韩国科学技术先进研究所

Accurate control of shield tunneling m requires a clear and profound understanding of their dynamic response characteristics. However, existing models predominantly rely on quasi-static frameworks, which fail to reveal the underlying dynamic behavior. This study develops an 18th-order state-space model of the shield-soil system to enable comprehensive dynamic response analysis of key state variables under diverse loading conditions and to assess system controllability. The proposed dynamic model is validated through analyses with both traditional quasi-static theories and field monitoring data. Results show that articulation center displacements and front/rear shield pitching angles are highly sensitive to vertical loads and segment buoyancy, exhibiting delayed responses and overshoot, which induce significant axis deviation and posture instability, especially in ultra-soft soils. Loads were categorized as controllable (jacking thrust) or uncontrollable (self-weight, segment buoyancy, soil–water pressures), with the front shield primarily governed by jacking. In contrast, the tail shield remains strongly affected by uncontrollable loads. Crucially, the proposed model reveals the inherent underactuation of the system, specifically highlighting the absolute uncontrollability of the rolling angles and roll rates, alongside the non-independent controllability of the pitch and yaw angles. To enhance system controllability, operational strategies are proposed, including optimizing shield configuration, employing fast-setting grout, and matching advance speed to regulate segment buoyancy, while structurally expanding the multi-dimensional controllability of the thrust system. These findings provide new theoretical insights into shield-soil dynamic interactions and practical guidance for real-time posture regulation and operational control in complex tunneling environments.

Natural load-bearing patterns such as leaf venation, trabecular bone, and spider webs achieve high stiffness per unit mass, yet classical topology optimizers rarely reach such geometries, and few let engineers express structural design intent through natural language. This work treats a frozen text-to-image diffusion model as a training-free source of design knowledge and distills it into the physics loop of density-based topology optimization via score distillation sampling, so that a text prompt becomes an explicit, machine-interpretable representation of engineer intent. The prompt-induced generative gradient and the finite element sensitivity are combined at every iteration, letting physics decide which prompt-induced features survive. In 245 primary SDS runs spanning four geometric domains and two physics regimes, 38 of 49 prompt--domain combinations achieved statistically significant compliance reductions (up to $-31.5\%$ mechanical and $-23.0\%$ thermoelastic), outperforming gradient-based baselines. Cross-domain morphological analysis identifies a recurring structural signature of improvement: in most domains the generative prior suppresses dead-end branches in the rib skeleton, with endpoint--compliance correlation $r = +0.56$ to $+0.99$. A Heaviside projection with $β$-continuation resolves a pronounced intermediate-density tendency in this diffusion--physics coupling ($42.6\%$ to $<3\%$), and an automated skeleton-based pipeline converts optimized density fields into \rev{candidate geometry ready for computer-aided design. By retargeting the generative prior across domains, loading conditions, and physics objectives through a change of text prompt, with each new problem's physics setup specified separately, the framework uses a pretrained generative model as a reusable, training-free prior for engineering design.

As an extreme vehicle dynamic behavior, the drift maneuver holds significant research value for enhancing the handling performance and active safety of vehicles under extreme operating conditions. However, maintaining the drift state of autonomous vehicles while ensuring accurate path following remains a significant challenge. Thus, this paper proposes an integrated control framework that incorporates a vehicle drift strategy into general path following, enabling drift tracking control for all-wheel-drive (AWD) vehicles. First, based on a nonlinear three-degree-of-freedom (3-DOF) vehicle dynamics model, the steady-state drift characteristics of the vehicle are analyzed, and different drift equilibrium points are determined. Subsequently, a hierarchical drift tracking controller is proposed based on Model Predictive Control (MPC): the upper-layer controller achieves path following functionality based on vehicle kinematics and drift equilibrium points, while the lower-layer controller derives the control inputs for the vehicle using the full-vehicle dynamics model and desired state derivatives. Lastly, a controller switching strategy is proposed to manage the transition into and out of the drift state, and its effectiveness and real-time performance are validated via simulations and hardware-in-the-loop (HiL) tests.

Abstract As a micro- and nanofabrication technique, maskless photolithography (MPL) eliminates static physical masks and instead utilizes computer-controlled light sources and optical systems to directly generate patterns. This significantly enhances process flexibility and design freedom and reduces production costs. It is an important technology that supports the field of advanced micro- and nanofabrication. This review systematically elaborates on the principles and equipment systems of MPL technology, introduces the development history and photoreaction mechanisms of different types of photoresponsive materials, and summarizes the application of MPL technology in micro- and nanofabrication. Finally, the prospects and future directions for the development of MPL technology are presented.

The steel industry, characterized by its broad applicability, relatively low production cost, and strategic importance, is one of the pillar industries underpinning modern society. Over the past few decades, steel enterprises have undergone substantial transformation in automation and digitalization, significantly improving production efficiency and operational performance. With the rise of Industry 4.0 and the growing demand for sustainable and coordinated development, there is an increasing need to further integrate information technologies with industrial processes to accelerate the realization of automated, digital, and intelligent manufacturing. Against this backdrop, this paper provides a comprehensive review of artificial intelligence applications in the intelligent development of steelmaking processes. It categorizes the existing studies into several major application areas, including quality prediction, operational optimization, process monitoring, and anomaly diagnosis. The paper also discusses current challenges, potential solutions, and future research directions. This review aims to clarify the evolving relationship between steelmaking processes and artificial intelligence technologies and to support the transformation of the steel industry toward greener, lower-carbon, and more intelligent production. Furthermore, a forward-looking perspective is presented for the construction of a novel steel manufacturing paradigm in the Industry 5.0 era, characterized by ecological resilience, digital twin capabilities, and closed-loop resource recycling.