人工神经网络在汽液相平衡常数预测中的应用

Application of Artificial Neural Network to Vapor-Liquid Equilibrium Constant Prediction

汽液相平衡常数的计算在化工分离过程中至关重要 ,传统方法参数众多 ,计算过程复杂 ,耗用机时多。本文在文献数据的基础上 ,首先利用过程模拟软件DesignⅡ对汽液相平衡常数进行了计算 ,然后使用三层BP神经网络及L -M算法对汽液相平衡常数进行预测 ,并以其他多种算法作为对比。结果表明 ,预测数据与实验数据吻合相当好 ,L -M算法运算速度明显快于其他算法 。

The calculation of vapor liquid equilibrium constant plays an important role in the chemical process of separation. There are many parameters in traditional methods, which are time consuming tasks, and the computation processes are quite complicated. In this paper, three layer BP neural network with Levenberg Marquardt algorithm is used to predict vapor liquid equilibrium constant. The predicted data are in good agreement with the experimental data. Various methods are used, the results indicate that Levenberg Marquardt algorithm is much faster than the other algorithms.