基于聚类算法的热试验加热器布局设计与优化  

The layout design and optimization of heating circuits in thermal test based on clustering algorithm

作　　者：龚喆 侯雅琴 李西园[2] 吴东亮[2] 邵华[2] 高庆华[2] 李琼[2] GONG Zhe;HOU Yaqin;LI Xiyuan;WU Dongliang;SHAO Hua;GAO Qinghua;LI Qiong(China Academy of Space Technology;Beijing Institute of Spacecraft Environment Engineering:Beijing 100094,China)

机构地区：[1]中国空间技术研究院 [2]北京卫星环境工程研究所,北京100094

出　　处：《航天器环境工程》2024年第4期421-429,共9页Spacecraft Environment Engineering

基　　金：北京市科技新星项目(编号:Z211100002121077)。

摘　　要：针对目前航天器热试验中热流模拟需求复杂、加热回路设计效率低以及不易量化等问题,文章以某型号中的一组热控模拟件为典型对象,通过有限元方法建立不同特征长度组件和薄膜电加热器组合所对应的温度不均匀度模型;提出基于温度不均匀度距离的聚类方法,并开展基于MATLAB的聚类仿真。与欧几里得距离聚类的对比表明,在同等条件下温度不均匀度距离算法可将温度不均匀度降低30%左右。此外,对比了聚类簇数、加热表面细分数及待加热表面数量等参数对聚类结果和温度不均匀度的影响,可为未来热试验中的加热回路布局设计与优化提供参考。At present,there exist issues in spacecraft thermal test involving intricate demands for thermal flux simulation,low efficiency in heating circuit design,and difficulty in quantification.Taking a set of thermal control simulation components of a mission as representative cases,temperature non-uniformity models corresponsive to the combination of various featured length components and thin film electric heaters were established through finite element method.A clustering method based on the temperature non-uniformity distance was presented.The clustering simulation was conducted based on MATLAB.Compared with the Euclidean distance clustering method,the results show that the temperature non-uniformity with the proposed algorithm can be reduced by about 30%under the same conditions.In addition,the influence of clustering number,heating surface fine fraction,and number of surfaces to be heated on the clustering results and temperature nonuniformity were compared.It may provide a reference for the layout design and optimization of heating circuits in the future thermal test.

关 键 词：航天器热试验 聚类算法 薄膜电加热器 热设计 温度不均匀度 

分 类 号：V474[航空宇航科学与技术—飞行器设计]