Bayesian Statistical Inference Method For Inverse Probiem Of Textile Material Design

With the progress of society and improvement of people’s living standards, theapplications and functions of textile materials have greatly increased. As the dailynecessity, the comfort requirements of clothing, including heat-moisture comfort,pressure comfort, and tactile comfort etc, have drawn a growing concern. Clothingcomfort is the core requirement of textile materials design, and thermal comfort isparticularly important. In the process of production, they are usually designed basedon human experience or experiment of human dress. In order to theoretically give thescientific guidance or technical support to textile materials, we study further themodel of Body-clothing-environment system on the basis of existing work.On the basis of the existence and uniqueness of solutions, this paper furtherstudies the stability, and obtain corresponding stability estimate. For the inverseproblem of parameter design of heat-moisture transfer, we first present the definitionsof single parameter decision as well as the multiple parameters decision. Then theinversion model based on bayesian inference is established. Finally, we numeriallysolve the inverse problem of textile material design by using Markov Chain MonteCarlo method.For the IPTMD, the common method is to construct an objective functional, andthen the original problem is put into an optimization problem. The optimizationproblem is usually solved by golden section method, Hooke-Jevees search algorithmand particle swarm algorithm, etc. This paper utilizes a new method, Markov chainMonte Carlo (MCMC) sampling method, to determine the parameters of IPTMD.Numerical simulation shows that this method can be used to tackle the inverseproblem with high accuracy. And hence it can provide theoretical basis and scientificguidance for textile materials design.The paper is divided into six chapters. The first chapter introduces the researchbackground and research progress. The second chapter formulates the mathematicalmodel of heat and moisture transfer of textile materials and gives its well-posednessresults. The third chapter discusses the inverse problem of thickness determination byBayesian theory of statistical inferences. In the forth chapter, the numerical simulationis presented for the parameters determination within single-layer textile in steady-state model. The fifth chapter performs the numerical simulation for bi-layer textilematerials. In the final, some conclusions and further research are concluded.