基于PCA的水质紫外吸收光谱分析模型研究

利用紫外光谱分析水中有机污染物已成为水质实时在线监测的重要方法之一，水样组分复杂且不稳定是影响其测量结果的主要因素。利用主成分分析法(PCA)结合欧氏距离分析水样紫外吸收光谱，对水样分类，效果良好。分别用主成分分析结合偏最小二乘法回归(PCA-PLSR)和直接利用多波长吸光度结合偏最小二乘法回归(MWA-PLSR)建立分析模型，并对比分析了不同浓度的COD标准液的实验数据。结果表明，采用第一、二主成分作为回归参数的PLSR模型的测量误差在5%以内，偏差最小。利用本文方法可同时实现水样分类和水质参数的精确定量。

Research on Water Quality Analysis Model with PCA Method and UV Absorption Spectra

Using the UV absorption spectrum to detect Organic pollutants content in water has become one of the most important methods for real-time online monitoring in the field of water quality inspection,however,the water complex and unstable compo-nents often bring much uncertain offset to the standard test.In this paper,water samples were classified firstly by analyzing UV absorption spectrum ranging from 200 nm to 400μm including the organic substances,through the way of combining principal component analysis (PCA)with Euclidean distance.In this paper,we compared the Principal component analysis combined with partial least squares regression (PCA-PLSR)and the direct multi-wavelength absorption models combined with partial least squares regression (MWA-PLSR),not only for the real water sample but also for the analysis of different concentrations of COD standard solution.The result indicates that the measurement errors of the PCA is less than 5%,it is the smallest by using the first and second principal components as regression parameters for PLSR.Using the methods above can simultaneously achieve to classify of water samples and to measure the concentration of water quality parameters more accurately.