基于GA-Bp算法的炼油化工机械性能建模与仿真研究

Modeling and Simulation of Refinery and Chemical Machinery Performance Based on GA-Bp Algorithm

传统的炼油化工机械性能模型仿真能力差,得到的结果准确性低。为了解决这一问题,基于GA-Bp算法建立了一种新的炼油化工机械性能模型,GA-Bp算法是同时具备Bp算法与遗传算法的优点,利用Bp算法良好的泛化和非线性映射特点与遗传算法具有的全局搜索能力的特点,结合两种算法的泛化能力、非线性映射能力、全局搜索能力写出了一种训练神经网络的算法(GA-Bp算法)。GA-Bp算法通过优化权值的范围,从而缩小对最优解的搜索范围,以此增加了训练神经网络的学习次数。实验结果表明,基于GA-Bp算法的炼油化工机械性能模型准确性更高。

The traditional oil refinery chemical machinery performance model has poor simulation ability and the obtained results have low accuracy.In order to solve this problem,a new performance model of refining and chemical machinery was established based on GA-Bp algorithm.GA-Bp algorithm has the annoying advantages of both Bp algorithm and genetic algorithm.It uses the good generalization and nonlinear mapping characteristics of Bp algorithm.With the characteristics of global search ability of genetic algorithm,combined with the generalization ability,non-linear mapping ability and global search ability of two algorithms,an algorithm for training neural networks(GA-Bp algorithm)was written.The GA-Bp algorithm reduces the area of the optimal solution space by optimizing the weights,and uses the Bp neural network algorithm to solve it again,thereby increasing the number of learning times for training the neural network.The experimental results show that the GA-Bp algorithm is more accurate in refining chemical machinery performance models.