摘要
研究用RBF神经网络建立乙醛氧化制醋酸氧化塔收率模型 ,并对RBFN的训练信息的 β值、Δ值的选取和在线学习等方面作了研究和改进 ,与改进前相比 ,性能明显提高。现场运行结果表明所建立的模型简单、精度高 ,能满足工程要求。
The yield model of oxidation tower in aldehyde oxidizing to acetic acid was built by RBF (Radial Basis Function) neural network in this paper, which studied and improved the selection of β and Δ of training information and modeling study online. It has proved that the model presented in this paper is simple, highly accurate and can meet industry demand.
出处
《北京化工大学学报（自然科学版）》
EI
CAS
CSCD
2001年第1期63-66,共4页
Journal of Beijing University of Chemical Technology(Natural Science Edition)
基金
中国石油天然气总公司资助!项目 (990 818 0 1 14)
