Atmospheric flow phenomena at length scales below 1 km strongly influence the efficiency of wind farms, heat and pollution management in cities, the health of local ecosystems, and flight dynamics of light aircraft and UAVs (uncrewed aerial vehicles). Dynamics at these scales are typically too large for traditional measurement techniques to capture and too small for regional climate models to predict. We thus seek to develop methods to observe atmospheric flows at scales of tens to hundreds of meters, using drones, drifters, and various types of particles to provide sparse Lagrangian data on the dynamics they encounter. This work aims to combine ideas from fluid mechanics, robotics, data assimilation, and data-driven modeling to bring scalable and inexpensive flow-mapping techniques to climate-critical engineering challenges. We plan to apply these efforts to problems in wind-farm siting, wildfire prevention and mitigation, regional ecology, and climate impacts on human settlements.

Status: Currently in planning stages, with some initial ideas being pursued at Princeton University in collaboration with Nick Conlin, Hannah Even, Nate Simon, Skywalker Li, Prof. Ani Majumdar, Prof. Eli Bou-Zeid, and Prof. Marcus Hultmark

Floating offshore wind turbines and kite-mounted airborne turbines can undergo large oscillations in the streamwise direction that may affect their efficiency and longevity in unsteady atmospheric flows. Similarly, traditional ground-fixed and tidal turbines often experience oscillations in the wind speed that may have similar effects to the oscillations of their mobile counterparts. To study these problems experimentally, I built a wind turbine on a traverse (see left) that can move back and forth in the streamwise direction at up to 2 m/s. Wind-tunnel experiments with this system have yielded analytical models for the torque and power extraction of a turbine as a function of its motion kinematics (Wei and Dabiri 2022), as well as for the flow properties upstream of the surging turbine (Wei and Dabiri 2023). We also found that surge motions can increase the time-averaged power extraction by as much as 6% over the stationary-turbine case. We conducted further measurements of the wake dynamics of surging turbines in a towing tank in the Rival Lab at Queen's University, finding that unsteady vortex dynamics in the wake can accelerate wake recovery by over 40% in some cases. We plan to investigate the synergies and tradeoffs between surge motions, power enhancements, beneficial wake dynamics, and fatigue loads in futher experiments, with a particular focus on floating offshore wind farms. These studies will guide the design and control of the next generation of wind-energy technologies for increased efficiency and longer life spans.

Status: Active February 2020 - present

Collaborators: Prof. John O. Dabiri and Noa Yoder (Graduate Aerospace Laboratories, Caltech); Prof. David E. Rival (TU Braunschweig); Frieder Kaiser, Adnan El Makdah, and JiaCheng Hu (Queen's University)

Vertical-axis wind turbines (VAWTs) are a promising alternative to traditional horizontal-axis turbines, due to their omnidirectional nature and ability to be placed in closely packed arrays for enhanced efficiency and power density. We have shown in wind-tunnel experiments that pairs of VAWTs show mutually enhanced performance due to beneficial aerodynamic interactions and near-wake vortical structures (Brownstein et al. 2019). We then studied the flow around full-scale straight- and helical-bladed VAWTs at the Field Laboratory for Optimized Wind Energy (FLOWE) using several GoPro cameras to track the motions of artificial snow particles in three dimensions (Wei et al. 2021). With these time-averaged 3D particle-tracking velocimetry (PTV) data, we were able to quantify vortical structures in the wakes of these turbines and infer the effects of turbine geometry on aerodynamic interactions in VAWT arrays.

Status: Active September 2018 - May 2020; ongoing collaboration with Prof. Di Yang (University of Houston)

PI: Prof. John O. Dabiri (Mechanical Engineering, Stanford University)

Collaborators: Ian D. Brownstein (XFlow Energy); Jennifer L. Cardona and Michael F. Howland (Stanford); Prof. Di Yang, Masoumeh Gharaati, and Shuolin Xiao (University of Houston)

Aircraft, flying animals, wind turbines, and other objects exposed to the atmosphere will encounter gust disturbances that affect their aerodynamic performance. These disturbances can be modeled with analytical transfer functions derived from potential-flow theory. The applicability of these simple models to real-world flow scenarios, however, is not obvious. We used an active grid in a wind tunnel at the University of Oldenburg to study the unsteady lift force on a NACA 0006 airfoil exposed to sinusoidal transverse gusts (see left). These experiments validated two models of the gust-airfoil interaction problem, the so-called Sears and Atassi functions, and showed that their accuracy held even in certain turbulent-flow situations that violated their modeling assumptions (Wei et al. 2019). In continued work, we are investigating the superposition of these models with similar models for airfoil pitch and plunge. We carried out two-dimensional large-eddy simulations (LES) using a modified version of the open-source Nalu code, which showed that the combined model for a pitching and plunging airfoil in a transverse gust does perform well when the gust and airfoil-oscillation frequencies are sufficiently separated. These analytical, experimental, and numerical results can help us build aircraft and structures that are more resilient to atmospheric gusts.

Status: Active September 2017 - July 2018; ongoing LES project with Omkar B. Shende

Collaborators: Johannes Kissing, Sebastian Wegt, Prof. Suad Jakirlić, and Prof. Cameron Tropea (TU Darmstadt); Tom Wester and Prof. Joachim Peinke (University of Oldenburg); Omkar B. Shende (Stanford University)

The generation of specific gust or turbulence characteristics in a wind tunnel is essential for experiments that seek to replicate and analyze unsteady flows in the atmosphere. This, however, is a difficult practical problem that requires creative solutions.

Active grids can afford the experimentalist many degrees of freedom for generating unsteady flows. Dr. Gregory Bewley designed an active grid with 129 independently controlled paddles to exert spatial control over turbulent flow correlations. Kevin Griffin and I created control algorithms that correlated the grid's motions across space and time, allowing us to modify the shape of the velocity correlation functions of the turbulent flows downstream of the grid. These results allowed the energy decay of turbulence with different correlations to be studied (Griffin et al. 2019).

To produce periodic gust disturbances, I reverse-engineered the Theodorsen function for the lift and moment on a pitching and plunging airfoil to develop a theory for gust generation with minimal interference from the airfoil wake. I tested this theory in wind-tunnel experiments with an oscillating airfoil and 2D particle-image velocimetry (PIV), showing that periodic gusts of controlled waveform shape could be produced based on my theoretical framework (Wei et al. 2019). This type of gust generator could be useful for facilities where a full-size active grid would be impractical to install.

Note: The image to the left is taken from smoke-wire flow visualizations in the TU Darmstadt wind tunnel of a purely plunging airfoil at high angle of attack.

Status: Active June 2015 - August 2015 (active-grid project); September 2017 - July 2018 (gust-generator project)

PIs: Dr. Gregory Bewley (Max Planck Institute for Dynamics and Self-Organization); Prof. Cameron Tropea (Institute for Fluid Mechanics and Aerodynamics, TU Darmstadt)

Collaborators: Kevin P. Griffin (Princeton University); Johannes Kissing (TU Darmstadt)

The mechanics of human vocalization and speech are very difficult to study in vivo. Therefore, in this study, we used a scaled-up and simplified model of human vocal folds in a water channel (see left), which allowed simultaneous 2D particle-image velocimetry (PIV) and pressure measurements to be taken of the glottal jet produced as the folds oscillate. A control-volume analysis demonstrated that unsteady effects are important for phonation, particularly at the beginning and end of each oscillation cycle. We also found that cycle-to-cycle variations based on recirculations downstream of the glottis when the folds are closed can lead to variations in the jet dynamics that may affect individual voice quality and color (Ringenberg et al. 2021). Lastly, we studied cases where the folds failed to close completely or moved asymmetrically, representing certain categories of vocal-fold pathologies, and showed how these altered kinematics affect sound production and voice strength (Wei et al. 2022).

Status: Active June 2017 - September 2017; ongoing work to finish a third manuscript

PI: Prof. Timothy Wei (University of Nebraska - Lincoln)

Collaborators: Hunter Ringenberg, Dylan Rogers, and Abigail Haworth (University of Nebraska - Lincoln); Mike Krane (Penn State)

For my undergraduate thesis, I constructed a streamwise-motion system for a bio-inspired propulsor in a water channel. I introduced a cyber-physical control scheme that moved the heaving and pitching propulsor in the streamwise direction according to the thrust it produced. We then studied the effect of the free-swimming condition, in which the propulsor oscillated in the streamwise direction according to its time-varying thrust output, on propulsive efficiency. We found no significant effect of the free-swimming condition on swimming performance, even for large-amplitude streamwise motions and intermittent ("burst-coast") swimming profiles (Van Buren et al. 2018). These results meant that studies of fish swimming done without a streamwise-motion system to simulate the free-swimming condition were equally valid, and could thus forego the additional complexity of such a system.

Status: Active June 2016 - June 2017

PI: Prof. Lex Smits (Mechanical and Aerospace Engineering, Princeton University)

Collaborators: Daniel Floryan and Dr. Tyler Van Buren (Princeton)