arXiv subjects

Bernardo Luiz R. Ribeiro

Publications and source records attributed to Bernardo Luiz R. Ribeiro.

5 recordsLinked to original sources

arXiv2022

A Machine Learning Approach to Classify Vortex Wakes of Energy Harvesting Oscillating Foils

A machine learning model is developed to establish wake patterns behind oscillating foils whose kinematics are within the energy harvesting regime. The role of wake structure is particularly important for array deployments of oscillating foils, since the unsteady wake highly influences performance of downstream foils. This work explores 46 oscillating foil kinematics, with the goal of parameterizing the wake based on the input kinematic variables and grouping vortex wakes through image analysis of vorticity fields. A combination of a convolutional neural network (CNN) with long short-term memory (LSTM) units is developed to classify the wakes into three groups. To fully verify the physical wake differences among foil kinematics, a convolutional autoencoder combined with k-means++ clustering is utilized and four different wake patterns are found. With the classification model, these patterns are associated with a range of foil kinematics. Future work can use these correlations to predict the performance of foils placed in the wake and build optimal foil arrangements for tidal energy harvesting.

physics.flu-dyn↗

arXiv2022

Leading edge vortex formation and wake trajectory: Synthesizing measurements, analysis, and machine learning

The strength and trajectory of a leading edge vortex (LEV) formed by a pitching-heaving hydrofoil (chord $c$) is studied. The LEV is identified using the $Q$-criterion method, which is calculated from the 2D velocity field obtained from PIV measurements. The relative angle of attack at mid-stroke, ${α_{T/4}} $, proves to be an effective method of combining heave amplitude ($h_0/c$), pitch amplitude ($θ_0$), and reduced frequency ($f^*$) into a single variable that predicts the maximum value of $Q$ over a wide range of operating conditions. Once the LEV separates from the foil, it travels downstream and rapidly weakens and diffuses. The downstream trajectory of the LEV has two characteristic shapes. At low values of ${α_{T/4}}$, it travels straight downstream after separating from the foil, while at higher values of ${α_{T/4}} $, an accompanying Trailing Edge Vortex (TEV) forms and the induced velocity generates a cross-stream component to the vortex trajectories. This behavior is accurately predicted using a potential flow model for the LEV and TEV. Supervised machine learning algorithms, namely Support Vector Regression and Gaussian Process Regression, are used to create regression models that predicts the vortex strength, shape and trajectory during growth and after separation. The regression model successfully captures the features of two vortex regimes observed at different values of ${α_{T/4}} $. However, the predicted LEV trajectories are somewhat smoother than observed in the experiments. The strengths of the vortex is often under-predicted. Both of these shortcomings may be attributed to the relatively small size of the training data set.

physics.flu-dyn↗

arXiv2021

A Machine Learning Approach to Classify Kinematics and Vortex Wake Modes of Oscillating Foils

Machine learning techniques have received attention in fluid dynamics in terms of predicting, clustering and classifying complex flow physics. One application has been the classification or clustering of various wake structures that emanate from bluff bodies such as cylinders or flapping foils, creating a rich diversity of vortex formations specific to flow conditions, geometry, and/or kinematics of the body. When utilizing oscillating foils to harvest energy from tidal or river flows, it is critical to understand the intricate and nonlinear relationship between flapping kinematics and the downstream vortex wake structure for optimal siting and operation of arrays. This paper develops a classification model to obtain groups of kinematics that contain similar wake patterns within the energy harvesting regime. Data is obtained through simulations of 27 unique oscillating foil kinematics for a total of 13,650 samples of the wake vorticity field. Within these samples three groups are visually labeled based on the relative angle of attack. A machine learning approach combining a convolutional neural network (CNN) with long short-term memory (LSTM) units is utilized to automatically classify the wakes into the three groups. The average accuracy on five test data subsets is 80% when the three visually labeled groups are used for classification. After analyzing the test subset with lowest accuracy, an update on the group division boundaries is proposed. With this update, the algorithm achieves an average accuracy of 90%, demonstrating that the three groups are able to discern distinct wake structures within a range of energy harvesting kinematics.

physics.flu-dyn↗

arXiv2021

Wake-foil interactions and energy harvesting efficiency in tandem oscillating foils

Oscillating foils in synchronized pitch/heave motions can be used to harvest hydrokinetic energy. By understanding the wake structure and its correlation with the foil kinematics, predictive models for how foils can operate in array configurations can be developed. To establish a relationship between foil kinematics and wake characteristics, a wide range of kinematics is explored in a two-foil tandem configuration with interfoil spacing from four to nine chord lengths separation and multiple interfoil phases. Using data from experiments and simulations, an in-depth wake analysis is performed and the mean velocity and the turbulent kinetic energy are quantified in the wake. With this energy quantification, the trailing foil efficiency is modified to account for the mean flow in addition to the energy transported by the coherent leading edge vortices (LEVs) shed from the leading foil. With the mean wake velocity, a predictive wake model is able to distinguish three regimes through analyzing trailing foil efficiency profiles and the strength of the primary LEV shed from the leading foil. Dividing the wake into regimes is an insightful way to narrow the range of foil kinematics and configurations and improve the energy harvesting in a two-tandem foil array.

physics.flu-dyn↗

arXiv2020

Vortex dynamics and Reynolds number effects of an oscillating hydrofoil in energy harvesting mode

The energy extraction and vortex dynamics from the sinusoidal heaving and pitching motion of an elliptical hydrofoil is explored through large-eddy simulations (LES) at a Reynolds number of $50,000$. The LES is able to capture the time-dependent vortex shedding and dynamic stall properties of the foil as it undergoes high relative angles of attack. Results of the computations are validated against experimental flume data in terms of power extraction and leading edge vortex (LEV) position and trajectory. The kinematics for optimal efficiency are found in the range of heave amplitude $h_o/c=0.5-1$ and pitch amplitude $θ_o=60^{\circ}-65^{\circ}$ for $fc/U_{\infty}=0.1$ and of $h_o/c=1-1.5$ and $θ_o=75^{\circ}-85^{\circ}$ for $fc/U_{\infty}=0.15$. Direct comparison with low Reynolds number simulations and experiments demonstrate strong agreement in energy harvesting performance between Reynolds numbers of $1000$ to $50,000$, with the high Reynolds number flows demonstrating a moderate $0.8-6.7\%$ increase in power compared to the low Reynolds number flow. In the high Reynolds number flows, the coherent LEV, which is critical for high-efficiency energy conversion, forms earlier and is slightly stronger, resulting in more power extraction. After the LEV is shed from the foil, the LEV trajectory is demonstrated to be relatively independent of Reynolds number, but has a very strong nonlinear dependence with kinematics. It is shown that the LEV trajectories are highly influenced by the heave and pitch amplitudes as well as the oscillation frequency. This has strong implications for arrays of oscillating foils since the coherent LEVs can influence the energy extraction efficiency and performance of downstream foils.

physics.flu-dyn↗