摘要
基于无监督机器学习技术和Spark云平台,将"无监督机器学习"与"支吊架布置设计"相结合,解决传统核电厂支吊架设计过程烦琐、费时耗力、出错率高、人力成本高等问题,解决"人工智能+云计算+支吊架设计"设计模式中训练样本集数量不足的问题,大大降低了支吊架布置设计过程中重复率高、迭代效率慢等问题,验证了低样本数量支吊架布置设计的可行性与准确性。
Based on the unsupervised machine learning technology and spark cloud platform,combining"unsupervised machine learning"with"support and hanger layout design",this paper solves the problems of cumbersome,time-consuming,high error rate and high labor cost in the traditional support and hanger design process of nuclear power plant,and solves the problem of insufficient training sample set in the design mode of"artificial intelligence+cloud computing+support and hanger design",which is greatly improved It reduces the problems of high repetition rate and slow iteration efficiency in the process of support and hanger layout design,and verifies the feasibility and accuracy of support and hanger layout design with low sample number.
作者
肖韵菲
黄捷
孙冠宇
高希龙
陈建国
文婷婷
文剑
XIAO Yunfei;HUANG Jie;SUN Guanyu;GAO Xilong;CHEN Jianguo;WEN Tingting;WEN Jian(Key Laboratory of Nuclear Reactor System Design Technology,China Nuclear Power Research and Design Institute,Sichuan 610213,China;Sichuan Electric Power Design Consulting Co.,Ltd.,Sichuan 610041,China)
出处
《电子技术（上海）》
2021年第1期58-61,共4页
Electronic Technology
