Factor Analysis Plays an Important Role in Quality Evaluation of Cashmere Knitting Yarn

In this paper, a prediction model of process parameters and performance in meltblown nonwovens is established by using the method of proximity support vector regression, and satisfactory results are obtained. Based on this, the experimental conditions are simulated by computer and the performance of the corresponding conditions is predicted. Optimum process conditions: The optimum process conditions of filtration efficiency are: temperature 270°C, DCD 15-19 cm, and quantitative 100 g/m2. The optimum process conditions for air permeability are temperature 270-300°C, DCD 20 cm, and basis weight 160 g. Near /m2. This parameter will have reference value for future experiments.

Yarn quality control is an important part of textile production. The quality of the yarn directly affects the quality of the final production process and the quality of the final product. In recent years, the industry scholars have achieved some certainty in the evaluation theory of yarn quality. Progress, such as the quantitative analysis of yarn strength, elongation at break, unevenness of strips, unevenness, unevenness, roughness, details, etc., using multivariate statistical theory Principal component analysis, hierarchical cluster analysis Laws and other methods to evaluate, the results often only stay in the analysis of a single index or a single index of the same type of extraction, and the comprehensive analysis of multiple indicators is less [1-2]. The common factor value in the actual problem is more The smaller the result is, the better the common factor is, and the better the result is, the more comprehensive the results obtained by the common factor are not easy to interpret. In order to effectively deal with the comprehensive statistics, this paper proposes the concept of negative indicators, combined with SPSS13. .0 statistical analysis software, established a comprehensive evaluation model of yarn quality.

1 Factor Analysis Principles and Calculation Procedures 1.1 Factor Analysis Principles Factor analysis methods mainly use the idea of ​​dimensionality reduction to ensure that indicators with certain correlations are represented by a factor to ensure that the evaluation results are easy to operate and To be objective and avoid duplicate information between indicators.
1.2 Factor Analysis Calculation Steps When factor analysis is used to solve practical problems, the following three steps are often taken:
(1) The standardized transformation of the collected raw data is to eliminate the influence of the dimension and the difference in the magnitude of the indicator. Let the observed value of n samples p indicators be X={xij}n×(pi=1, 2,...,n;j=1,2,...,p), then the normalized index matrix is ​​Z={Zij}n×p, Zij= , (i=1,2,...,n;j=1,2,...,p). (2) Find the eigenvalues ​​and eigenvectors of the correlation matrix R, determine the number of factors and the contribution weight of each index to each factor. The introduction of the correlation matrix R = (rjk) p × p = 1n-1ZZT, obviously, R is a symmetric matrix and the elements on the main diagonal are 1, the eigenvalues ​​of the correlation matrix R in descending order of λ1 ≥ λ2 ≥...≥λp≥0, the corresponding feature vector u1,u2,...,up composes the vector matrix U:U=(u1,u2,...,up)p×p, saying that A=(姨λ1u1,姨λ2u2,... , 姨λpup) is the factor load matrix.
(3) Find the common factor to determine the comprehensive statistics and calculate the comprehensive score. According to the Thomson regression estimation method, the expression of the i-th common factor is known as Fi=βi1z1+βi2z2+...+βipzp, (i=1,2,..., k) Where: (βi1, βi2, ..., βip) is the ith row element of the matrix ATR-1, then the integrated statistics Where k is the number of eigenvalues ​​of λi>1.
2 Data collection and processing 2.1 Experimental methods Commonly used indicators affecting yarn quality are average elongation, minimum elongation, strength, dry CV value, thick section, details, and neps, so this article uses YG136 dry uniformity respectively. The above indicators were tested on 50 randomly selected batches of 38.46 tex cashmere knitting yarns, such as instrument, Y331, and YG029F benchtop automatic strength meter. The data are shown in Table 1.
2.2 Data Processing and Analysis In order to find out the relationship between the quality of the yarn and the various variables for comprehensive scoring, SPSS [3] is used for data analysis. The results are shown in Table 2 to Table 6. You can see from Table 2. The KMO statistic is 0.617, which is less than 0.7, indicating that the overlap of the information between the variables may not be particularly high, and the factor analysis model that may be made is not perfect; however, if the significance level of the Bartlett's sphere test is less than 0.05, each The independent assumption of variable, that is to say that the variables have a high correlation, so this method is worth trying

Table 1 Cashmere knitting yarn index value

As can be seen from Table 3, the common factors that can be extracted from the original information contained in almost all variables are represented by more than 80%, so the extracted common factors have strong explanatory power to each variable.

SPSS only extracted the first three common factors, and the first three factors have a rate of 82.384%, which is enough to describe the quality of the yarn. It can be seen from Table 5 that the first common factor reflects the yarn from the dry CV value. Quality is good or bad, so the naming sub-element; the second common factor reflects the quality of the yarn in terms of the average elongation and the minimum elongation. Therefore, the strength is determined by the common factor to reflect the yarn quality from the details. , for the detail factor. From Table 6 we can see that the expression of each common factor is:
F1=-0.069 average elongation +0.02 minimum elongation
0.043 Intensity + 0.38 negative dry CV +
0.394 negative thick-0.014 negative detail +
0.355 negative nep
F2 = 0.386 average elongation + 0.348 minimum elongation
0.406 Strength + 0.02 Negative Dry CV Value -
0.008 Negative Thickness + 0.033 Negative Details - 0.1 Negative
F3=0.032 average elongation -0.157 minimum elongation
0.237 Intensity + 0.078 Negative Dry CV Value -
0.012 Negative Coarse + 0.92 Negative Details - 0.09 Negative Covariance contribution rate for each common factor as a weight, combined with statistics


It can be seen that the relationship between the comprehensive statistics F and each index is:



(factor score coefficient matrix) T × (average elongation,
Minimum elongation, strength, negative dry CV value, negative thick section,
Negative details, negative neps) T=0.172 898 Average elongation +
0.140 746 minimum elongation + 0.118 692 intensity +
0.185 378 negative dry CV + 0.193 443 negative thick +
0.148 798 negative detail + 0.185 511 negative nep =
0.172 898 average elongation +0.140 746 minimum elongation +
0.118 692 Intensity - 0.185 378 dry CV -
0.193 443 thick section -0.148 798 detail -
0.185 511 Neps

Substituting the data in Table 1 into the above formula yields sample No. 39 with F = 1.19, which has the largest value, indicating the best quality; and sample No. 23 with F = -1.73, which has the smallest value and is inferior to this sample.

Each process parameter randomly generates 10,000 random numbers. The range of random numbers is as follows: temperature 220 to 300°C, DCD 5 to 30 cm, and quantification 0 to 500 g/m2. Part of the simulation results of filtration efficiency and air permeability The optimum process conditions are shown in Table 4. (For limited space, only a maximum of 2 decimal places are reserved.

Gym tight wear

Guangzhou LIDONG Garment Industrial Co., Ltd. , https://www.lidonggarment.com