遗传神经网络分光光度法同时测定低合金钢中钛、钼和钨  

Simultaneous Spectrophotometric Determination of Titanium,Molybdenum and Tungsten in Low Alloy Steel by Applying the Aiogrithm of Genetic Artificial Neural Network

作　　者：窦成[1] 李井会[1] 陈平[1] 褚宁[1] 唐艳秀[2] 张国民[2] 单桂艳[2] 

机构地区：[1]鞍山科技大学化工学院,鞍山114044 [2]鞍山钢铁集团公司,鞍山114021

出　　处：《理化检验（化学分册）》2008年第11期1074-1076,共3页Physical Testing and Chemical Analysis(Part B:Chemical Analysis)

摘　　要：构造遗传算法的多组分适应度函数,应用遗传算法自适应概率搜索能力优化神经网络结构,使网络结构和参数与输入数据达到最优匹配,建立用于多组分同时测定的遗传神经网络。在钛(钼、钨)-二溴羟基苯基荧光酮-乳化剂OP同时测定显色体系中,钛、钼和钨配合物的表观摩尔吸光率分别为1.03×10~5,1.31×10~5,1.21×10~5L·mol^(-1)·cm^(-1)。应用遗传神经网络(GA-ANN)分光光度法同时测定低合金钢标准样品中钛、钼和钨。On the building of multicomponent adaptive functions, optimal matching of network structure and parameters as well as the input data was achieved by applying the genetic algorithm to optimized the artificial neural network structure with adaptive probability searching ability, on the base of which the algorithm of genetic artificial neural network （GA-ANN） for simultaneous spectrophotometric determination of 3 components （Ti, W and Mo） was established. The color reagent used in the determination was dibromohydroxyphenylfluorone （DBHPF）, and values of molar absorptivity of complexes of DBHPF with Ti（Ⅵ）, Mo（Ⅵ ） and W（Ⅵ ） found were 1.03 × 10^5, 1.31 ×× 10^5 , 1.21 × 10^5 L · mol^-1 · cm^-1 respeetively. The proposed method was applied to the determination of Ti, Mo and W in a CRM of low alloy steel giving results in consistency with the certified values.

关 键 词：遗传神经网络 分光光度法 钛 钼 钨 二溴羟基苯基荧光酮 低合金钢 

分 类 号：O657.31[理学—分析化学]