Search for: All records

Nonlinear aeroelastic limit-cycle oscillations (LCOs) have become an area of interest due to both detrimental effects on flying vehicles and use in renewable energy harvesting. Initial studies on the interaction between aeroelastic systems and incoming flow disturbances have shown that disturbances can have significant effects on LCO amplitude, with some cases resulting in spontaneous annihilation of the LCO. This paper explores this interaction through wind-tunnel experiments using a variable-frequency disturbance generator to produce flow disturbances at frequencies near the inherent LCO frequency of an aeroelastic system with pitching and heaving degrees of freedom. The results show that incoming disturbances produced at frequencies approaching the LCO frequency from below produce a cyclic growth-decay in LCO amplitude that resembles interference between multiple sine waves with slightly varying frequencies. An aeroelastic inverse technique is applied to the results to study the transfer of energy between the pitching and heaving degrees of freedom as well as the aerodynamic power moving into and out of the system. Finally, the growth-decay cycles are shown to both excite LCOs in an initially stationary wing and annihilate preexisting LCOs in the same wing by appropriately timing the initiation and termination of disturbance generator motion. 

Due to COVID-19, engineering summer camps offered by North Carolina State University (NCSU) shifted to a virtual format for the summer of 2021 and required a new curriculum to be designed with an emphasis on providing a hands-on experience in a virtual environment. The Department of Mechanical and Aerospace Engineering created a curriculum which included some hands-on activities used in previous, in-person camps, a homebuilt wind tunnel used to demonstrate aerospace fundamentals, and a popular engineering game used as a teaching tool to explain astronautics concepts. Each week-long camp was conducted via Zoom and led by a team consisting of a NCSU graduate student, three undergraduate students, and a faculty advisor. Anonymous student feedback following the completion of the camps showed overwhelmingly positive results with a majority of students showing interest in pursuing an engineering degree with multiple students expressing interest in attending NCSU 

Predicted rapid increases in urbanization in the face of accelerating biodiversity loss underscores the need for urban development that promotes, rather than displaces, native plants and animals. One approach for increasing urban biodiversity is through the development of “green infrastructure”. Although research has explored urban-rural gradients and the overall value of urban green infrastructure, few studies have investigated the habitat value for wildlife of different types of urban greenspace. Here, we use a well-established metric in ecology, giving up-densities (GUDs), to compare foraging costs for a common urban wildlife species, the eastern gray squirrel (Sciurus carolinensis), among three green infrastructure categories: municipal parks, college campuses, and residential yards. We found that GUDs for gray squirrels did not differ significantly among location categories after controlling for proximity to roads, but proximity to roads was associated with significantly higher GUDs in all locations. In an explicit test, we also found that both proximity to roads and traffic volume were associated with higher GUDs. We also found that maximum distance from roads was significantly higher for campuses and parks than for residential yards, indicating a greater proportion of the area of campuses and parks is “away from roads” compared to residential yards. Our results indicate that vehicle traffic may contribute significantly to an “urban landscape of fear” for gray squirrels and suggest that campus and park configurations that reduce road effects could improve habitat quality for squirrels and possibly other animals. 

The interaction between upstream flow disturbance generators and downstream aeroelastic structures has been the focus of several recent studies at North Carolina State University. Building on this work, which observed the modulation of limit cycle oscillations (LCOs) in the presence of vortex wakes, this study examines the design and validation of a novel disturbance generator consisting of an oscillating cylinder with an attached splitter plate. Analytical design of the bluff body was performed based on specific flow conditions which produced LCO annihilation in previous studies. Computational fluid dynamics simulations and experimental wind tunnel tests were used to validate the ability of the new disturbance generator to produce the desired wake region. Future work will see the implementation of this novel design in conjunction with aeroelastic structures in an effort to modulate and control LCOs, including the excitation and annihilation thereof. 

Recent works apply Graph Neural Networks (GNNs) to graph matching tasks and show promising results. Considering that model outputs are complex matchings, we devise several techniques to improve the learning of GNNs and obtain a new model, Stochastic Iterative Graph MAtching (SIGMA). Our model predicts a distribution of matchings, instead of a single matching, for a graph pair so the model can explore several probable matchings. We further introduce a novel multi-step matching procedure, which learns how to refine a graph pair’s matching results incrementally. The model also includes dummy nodes so that the model does not have to find matchings for nodes without correspondence. We fit this model to data via scalable stochastic optimization. We conduct extensive experiments across synthetic graph datasets as well as biochemistry and computer vision applications. Across all tasks, our results show that SIGMA can produce significantly improved graph matching results compared to state-of-the-art models. Ablation studies verify that each of our components (stochastic training, iterative matching, and dummy nodes) offers noticeable improvement. 

Periodic upstream flow disturbances from a bluff body have recently been shown to be able to modulate and annihilate limit cycle oscillations (LCOs) in a downstream aeroelastic wing section under certain conditions. To further investigate these phenomena, we have implemented a controllable wind tunnel disturbance generator to enable quantification of the parameter ranges under which these nonlinear interactions can occur. This disturbance generator, consisting of a pitch-actuated cylinder with an attached splitter plate, can be oscillated to produce a von Karman type wake with vortex shedding frequency equal to the oscillation frequency over a range of frequencies around the natural shedding frequency of the cylinder alone. At vortex shedding frequencies away from the LCO frequency of the wing, forced oscillations were observed in the wing, but the wing did not enter self-sustaining LCOs. However, when disturbances were introduced near the LCO frequency, the initially static downstream wing entered self-sustaining oscillations in the presence of the incoming vortices, and these LCOs persisted when the disturbance generator was stopped. Annihilation of the wing LCOs was also observed disturbance vortices were introduced upstream of the wing in LCO. 

Abstract Motivation As experimental efforts are costly and time consuming, computational characterization of enzyme capabilities is an attractive alternative. We present and evaluate several machine-learning models to predict which of 983 distinct enzymes, as defined via the Enzyme Commission (EC) numbers, are likely to interact with a given query molecule. Our data consists of enzyme-substrate interactions from the BRENDA database. Some interactions are attributed to natural selection and involve the enzyme’s natural substrates. The majority of the interactions however involve non-natural substrates, thus reflecting promiscuous enzymatic activities. Results We frame this ‘enzyme promiscuity prediction’ problem as a multi-label classification task. We maximally utilize inhibitor and unlabeled data to train prediction models that can take advantage of known hierarchical relationships between enzyme classes. We report that a hierarchical multi-label neural network, EPP-HMCNF, is the best model for solving this problem, outperforming k-nearest neighbors similarity-based and other machine-learning models. We show that inhibitor information during training consistently improves predictive power, particularly for EPP-HMCNF. We also show that all promiscuity prediction models perform worse under a realistic data split when compared to a random data split, and when evaluating performance on non-natural substrates compared to natural substrates. Availability and implementation We provide Python code and data for EPP-HMCNF and other models in a repository termed EPP (Enzyme Promiscuity Prediction) at https://github.com/hassounlab/EPP. Supplementary information Supplementary data are available at Bioinformatics online.