System and method for fibrogram fiber quality evaluation

Patent metadata

Field	Value
Publication number	US-11579137-B2
Application number	US-201816763924-A
Country	US
Kind code	B2
Filing date	Nov 13, 2018
Priority date	Nov 13, 2017
Publication date	Feb 14, 2023
Grant date	Feb 14, 2023

Abstract

Disclosed is a system and method for extraction of information of within sample distribution of fiber quality from high-volume instrument (HVI) fibrogram to better predict yarn quality than the standard HVI output. The present invention allows for information on fiber quality to be obtained while avoiding testing samples with more expensive techniques. The disclosed system and method extracts HVI data for collecting a respective set of initial fibrograms from a set of fiber samples and representing them as a distance matrix to form a matrix of transformed fibrogram data, said matrix of transformed fibrogram data comprising a vector of scores to represent each sample and thereafter explaining variation in yarn quality by extracting all of the information available from the fibrogram.

First claim

What is claimed is:

1. A system for estimating unknown fiber quality of a fiber sample, comprising:
a. a high volume instrument (HVI) for collecting a respective set of fiber samples from the HVI and representing the respective set of fiber samples as a fibrogram having a distance matrix; and
b. a computer for:
i. decomposing a total variation captured in said distance matrix to form a matrix of transformed fibrogram data, said matrix of transformed fibrogram data comprising a vector of scores to represent each sample;
ii. mutually exclusively partitioning a total multivariate space captured by the initial fibrograms into a set of independent variables from the HVI; and
iii. explaining variation in yarn quality by regressing each yarn quality parameter over said vector of scores obtained from the matrix of transformed fibrogram data.

2. The system of claim 1 , wherein the distance matrix is based on the Euclidean distance metric between initial fibrograms.

3. The system of claim 1 , wherein the distance matrix is based on the chi-squared distance metric between initial fibrograms.

4. The system of claim 1 , wherein the distance matrix is based on the Bray-Curtis dissimilarity metric between initial fibrograms.

5. The system of claim 1 , wherein the distance matrix is based on a correlation matrix between initial fibrograms.

6. The system of claim 1 , wherein the decomposition may comprise singular value decomposition.

7. The system of claim 1 , wherein the vector of scores further comprises at least one score representing the largest source of total variation among all initial fibrograms.

8. The system of claim 1 , wherein the vector of scores further comprises at least two scores representing a first score representing the largest source of total variation among all initial fibrograms, and a second score representing the second largest source of total variation among all initial fibrograms.

9. A method of estimating unknown fiber quality of a fiber sample, comprising:
a. collecting a respective set of initial fibrograms comprising a set of fiber samples from a high volume instrument (HVI) and representing the set of fiber samples as a distance matrix;
b. decomposing a total variation captured in said matrix form from the distance matrix to form a matrix of transformed fibrogram data, said subsequent matrix of transformed fibrogram data comprising a vector of scores to represent each sample;
c. mutually exclusively partitioning a total multivariate space captured by the initial fibrograms into a set of independent variables from the HVI; and
d. explaining variation in yarn quality by regressing each yarn quality parameter over said vector of scores obtained from the transformed fibrogram data.

10. The method of claim 9 , wherein the distance matrix is based on the Euclidean distance metric between initial fibrograms.

11. The method of claim 9 , wherein the distance matrix is based on the chi-squared distance metric between initial fibrograms.

12. The method of claim 9 , wherein the distance matrix is based on the Bray-Curtis dissimilarity metric between initial fibrograms.

13. The method of claim 9 , wherein the distance matrix is based on a correlation matrix between initial fibrograms.

14. The method of claim 9 , wherein the decomposition may comprise singular value decomposition.

15. The method of claim 9 , wherein the vector of scores further comprises at least one score representing the largest source of total variation among all initial fibrograms.

16. The method of claim 9 , wherein the vector of scores further comprises at least two scores representing a first score representing the largest source of total variation among all initial fibrograms, and a second score representing the second largest source of total variation among all initial fibrograms.

17. The method of claim 9 , wherein the vector of scores further comprises more than two scores representing ranked scores representing the source of total variation among all initial fibrograms from largest to smallest.

18. The method of claim 9 , wherein the fiber quality further comprises quality parameters selected from a group consisting of: breaking force, work-to-break, elongation at break, yarn irregularity (CVm %), thin places, thick places, neps, hairiness, imperfection index, and combinations thereof.

19. The method of claim 9 , further comprising performing said determining each fiber quality parameter on a computer.

20. The method of claim 9 , wherein the vector of scores further comprises more than two scores representing ranked scores representing the source of total variation among all initial fibrograms from largest to smallest.

21. The method of claim 9 , wherein the fiber quality further comprises quality parameters selected from a group consisting of: breaking force, work-to-break, elongation at break, yarn irregularity (CVm %), thin places, thick places, neps, hairiness, imperfection index, and combinations thereof.

Assignees

• Univ Texas Tech System

Inventors

• Kelly Brendan
• Hequet Eric
• Sayeed Md Abu
• Hinds Zach

Classifications

• G01N33/362 Primary

  Material before processing, e.g. bulk cotton or wool · CPC title

• D01H13/32

  Counting, measuring, recording or registering devices · CPC title

• D01G99/00 Primary

  Subject matter not provided for in other groups of this subclass · CPC title