可视化分析用于油茶籽油提取工艺优化及成分分析-optimization of extraction process and composition analysis of camellia seed oil by visual analysis.docxVIP

可视化分析用于油茶籽油提取工艺优化及成分研究摘 要本文采用超临界 CO2 提取油茶籽中的油茶籽油，并对其工艺过程和油茶 籽油主要成分进行了较为系统的研究。将油茶籽去壳，烘干，粉碎，过筛备 用。称取一定量的油茶籽粉末置于 1L 的萃取釜中，用超临界 CO2 提取油茶 籽粉中的油茶籽油。选择萃取压力、萃取温度、萃取时间、分离 I 压力为影 响因素，用课题组自主提出的多因素多水平可视化设计法（m2VD）安排试 验。超临界 CO2 提取油茶籽油的工艺研究，以提取率为试验指标，用多因素 多水平可视化分析方法（m2VA）对试验影响因素与试验指标的关系进行分 析，找出最佳工艺范围：萃取压力为 25～30MPa、萃取温度为 30～35℃、 萃取时间为 110～140min、分离 I 压力为 4～5M Pa。对提取出的油茶籽油进行甲酯化，采用气相色谱—质谱法测定其成分及 含量，结果得出油茶籽油中的主要单不饱和脂肪酸为油酸，多不饱和脂肪酸 为亚油酸，饱和脂肪酸为棕榈酸和硬脂酸。用多因素多水平多目标可视化分 析方法分析研究了四个工艺因素（萃取压力、萃取温度、萃取时间、分离 I 压力）对油酸、亚油酸、硬脂酸和棕榈酸含量之间关系的影响。找出了得 到高含量不饱和脂肪酸和较低含量饱和脂肪酸的优势工艺区间：萃取压力低于 30MPa，萃取温度小于 50℃，萃取时间 100～150min，分离压力 4～6MPa。 该工艺区间与得到较高提取率的最佳工艺区间一致性较好。在优势工艺区间内选取：萃取压力 27MPa、萃取温度 32℃、萃取时间 120min、分离 I 压力 4.5MPa，进行验证试验，得到油茶籽油的提取率为 51.9%， 超过了设计试验的最高数值 45.02%。并检测出其中的油酸含量为 68.49%， 亚油酸含量为 8.12%，棕榈酸含量为 12.49%，硬脂酸含量为 4.45%。证明了 多因素多水平多目标可视化设计与分析方法的有效性，可以为工业化生产油 茶籽油提供工艺指导，具有重要的实践意义。关键词：油茶籽油，超临界 CO2 提取，多因素多水平多目标可视化分析 方法，脂肪酸Study of Extraction Process Optimization of Camellia Seed Oil and its Compositions with Visual AnalysisAbstractThe experiments of extracting camellia seed oil from camellia seed with supercritical carbon dioxide technology were carried out. The physic-chemical rule of the process and main compositions of camellia seed oil were systematically investigated. The camellia seed shells were dislodged. And camellia seed were dried, crushed and sieved. Then powder of camellia seed was stored for using. Some certain volume powder was weighted and put into the extraction vessel of 1 liter to extract camellia seed oil with supercritical carbon dioxide. Four factors (extracting pressure, extracting temperature, extracting time and separating temperature) were chosen as influencing ones and the multifactor multilevel Visual Design method (m2VD) proposed by our research group wasapplied to arrange the experiments. The experiment target was the extraction rate in the process research of supercritical carbon dioxide technology. This paper used our multifactor and multilevel Visual Analysis method (m2VA) to ana