人工神经网络光度法同时测定5组分染料混合物

Application of Artificial Neural Network to Simultaneous Spectrophotometric Determination of Five Components of Dyestuff Mixture

应用人工神经网络原理 ,以快速BP算法 ,对紫外可见吸收光谱严重重叠的 5组分的染料溶液同时进行含量测定。在 2 0 0～ 5 80nm的范围内 ,以 10个特征波长处的吸收值作为网络特征参数 ,通过网络训练 ,甲基紫、酸性红B、酸性橙Ⅱ、酸性嫩黄、碱性桃红的相对标准偏差为 0 .2 2 %～ 4 .80 % ;5种成分的回收率在98.3%～ 10 3%之间。实验表明 ,该算法速度快 ,预测结果准确 ,并用该方法定量测定光解废水中 5组分混合染料。

By means of artificial neural network and back-propagation train algorithm at speed, a method for the simultaneous spectrophotometric determination of five-components of dyestuff mixture was proposed. In the range of 200 similar to 580 nm, the absorbances at 10 characteristic wavelengths were taken as characteristic parameter of artificial neural network. The RSD of the method for the determination of methyl violet, acidic red B, acidic orange II, acidic light yellow and safranine T is between 0.2 % - 4.8 %. The recovery of the results is between 98.3% similar to 103%. The results are better in training speed and the accuracy. In conclusion, the artificial nerve network combined with spectrophotometry is a good method for resolving and determinating the five-components of dyestuff mixture in waste water.