2烟草工业舒芳誉《基于多变异分析的烟支物理质量影响因素研究》(已经排版稿).docx

“已发表”基于多变异分析的烟支物理质量影响因素研究舒芳誉，罗靖，马明旻厦门烟草工业有限责任公司，福建省厦门市海沧新阳工业区新景路1号 361022摘要：为提升烟支物理质量的均质化水平，运用嵌套析因设计，考察分析了卷制生产过程中5个可能的变异来源（班别、班次、卷烟机台、烟丝批次和抽样时间）对烟支物理质量稳定性的影响。结果表明：①卷烟机台和烟丝批次是显著影响烟支物理指标、样本方差和样本广义方差的变异源；②在烟丝物理质量一定的情形下，同一卷烟机台对烟支物理指标的影响不随抽样时间变化；③班别、班次不是显著影响烟支物理指标、样本方差和样本广义方差的变异源。因此，为了使卷制过程烟支物理质量均匀、稳定，要提升烟丝批次内和批次间的物理质量稳定性，提升同型卷烟机间工作状态的一致性。关键词：嵌套析因设计；多变异分析；烟支物理质量；样本方差；样本广义方差中图分类号：TS412文献标识码：B 文章编号：DOI：10.16135/j.issn1002–0861tudy on Influencing Factors of Physical Quality of Cigarette Based on Multi-variation AnalysisSHU Fangyu, Luo Jin, and MA MingmingXiamen Tobacco Industrial Co.,,Ltd., 361022,Fujian,ChinaAbstract: In order to improve the consistency of physical quality of cigarettes, the effects of five factors(operator, shift, cigarette maker, batch of cut filler and sampling time) on the stability of physical qualityof cigarettes were analyzed with nested factorial design. The results showed that: 1) Cigarette maker andbatch of cut filler were the main variables affecting the physical indices, sample variance and samplegeneralized variance. 2) For the cut filler of constant physical quality, the influence of the same cigarettemaker on cigarette physical indices did not vary with sampling time. 3) Operator and shift were not thevariation sources affecting the physical indices, sample variance and sample generalized variancesignificantly. Therefore for pursuing higher consistency and stability of cigarette physical quality, it isessential to promote, the intra- and inter-batch consistency of physical quality of cut filler and theconsistency of working states of cigarette makers.Keywords: Nested factorial design; Multi-variation analysis; Physical quality of cigarette; Sample variance;Sample generalized variance多变异分析(Multi-variation Analysis，MVA)是一种分析工序或过程质量特征值变异规律的方法[1]，也是一种嵌套设计(Nested design)试验方法。近年来，科研人员不仅提出了工序质量分析与控制中的多变异来源的分析方法[2–7]，而且还将方法应用于对休哈特控制图的改进[8–11]，提升了多变异分析方法的实用性。同时，何桢等[12]将方法运用于机械制造的轴加工过程，结果表明它能够克服传统休哈特控制图控制多变异来源的不足；史金飞等[13]在晶圆生产中的实际应用，以及仲高艳[14