摘要
传统的卫星用飞轮轮体通常使用一体加工的方式,由整块材料加工而成,具有材料消耗大,轮体结构偏重,备料周期长,结构优化水平受加工能力限制较大、不易进行模块化生产等问题。为了达到优化传统的卫星用飞轮轮体结构,提高轮体性能,获得高质量惯性比的目的,本文提出了一种轻质化、高可靠性的焊接飞轮研制方法。通过BP神经网络对焊接参数进行非线性拟合,使用模拟退火的粒子群算法对焊接参数进行寻优。实验结果表明该方法制造的飞轮在力学试验、微振动、重量等方面均优于传统飞轮。
IIn order to optimize the structure of the wheel body and improve the performance of the wheel body.A lightweight and high reliability welding flywheel development method was proposed.The welding parameters were fitted nonlinearly by the Bp neural network.the particle swarm optimization algorithm simulated annealing was used to optimize the welding parameters.The experimental results show that the flywheel manufactured by this method was better than the traditional flywheel in terms of mechanical test,micro-vibration and weight.
作者
申天宇
郑少帅
季文玮
张子维
王洁
SHEN Tian-yu;ZHENG Shao-shuai;JI Wen-wei(Shanghai Institute of Aerospace Control Technology,Shanghai 201109;Shanghai Inertial Technology Research Center,Shanghai 201109)
出处
《航空精密制造技术》
2024年第5期44-46,57,共4页
Aviation Precision Manufacturing Technology
关键词
模拟退火算法
粒子群算法
BP神经网络
simulated annealing algorithm
particle swarm optimization
backpropagation neural network