专题编号与专题名:新理论、新方法、新技术(10.5)
宏微跨尺度空间粉尘超高速撞击复杂航天器精确源定位方法研究
卢艺博
1; 马可宸
1; 康妍
1; 师湫航
1; 曹武雄
2*; 李丽芳
3*
- 哈尔滨工业大学 未来技术学院,哈尔滨 150001
- 哈尔滨工业大学 航天学院空间碎片高速撞击研究中心,哈尔滨 150001
- 哈尔滨工业大学 空间环境与物质科学研究院,哈尔滨 150001
摘要:面对地球轨道上亿万颗行星际尘埃和空间碎片构成的超高速撞击威胁,其中撞击速度和粒径的广泛分布使得来源难以确定。针对此问题,本研究提出了一种利用超声传感阵列原位感知辨识跨尺度颗粒撞击定位源的新方法。本方法利用超声传感阵列网络获取其撞击航天器产生的声发射波信号,结合人工智能方法获取撞击源的精确位置。本研究以空间站复杂舱段结构为研究对象,通过数值模拟和低速撞击实验获取超高速撞击声发射波信号,结合信号分析和数据处理构建定位样本数据库,并基于深度学习开发小样本快速增量学习和迁移学习方法,以实现样本快速筛选、缺失补偿和数量增强;在此基础上,结合集成加权回归开发计算效率较高的撞击源精确定位方法,并通过实际舱体局部结构模拟件的超高速撞击实验验证算法的精度及普适性。该定位方法计算效率较高,且不受材料属性(非均匀厚度、材料非各向同性)、结构不连续(凸起、孔洞)等导致的超声波显著扰动和传播路径复杂化的影响;未来该算法可进一步从局部模拟件扩展应用到整个舱体结构,实现在轨更新和撞击精确源定位。本方法简洁可靠,预期将广泛应用于卫星、空间站、空间探测器等各类航天器,为它们的在轨安全运行提供坚实保障。
第一作者简介: 卢艺博(2004—),男,本科生,研究方向:行星际尘粒碰撞感知与探测技术. E-mail:648255593@qq.com
*通信作者简介:
曹武雄(1989—),男,助理研究员,“行星际尘粒撞击感知及致损评估”,E-mail: wuxiong.cao@hit.edu.cn
李丽芳(1981—),女,教授,研究方向:行星多因素环境模拟与试验技术,空间太阳能系统,E-mail:lilifang@hit.edu.cn
Accurate Localization of Space Debris and Interplanetary Dust Particles Hypervelocity Impact- induced Damage on Complex Spacecraft
Abstract
Faced with the threat of hypervelocity impact (HVI) of billions of interplanetary dust particles and space debris in Earth orbit, ultrasonic sensor network is utilized to acquire HVI-induced acoustic emission(AE) wave signals on spacecraft in this study, combining with the use of artificial intelligence methods to obtain the accurate location of HVI, so as to provide support for the safety of the on-orbit spacecraft. This study focuses on the complex module structure of the space station, and acquires the HVI-induced AE wave signals through numerical simulation and low-speed impact experiments. then a localization sample database is built by combining signal analysis and data processing. In addition, deep learning-based small-sample fast incremental learning and transfer learning methods are developed to realize fast screening, missing compensation and quantity enhancement of data samples. On this basis, a computationally efficient impact source location algorithm is proposed by combining the integrated weighted regression, and verified through HVI experiments on simulated parts of the actual spacecraft. The proposed localization method has high computational efficiency and is not affected by the significant ultrasonic wave perturbation and propagation path complexity caused by material properties (non-uniform thickness, non-isotropic material) and structural discontinuities (bumps, holes), etc. In the future, this algorithm can be applied to the whole spacecraft structure to realize the on-orbit updating and accurate localization of HVI.
发表评论