each second initial cluster center, and classify each batch of fabrics into each second initial cluster center to obtain several second intermediate clusters, where the classification condition is that the consistency of operation parameters between each batch of fabrics in the second intermediate cluster and the cluster center of the second intermediate cluster is the largest; (3) Calculate the mean value of the consistency of operation parameters between each fabric in each second intermediate cluster and other fabrics, and record the fabric with the largest mean value as the new second cluster center; (4) Repeat steps (2) and (3) until no new second cluster centers are generated. Then, record the batch set of fabrics in the cluster corresponding to each currently obtained second cluster center as a same-parameter fabric set, and obtain several same-parameter fabric sets.

7. The digital prediction method for the air permeability of a wool knitted fabric according to claim 6, wherein The specific steps for obtaining the consistency of operation parameters are as follows: In the nth set of fabrics of the same type, the operation parameter consistency β between the p'-th batch of fabrics and the q'-th batch of fabrics n,p′,q′ is calculated as follows: where γ4 is a preset fourth hyperparameter; K is the number of operation parameters, H n,p′,k is the k-th operation parameter of the p'-th batch of fabrics in the n-th set of fabrics of the same type, H n,q′,k is the k-th operation parameter of the q'-th batch of fabrics in the n-th set of fabrics of the same type; ED() is the Euclidean distance function.

8. The digital prediction method for the air permeability of a wool knitted fabric according to claim 1, characterized in that, The specific steps for obtaining the optimal air permeability index are as follows: If the air permeability index measurement coefficient is greater than or equal to 1 / 2 of the maximum air permeability index of all fabrics in the same-parameter fabric set, record the air permeability index of each fabric as the optimal air permeability index of each fabric; If the air permeability index measurement coefficient is less than 1 / 2 of the maximum air permeability index of all fabrics in the same-parameter fabric set, record the fabrics in the same-parameter fabric set with an air permeability index less than 1 / 2 of the maximum air permeability index as the to-be-repaired fabrics in the same-parameter fabric set, record the sum of the air permeability index of the to-be-repaired fabrics in the same-parameter fabric set and 1 / 2 of the maximum air permeability index as the optimal air permeability index of each to-be-repaired fabric, and record the air permeability index of the other fabrics in the same-parameter fabric set except the to-be-repaired fabrics as the optimal air permeability index of each fabric.

9. The digital prediction method for the air permeability of a wool knitted fabric according to claim 8, wherein The specific steps for obtaining the air permeability index measurement coefficient are as follows: Record the difference between the maximum air permeability index and the minimum air permeability index of all fabrics in each same-parameter fabric set as the air permeability index measurement coefficient in the same-parameter fabric set.

10. A digital prediction system for the air permeability of a wool knitted fabric, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for digitally predicting the air permeability of a wool knitted fabric according to any one of claims 1-9.

Images