【作者】 许建梅；

【导师】 白伦；

【作者基本信息】 苏州大学 ， 纺织工程， 2007， 博士

【摘要】 我国目前使用的生丝检验标准中,检验指标主要为纤度偏差、匀度、清洁、洁净。但是随着检测技术的进步,电子检验手段的推广,国外生丝消费方面对于生丝电子检验的需求不断加强。自从1995年国际丝绸协会1995年发表《95`生丝便览》,提出了以生丝纤度变异系数(CV系数)为主要质量检验指标的提案以来,生丝电子检验技术的影响进一步扩大,生丝生产与需求方面因为所采用的标准不一致而发生的各种贸易纠纷日趋激化,并且直接造成了像中国这样生丝出口总额占世界80%的生丝出口大国的利益损失。可以说生丝电子检验标准的制定不但事关国家利益,而且决定了生丝贸易的国际竞争成败。为了构建科学合理高效的国际生丝质量评价与贸易的平台,争取在国际化生丝检验标准研制中的主导权和发言权,以保护在国际贸易中具有最大生丝出口份额的我国丝绸产业的利益,积极开展生丝电子检验标准相关理论研究,具有不同寻常的意义。在此背景下,在我国有关部门支持下,本课题就生丝电子检验标准研发的基础理论问题,也是国际生丝检验标准研究中各国共同关注的生丝纤度变异系数为质量指标的生丝检验分级标准相关理论问题,展开了系统的研究。为了能够制定出科学合理的生丝电子检验标准体系,本文对生丝纤度变异系数质量指标的分布及其特性进行了理论分析,同时结合大量的乌斯特条干均匀度仪实测数据,以生丝抽样分级理论为指导,得到了以生丝纤度变异系数为质量指标的分级标准的制定方法及分级方案,并通过模拟实验对分级方案进行了实验验证。论文的主要研究内容及结果如下:1、对具有相关性的正态总体的随机样本的变异系数分布进行了研究,给出了这种情况下的变异系数抽样分布及其数字特征,并详细分析了正态总体特征参数及子样容量对变异系数抽样分布形态及其数字特征的影响。变异系数抽样分布的导出不仅是对概率论中变异系数抽样分布理论作了一个很有意义的补充与完善,而且还为后面以纤度变异系数为质量指标的分级标准的制定提供了理论基础。2、基于独立样本的变异系数分布,根据生丝检验中生丝纤度样本值是连续变化的特点,对具有相关性的正态总体的连续样本的变异系数分布进行了研究。利用等效样本的方法,推导了生丝电子检验中抽样丝段的变异系数分布及其数字特征。并通过模拟实验验证了该等效分布能够很好的描述实际抽样检验中样本变异系数的分布,同时探讨了各个影响因素对等效所产生的误差的影响。3、探讨了生丝电子检验中样本容量问题。利用模拟试验的方法探索了合理的样本数量范围。在本研究采用的检验体系内生丝检验的抽样丝长采用800-1000米,检验筒子数采用30-40个时,检验结果相对误差较小而且检验效率比较高;从而提出了较合理的抽样方法。4、利用生丝纤度偏差比公式,分析了不同检验丝长的生丝纤度变异系数之间的关系,利用纤度偏差比来估计反映生丝纤度波动特性的两个参数(拟似周期T和衰减率λ)的值,并给出了不同检验丝长的变异系数之间的关系。5、根据上面研究得到的变异系数抽样分布和抽样方法,对不同检验丝长的生丝纤度变异系数的等级设定理论及分级方法进行了研究。通过理论分析与实测数据检验,给出了各不同检验丝长变异系数的抽样分布,从而得到了纤度变异系数的分级精度表达式及等级界限值的计算公式。最后提出不同规格生丝的分级方案,并通过模拟实验验证了分级方案的合理性和精确性。6、对卷装间条干均匀度的差异,即变异系数的波动特性及其分级问题进行了研究。给出了CV系数的变异系数的近似分布,及其分级精度的数学表达式,从而提出了分级方案,并通过模拟实验对该分级方案进行了验证。更多还原

【Abstract】 At present in our country the testing indexes in the testing standards for raw silk mainly include size variance,evenness,cleanness,and neatness.But with the popularization of the electronic testing means,overseas customers are increasingly inclined to the electronic testing and the electronic testing standards Raw Silk Enchiridion 1995 developed by the International Raw Silk Association in 1995.Thus all kinds of trade fightings become more and more violent due to different testing standards used by both sides,and these fightings cause great loss of our country whose output of raw silk is 80% of the total international trade.Therefore,the development of the electronic testing standards for raw silk is not only important to the national interests,but also determines the success or failure in the international trade of raw silk.In order to construct an international platform for the quality appraisal and trade of the raw silk more efficiently and more rationally,to get a dominate authority in the research of the international testing standards for raw silk,to protect the interest of our silk industry,it is very significant to make the study on the grading theory of the electronic testing for raw silk.In such circumstance,under the support of the governmental departments of our country,we study the basic theories in the research of the electronic testing standards for raw silk size;make a systematic study on the grading theories of the size coefficient of variation in the electronic testing for raw silk.In order to make more rational and scientific electronic standard system for raw silk, this paper studies the distribution of the C.V.of the raw silk size and its characteristics. According to the sampling and grading theory of the raw silk,the paper also proposes the method to make a grading standard with the C.V.of the raw silk size as the quality index and the corresponding grading scheme on the basis of large amount of testing data of all kinds of raw silk by Uster Tester,verifies the grading scheme by simulation testing.The main work and results of the paper are as follows:1.The paper studies the distribution of the C.V.of random sample from a correlated normal population,gives its sampling distribution and its characteristics,and analyzes the effects of the characteristics of the normal population and the sample size on the shape of the sampling distribution and the characteristics of the C.V..And the study of the sampling distribution of the C.V.will not only be a very significant supplement and perfection to the theory of the sampling distribution of the C.V.in the field of probability,but affords theoretical basis for the development of the grading standard for the C.V.of the raw silk size.2.Based on the distribution of the C.V.of random sample,this paper studies the distribution of the C.V.of the continuous sample from the correlated normal population for the sample value of the raw silk size in the raw silk testing is continuously changed. Using equivalent sample,this paper gives the distribution of the C.V.of the sampling silk segment from the correlated normal population and its characteristics.And the corresponding simulation testing proves that the equivalent distribution has a good representation to the distribution of the C.V.of the sample in actual sampling testing.At the same time the paper also discusses the effects of all kinds of affecting factors on the error caused by the equivalent method.3.The problem of sample size in the electronic raw silk testing is fully discussed.By simulation,the rational range of the sample size is studied.In the testing system adopted by this study,the relative testing error will be smaller and the testing efficiency will be higher when the sampling length of the silk segment varies between 800 and 1000 meters, the number of the sampling bobbins varies between 30 and 40,thus we can get a rational sampling method.4.Using the ratio formula of the size variance of the raw silk,the paper analyzes the relation between the C.V.of the raw silk size with different testing length,studies how to estimate the two parameters(mimetic period T and attenuation rateλ) of the size fluctuation of raw silk,gives the relation between the coefficients of variation of the silk segments with different testing length.5.According to the above sampling distribution of the C.V.and the sampling method, this paper studies the grade making theory and the grading method of the C.V.of the raw silk size.By theoretical analysis and the actual testing data,the sampling distributions of the C.V.of silk segments with different testing length are given,and the calculating formulas of the grading precision and the grading boundaries are also given.Finally the paper proposes the grading schemes of raw silk with different specification,and verifies the rationality and accuracy of the grading scheme by simulation.6.The paper also studies the characteristics and the grading of the difference of the size evenness among the sampled bobbins,namely,the fluctuation of the C.V..Thus the paper gives the approximately distribution of the coefficient of variation of the C.V.and the expression of its grading precision,proposes its grading scheme,and verifies the rightness of the grading scheme by simulation.更多还原