摘要
基于人工神经网络(ANN)的皮尔斯电子枪初值设计方法,不但能有机地结合迭代综合法与非迭代综合法的优点,还能把前人体现在实验数据中的设计经验通过训练样本融入到其中,从而设计出很接近实验值的新电子枪。在有较多电子枪实验数据的情况下,ANN法的优点尤为显著,经20多支电子枪的检测结果表明:ANN法的设计值比综合法的计算值更接近实验结果。
Artificial neural network(ANN) has been successfully used in the design of Pierce electron gun. The newly-developed design technique not only combines the strengths of both the iterative and non-iterative synthesis methods but also makes use of the existing data and past exiperience in previous designs of Pierce electron gun. The numerically simulated results agree well with the test results of over 20 Pierce electron guns, showing that ANN technique works better than the synthesis methods.
出处
《真空科学与技术学报》
EI
CAS
CSCD
北大核心
2007年第5期380-385,共6页
Chinese Journal of Vacuum Science and Technology
基金
国防科技基础研究基金资助项目(No.51440030203DZ02)
关键词
人工神经网络(ANN)
综合法
皮尔斯电子枪
阴极半锥角
射程
Artificial neural network, Synthesis technique, Pierce electron gun, Convergence half angle, Gunshot