Construction Of The Grade Evaluation Model Of Heterogeneous Fiber Content

The textile machinery industry in the "13th Five-Year"plan proposes further research on the evaluation to carry out intelligent textile testing and data mining.The content of foreign fibers is an important indicator for assessment of cotton quality.With digital processing techniques as a means to carry out foreign fiber content testing and rating studies.This project focuses on the determination of the characteristic parameters of foreign fiber content,analyzing the correlation degree and optimizing the parameters.Study the relationship among the grade parameters of foreign fiber conten and establish the evaluation model of foreign fiber content.The problem of the weight index in foreign fiber content is single,incomplete and inaccurate.In this paper proposes a multi-index foreign fiber content grade parameters system to solve the problem.The content index of the foreign fiber are obtained by analyzing the cotton yarn forming process,the kind of foreign fiber,the foreign fiber content and the yarn defect model,and the relationship among the properties and content of the foreign fiber.Detection content of foreign fiber content and the grade evaluation content index are determined.According to the characteristic of multiple content index information,the correlation degree among the index elements is analyzed.The rough set theory algorithm based on genetic algorithm and Johnson algorithm is used to reduce the dimension of the index system space.The index system of diameter,length,area,weight,toughness and dyeability is proposed.The index value interval is graded according to the four content grades specified in the national standard.Use the maximum entropy criterion to initialize the index value interval.Use multi-dimensional optimization to improve interval.The grade interval of each index is established.Aiming at the problems of low efficiency and large error of artificial foreign fiber content detection,this paper proposes an integrated algorithm of subtractive clustering,fuzzy clustering algorithm and validity function to realize fast online detection of foreign fiber content grade.The clustering center of the foreign fiber content is selected as the input of fuzzy clustering algorithm by subtractive clustering.Determine iterative process of fuzzy clustering algorithm.According to evaluation algorithm,the weight of content index is integrated into fuzzy clustering algorithm.The iterative formula of the weight is obtained by the objective function to meet the weight of the actual and evaluation methods.The evaluation result is judged by the validity index.In order to avoid the fuzzy clustering algorithm falling into the local optimal solution,the competitive learning principle is proposed.Optimization of the fuzzy partitioning strategy in the algorithm is used to improve the convergence speed and accuracy.By comparing the experimental results with the fuzzy comprehensive evaluation method,the error rate is 5%.It is verified that foreign fiber content grade evaluation model gets a promising result.Study on evaluation of foreign fiber content grade based on information fusion improves the accuracy of foreign fiber content grade.Achieve digital and intelligent of evaluation and the foreign fiber content detection.It can be applied to the textile industry related fields,with a certain theoretical and practical value.