样本质量对近红外法预测远红外纤维含量的影响

传统的纤维定量方法不能够满足远红外纤维与其基体普通纤维之间定量分析的问题，由此提出了近红外光谱分析技术。选用常见的远红外涤纶、远红外丙纶、涤纶和丙纶作为研究对象。利用人为自主混合取样的方法，采用偏最小二乘法(PLS)，结合样本重量对所建模型的影响，建立了有关远红外纤维含量的远红外涤纶/涤纶，远红外丙纶/丙纶两个校正模型，发现两个模型都能达到比较好的预测效果。其结果：远红外涤纶/涤纶定量校正模型远红外涤纶含量的平均相对误差为2.03%；远红外丙纶/丙纶定量校正模型远红外丙纶含量的平均相对误差为3.76%。研究表明近红外光谱分析技术可以用于远红外纤维的定量分析。

Influence of sample weight on prediction of far-infrared fiber content based on near-infrared spectrodscopy

The traditional methods cannot meet the far-infrared and matrix ordinary fiber quantitative analysis, so the Near-infrared spectrum analysis technique is presented. Choose the common far-infrared polyester fiber, far-infrared polypropylene, polyester and polypropylene fiber as the research object.In this reach,combined with the sample weight affect the model,the far-infrared polyester/polyester and far-infrared polypropylene/polypropylene models which can achieve good prediction results were established by manual sampling and PLS method.Results show that the average relative errors of the far-infrared polyester and far-infrared polypropylene content were 2.03% and 3.76%.The research shows that the Near-infrared spectrum technique can be used as a method of quantitative analysis for far-infrared fiber.