Overview
The goal of this research was to pioneer a cyber-physical systems (CPS) approach to the optimal design of structures subject to wind hazards. The approach brings together numerical optimization algorithms to drive the design and boundary layer wind tunnel modeling to evaluate the fitness of each design candidate. Wind tunnels provide accurate turbulent flow conditions and avoid the uncertainties and computational burden associated with computational fluid dynamics, creating an efficient and accurate framework to explore candidate designs. The research also led to the first mechatronic models for wind tunnel testing, i.e., models that can automatically adjust their dynamic and aerodynamic properties, seeking the optimal design. Wind tunnel experiments and cyber-infrastructure developments were conducted at the University of Florida's NSF-sponsored Natural Hazard Engineering Research Infrastructure (NHERI) Experimental Facility (EF). 

The research began with a proof-of-concept for CPS optimization built around a low-rise rigid (pressure-tapped) building model with a parapet wall. The height of the parapet wall was physically adjusted using stepper-motors. A coordinating computer executed an optimization algorithm to evaluate candidate designs (i.e., parapet heights) and guided the physical model to the optimal solution. The experiments centered on achieving the optimum parapet height that reduced both components and cladding (C&C) and main wind force resisting system (MWFRS) loads. 

The CPS approach was then applied to an aeroelastic model of a tall building with adjustable stiffness and aerodynamic features. The traditional design process for tall buildings requires a lengthy collaboration between designers and wind tunnel operators to achieve a cost-effective solution, which may include the construction of multiple building models. In this research, candidate tall building designs were generated by physically adjusting the dynamic and aerodynamic properties of a multi-degree-of-freedom aeroelastic model. The specimen was equipped with an actuation system consisting of (1) a series of variable stiffness devices that enabled precise control of the lower natural frequencies and model shapes and (2) a series of individually controllable slotted fins that enabled precise modifications of the aerodynamic shape. The specimen was instrumented with accelerometers and displacement transducers to capture wind-induced building responses. Results show that the CPS framework can reliably drive the aeroelastic specimen to an optimal solution while meeting user-specified performance requirements (e.g., acceleration and drift constraints under a given return period wind event). 

Key Outcomes
The key outcome of this research was a framework for CPS design that supplements or replaces laborious trial-and-error design approaches. Current design approaches often require extensive iterations and communication burden between the architects and structural engineers and do not exhaustively explore a wide range of design alternatives. The use of scaled building models with physically adjustable properties (e.g., geometry, stiffness, damping, etc.) allows optimum designs to be attained faster than conventional methods and eliminates the need to design, reconstruct new models, and perform additional wind tunnel tests. The research demonstrated the framework using both gradient-based and heuristic optimization algorithms to drive low-rise and tall building models to their optimal configurations in the wind tunnel. The impact of choices made in the design process (e.g., objectives and constraints) on the results were explored, providing feedback needed to prioritize competing objectives from multiple stakeholders.

Additionally, in the last two decades, design trends for civil infrastructure have shifted from prescriptive code procedures to performance-based methods. The CPS approach to optimal design meshes well with performance-based design methods. Design that is driven by optimization algorithms can use performance metrics as constraints while seeking lighter, taller, and more cost-effective designs in the face of strong winds. By including feedback from stakeholders throughout this research, the approach developed ensures that the parameters, constraints, and performance objectives important to wind and structural engineering are considered and appropriately balanced. 

Intellectual Merits
This research advanced the fields of wind and structural engineering by: (1) combining the strengths of experimental testing and numerically-driven optimization, (2) advancing the development and application of heuristic optimization algorithms in a practical engineering setting, including features to overcome the challenges of noisy data from experiments, (3) discovering new design and detailing features to achieve cost-effective and efficient civil infrastructure under wind hazards through broad search algorithms and high-fidelity testing, and (4) creating a system for satisfying demanding performance requirements extending well beyond prescriptive code provisions (e.g., performance-based design).

Broader Impacts
The broader impacts met by the project included: (1) more sustainable use of resources to build structures resilient to wind loading under a range of hazard levels through design optimization, (2) enhanced capabilities of laboratories worldwide through the development of new cyber-infrastructure and new applications of cyber-physical systems, (3) increased scientific literacy and public engagement through outreach activities, (4) inclusion of multiple stakeholders and experts in wind engineering when developing the CPS framework, and (5) the participation underrepresented groups in STEM fields, including within the project team. 

Last Modified: 11/27/2019
Modified by: Brian M Phillips