速差动力学分析法同时测定铁、锌、铜

Kinetic Spectrophotomatric Determinations of Mixtures of Iron, Zinc and Copper Using Artificial Neural Networks

将化学计量学方法引入速差动力学分析方法中，在不预知动力学模型参数（速率常数）的情况下，用人工神经网络（ＡＮＮ）依据铁、锌、铜的ＥＧＴＡ配合物与ＰＡＲ置换反应的速度差异，对其三组分混合体系进行了同时测定．并对人工神经网络和偏最小二乘法对多波长、多时间点的三维量测模型的解析能力进行了比较，结果表明前者总体上优于后者．混合体系中铁、锌、铜测定的相对标准偏差分别为１．６３％，３．２９％和４．４１％．本法还被用于饲料添加剂中微量元素的测定。

In this paper, chemometric approaches were applied to the simultaneous kinetic analysis of ternary mixtures of iron, zinc and copper. The analysis of these metals was based on the displacement reaction of their EGTA complexes by PAR with the differential reaction rates. The measurement data were then processed by artificial neural networks(ANN), giving a relative standard error of 163%, 329% and 441% for iron, zinc and copper, respectively. The proposed method was also applied to the analysis of feed additive samples with satisfactory results as compared with that obtained by ICPAES.