基于SIMCA模型的纸浆种类近红外光谱鉴别

Identification of Pulps by Near Infrared Spectroscopy Based on SIMCA Model

关键词:  近红外光谱  SIMCA  纸浆种类  模式识别

Key Words:near infrared spectroscopy  soft independent modeling of class analogy (SIMCA)  raw materials of pulp  pattern recognition

摘要:研究了近红外光谱分析与SIMCA模式识别技术相结合鉴别纸浆种类的方法。收集了桉木浆、麦草浆（漂白与未漂白）、棉浆、湿地松浆等4种纸浆共90个样品，抄成不同定量纸样并采集其近红外光谱，其中部分桉木浆添加化学助剂。选择60个样品作为训练集建立SIMCA类模型，剩余30个样品用于模型检验。研究结果表明，建立的模型能完全正确识别各类纸浆，且不受抄纸定量和添加化学助剂因素的影响，为快速无损鉴别纸浆种类提供了一种准确可靠的方法。

Abstract:A method was developed for identification of paper fibers with near infrared spectroscopy (NIRS) technology and soft independent modeling of class analogy (SIMCA) pattern recognition. Ninety samples of four kinds of pulp including eucalyptus pulp, wheat straw pulp（bleached and unbleached）, cotton pulp and slash pine pulp were collected, which were prepared to the sheet samples with various basis weight, and some of eucalyptus pulp were added chemical additives. The NIRS of samples were measured. Sixty samples were chosen as calibration set to establish SIMCA class model and thirty samples as predicting set to verify the model. Results showed that every kind of pulp could be successfully identified by the established model without any influence of basis weight and chemical additives, which provides the rapid and accurate identification for raw materials of pulps.