摘要
丙烷氨氧化制丙烯腈的催化剂中以 V Sb Al O X 体系最有前景,本文在优化催化剂配方的基础上,利用基于 B P算法的多层前向神经网络对其进行工艺条件的辅助优化设计。优化设计的结果是:丙烷转化率为85.2% ,丙烯腈选择性 69.2% ,丙烯腈的单收达到 58.9% 。结果表明利用神经网络的良好性能来辅助催化剂设计是十分有效的方法.
In the paper, a neuralnetw ork ( N N) aided technique, in w hich the N N isa feed forw ard netbased on Back propagation algorithm , w asused to optim ize the reactionconditions for propane am m oxidation to acrylonitrile w ith respect to the catalysts devel oped w ith optim izcd design technology. The results of optim ization show ed that the pro posed technique w as available to design catalysts for propane am m oxidation to acryloni trile, the yield ofacrylonitrile for the best catalystw as up to 58.9% .
出处
《化学反应工程与工艺》
CAS
CSCD
北大核心
1999年第3期282-287,294,共7页
Chemical Reaction Engineering and Technology
基金
国家自然科学基金