人工网络法计算累积稳定常数和化学形态

Calculation of Accumulative Stabilization Constant and Chemical Speciations of Lead by Artificial Neural Network

用前馈线性网络法的不同算法计算了水体系中Ｐｂ２＋与ＯＨ－之间的累积稳定常数，并比较了各种算法的特点，根据计算得到的累积稳定常数，用反馈网络法计算了水溶液中Ｐｂ２＋各种化学形态的浓度，△Ｅｐ随ｐＨ变化的模拟曲线，讨论了影响模拟结果的原因，结果表明，测量误差、在４＜ｐＨ＜９时ＣＯ２－３对Ｐｂ２＋形态分布有明显的影响，在实验误差范围内。

Linear Feedforward Neural Network model is used to calculate the Accumulative Stabilization Constants between Pb 2+ and OH - by three different algorithms.Using the data obtained,the concentrations of different chemical speciations in aqueous solution for Pb 2+ and simulation curve of ΔE p change with pH are calculated using the feedback Neural Network.It shows that the measuring errors and CO 2- 3 within 4<pH<9 have great effect on the results of simulation while effect of the temperature can be omitted within the experimental error.