【摘要】 使用改进的BP算法 ,避免了可能产生的麻痹现象 ,对几处酸及烷基磷酸酯及深井水样的电位滴定曲线进行拟合 ,用少量的实验数据恢复了大量的体系信息 ,并用曲线一次微商极大值确定电位滴定终点。更多还原

【Abstract】 A three\|layer artificial neural network is used to fit the curve of the potentiolmetric titration. The paralysis in the procedure of training of ANN has been avoided with the improved delta(δ) of output layer.A great amount of information of the system is obtained from a small number of experimental data point,The end\|points determined by the first derivative of those data picked from the fitted curves of the titration of Chloride acid,Acetic acid, Phosphorous acid and Dodecyl phosphate, underground water with Sodium hydroxide are accurate.更多还原