Cyber-Physical Systems Approach to the Optimal Design of Structures for Wind Hazards

本研究的目标是开创一种网络物理系统（CPS）方法，用于对受风危害的结构进行优化设计。CPS方法将在美国国家科学基金会支持的佛罗里达大学自然灾害工程研究基础设施设施中进行风洞测试，并结合计算机增强设计，以比纯实验或纯计算方法更快、更有信心地产生最佳结构设计。实验风洞测试在稳定风荷载、阵风和复杂风-结构相互作用下的建筑和桥梁设计的开发和评估中提供了无与伦比的准确性。同时，计算优化方法能够快速创建和评估竞争设计，以最好地满足指定的目标。CPS科学的进步可以将物理风洞测试无缝集成到计算机驱动的设计和优化中。CPS方法可以补充或取代费力的试错设计方法，这通常需要大量的迭代和建筑师和结构工程师之间的沟通负担，并且不能详尽地探索广泛的设计替代方案。该项目将提高在面对风灾害时建造更坚固、更轻、更有弹性的结构的能力。同时，通过在设计方法中直接权衡成本效益，选择的设计将更可持续地利用资源，最终有更好的建造机会。将成立一个利益相关者小组，以确保在该方法中考虑并适当平衡来自各种学术、工业和政府组织的与风力工程相关的参数、约束和性能目标。此外，项目外展活动将提高科学素养和公众对风害和工程解决方案的认识，同时包括科学、工程、技术和数学（STEM）领域代表性不足的群体直接参与研究。通过将实验风洞测试的可靠性与基于计算的优化技术的效率相结合，本研究将推进风工程的理论、研究和实践。CPS方法将由高性能计算机指导，实现优化算法，而每个候选解决方案将通过网络边界层风洞的实验测试进行快速评估。这种方法将优化几何（例如，形状和孔隙度）和结构（例如，刚度和阻尼）的比例结构模型的性能。在每次风洞测试之前，这些特性将被迅速调整。一台联网的超级计算机将监控来自传感器的反馈信息，应用优化技术（通过有限元分析增强），并为下一次物理测试确定新的结构配置。目标将是用户定义的（例如，在约束条件（例如，满足漂移、加速和占用的要求）内最小化重量或基础剪切）。这项研究将通过以下方式推动风力和结构工程领域的发展：(1)结合高保真实验测试和数值驱动优化的优势；(2)在实际工程环境中推进元启发式优化算法的开发和应用；(3)发现新的设计和细节特征，以实现风灾害下具有成本效益的民用基础设施；(4)创建满足性能要求的系统，例如基于性能的设计。

The goal of this research is to pioneer a cyber-physical systems (CPS) methodology for the optimal design of structures subjected to wind hazards. The CPS approach will combine wind tunnel testing at the NSF-supported Natural Hazards Engineering Research Infrastructure facility at the University of Florida with computer-augmented design to produce optimal structural designs faster and with greater confidence than purely experimental or purely computational methods. Experimental wind tunnel testing provides unparalleled accuracy in the development and evaluation of building and bridge designs under steady wind loads, gusts, and complex wind-structure interaction. At the same time, computational optimization methods enable the rapid creation and evaluation of competing designs to best meet specified objectives. Advances in the science of CPS can lead to seamless integration of physical wind tunnel testing into computer-driven design and optimization. The CPS approach can supplement or replace laborious trial-and-error design approaches, which often require extensive iterations and communication burden between the architects and structural engineers and do not exhaustively explore a wide range of design alternatives. This project will advance the capability to build stronger, lighter, and more resilient structures in the face of wind hazards. At the same time, by weighing cost-effectiveness directly in the design approach, selected designs will make more sustainable use of resources and ultimately have a better chance of being constructed. A stakeholder group will be formed to ensure that the parameters, constraints, and performance objectives relevant to wind engineering from various academic, industrial, and governmental organizations are considered and appropriately balanced in the approach. Additionally, project outreach activities will increase the scientific literacy and public awareness of wind hazards and engineering solutions while including the participation of underrepresented groups in science, engineering, technology, and mathematics (STEM) fields directly in the research.This research will advance theory, research, and practice in wind engineering by combining the reliability of experimental wind tunnel testing with efficiency of computational-based optimization techniques. The CPS methodology will be directed by a high performance computer, implementing optimization algorithms, while each candidate solution will be rapidly evaluated through experimental testing in a networked boundary layer wind tunnel. This methodology will optimize geometric (e.g., shape and porosity) and structural (e.g., stiffness and damping) properties of scaled structural models. The properties will be rapidly adjusted prior to each scaled duration wind tunnel test. A networked supercomputer will monitor feedback information from sensors, apply optimization techniques (augmented by finite element analysis), and determine a new structural configuration for the next physical test. Objectives will be user-defined (e.g., minimize weight or base shear) within constraints (e.g., meeting requirements for drift, acceleration, and occupancy). This research will advance the fields of wind and structural engineering by: (1) combining the strengths of high-fidelity experimental testing and numerically-driven optimization, (2) advancing the development and application of meta-heuristic optimization algorithms in a practical engineering setting, (3) discovering new design and detailing features to achieve cost-effective civil infrastructure under wind hazards, and (4) creating a system for satisfying performance requirements, e.g., for performance-based design.