CN104715136B - A kind of method of overall merit spinning process level - Google Patents

Abstract

The invention discloses a kind of method of overall merit spinning process level, it is first by detecting index and the parameter of main expression spinning material performance by professional equipment instrument, use mathematical statistics analysis means to draw and express spinning material each principal component numerical value of different nature, and the comprehensive evaluation value that further obtains expressing spinning material bulk property; The comprehensive evaluation value that obtains yarn quality by same method, by after yarn quality comprehensive evaluation value and the standardization of spinning material comprehensive evaluation value, subtracts each other and obtains the comprehensive evaluation value of standardized spinning process level. The present invention's spinning assessment of levels standard that objectifies, simplifies evaluation procedure and method.

Description

A kind of method of overall merit spinning process level

Technical field

The invention belongs to spinning detection technique field, be specifically related to a kind of method of overall merit spinning process level.

Background technology

In scientific experimentation or production practices, usually need various various yarn qualities of different nature and spinning material to enter at presentRow relatively. But the each individual event character to yarn quality and spinning material is measured, often exist each character height different, instituteThe conclusion obtaining inconsistent situation. Run into the every test index parameter values size of yarn quality, spinning material conflictingWhen situation, engineering and scientific research personnel can only be by rule of thumb and the quality of feeling to judge yarn qualities, and yarn quality evaluation of estimateDirectly affect spinning process assessment of levels value with the evaluation of estimate of spinning material, thereby need a kind of simple effective method overall meritSpinning process level.

Summary of the invention

The present invention, by using computer software to arrange operational data, by each detection numerical value of spinning material, yarn quality, changesBe counted as each principal component for being expressed as spinning material, yarn quality, former each detection that each principal component has comprised different proportionThe numerical value of item index, and the contribution proportion of yarn quality, spinning material being expressed by each main composition, linear expression yarn quality,The comprehensive evaluation value of spinning material, the evaluation of estimate of yarn quality deducts the evaluation of estimate of spinning material, must evaluate spinning process levelEvaluation of estimate.

Its technical solution comprises:

A method for overall merit spinning process level, comprises the following steps successively:

Step 1, first, is used correlation test instrument to detect the each performance indications of spinning material, obtains detecting numerical value; Then to detectingNumerical value is compiled; Finally, reduced data is inputted to data processing unit and processes, its processing method comprise withLower sub-step:

A, collection p dimension random vector X=(x₁,x₂,...,x_p)＇，X_i＝(x_i1,x_i2,...,x_ip) ＇, wherein, i=1,2 ... n, n is sampleNumber of times and the n of observation > p, form sample battle array, to sample array element according to carrying out standardized transformation as shown in the formula (1), (2), (3),Obtain standardization battle array,

$z_{ij}=\frac{x_{ij}-{\stackrel{\‾}{x}}_j}{s_j},---\left(1\right)$

${\stackrel{\‾}{x}}_j=\frac{{\mathrm{\Σ}}_{i=1}^n{x}_{ij}}{n},---\left(2\right)$

$s_j^2=\frac{{\mathrm{\Σ}}_{i=1}^n{(x_{ij}-{\stackrel{\‾}{x}}_j)}^2}{n-1},---\left(3\right)$

In above formula (1), (2), (3), i=1,2 ..., n; J=1,2 ..., p_，Wherein, p=9, x₁，x₂，...x_pGeneration successivelyThe principal length of table spinning material, quality length, short fiber content, maturity, fiber strength, horse value, percentage of impurity, fiber moisture regainRate and raw cotton fault;

B, according to formula (4), to standardization battle array, ask correlation matrix;

$R={\[r_{ij}\]}_p{x}_p=\frac{Z^{T}Z}{n-1},---\left(4\right)$

Wherein, $r_{ij}=\frac{{\mathrm{\Σz}}_{kj}\·z_{ki}}{n-1},$ i,j＝1,2,...p；

The characteristic equation ︱ R-λ I of c, solution sample correlation matrix R_P︱=0 obtains p characteristic root, determines principal component, according to formula (5)Determine m value, the utilization rate of information is reached more than 55%, to each λ in formula (5)_j, j=1,2 ..., m, Rb=λ solves an equation_jb,Obtain unit character vector b_j ⁰；

D, according to formula (6), the target variable after standardization is converted to principal component,

U_ij＝z_i ^Tb_j ⁰,j＝1,2,...,m；(6)

U₁Be called first principal component, U₂Be called Second principal component, ..., U_pBe called p principal component;

E, m principal component carried out to overall merit,

M principal component is weighted to summation, obtains final evaluation of estimate, the variance contribution ratio that flexible strategy are each principal component;

Step 2, first, is used correlation test instrument to detect yarn quality property indices, obtains detecting numerical value; Then,The detection numerical value obtaining is compiled; Finally, reduced data is inputted to data processing unit and process, compriseFollowing sub-step:

F, collection p dimension random vector X=(x₁,x₂,...,x_p)＇，X_i＝(x_i1,x_i2,...,x_ip) ＇, wherein, i=1,2 ... n, n is sampleNumber of times and the n of observation > p, form sample battle array, to sample array element according to carrying out standardized transformation as shown in the formula (7), (8), (9),Obtain standardization battle array,

$Z_{ij}=\frac{x_{ij}-{\stackrel{\‾}{x}}_j}{s_j},---\left(7\right)$

${\stackrel{\‾}{x}}_j=\frac{{\mathrm{\Σ}}_{i=1}^n{x}_{ij}}{n},---\left(8\right)$

$s_j^2=\frac{{\mathrm{\Σ}}_{i=1}^n{(x_{ij}-{\stackrel{\‾}{x}}_j)}^2}{n-1},---\left(9\right)$

In above formula (7), (8), (9), i=1,2 ..., n; J=1,2 ..., p_，Wherein, p=24, x₁，x₂，...x_pFor describingThe variable of yarn quality, it represents weight CV%, deviation of weight, yarn regain %, powerful CV%, fracture strength, fracture successivelyPowerful, minimum brute force, elongation at break, elongation CV%, the dry CV% of bar, the dry CVb of bar, details-50%, slubbing+50%, cotton knot+ 200%, filoplume, filoplume CV%, cotton knot, cotton assorted, the twist, uneven twist, twist factor, empty ingot rate, instantaneous end breakage rate, cylinder yarnRegain %;

G, according to formula (10), to standardization battle array, ask correlation matrix;

$R={\[r_{ij}\]}_p{x}_p=\frac{Z^{T}Z}{n-1},---\left(10\right)$

Wherein, $r_{ij}=\frac{{\mathrm{\Σz}}_{kj}\·z_{ki}}{n-1},$ i,j＝1,2,...p；

The characteristic equation ︱ R-λ I of h, solution sample correlation matrix R_P︱=0 obtains p characteristic root, determines principal component, according to formula (11)Determine m value, the utilization rate of information is reached more than 55%, to each λ in formula (11)_j, j=1,2 ..., m, Rb=λ solves an equation_jb,Obtain unit character vector b_j ⁰；

K, according to formula (12), the target variable after standardization is converted to principal component,

U_ij＝z_i ^Tb_j ⁰,j＝1,2,...,m；(12)

U₁Be called first principal component, U₂Be called Second principal component, ..., U_pBe called p principal component;

L, m principal component carried out to overall merit,

M principal component is weighted to summation, obtains final evaluation of estimate, the variance contribution ratio that flexible strategy are each principal component;

The evaluation of estimate of the yarn quality that step 3, above-mentioned steps two obtain deducts the evaluation of estimate of the spinning material of step 1 acquisition,Obtain the initial evaluation value of spinning process level; By after yarn quality comprehensive evaluation value and the standardization of spinning material comprehensive evaluation value, phaseSubtract and obtain the comprehensive evaluation value of spinning process level.

The useful technique effect that the present invention brings:

First the present invention carries out integrated treatment by the parameter of all technical that spinning material laboratory of spinnery is recorded and dividesAnalyse, filter out index and the parameter of main expression spinning material performance by computer computing, draw spinning material performanceThe comprehensive evaluation index of evaluating, uses mathematical statistics analysis means to draw and expresses spinning material each number of principal components of different natureBe worth, and further obtain the evaluation of estimate of the comprehensive evaluation value acquisition spinning material of expressing spinning material bulk property; Then by rightThe quality of spinnery's yarn is carried out integrated treatment analysis by the parameter of all technical that records in laboratory, by byComputing filters out index and the parameter of main expression yarn quality performance, draws combining that yarn quality performance is evaluatedClose evaluation index, use mathematical statistics analysis means to draw and express spinning material each principal component numerical value of different nature, and furtherObtain expressing the comprehensive evaluation value of yarn quality bulk property; Finally deduct the evaluation of spinning material by the evaluation of estimate of yarn qualityValue, the evaluation of estimate of acquisition spinning process level.

The present invention's spinning assessment of levels standard that objectifies, simplifies evaluation procedure and method.

Brief description of the drawings

Below in conjunction with accompanying drawing, the present invention is done to further clear, complete explanation:

Fig. 1 is the rubble figure that in the present invention, spinning material calculates the principal component initial characteristics value of gained;

Fig. 2 is the rubble figure that in the present invention, yarn quality calculates the principal component initial characteristics value of gained.

Detailed description of the invention

The invention discloses a kind of method of overall merit spinning process level, in order to make advantage of the present invention, technical scheme moreClear, clear and definite, below in conjunction with specific embodiment, the present invention is done to further clear, complete explanation.

The inventive method specifically comprises the following steps:

Step 1, service test equipment and instrument detect spinning material property indices, obtain detecting numerical value, use Y111AFiber type length analyzer records principal length, quality length, the short fiber content of spinning material; Use Y145C type mic value to surveyDetermine instrument and Y175 type cotton fiber pneumatic tester records: maturity, horse value; Use YG041 type raw material impurity analytical engine and JA1003 type serial analysis electronic balances record percentage of impurity, raw cotton fault; Use YG001B type electronic mono-fiber brute forceInstrument records fiber strength; Use YG202 type weaving moisture apparatus to record fiber regain;

Then, use Microsoft Excel to compile data, rejecting abnormalities value, fills the value of saving;

Finally, the data input data processing units (computer) of collecting are processed, processing method is as follows:

Step 1, collection p dimension random vector X=(x₁,x₂,...,x_p)＇，X_i＝(x_i1,x_i2,...,x_ip) ＇, wherein, i=1,2 ... n, n isNumber of times and the n of sample observation > p, form sample battle array, to sample array element according to carrying out standardization change as shown in the formula (1), (2), (3)Change, obtain standardization battle array,

$z_{ij}=\frac{x_{ij}-{\stackrel{\‾}{x}}_j}{s_j},---\left(1\right)$

${\stackrel{\‾}{x}}_j=\frac{{\mathrm{\Σ}}_{i=1}^n{x}_{ij}}{n},---\left(2\right)$

$s_j^2=\frac{{\mathrm{\Σ}}_{i=1}^n{(x_{ij}-{\stackrel{\‾}{x}}_j)}^2}{n-1},---\left(3\right)$

In above formula (1), (2), (3), i=1,2 ..., n; J=1,2 ..., p_，Wherein, p=9, x₁，x₂，...x_pGeneration successivelyThe principal length of table spinning material, quality length, short fiber content, maturity, fiber strength, horse value, percentage of impurity, fiber moisture regainRate and raw cotton fault;

Step 2, according to formula (4), to standardization battle array, ask correlation matrix;

$R={\[r_{ij}\]}_{p}xp=\frac{Z^{T}Z}{n-1},---\left(4\right)$

Wherein, $r_{ij}=\frac{{\mathrm{\Σz}}_{kj}\·z_{ki}}{n-1},$ i,j＝1,2,...p；

Above-mentioned concrete as embodiment 1, first table (1) is tried to achieve data normalization matrix, tries to achieve " phase relation by table (2)Matrix number ";

The characteristic equation ︱ R-λ I of step 3, solution sample correlation matrix R_P︱=0 obtains p characteristic root, determines principal component, as table 3,Table 4, table 5, shown in Fig. 1, (5) determine m value according to the following formula, m value is used for for determining the number of principal component calculated, makesThe utilization rate of information reaches more than 55%, to each λ in formula (5)_j, j=1,2 ..., m, Rb=λ solves an equation_jb, obtain unit character toAmount b_j ⁰；

Step 4, according to formula (6), the target variable after standardization is converted to principal component, as shown in table 9, table 10,

U_ij＝z_i ^Tb_j ⁰,j＝1,2,...,m；(6)

U₁Be called first principal component, U₂Be called Second principal component, ..., U_pBe called p principal component;

Step 5, m principal component carried out to overall merit,

M principal component is weighted to summation, obtains final evaluation of estimate, the variance contribution ratio that flexible strategy are each principal component, as table 5Shown in.

By being that n=159 tests to the routine product of spinnery 159, draw the technical indicator parameters such as principal length, use systemMeter method is carried out computing by computer, draws following result, and as shown in table 1-table 11, descriptive statistics initial data is as table 1Shown in:

Table 1 descriptive statistics

n＝159

Table 2 correlation matrix

Table 3KMO and bartlett's test

KMO statistic numerical value is greater than 0.9 o'clock best results, more than 0.7 can accept, and below 0.5, should not do factorial analysis, thisIn example, 0.610 still can accept, and the conspicuousness value of the sphericity inspection of Bartlett is less than 0.01, negates that correlation matrix is thusThe null hypothesis of unit matrix, thinks and between each variable, exists significant correlation, the conclusion phase drawing with table 2 correlation matrixSymbol.

Table 4 communality

Table 5 population variance is explained

Table 6 becomes sub matrix

(a. has extracted 2 compositions).

The postrotational one-tenth sub matrix of table 7

(spinning solution: Kaiser standardization varimax, a. restrains after being rotated in 3 iteration).

Table 8 composition transformation matrix

Assembly	1	2
			1	.774	.633
2	-.633	.774

(spinning solution: Kaiser standardization varimax).

Table 9 composition score coefficient matrix

Table 10 composition score covariance matrix

Assembly	1	2
			1	1.000	.000
2	.000	1.000

(spinning solution: Kaiser standardization varimax).

Table 11 some numerical results data

n	FAC1_1	FAC2_1	ZF
				1	-2.98594	1.85022	-0.67
2	-2.98594	1.85022	-0.67
				3	-2.98594	1.85022	-0.67
4	-2.96336	1.81251	-0.67
				5	-2.91818	1.73709	-0.68
6	-2.87301	1.66167	-0.69
				7	-2.35167	1.59027	-0.48
8	-2.30197	1.50731	-0.49
				9	-2.16273	1.3497	-0.48
10	-2.16273	1.3497	-0.48
				11	-2.08933	1.24963	-0.48
12	-2.08933	1.24963	-0.48
				13	-1.65627	0.92966	-0.41
14	-1.65627	0.92966	-0.41
				15	-1.52904	0.65404	-0.45
16	-1.40656	0.45464	-0.47
				17	-1.36476	0.27805	-0.51
18	-1.33391	0.17754	-0.53
				19	0.14506	-1.35617	-0.43
20	0.14506	-1.35617	-0.43
				21	0.14506	-1.35617	-0.43
22	0.14506	-1.35617	-0.43
				23	0.14506	-1.35617	-0.43

According to the data in table 9 (composition score coefficient matrix), calculate respectively the value of 2 principal components, that is:

FAC1_1=0.233* principal length+0.287* quality length-0.228* short fiber content-0.082* maturity+0.151* fiber is strongPower+0.053* horse value+0.006* percentage of impurity-0.183* fiber regain+0.296* raw cotton fault;

FAC2_1=-0.023* principal length-0.114* quality length+0.381* short fiber content+0.328* maturity+0.094* fiber is strongPower+0.180* horse value+0.219* percentage of impurity-0.023* fiber regain-0.258* raw cotton fault;

According to table 5 (population variance explanation), the numerical value of " variance percentage " of " rotation load quadratic sum " the inside, accumulative total variancePercentage is 81.049%, calculates spinning material quality comprehensive and obtains score value:

ZF＝44.81％*FAC1_1+36.239％*FAC2_1。

The property indices of step 2, service test equipment and instrument detection line, obtains detecting numerical value, uses YG086 type threadYarn gauging machine and Y2101 type weaving electronic scale record weight CV%, deviation of weight; Use YG202 type weaving moisture apparatus to surveyObtain yarn regain %, cylinder yarn regain %; Use YG171 type yarn filoplume tester to record filoplume, filoplume CV%; UseY331A type yarn twist meter records the twist, uneven twist, twist factor; Use YG023B-II type automatically singleYarn strength machine records powerful CV%, fracture strength, ultimate strength, minimum brute force, elongation at break and elongation CV%; Use recordsYG133B/M fiber strand evenness tester record the dry CV% of bar, the dry CVb of bar, details-50%, slubbing+50%, cotton knot+200%,Cotton knot; Use YG072A type yarn defect analyzer to record cotton assorted; Collect and calculate by touring counting: empty ingot rate, instantaneous end breakage rate;

Then, use Microsoft Excel to compile data, rejecting abnormalities value, fills the value of saving;

Finally, the data input data processing units (computer) of collecting are processed, processing method is as follows:

Step 1, collection p dimension random vector X=(x₁,x₂,...,x_p)＇，X_i＝(x_i1,x_i2,...,x_ip) ＇, wherein, i=1,2 ... n, n isNumber of times and the n of sample observation > p, form sample battle array, to sample array element according to carrying out standardization change as shown in the formula (1), (2), (3)Change, obtain standardization battle array,

$z_{ij}=\frac{x_{ij}-{\stackrel{\‾}{x}}_j}{s_j},---\left(1\right)$

${\stackrel{\‾}{x}}_j=\frac{{\mathrm{\Σ}}_{i=1}^n{x}_{ij}}{n},---\left(2\right)$

$s_j^2=\frac{{\mathrm{\Σ}}_{i=1}^n{(x_{ij}-{\stackrel{\‾}{x}}_j)}^2}{n-1},---\left(3\right)$

In above formula (1), (2), (3), i=1,2 ..., n; J=1,2 ..., p_，Wherein, p=24, x₁，x₂，...x_pFor describingThe variable of yarn quality, it represents weight CV%, deviation of weight, yarn regain %, powerful CV%, fracture strength, fracture successivelyPowerful, minimum brute force, elongation at break, elongation CV%, the dry CV% of bar, the dry CVb of bar, details-50%, slubbing+50%, cotton knot+ 200%, filoplume, filoplume CV%, cotton knot, cotton assorted, the twist, uneven twist, twist factor, empty ingot rate, instantaneous end breakage rate, cylinder yarnRegain %;

Step 2, according to formula (4), to standardization battle array, ask correlation matrix;

$R={\[r_{ij}\]}_{p}xp=\frac{Z^{T}Z}{n-1},---\left(4\right)$

Wherein, $r_{ij}=\frac{{\mathrm{\Σz}}_{kj}\·z_{ki}}{n-1},$ i,j＝1,2,...p；

Above-mentioned concrete as embodiment 1, first table (12) is tried to achieve data normalization matrix, tried to achieve by table (13) " relevantCoefficient matrix ";

The characteristic equation ︱ R-λ I of step 3, solution sample correlation matrix R_P︱=0 obtains p characteristic root, determines principal component, as table14, table 15, table 16, shown in Fig. 2, determines m value according to formula (5), the utilization rate of information is reached more than 55%, to formula (5)In each λ_j, j=1,2 ..., m, Rb=λ solves an equation_jb, obtain unit character vector b_j ⁰；

Step 4, according to formula (6), the target variable after standardization is converted to principal component, as shown in table 20, table 21,

U_ij＝z_i ^Tb_j ⁰,j＝1,2,...,m；(6)

U₁Be called first principal component, U₂Be called Second principal component, ..., U_pBe called p principal component;

Step 5, m principal component carried out to overall merit,

M principal component is weighted to summation, obtains final evaluation of estimate, the variance contribution ratio that flexible strategy are each principal component, as tableShown in 16.

By being that n=159 tests to the routine product of spinnery 159, draw the technical indicator parameters such as weight CV%, use systemMeter method is carried out computing by computer, draws following result, and as shown in table 12-table 22, descriptive statistics initial data is as table 12Shown in:

Table 12: descriptive statistics

Table 13 correlation matrix

Table 14:KMO and bartlett's test

KMO statistic numerical value is greater than 0.9 o'clock best results, more than 0.7 can accept, and below 0.5, should not do factorial analysis, thisIn example, 0.753 can accept. The conspicuousness value of the sphericity inspection of Bartlett is less than 0.01, and negative correlation matrix is single thusThe null hypothesis of bit matrix, thinks and exists significant correlation between each variable, the conclusion drawing with table 2 correlation matrix conforms to.

Table 15

Table 16 population variance is explained

Table 17 becomes sub matrix

The postrotational one-tenth sub matrix of table 18

Table 19 composition transformation matrix

Table 20 composition score coefficient matrix

Table 21 composition score covariance matrix

According to the data in table 9 (composition score coefficient matrix), calculate respectively the value of 6 principal components:

FAC1_1=0.087* weight CV%-0.045* deviation of weight+0.071* yarn regain-0.009* yarn brute force+...+0.161* yarn cylinder returnsTide rate;

FAC6_1=-0.088* weight CV%-0.117* deviation of weight+0.109* yarn regain-0.306* yarn brute force+...-0.114* yarn cylinderRegain;

According to table 5 (population variance explanation), the numerical value of " variance percentage " of " rotation load quadratic sum " the inside, accumulative total variancePercentage is 69.033%, as shown in table 11, calculates that yarn qualities is comprehensive must be worth:

ZF＝16.162％*FAC1_1+15.544％*FAC2_1+13.675％1*FAC3_1+8.324％*FAC4_1+7.856％*FAC5_1+7.472％*FAC6_1

Table 22: some numerical results data

The evaluation of step 3, spinning process level, the ZF establishing in step 1 is ZFo, the ZF in step 2 is ZFyarn, spinning process waterThe initial value of flat comprehensive evaluation value is ZFp,

ZFp_i＝ZFyarn_i-ZFo_i；i＝1,2,...,n；

ZFp_iValue larger, illustrate that spinning process level is higher.

By ZFo and ZFyarn standardization (formula is the same), the comprehensive evaluation value of spinning process level is ZZFp,

ZZFp_i＝ZZFyarn_i-ZZFo_i；i＝1,2,...,n

ZZFp_iValue larger, illustrate that spinning process level is higher.

Table 23: descriptive statistics

Table 24 some numerical results data

The inventive method is used for evaluating spins spinning process level, has the advantages such as objective, efficient, simple.

Claims (1)

1. a method for overall merit spinning process level, is characterized in that: comprise the following steps successively:

Step 1, first, is used correlation test instrument to detect the each performance indications of spinning material, obtains detecting numerical value, wherein, and spinningEach performance indications of raw material refer to principal length, quality length, short fiber content, maturity, fiber strength, horse value, percentage of impurity,Fiber regain, raw cotton fault, weight CV%, deviation of weight, yarn regain %, powerful CV%, fracture strength, ultimate strength,The dry CV% of minimum brute force, elongation at break, elongation CV%, bar, the dry CVb of bar, details-50%, slubbing+50%, cotton knot+200%,Filoplume, filoplume CV%, cotton knot, cotton assorted, the twist, uneven twist, twist factor, empty ingot rate, instantaneous end breakage rate and cylinder yarn regain %;Then compile detecting numerical value; Finally, reduced data is inputted to data processing unit and process, its processingMethod comprises following sub-step:

A, collection p dimension random vector X=(x₁,x₂,...,x_p)＇，X_i＝(x_i1,x_i2,...,x_ip) ＇, wherein, i=1,2 ... n, n is sampleNumber of times and the n of observation > p, form sample battle array, to sample array element according to carrying out standardized transformation as shown in the formula (1), (2), (3),Obtain standardization battle array,

$z_{ij}=\frac{x_{ij}-{\stackrel{\‾}{x}}_j}{s_j},---\left(1\right)$

${\stackrel{\‾}{x}}_j=\frac{{\Σ}_{i=1}^n{x}_{ij}}{n},---\left(2\right)$

$s_j^2=\frac{{\mathrm{\Σ}}_{i=1}^n{(x_{ij}-{\stackrel{\‾}{x}}_j)}^2}{n-1},---\left(3\right)$

In above formula (1), (2), (3), i=1,2 ..., n; J=1,2 ..., p, wherein, p=9, x₁，x₂，...x_pGeneration successivelyThe principal length of table spinning material, quality length, short fiber content, maturity, fiber strength, horse value, percentage of impurity, fiber moisture regainRate and raw cotton fault;

B, according to formula (4), to standardization battle array, ask correlation matrix;

$R={\[r_{ij}\]}_p{x}_p=\frac{Z^{T}Z}{n-1},---\left(4\right)$

Wherein,i,j＝1,2,...p；Z^TFor the transposed matrix of Z;

The characteristic equation ︱ R-λ I of c, solution sample correlation matrix R_P︱=0 obtains p characteristic root, determines principal component, λ whereinRefer to the characteristic root of correlation matrix R, I_PRefer to p rank unit matrix, determine m value according to formula (5), the utilization rate of information is reachedMore than 55%, to each λ in formula (5)_j, j=1,2 ..., m, Rb=λ solves an equation_jb, obtain unit character vector b_j ⁰；

D, according to formula (6), the target variable after standardization is converted to principal component,

U_ij＝z_i ^Tb_j ⁰,j＝1,2,...,m；(6)

U₁Be called first principal component, U₂Be called Second principal component, ..., U_mBe called m principal component;

E, m principal component carried out to overall merit,

M principal component is weighted to summation, obtains final evaluation of estimate, the variance contribution ratio that flexible strategy are each principal component;

Step 2, first, is used correlation test instrument to detect yarn quality property indices, obtains detecting numerical value; Then,The detection numerical value obtaining is compiled; Finally, reduced data is inputted to data processing unit and process, compriseFollowing sub-step:

F, collection p dimension random vector X=(x₁,x₂,...,x_p)＇，X_i＝(x_i1,x_i2,...,x_ip) ＇, wherein, i=1,2 ... n, n is sampleNumber of times and the n of observation > p, form sample battle array, to sample array element according to carrying out standardized transformation as shown in the formula (7), (8), (9),Obtain standardization battle array,

$Z_{ij}=\frac{x_{ij}-{\stackrel{\‾}{x}}_j}{s_j},---\left(7\right)$

${\stackrel{\‾}{x}}_j=\frac{{\mathrm{\Σ}}_{i=1}^n{x}_{ij}}{n},---\left(8\right)$

$S_j^2=\frac{{\mathrm{\Σ}}_{i=1}^n{(x_{ij}-{\stackrel{\‾}{x}}_j)}^2}{n-1},---\left(9\right)$

In above formula (7), (8), (9), i=1,2 ..., n; J=1,2 ..., p, wherein, p=24, x₁，x₂，...x_pFor describingThe variable of yarn quality, it represents weight CV%, deviation of weight, yarn regain %, powerful CV%, fracture strength, fracture successivelyPowerful, minimum brute force, elongation at break, elongation CV%, the dry CV% of bar, the dry CVb of bar, details-50%, slubbing+50%, cotton knot+ 200%, filoplume, filoplume CV%, cotton knot, cotton assorted, the twist, uneven twist, twist factor, empty ingot rate, instantaneous end breakage rate, cylinder yarnRegain %;

G, according to formula (10), to standardization battle array, ask correlation matrix;

$R={\[r_{ij}\]}_p{x}_p=\frac{Z^{T}Z}{n-1}---\left(10\right)$

Wherein,i,j＝1,2,...p；

The characteristic equation ︱ R-λ I of h, solution sample correlation matrix R_P︱=0 obtains p characteristic root, determines principal component, and λ is whereinRefer to the characteristic root of correlation matrix R, I_PRefer to p rank unit matrix, determine m value according to formula (11), make in formula (11) not etc.The numerical value in formula left side reaches more than 55%, to each λ in formula (11)_j, j=1,2 ..., m, Rb=λ solves an equation_jb, obtain unit characterVector b_j ⁰，

K, according to formula (12), the target variable after standardization is converted to principal component,

U_ij＝z_i ^Tb_j ⁰,j＝1,2,...,m；(12)

U₁Be called first principal component, U₂Be called Second principal component, ..., U_mBe called m principal component;

L, m principal component carried out to overall merit,

M principal component is weighted to summation, obtains final evaluation of estimate, the variance contribution ratio that flexible strategy are each principal component;

The evaluation of estimate of the yarn quality that step 3, above-mentioned steps two obtain deducts the evaluation of estimate of the spinning material of step 1 acquisition,Obtain the initial evaluation value of spinning process level; By after yarn quality comprehensive evaluation value and the standardization of spinning material comprehensive evaluation value, phaseSubtract and obtain the comprehensive evaluation value of spinning process level.