摘要
针对分子蒸馏过程多变量、非线性、内部机理复杂、建模困难等问题,基于神经网络自学习、自适应及强非线性映射能力,提出了改进的BP神经网络产品纯度预测模型,深入探讨了神经网络在分子蒸馏过程中的应用。实验证明所提出的模型可以用来预测产品纯度。
The molecular distillation process has the features of multiple variables , nonlinearity , complex internal mechanism and difficult modeling . Based on self-learning , self-adapt and strong nonlinear mapping properties of the modified BP neural network ,a estimation model for the product purity is put forward and applied into the molecular distillation process .The experiments verify that the model is suitable for the product purity estimation .
出处
《长春工业大学学报》
CAS
2013年第5期488-492,共5页
Journal of Changchun University of Technology
基金
吉林省科技厅基金资助项目(20101505
20111819)
关键词
分子蒸馏
预测模型
BP神经网络
产品纯度
molecular distillation
estimation model
BP neural network
product purity
