【作者】 张凯；

【导师】 庞宝君；

【作者基本信息】 哈尔滨工业大学 ， 固体力学， 2015， 博士

【摘要】 随着人类航天事业的发展,日趋恶化的空间碎片环境对航天器的安全运行构成严重威胁。作为应对措施,主要采取减缓、规避、防护等措施保护航天器的安全。但是,受技术经济条件的制约,空间碎片仍然存在对在轨运行航天器撞击的威胁,航天器特别是载人航天器对在轨感知技术具有迫切的应用需求。基于声发射的空间碎片撞击在轨感知技术对撞击事件的发生、撞击位置、损伤模式及损伤程度进行实时监测具有较好的应用前景。已有研究成果主要实现了如何判别铝合金平板结构空间碎片撞击事件发生与否,撞击源定位,但对撞击损伤进行模式识别及损伤程度评估尚不完善。而且,作为在轨感知技术的重要应用对象,载人航天器密封舱通常是铝合金加筋板,并在关键部位加装防护屏,其撞击声发射信号特性及信号在结构中的传播规律较之平板更为复杂,相关研究工作较少,需要深入开展研究工作。基于以上背景,本文以球形弹丸模拟空间碎片,分别撞击铝合金单层板和铝合金双层板结构,针对超高速撞击声发射信号特性、撞击源定位、损伤模式识别及损伤程度评估等问题开展研究。对于铝合金单层板结构,分析了弹丸速度、尺寸及靶板厚度对声发射信号特性的影响,以及加筋板筋体尺寸对信号特性及传播规律的影响;对于铝合金双层板结构,分析了弹丸撞击参数及双层板结构参数对二次碎片云撞击声发射信号特性的影响;在此基础上,分别提出单层板结构以及双层板结构撞击源定位及损伤模式识别方案。本文主要研究内容如下:第一,分别针对铝合金平板、铝合金加筋板及铝合金双层板结构等靶板样件,给出了超高速撞击声发射信号获取方案。搭建了超高速撞击实验平台及信号采集系统,用于模拟空间碎片超高速撞击载人航天器典型结构,获取超高速撞击损伤效应并采集声发射信号。采用AUTODYN有限元软件利用SPH算法仿真超高速撞击现象,建立有限元模型并获取了基于数值仿真的超高速撞击声发射信号。通过比较典型地面模拟实验及数值仿真结构中靶板损伤情况以及声发射信号特征,验证了数值仿真的有效性。第二,建立单层板结构撞击声发射信号特征参数与弹丸撞击参数、靶板几何参数之间的关系。通过地面模拟实验与数值仿真,基于板波理论及小波变换分析了弹丸撞击铝合金平板声发射信号特性,获得了弹丸初始速度、弹丸直径及靶板厚度对撞击声发射信号特性的影响规律;根据数值仿真及断铅实验,发现加筋板中筋体对信号S0模态波传播速度影响较小,在利用S0波到达时刻定位时,可将加筋板视为各向同性,但信号经过筋体后其幅值将衰减,其中高频成分与低频相比信号幅值衰减更快。第三,建立双层板结构撞击声发射信号特征参数与弹丸撞击参数、双层板几何参数之间的关系。利用超高速撞击实验,对二次碎片云撞击铝合金后板损伤特性、声发射信号特性及其之间的关系进行了研究。基于碎片云撞击声发射信号特征,利用小波包技术并引入能量熵原理构建碎片云撞击声发射信号特征参数小波包能量熵。研究结果表明,弹丸初始速度、前板厚度、弹丸直径等参数是影响后板损伤程度的重要因素;碎片云撞击声发射信号特征与后板的损伤程度及损伤区域密切相关;碎片云撞击声发射信号小波包能量熵值可以表征弹丸破碎程度进而实现对弹丸初始速度进行估计。第四,分别提出弹丸撞击铝合金加筋板、碎片云撞击铝合金平板的定位方法。根据铝合金加筋板超高速撞击实验,分析了声发射信号到达时刻判定方法对波速的影响。研究结果表明,将信号的第一峰值时刻作为到达时刻,可以减小由于传统阈值法引起的波速差异,利用S0模态波到达时刻及平板的定位方法可实现加筋板超高速撞击源定位;通过对碎片云形态及运动特性的分析,确定利用时差定位算法可定位出二次碎片云撞击损伤中心,根据碎片云撞击信号特性,建立了以信号高频能量比例为特征参数的预测碎片云损伤区域经验公式。第五,提出一种基于神经网络的单层板结构超高速撞击损伤模式识别方案。基于超高速撞击实验及数值仿真,分析了可用于损伤模式识别及弹丸直径估计的声发射信号特征参数;在利用数值仿真实验对铝合金平板结构损伤模式识别网络及弹丸直径估计网络进行可行性分析的基础上,基于地面模拟实验分别搭建了损伤模式识别神经网络及用于估计弹径的神经网络,进而实现弹丸撞击铝合金单层板结构损伤模式识别及损伤程度评估。第六,提出基于防护结构撞击极限曲线的后板损伤模式识别方法,给出双层板防护结构超高速撞击损伤模式识别方案。对于典型防护结构,如通过感知系统能够实现对弹丸直径和速度的估计,根据撞击极限曲线可实现对防护结构后板是否失效进行判别。铝合金双层板结构的损伤模式识别方案为:根据弹丸撞击前板声发射信号,利用弹径估计网络对弹丸直径进行估计;根据碎片云撞击后板声发射信号小波包能量熵值,对弹丸初始速度进行估计;结合双层板结构撞击极限曲线,实现损伤模式识别及损伤程度评估。综上,本文系统的研究了空间碎片撞击航天器典型结构的声发射信号特性,并详细地讨论了弹丸撞击单层板结构及双层板结构源定位问题及损伤模式识别问题,对发展基于声发射的空间碎片撞击在轨感知技术具有一定的参考价值。更多还原

【Abstract】 With the development of human spacecraft, space debris became a threat to the spacecraft. As a response, people have developed protection programmes such as mitigation, active dodge and protect from space debris. However, due to technical and economic conditions, the threat from space debris impact still exists on the spacecraft, space debris impact sensing technology had immediate application equirements for manned spacecraft. The on-orbit sensing technology based on acoustic emission had application prospect for monitoring event moments, source location, damage pattern recognition and damage degree evaluation. The existing research results achieved how to determine the impact event occurred and source location in aluminum alloy plate, but pattern recognition and damage degree evaluation had not be achieved. And as the important application of on-orbit sensing technology, manned spacecraft capsules were usually aluminum stiffened plates and the important parts were added protective shield. Compared with plate, the characteristics of acoustic emission impact signals and signal propagation in the structure were more complicated. The relevant research work were less and needed to be further studied.Under such background, spherical projectile were used to simulate space debris, respectively impact single-layer aluminum alloy plate and dual-wall structure, aimed at the questions such as the characteristics of acoustic emission signals caused by hypervelocity impact, source localization, damage pattern recognition and damage degree evaluation. For aluminum alloy single-layer structure, the effect of projectile velocity, size and plate thickness on the impact acoustic emission signal characteristics were analyzed, the size of stifferners effect on signal propagation law and characteristics were analyzed. For aluminum alloy dual-wall structure, the effect of projectile impact parameters and dual-wall structural parameters on the characteristics of the secondary debris cloud impact signal were investigated. On this base, the schemes of source location and damage pattern recognition for single layer plate and dual wall structure were investigated. This paper mainly includes the following:First, aluminum alloy plate, aluminum alloy stiffened plate and aluminum alloy dual-wall structure were designed as targets, hypervelocity impact experiments platform and signal acquisition system were designed, the hypervelocity impact damage effect and signals were obtained. By using SPH simulation of hypervelocity phenomena, finite element models were set up to obtain acoustic emission signal. By comparing experimental and simulation of damage and the impact of the target signal waveforms, the effectiveness of numerical simulation experiments was verifed.Secondly, the relationship between the characteristics parameters of acoustic emission signal and the projectile impact parameters, target plate geometric parameters were established. Through experimental and numerical simulation of hypervelocity impact experiments, the characteristics of acoustic emission signal caused by projectile hypervelocity impact aluminum alloy plate were analysed based on lamb wave theory and wavelet transform, the influence on the characteristics of acoustic emission signal by projectile initial velocity, diameter and plate thickness were discussed. Based on the numerical simulation and lead break experiment, it were found that the stiffeners had little effect on the wave propagation velocity of S0 mode. But the signal amplitude were attenuation when the signal through the stiffeners, high frequency components of signal amplitude attenuation faster than low frequency.Thirdly, the relationship between the characteristics parameters of acoustic emission signal and projectile impact parameters and dual-wall structural parameters were established. The characteristics of damage and its acoustic emission signal caused by the secondary debris cloud impact plate, and the relationship between them were investigated. According to the acoustic emission signal frequency distribution, based on wavelet packet and entropy principle, the characteristic parameter wavelet packet energy entropy were built. The researches showed that the initial velocity of projectile, bumper thickness and diameter of projectile had influence on the degree of damage, the characteristics of acoustic emission signal closely related to the degree of damage and the damage area. The wavelet packet energy entropy of debris cloud impact signal could be used to show the degree of projectile fragmentation, and then to estimate projectile velocity.Fourthly, location methods for projectile impact aluminum stiffened plate and debris cloud impact plate were proposed. According to the characteristics of acoustic emission signals, the advantages and disadvantages of signal arrival time determination were analyzed. Experimental results showed that in comparison with the the traditional threshold method, the first peak were used as the arrival time of signal could reduce the error of wave velocity. In addition, debris cloud impacted localization programme were given by experimental and theoretical analysis. Using time difference of arrival location algorithm debris cloud impact damage centre could be given, and debris cloud damaged area were predicted by the establishment of empirical equation.Fifthly, the damage pattern recognition method based on neural network for projectiles hypervelocity impact single layer plate were proposed. Based on hypervelocity impact experiment and numerical simulation, the feature parameters used for damage pattern recognition and estimating the diameters of the projectile were analyzed. The neural network used for damage pattern recognition and estimating the diameters of the projectile were built up. The pattern recognition and evaluation of damage degree for projectile hypervelocity impact single layer plate were achieved.Finally, the damage pattern recognition method based on ballistic limit curve for projectiles hypervelocity impact spacecraft with shield were proposed, the schemes for dual-wall structural were given. Acorrding to the acoustic emission signal on the bumper, the diameter of the projectile could be estimated by using neural network, acorrding to the debris cloud impact signals on the target, the velocity of the projectile could be estimated through the signal wavelet packet energy entropy, combined with the impact limit curve of dual-wall structure, the pattern recognition and evaluation of damage degree were achieved.This paper studied the characteristics of acoustic emission signals caused by space debris impacted the typical structure of a spacecraft bulkhead. And discussed in detail source localization and damage pattern recognition process of projectile impact single-layer and dual-wall structure. The results had showed great value for the development of acoustic emission on orbit monitoring technique.更多还原