Research And Prediction In Moisture-heat Transmission Performance Of Bast Fiber Fabric Based On Neural Network

The moisture-heat transmission performance of fabric is one of the most important assessments for the comfort of clothing. Previous researchers put lots of effects on the moisture-heat transmission performance of the fabric and clothing which is based on original concepts and rights, test measurements, evaluation of the index, types of fibers, and the mechanism of thermal comfort. However, there is still many theoretical and practical issues remain to be solved. Studies on the moisture-heat transmission of fabric, there are some factors such as natural fiber, yarn and fabric structure. These factors influence the moisture-heat transmission of fabrics, as well as create a complex non-linear relationship between themThis article mainly states the process of wearing clothes through moisture-heat transmission mechanism. Moreover, adjusted formulae and calculations have been given for the moisture-heat transmission among the body, clothing and environment which are derived from the theory of the moisture-heat transmission of fabrics. First of all, the fabric has been considered as solid for moisture-heat transmission research, then it has been analyzed as a porous material. However, there is a great calculation error, since the research is based on an assumption that each unit of the fabric is uniformly arranged. In accordance with fractal theory, it has been given an accurate description of the geometrical structure of the fabric and recognition of apparent factual characteristics. Furthermore, it overcomes the deficiency in the theory of coefficient of heat conductivity from the classical geometric method in the normal environmental condition. Based on the experimental analysis, the result of value of moisture permeability is more close to the measured value.The improvement of the model YG-601 moisture permeability box is more reliable than microclimate box related to accurate data testing, better repeatability and static data comparison. There are 25 types of samples being chosen for comparison and analysis according to the research objectives in which 18 types are pure ramie fabrics or blend fiber fabrics and the rests are bamboo fiber fabrics. There are some differences among warp and weft counter, yarn number, fabric thickness, gram per square meter and weave structure. These data are essential to intensive research of the moisture-heat transmission performance of bast fiber fabrics and prediction of comfort use. As a result, it gives basic assurance for the validity and feasibility of the outcome.This article also analyses that the moisture-heat transmission performance could be predicted based on neural network technology, discussed the basic method to build the prediction model of neural network, in conjunction with the example to testing the feasibility of neural network model, and providing further questions needed to be solved. Relating to the example of the prediction of moisture-heat transmission performance shows that building the prediction of neural network model is reasonable, unless providing the representative samples for learning and training. The prediction results fully satisfy the requirement of precision. There are many merits of the model such as easy and quick procedure, simply correct the mistakes, easy to use, and flexible.In conclusion, this article uses grey system theory, fuzzy similarity and priority treatment, and the research of neural networks for moisture-heat and comfort of bast fiber fabric. It provides new evaluation for the moisture-heat and comfort of fabric; improving the analysis and process methods in order to get more accurate and objective data. The assessment is a vague concept. Therefore, providing five subset sums for judgement, there is a fuzzy boundaries between two neighbouring status.