摘要
在对石油射孔枪结构进行有限元分析的基础上,利用BP神经网络建立射孔枪结构设计参数盲孔处最大应力与盲孔深度、盲孔直径的全局性映射关系,获得遗传算法求解结构优化问题所需的目标函数值,用改进的遗传算法进行射孔枪结构优化设计。结果表明,基于神经网络和遗传算法的优化技术应用在射孔枪结构优化设计中有效、合理。提出的优化技术为工程领域中复杂、多变量,尤其是设计目标无法或难以表示成设计变量显函数的优化问题求解,提供了新的思路和技术手段。
This paper first analyzed the frame structure of oil perforating gun with finite element method. Then a non -linear mapping function was constructed within BP neural networks from maximum stress at blind hole to blind hole depth and diameter calculated. The objective function values used for solving the structural optimization problems with genetic algorithm were obtained, modified genetic algorithm was used to optimize perforating gun structure. The results show that application of this optimal technology based on neural networks and genetic algorithm to perforating gun structural design is effective and reasonable. The technology provides a new thought and method for solving those optimization problems existing in complex structure with multi-variables especially when the optimization object can hardly be explicitly expressed as the function of desigu variables.
出处
《石油钻采工艺》
CAS
CSCD
北大核心
2009年第1期47-50,共4页
Oil Drilling & Production Technology
基金
国家技术研究发展计划(863)计划(编号:2006AA098326)
吉林省科技发展计划项目(编号:20060535)部分内容
关键词
射孔枪
结构优化
神经网络
遗传算法
perforating gun
structural optimization
neural network
genetic algorithm