Institute Of Mechanics,Chinese Academy of Sciences
Not a member yet
33838 research outputs found
Sort by
Gravitational wave signal denoising and merger time prediction with a deep neural network
No full textThe mergers of massive black hole binaries could generate rich electromagnetic emissions, which allow us to probe the environments surrounding these massive black holes and gain deeper insights into the high energy astrophysics. However, due to the short timescale of binary mergers, it is crucial to predict the time of the merger in advance to devise detailed observational plans. The overwhelming noise and slow accumulation of the signal-to-noise ratio in the inspiral phase make this task particularly challenging. To address this issue, we propose a novel deep neural denoising network in this study, capable of denoising a 30-day inspiral phase signal. Following the denoising process, we perform the detection and merger time prediction based on the denoised signals. Our results demonstrate that, for a 30-day inspiral phase data with a signal-to-noise ratio between 10 and 50 occurring no more than 10 days before the merger, our absolute prediction error for the merger time is generally within 24 h
Evolution of surface and subsurface properties of Zr-4 alloy under the combined effects of laser heat treatment and laser gas nitriding: Microscopic morphology, chemical composition, and mechanical behavior
No full textZr-4 alloy, extensively employed as cladding in nuclear reactor fuel rods, stands as the foremost line of defense for reactor safety. Given its susceptibility to defects such as micro-pores during casting and its long-term exposure to extreme conditions during practical service processes, enhancing the surface and subsurface properties of Zr-4 alloy is essential. Herein, a nanosecond pulsed laser is used to irradiate the Zr-4 alloy in a nitrogen environment to investigate the evolution of its surface and subsurface properties under the combined effects of laser heat treatment and laser gas nitriding. The experimental results reveal a remarkable augmentation in surface hardness and scratch resistance of the Zr-4 alloy after laser irradiation. Notably, the hardness of the laserirradiated surface obtained with a laser power of 23.0 W and a scanning speed of 150 mm/s reaches 19.10 GPa, which is more than 7 times higher than that of the untreated surface (2.37 GPa). These improvements are attributed to the synergistic effects of the introduction of hard ZrN phase, the reduction of subsurface porosity, and grain refinement. This study demonstrates a promising approach for significantly improving the mechanical properties of Zr-4 alloy, holding considerable practical significance for its applications in the nuclear energy industry and other relevant fields
High-throughput investigation of temperature-dependent creep mechanisms in Mg-Zn alloys using nanoindentation
No full textA novel high-throughput methodology is proposed to investigate the influence of Zn alloying and temperature on the creep mechanisms of Mg through a combination of a Mg/Mg-Zn diffusion couple and advanced nanomechanical testing. Systematic nanoindentation creep tests were conducted at room temperature (RT) and elevated temperatures (100 degrees C, 200 degrees C and 300 degrees C) within individual grains in a Mg/Mg-Zn diffusion couple with Zn content ranging from 0 to 0.8 at%. Electron backscattered diffraction-assisted trace analysis was employed to characterize the activated deformation mechanisms, including twinning and dislocation slip. The activation volume and creep activation energy were analyzed to elucidate the fundamental creep mechanisms as functions of temperature and Zn content. At the low temperature regime (RT-100 degrees C), the creep activation energy was approximate to 80 kJ mol(-1), indicating that twinning dominated deformation mechanisms. At the high temperature regime (>200 degrees C), all alloys converge to a relatively high activation energy, of the order of 150 kJ mol(-1). This is in good agreement with the activation energy reported for cross-slip deformation mechanisms. The results indicate that the derived activation energy aligns with the transition in creep deformation mechanisms, shifting from twin boundary migration to cross-slip. Moreover, Zn alloying significantly enhances creep resistance, particularly within the temperature range of RT to 200 degrees C
Higher strength of double-double over traditional quad laminates in equal-stiffness design of single lap joint
No full textDouble-Double (DD) is a novel laminate design concept that not only reduces design variables and simplifies the manufacturing process but also exhibits unique homogenization properties. While theory suggests that for any QuaDriaxial (QUAD or QD) laminate, DD can provide a corresponding laminate configuration with equivalent stiffness, the mechanical behavior of their Single-Lap Joint (SLJ) still requires further investigation. In this study, a DD SLJ equivalent to the quasi-isotropic QD SLJ commonly used in fuselage structures is designed by ensuring equivalence in the in-plane stiffness matrix. While the experimental results show nearly identical in-plane stiffness, the tensile strength of the DD SLJ is approximately 13 % higher than that of the Quad SLJ. Additionally, DD SLJ exhibit reduced out-of-plane deformation compared to Quad SLJ. Both theoretical and numerical analyses are conducted to elucidate the underlying mechanisms responsible for the superior mechanical behavior of DD SLJ. The findings indicate that the observed differences are primarily due to variations in bending stiffness, despite identical in-plane stiffness matrices. The results further demonstrate that the smaller bending deformation of DD laminates leads to reduced interfacial peeling forces, making DD SLJ less susceptible to failure. Finally, a comparison of different staggered ways of DD laminates indicated no significant differences in strength or stiffness. The study demonstrates that DD SLJ, designed with equivalent in-plane stiffness matrices, not only achieve the same tensile stiffness as QD SLJs but also exhibit higher strength. This highlights the great potential of DD laminates for application in the single-lap joints of aircraft structures
Auto-resolving the atomic structure at van der Waals interfaces using a generative model
No full textThe high-resolution visualization of atomic structures is significant for understanding the relationship between the microscopic configurations and macroscopic properties of materials. However, a rapid, accurate, and robust approach to automatically resolve complex patterns in atomic-resolution microscopy remains difficult to implement. Here, we present a Trident strategy-enhanced disentangled representation learning method (a generative model), which utilizes a few unlabelled experimental images with abundant low-cost simulated images to generate a large corpus of annotated simulation data that closely resembles experimental results, producing a high-quality large-volume training dataset. A structural inference model is then trained via a residual neural network which can directly deduce the interlayer slip and rotation of diversified and complicated stacking patterns at van der Waals (vdW) interfaces with picometer-scale accuracy across various materials (e.g. MoS2, WS2, ReS2, ReSe2, and 1 T'-MoTe2) with different layer numbers (bilayer and trilayers), demonstrating robustness to defects, imaging quality, and surface contaminations. The framework can also identify pattern transition interfaces, quantify subtle motif variations, and discriminate moir & eacute; patterns that are difficult to distinguish in frequency domains. Finally, the high-throughput processing ability of our method provides insights into a vdW epitaxy mode where various thermodynamically favorable slip stackings can coexist
Comparative study on in-situ testing methods for undrained shear strength of cohesive soil
No full text随着全球能源需求的持续增长，浅海油气资源逐渐枯竭，海洋工程向深水及 超深水领域拓展，准确评估海洋土体的力学特性成为保障工程安全的关键。其中， 黏性土的不排水抗剪强度直接影响海底基础、管道和锚固系统的稳定性，但由于 不同原位测试方法（十字板剪切试验（VST）、圆锥贯入试验（CPT）、T-bar贯入 试验（TPT）和球型贯入试验（BPT））在实际应用中存在显著差异。这些差异一 方面源于测试方法的原理不同，另一方面则与土体的率效应、软化效应以及破坏 模式的复杂性密切相关。因此，如何统一评估这些方法的可靠性并揭示其内在关 联，成为工程实践中亟待解决的问题。为此，本研究使用耦合欧拉-拉格朗日(CEL) 方法进行了一系列大变形有限元分析，通过Fortran语言编写用户自定义子程序 （VUSDFLD）实现了修正的Tresca模型，引入了应变软化和原位剪切强度的应 变率效应。通过一系列灵敏度分析验证的数值模型具有高可靠性，准确捕捉了不 同测试方法的测试过程中的响应。结果表明，在十字板剪切试验的模拟中，随着 转动角度的增大等效塑性应变区域逐渐从方形发展为圆形。对于贯入试验（CPT、 TPT 和BPT），模拟结果揭示了不同的破坏模式：BPT和TPT在贯入过程中依次 经历开放空腔、封闭空腔和全流三个阶段，而CPT的塑性变形区域相对较小， 但是影响范围内产生了相对较大的等效塑性应变，这些差异直接影响了各测试方 法的解译结果。通过对不同测试方法的横向对比，研究发现，在不考虑软化效应 和率效应的情况下，BPT和TPT在土体强度较高时难以形成全流机制，其解译值比初始抗剪强度低约15%，CPT和VST的阻力系数在很大程度上不受剪切强 度的影响。引入率效应和软化效应后，TPT、BPT和VST的解译强度普遍降低 4%-5%，但CPT的解译强度(sut)却反常地增加了13.5%。此外，在单独考虑应变 软化的影响时，在不同的软化参数条件下，CPT的强度折减比NS/N0（76%-100%） 明显高于TPT、BPT和VST的强度折减比NS/N0（65%-88%），证明了CPT在所 研究的四种测试方法中受应变软化的影响最小</p
Multi-scenario dynamic load identification based on CNN and LSTM networks
No full text动态载荷识别在结构优化设计、健康监测与安全维护等领域具有重要的工程应用价值。相比于静态载荷，动态载荷具有时间依赖性和频率特性，并且结构动态响应涉及惯性、阻尼和弹性效应的耦合，从而大幅增加了载荷识别的难度。传统动态载荷识别方法通常高度依赖于精确的结构模型，在处理非线性系统和复杂载荷形式时精度不足，难以满足日益严苛的工程结构要求。近年来，人工智能技术蓬勃发展，其中神经网络凭借其强大的非线性拟合能力、自适应学习能力和特征提取能力，避免了对结构模型的依赖，为突破传统方法的局限提供了新途径。基于上述背景，本文系统开展了从理论模型到实际工程应用的递进式研究，依次针对线性系统分布载荷识别、非线性系统随机载荷识别以及机翼结构的突风载荷识别三类问题，逐步深入探讨了在不同动态载荷识别场景下神经网络的构建与优化策略，并通过模型性能对比分析，验证了方法的有效性。主要研究内容如下：
首先，针对线性系统的分布载荷识别问题，以悬臂梁结构为研究对象，在频域内开展了基于卷积神经网络（CNN）的载荷幅值识别研究。本文提出以三种典型载荷分布形式（直线型、斜线型和抛物线型）作为基向量，通过其线性组合近似复杂载荷分布。采用振型叠加法获得结构在典型分布载荷作用下的稳态响应，构建了训练样本。考虑到结构响应在固有频率附近呈现明显的局部峰值特征，且CNN能够高效提取局部特征并保持数据的空间关联性，进一步建立了1D-CNN模型，实现了从单一测点的频率-稳态加速度幅值到分布载荷幅值的精准识别，突破了传统方法对测点数目的限制。结果表明，该模型在典型载荷识别中的平均相对误差（MRE）低至0.077%，并对未参与训练的叠加型载荷也表现出良好的泛化性能，验证了基向量建模策略的有效性。
然后，针对非线性系统的随机载荷识别问题，以Duffing系统为研究对象，开展了基于长短时记忆（LSTM）网络的载荷时间历程识别研究。通过Runge-Kutta法求解了系统在随机载荷作用下的响应，并构建数据集。由于数据具有时序依赖性，而LSTM网络能够有效捕捉序列数据中的动态变化信息，因此初步使用典型LSTM网络进行识别，发现其载荷幅值预测精度不足（MRE达276.845%）。进一步考虑到载荷作用具有滞后效应，需要结合一定时间窗口内的响应数据来充分反映载荷在时域上的影响范围，因此提出了滑窗LSTM网络。基于多重尺度法分析了窗口长度的物理意义（即位移的渐进衰减过程），并通过网格搜索确定了最优窗口长度。结果表明，滑窗LSTM网络在最优窗口长度下的识别精度显著提升（MRE降低至5.494%），远优于直接求逆法（MRE为807.314%），证明了该方法在处理非线性载荷识别问题中的优势。
最后，研究进一步拓展到更为复杂的实际工程应用场景，以双三角翼结构为研究对象，开展了机翼突风载荷时间历程的识别研究。基于有限元数值模拟，获得了双三角翼在突风作用期间的强迫振动响应及突风结束后的自由振动响应，构建了同时涵盖两种振动模式的数据形式。针对强迫振动与自由振动阶段响应特征差异显著的复杂数据，分别构建了滑窗LSTM网络、CNN-LSTM网络以及引入物理信息约束的PI-LSTM网络。结果表明，滑窗LSTM网络在两个阶段的识别精度均存在明显瓶颈；CNN-LSTM网络借助卷积模块的特征提取机制，显著提升了识别性能；PI-LSTM网络则通过裁剪自由振动阶段数据，并引入掩码机制和自定义损失函数，使网络专注学习载荷作用阶段的特征，完全消除了自由振动阶段的预测误差。上述结果充分体现了特征提取机制与物理信息约束在复杂动态载荷识别任务中的关键作用。</p
Study on Aerodynamic Configuration Design of Spacecrafts in Ultra Low Earth Orbit
No full text极低轨飞行器凭借其高分辨率对地观测能力和低延迟通信优势，在遥感、应急响应和全球互联等领域展现出巨大应用潜力。然而，与常规轨道不同，极低轨道（ULEO）的大气密度随轨道高度降低呈指数增长，导致飞行器气动阻力显著增加，进而引发快速轨道衰减。传统卫星依靠电推进或化学推进维持轨道高度，但在极低轨道环境下，气动阻力的增加使推进工质消耗急剧上升，工质携带量成为制约飞行器在轨寿命的关键因素。因此，本论文聚焦于极低轨飞行器的气动构型设计，旨在通过优化气动外形，缓解工质消耗激增问题，延长飞行器的任务时间。论文主要研究内容和结论如下。
首先，基于直接模拟蒙特卡洛（DSMC）方法，分析GOCE、SLATS和力星一号在极低轨道环境下的气动阻力特性，重点关注轨道高度、壁面适应系数和姿态角对气动阻力的影响。研究发现：稀薄气体环境中飞行器摩阻分量不可忽视，特别是SLATS卫星，其大尺寸太阳能帆板结构使摩阻超过压阻。随着轨道高度降低，大气密度指数增长导致气动阻力急剧增加，但阻力系数降低。壁面适应系数对气动特性影响显著，适应系数增大时，摩阻分量上升而压阻分量下降，总阻力的变化趋势与外形有关。姿态角增大时，阻力分量及阻力系数均单调递增，其中俯仰角的影响大于侧滑角。
其次，针对吸气式电推进（ABEP）飞行器多参数强耦合的特点提出一种构型设计方法，以推阻平衡为前提，以工质平衡和能量平衡为约束，定义能量工质无量纲化积RP×Rm，以最大化RP×Rm为优化目标，分析圆柱构型下影响飞行器飞行的主要因素。研究发现：在圆柱最优构型下，长径比(l/d)opt仅与捕集率和面元阻力系数相关，且摩阻与压阻相等。基于现有组件性能参数，圆柱构型的轨道高度可行下限为186 km（对应大气密度4.01×10-10 kg/m3），而任意截面形状的柱体飞行器在截面扁平度（Rs/R//）较大时，RP×Rm增大，飞行可达轨道下限降低，极限高度为172 km（对应大气密度6.58×10-10 kg/m3）。
最后，以圆柱加翼为基准构型进行尺寸优化，确定最优几何参数，并通过尾部添加后锥角的减阻结构设计进一步提升性能。随后开展两种不同构型的极低轨飞行器模型地面风洞测力实验，结果表明，优化后模型阻力较原始模型减少约1.8 mN，减阻率达11.0%，验证了尾部锥角减阻设计的有效性。此外，通过采用DSMC方法针对不同壁面适应系数下的阻力特性进行了数值模拟，进一步分析壁面适应系数对尾部锥角减阻效果的影响。</p
Towards fatigue-resistant steels: Interfacial fatigue crack mechanisms in complex inclusions revealed by TEM and atomic simulations
No full textDespite the significance of inclusions on fatigue properties, a thorough and physics-based understanding of the failure behavior of complex inclusions is still missing due to experimental limitations. An atomic-scale simulation approach is proposed for investigating the interfacial cracking behavior of Ca2Al2SiO7/CaS complex inclusion in steel under tensile-compressive cyclic loading conditions. The results show that regular parallel lattice distortions are induced in CaS to generate stable interfacial structures. Interestingly, these distortion rows will gradually disappear under compression and eventually deteriorate the distortions adjacent to the interfacial region, which become the crack nucleation sites under tension. This study provides an in-depth understanding of the fatigue mechanism of modern steels with complex inclusions and potentially offers a physics-based bottom-up method to determine the fatigue life of materials