摘要
阐述了神经网络用于过程建模目前存在的问题,提出了改进的方法,并将此应用到了大型炼油厂脱蜡过程的建模中。通过比较改进前后的BP算法仿真结果可以看出,改进后的算法在网络的训练收敛速度上有了较大提高。
The problem is presented in BP algorithm of the neural network for the processmodelling. A fast BP learning algorithm and BFGS (Broyden, Fletcher, Goldstein and Shanno)algorithm are discussed. Simulation demonstrates that the results are satisfactory in the dewaxingProcess whon this method is used.
出处
《青岛化工学院学报(自然科学版)》
1996年第4期387-391,共5页
Journal of Qingdao Institute of Chemical Technology(Natural Science Edition)
关键词
过程建模
BP算法
建模
神经网络
neural network
fast BP learning algorithm
modelling