【作者】 张翠；

【导师】 薛天宇；

【作者基本信息】 北京服装学院 ， 机械电子工程， 2008， 硕士

【摘要】 计算机技术和网络技术的飞速发展,使得繁琐的服装设计知识元素的获取变得更为简易。作为影响服装美感的重要因子,面料悬垂性能的优劣直接关系着服装的设计决策。根据测试数据建立悬垂性能指标自动预测和评价的系统,对生产与科研单位研制新型纤维及改进织物设计、提高织物悬垂性能和美感度是十分必要的。针对现有悬垂性能预测系统与评价体系的现状与需求,本文围绕悬垂性能的预测及评价过程,结合服装纺织专业知识以及软件工程的设计思想,论述了系统的总体架构分析设计、后台数据库建立、功能算法的实现,以及系统的详细设计过程和系统实现过程等。本文探讨了悬垂性能指标预测以及模糊聚类评价两个功能算法。利用神经网络预测算法对50组织物试样数据进行学习和预测,最终验证了算法的实用性,确定了基于有动量和自适应学习速率梯度下降方法的BP预测算法,并实现从织物规格参数到悬垂性能参数的非线性预测;在模糊聚类评价系统中,着重讨论了FCM聚类算法无法避免噪声点影响,引入基于FPCM的模糊聚类算法,抑制了各种不利因素产生的误差。实测数据证明,聚类算法的准确度较高,聚类分析结果与主观评价能够保持良好的一致性。基于B/S(Browser / Server)结构的织物悬垂性能预测与评价系统的建立将大大简化织物悬垂性的测试与评价工序,实现了纤维、纱线及织物相关性能与织物悬垂性能指标的查询,以及从已知织物组织结构参数到悬垂性能指标的预测,最终实现对悬垂性能优劣的客观全面评价。更多还原

【Abstract】 With the rapid development of computer and network technology, the element knowledge of fashion design has become easier to obtain. As an important impact on the aesthetic factor, the performance of fabric drape is directly related to the fashion design decision. According to the test data to establish the system which can predict and evaluate fabric drape indexes automatically, it is very necessary to develop new types of fibers, improve the fabric design , enhance fabric drape performance and aesthetic value for manufactory and scientific research units.Based on current situation and demand of the existing fabric drape test and evaluation system, combined with garment and textiles professional knowledge and software engineering design idea, this paper which takes the process of fabric drape prediction and evaluation as center, discusses the overall architecture analysis and design of system, the establishment of background database, the implementation of function algorithms, as well as the detailed design process of system and the process of system implementation.This article discusses both algorithms of the drape performance prediction and fuzzy clustering evaluation. The BP prediction algorithm based on additional momentum and adaptive learning rate was determined by training and predicting 50 groups of fabric sample data. And finally, the results prove the validation of the algorithms and implement the non-line parameters prediction from fabric specifications to the drape performance; In the system of fuzzy clustering evaluation, this article emphatically discusses on the noise point of the FCM fuzzy clustering algorithm, introduces FPCM fuzzy clustering algorithm to inhibit errors aroused from a variety of unfavorable factors. Measured data shows that the accuracy of prediction algorithm is higher, and cluster analysis has a good consistency with the results of subjective evaluation.The establishment of Fabric Drape Performance Prediction and Evaluation System based on B/S(Browser / Server)architecture will simplify the process of drape testing and evaluation consumedly, realizes the query of fibers、yarn and fabric related performance, as well as the prediction from fabric organization structure parameters to fabric drape indexes and finally realizes fabric drape objectively and completely evaluated.更多还原