模糊聚类-RBF神经网络法同时测定铁、镍、钒、铜、钴

Simultaneous Determination of Iron,Nickle,Vanadium,Copper and Cobalt with Fuzzy Cluster-Radial Basis Function Neural Networks

将模糊聚类分析与RBF神经网络法相结合，对模拟原油中吸收光谱严重重叠的重金属多组分体系进行解析，较好地解决了光度分析计算中校准模型的优化问题，提高了分析结果的准确度。用此方法研究铁、镍、钒、铜、钴-乳化剂OP-5-Br—PADAP螯合显色体系，计算表明模拟样品各组分的回收率为94％～105．5％，油样测定结果的相对标准偏差在1．85％～4．3％之间。

The over lapping spectra of the multicomponent system of heavy metals in simulative samples of crude oil have been investigated by using the algorithm, which united the fuzzy cluster and radial basis function neural networks （RBF）. The optimization of the calibration sets in the calculation spectrophotometry analysis have been resolved, and the precision of the results has been improved. Applied this method to the preliminary study of the color reactions of Fe（ Ⅲ ） ,Ni（ Ⅱ ） ,V（ Ⅴ ） ,Cu（ Ⅱ ） and Co（ Ⅱ ） with 5-Br-PADAP in the presence of OP, the recovery test with five simulated samples gave results of recovery in the range of 94-105.5% and the relative standard deviation of five ions in crude oil samples were in the range of 1.85 % -4.3 % .