人工神经网络吸光光度法同时测定钛和铌

确立了钛 4,5-二溴邻硝基苯基荧光酮 - CTMAB显色体系 ,在此基础上确立了钛和铌同时测定的显色体系 ,应用三层 ANN- BP网络解析钛和铌吸收光谱 ,吸光光度法同时测定铌和钛。方法选择性好 ,表观摩尔吸光系数 εTi557=1 .2 4× 1 0 5L· mol-1· cm-1,εNb54 0 =1 .63× 1 0 5L· mol-1·cm-1。对合金钢中铌和钛进行了同时测定 ,钛和铌的平均相对误差分别为 1 .30 %和 1 .0 8%。使用改进的 BP算法 ,避免了可能产生的麻痹现象。提出了便于网络参数选择的收敛评价函数。

SIMULTANEOUS SPECTROPHOTOMETRIC DETERMINATION OF TITANIUM AND NIOBIUM USING ARTIFICIAL NEURAL NETWORKS

In this paper a three-layer artificial neural networks was applied to the simultaneous determination of titanium and niobium by spectrophotometry. Color reactions of Ti( Ⅳ ) and Nb( V ) with 4,5-dibromo-o-ni-trophenyl-fiuorone in the presence of CTMABwas studied. The apparent molar absorptivities of the complexes of titanium and of niobium were found to be 1.24×105L·mol-1·cm-1 and 1.63×105L·mol-1·cm-1 respectively.A method was established for the simultaneous determination of titanium and niobium in steel with RSD's for Ti and Nb to be 1.3%and 1.08% respectively. The paralysis in the procedure of training about ANN was avoided with the improved backpropagation algorithms. The criterion function convenient for selecting of the parameters about the training of networks was also suggested.