基于BP神经网络与遗传算法的涡轮安装角优化  

Optimization Design for Installation Angle of Turbodrill Blades Based on BP Neural Network and Genetic Algorithms

作　　者：谭春飞[1] 于瑞丰 郝明钊 刘彬 闫文双 

机构地区：[1]中国石油大学(北京)油气资源与探测国家重点实验室

出　　处：《石油机械》2018年第2期1-4,共4页China Petroleum Machinery

基　　金：国家科技重大专项"复杂结构井;丛式井设计与控制新技术"(2017ZX05009-003)

摘　　要：为了进一步提高涡轮钻具水力性能,采用FINE/Design 3D平台,基于BP神经网络与遗传基因算法对某型号?178 mm涡轮钻具转子叶型安装角进行优化设计,优化目标为水力效率与扭矩。对优化过程中样本库进行分析,总结了?178 mm涡轮钻具水力性能随转子安装角改变的变化规律。经过优化设计,单级涡轮输出扭矩提升了3.06 N·m,水力效率提高了8.29%。流场分析结果表明:优化后叶片性能的提高主要源于抑制了转子叶型吸力面的二次流发展,以及增大了压力面与吸力面间的压差。研究方法可为今后涡轮叶片的优化设计提供新的思路。In order to further improve the hydraulic performance of turbodrill, the FINE / Design 3D platform was used to optimize the rotor blade installation angle for a 0178 mm turbodrill based on BP neural network and ge-netic algorithm. The optimized target was hydraulic efficiency and torque. In the process of optimization, the sam-ple data base was analyzed. The variation of hydraulic performance of 0178 mm turbodrill with the change of rotor installation angle was summarized. Through the optimized design, the single-stage turbine had an increased output torque of 3. 06 N ？ m and an increased hydraulic efficiency by 8. 29%. The flow field analysis shows that the im-provement of the blade performance after optimization is mainly due to the suppression of the secondary flow devel-opment of the rotor blade suction surface and the increase of the pressure difference between the pressure surface and the suction surface. The research method provides a new idea for the future optimal design of turbine blades.

关 键 词：BP神经网络 遗传算法 涡轮钻具 安装角 优化设计 

分 类 号：TE921[石油与天然气工程—石油机械设备]