人工神经网络X射线荧光光谱法测定钢中酸溶铝  

XRFS DETERMINATION OF ACID-SOLUBLE ALUMINUM IN STEEL WITH ARTIFICIAL NEURAL NETWORK

作　　者：董林[1] 李井会[1] 朱晓明[1] 高礼让[1] 朱秀慧[1] 宋立伟[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

出　　处：《理化检验（化学分册）》2004年第10期603-604,607,共3页Physical Testing and Chemical Analysis(Part B:Chemical Analysis)

摘　　要：　建立了一种人工神经网络-X射线荧光光谱法测定钢中酸溶铝的方法,用X射线荧光光谱法测定低合金钢中总铝值,应用所建立的ANN-BP网络模型,输入总铝含量直接预测出酸溶铝含量。同时使用改进的BP算法,避免了神经网络学习中可能产生的麻痹现象。该方法用于钢中酸溶铝的测定,结果满意。Total aluminum content in low alloy steel was determined by XRFS, and by applying the ANN-BP network model, the content of acid-soluble aluminum was calculated from the data of total aluminum content input. The application of the improved BP algorithm, paralysis which may happen in the training procedure of ANN, was effectively avoided. In its practical application, 17 standard samples were used as training collection, 6 standard samples were used as a collection of prediction, and 7 low alloy steel samples were taken to be analyzed by the proposed method. Results obtained are in conformity with those obtained by the chemical method.

关 键 词：X射线荧光光谱法 人工神经网络 全铝 酸溶铝 

分 类 号：O657.31[理学—分析化学]