Announcements

Author: Nikhil Oberoi

Title: MULTI-FIDELITY PARAMETRIC SENSITIVITY FOR LARGE EDDY SIMULATION

Committee members:

Professor Johan Larsson, Chair/Advisor
Professor Christoph Brehm, Dean’s representative
Professor Kenneth Kiger
Professor Arnaud Trouvé
Professor Jacob Wenegrat

Location: EGR-2164

Date and time: March 28, 2.00 pm

Abstract: Designing engineering systems involving fluid flow under uncertainty or for optimality often requires performing many computational fluid dynamics (CFD) calculations. For low-fidelity turbulence modeling simulations such as Reynolds-averaged Navier-Stokes (RANS), such a framework has been established and is in use. However, for high-fidelity turbulence resolving simulations such as large eddy simulations (LES), the relatively high computational cost of even a single calculation hinders the development of such a framework. The overarching goal of this work is to aid LES in becoming a usable engineering design tool.

In this thesis, a computationally affordable approach to estimate parametric sensitivities of engineering relevant quantities of interest in an LES is explored. The method is based on defining a RANS problem that is constrained to reproduce the LES mean flow field. The proposed method is described and assessed for a shock/boundary layer interaction problem, where the shock angle and wall temperature are considered variable or uncertain. In the current work, a proof-of-concept of the proposed method is demonstrated. The method offers qualitative improvements to the sensitivity prediction of certain flow features as compared to standalone RANS simulations, while using a fraction of the LES cost. Different cost functions to infer auxiliary RANS variables are also examined and their influence on the sensitivity estimation is assessed. Overall, the results serve as an important proof-of-concept of the method and suggests the most promising path for future developments.

Author: Vishal S Sivasankar

Title: Simulation of Polymeric Drop Dynamics: Effect of Photopolymerization, Impact Velocity, and Multi-material Coalescence

Day/Time: March 29, 11:00 am | Location: EGR 2164 (DeWalt Seminar Room

Committee Members:

Dr. Siddhartha Das, Chair/Advisor

Dr. Abhijit Dasgupta

Dr. Amir Riaz

Dr. Eleonora Tubaldi

Dr. Peter Kofinas, Dean’s Representative

Author: Xiaotian Xu

Date: Tuesday, March 28th, 2023, at 1:00 pm

Location: EGR-1131B

Zoom (for the Public Presentation):https://umd.zoom.us/j/7739971299?pwd=L3Z4T1dNTlVUT2kxU241ekwxS2RpZz09

Advisory Committee:
Professor Yancy Diaz-Mercado, Chair/Advisor 
Professor Nikhil Chopra 
Professor Hosam K. Fathy 
Professor Jin-Oh Hahn 
Professor Erick Rodriguez-Seda, Special Member
Professor Derek A. Paley, Dean’s Representative

Title of Dissertation: Multi-Agent Spatial Coordination Via Time-Variations In Coverage Control

Abstract:

Coverage control of multi-agent systems (MASs) spatially spreads out a group of agents to form a configuration over a domain of interest. This research investigates the two fundamental elements embedded in coverage control, i.e., the time-varying density function and the time-varying domain, and how these can be leveraged to achieve collaborative controls of MASs. We focus on three problems: first, we abstract a robotic swarm, so it can be controlled as a whole where the robotic team adaptively finds the suitable spatial configuration; second, such abstraction of a MAS is extended to a higher-dimensional embedding for an interactive multi-agent aerial cinematography application; and third, a multi-objective formulation is developed to spatially distribute a MAS and take advantage of its collective effort to persistently cover a space.

In contrast to the coverage with time-varying densities which has been actively studied, we address the coverage control over time-varying domains in the first problem, so the control of a MAS, in terms of its position, scale, shape, etc., is enabled and is simplified into manipulating the domain to be covered directly. The agents coordinate themselves to accommodate the evolution of the domain, even when the domain is evolving fast. A MAS control algorithm, named Swarm Herding, which is built upon the proposed control mechanism is implemented. In pursuit of this approach, contributions are made to the problem of coverage control over time-varying convex and non-convex domains for abstracting the swarm and synthesizing the specialized controllers for every agent in the swarm.

In the second, the abstraction is extended to a hemispherical manifold under the geodesic metric, and it is employed to enable an interactive motion coordinator for multi-robot aerial cinematography. The emphases are on collaborative behavior for multiple unmanned aerial vehicles (UAVs), tracking of a dynamic target, and real-time interaction for aesthetic cinematography objectives. Contributions are made in the design of a distributed interactive framework to provide high-level position instructions for a group of UAVs which addresses the gap in the “one-pilot-many-robot” feature.

In the third problem, a multi-objective coverage control of MASs is formulated to take advantage of the collective effort of a team of mobile sensors to persistently explore a domain of interest. In addition to the standard locational coverage objective, a new perceptional coverage objective is introduced to drive agents around in the domain to gain information. The collaboration between agents is defined not only in terms of exchanging the knowledge of the domain as in previous work but also in terms of inter-agent motion coordination which reduces redundant visits to certain locations by agents. Contributions are made with respect to information exchange with performance guarantees, multi-objective coverage control of MASs with time-varying state-dependent density functions, and analysis of the effects of the multiple objectives.

Author: Sai Etha

Dissertation title: Molecular-scale exploration of interactions between drops and particles with a polymeric layer
Date, time and venue: 3/16/23 from 3-5pm in EGR 2164

Committee Members: 

Dr. Siddhartha Das (Chair)

Dr. Taylor J Woehl (Dean’s Representative)

Dr. Pratyush Tiwary

Dr. Amir Riaz

Dr. Avik Dutt

Abstract: Surface-grafted polymer molecules have been extensively employed for surface modifications as they ensure changes to the inherent physical/chemical properties of surface. Bottom-up surface processing with well-defined polymeric structures becomes increasingly important in many current technologies. Polymer brushes, which are polymer molecules grafted to a substrate by its one end at close enough proximity (thereby ensuring that they stretch out like the “bristles” of a toothbrush), provide an exemplary system of materials capable of achieving such a goal. In particular, producing functional polymer brushes with well-defined chemical configurations, densities, architectures, and thicknesses on a material surface has become increasingly important in many fields.

In my dissertation, I employ Molecular Dynamics (MD) simulations to study the interplay of interactions between nanoparticles (NPs), solvent drops and polymer grafted surfaces under various system conditions. This study will help us to understand (1) the wetting dynamics of brush grafted surfaces and the associated brush conformational changes, (2) polymer-insoluble solvophilic NP assembly in brush grafted surfaces and the steric interactions driven establishment of direct contacts between a NP and a polymer layer (highly phobic to the NP), and (3) microphase separation and distillation-like behavior of grafted polymer bilayers interacting with a binary liquid mixture, and the resulting nanofluidic valving behavior of swollen polymer bilayers in a weak interpenetration regime.

Author: Namkyoung Lee

Date: Friday, March 17th, 2023, at 10:30 am

Location: EGR-2164

Zoom: https://umd.zoom.us/meeting/register/tJEkfuCrrzwoH9aKI60Y6SIXknB_NnBugyuM?_x_zm_rtaid=6SbQ52e7TFKsE1qdZ94tGg.1678212040878.bfb2c061cac0aca9eb0188a0a320b8cd&_x_zm_rhtaid=709

Committee Members:

Dr. Michael Pecht, Chair / Advisor
Dr. Michael H. Azarian, Co-chair / Co-advisor
Dr. Yunfeng Zhang, Dean’s Representative
Dr. Balakumar Balachandran
Dr. Mark Fuge
Dr. Gregory W. Vogl

Title of Thesis: Interpretable and Speed Adaptive Convolutional Neural Network for Prognostics and Health Management of Rotating Machinery

Abstract:

Faulty rotating machines exhibit vibrational characteristics that can be distinguished from healthy machines using prognostics and health management methods. These characteristics can be extracted using signal processing techniques. However, these techniques require certain inputs, or parameters, before the desired characteristics can be extracted. Setting the parameters requires skill and knowledge, as they should reflect the component geometries and the operational conditions. Using convolutional neural networks for diagnosing faults on a rotating machine eliminates the need for parameter setting by replacing signal processing with mathematical operations in the networks. The parameters that affect the outcomes of the operations are learned from data during the training of the neural networks. The networks can capture characteristics that are related to the health state of a machine, but their operations are not interpretable. Unlike signal processing, the internal operations of the networks have no constraints that guide the networks to transform vibrations into certain information, that is, vibrational characteristics. Without the constraints, there is no basis for understanding the characteristics in terms that can be associated with the physics of failure. The lack of interpretability impedes the physical validation of vibrational characteristics captured by the networks.
This dissertation presents a method for changing the internal operations of a convolutional neural network to emulate a specific type of signal processing known as envelope analysis. Envelope analysis demodulates vibrations to extract vibrational signatures associated with mechanical impact on a defective rolling component. An understanding of envelope analysis, along with knowledge of the geometries of machine components and operational speeds, allows for a physical interpretation of the signatures. The dissertation develops speed adaptive convolutional layers and a rotational speed estimation algorithm to identify defect signatures whose frequency components change as the speed changes. The characteristics that are captured by the developed convolutional neural network are verified through a feature selection process that is designed to filter out physically implausible features. Case studies on three different systems demonstrate the feasibility of using the developed convolutional neural network for the diagnosis.