基于近红外技术的苎麻化学成分快速定量分析研究

【作者】 姜伟；

【导师】 韩光亭；

【作者基本信息】 青岛大学 ， 纺织工程， 2009， 硕士

【摘要】 苎麻作为植物纤维原料在纺织生产中占有很大的比重,而把苎麻生产成纺织可用纤维则需要经过脱胶工序,准确和快速对苎麻原料的化学成分的定量分析则是建立合理脱胶工艺的基础。同时,在农业领域中快速预测苎麻化学成分含量也对苎麻品种的优选有重要的指导作用。为此我们研究并改进了天然植物纤维原料的化学成分定量分析方法,然后使用AOTF近红外光谱仪扫描苎麻近红外光谱,并采用多元校正方法结合使用改进方法得到的苎麻化学成分含量,建立了苎麻各化学成分含量的预测模型,从而实现了苎麻部分化学成分快速预测,且对其它不可预测成分分析方法提出了改进措施,以在进行进一步工作后达到准确预测的目的。针对植物纤维原料不同成分的特点在GB/T5889-86《苎麻化学成分定量分析方法》的基础上提出了多种定量分析方案,通过实验对其优选并确立了相对准确的化学定量分析方法。最终确定改进方案为:采用原料粉碎后用分样筛布包裹的方式,使用四氯化碳乙醇提取,然后把提取液蒸馏,取残液烘干称重测定脂蜡质含量;测定脂蜡质成分后的样品,利用沸煮法测定样品的水溶物含量,其次利用草酸铵测定样品果胶含量,再次利用氢氧化钠测定样品的半纤维素含量和部分木质素含量B,最后利用醋酸和亚氯酸钠测定木质素含量A,其中总木质素含量采用硫酸法测定,从而计算或者称量得到半纤维素和纤维素的含量。实验证明此方法对植物纤维原料各化学成分测定比较准确,且具有普遍适用性。采用改进方法对60种化学成分含量差异较大的苎麻原料进行化学定量分析,使用AOTF近红外光谱仪对分析后的苎麻原料进行扫描,使用一阶导数对光谱进行预处理,通过PLS1方法对扫描光谱和化学分析结果的分析优化得到了苎麻化学成分近红外预测模型,并选取了几种样品作为预测样品集进行预测工作,使用统计思想对预测后的结果进行检验来验证模型的预测能力。结果显示:对于半纤维素,纤维素,含胶率等大量成分,已建立准确的预测模型,可以进行苎麻半纤维素、纤维素和含胶率成分快速准确的预测工作;对于水溶物、果胶、灰分等常量成分,已经建立了可以进行预测的校正模型,但预测仍具有一定的误差,仍需要增加校正样品集数量来提高预测的精度;由于分析方法和样品数量的限制,对于木质素等含量较少的成分,虽建立了稳健的校正模型,但预测能力较差,需要使用精确分析方法和增加校正样品集的数量来建立可以进行预测的模型。更多还原

【Abstract】 As a plant fiber material,ramie is widely used in textile industry.Ramie must be degummed before it can be used as textile fiber.Accurate and fast quantitative analysis in the chemical composition of raw ramie is a rational basis of establishing degumming process.Meanwhile,fast forecasting of chemical composition and content of ramie has important guiding role in ramie varieties optimization in the field of agricultural. Therefore,quantitative chemical composition analysis method of natural plant fiber was studied and improved and near infrared spectrum of ramie was tested by AOTF near infrared spectrometer.After that,the prediction model of chemical compositon content of ramie was established using multivariate calibration method combined with the chemical composition content which analysised by improved method.Thereby carry out the rapid prediction of certain chemical composition content of ramie and put forward improved measures for other unpredictable composition analysis methods in order to accurate predict these compositions in the further work.Put forward several quantitative analysis methods in order to modify GB/T5889-86 "ramie chemical composition of quantitative analysis method",a relative accuracy of chemical quantitative analysis method has been optimized by experiment.The optimised method was as follows:firstly,wrap the crushed plants in the bolting cloth,extracte lipid with Carbon tetrachloride & alcohol in Soxhletsextractor,distill the extracted liquid,then have the residue dried and weighed as wax content;secondly,water soluble matters content was analyzed in boiling water,content of pectin was determined using ammonium oxalate,the total content of hemicelluloses and partial liguin was tested using sodium hydroxide and hemicelluse content can be calculated after delignification;at last, content of cellulose was obtained after the residue was treated in acetic acid and sodium chlorite aqua to remove lignin.Liguin content has been determined by sulphuric acid.The result shows that this method is accurate and it has wide application in chemical composition analysis of plants.Chemical compositions of 60 ramies were quantitatively analyzed using the above optimized method.At the same time,near infrared spectrums of this 60 samples were scanned by AOTF near infrared spectrometer and the obtained spectrums were first derivated.The prediction model of chemical compositon content of ramie was established by PLS1 according to the result of chemical analysis and near infrared spectrums.The prediction model was validated by several samples.The results show that:those hemicelluloses、cellulose and gun content which content are more than 10%has establisted the fast and accurate predict model,they can predict the hemicelluloses、cellulose and gun content fast and accurate;the macroconstituent such as water soluble matters、pectin and ash have established prediction models which can be do the predicting work,but a little difference has exit too,more samples can help to increase the accuracy of the predict model;the microconstituents such as wax and ligin has established stable models but poor prediction ability due to inaccuracy chemical analysis and less number of samples,more accurate chemical analysis method and more samples are needed to establish a accurate predition model.更多还原