Introduction

Texas ScholarWorks was established to provide open, online access to the products of the University's research and scholarship, to preserve these works for future generations, to promote new models of scholarly communication, and to help deepen community understanding of the value of higher education.

UT Tower and campus image credit: Earl McGehee, CC-BY, https://www.flickr.com/photos/ejmc/7452145850

Communities in TSW

Select a community to browse its collections.

Now showing 1 - 5 of 5

Conference Proceedings and Journals 5821
Student Works 29269
UT Communities 45961
UT Electronic Theses and Dissertations 34212
UT Faculty/Researcher Works 4029

Recent Submissions

Operational disruption scheduling
(2026-05) Sahoo, Aditya ; Bickel, J. Eric; Hasenbein, John J.
Hospitals are some of the most critical institutions that ought to be kept running, irrespective of any disruptive events that may or may not have a direct impact. The University of Texas MD Anderson Cancer Center, located in Houston, Texas is one of the largest cancer hospitals in the world, treating thousands of patients every year and leading cancer research with cutting-edge infrastructure. Given its size, any major disruptive event is expected to bring operations to a standstill. Our team at The University of Texas at Austin was tasked to recommend operational strategies in terms of deploying additional operating room (OR) capacity on weekends to clear the backlog of impacted patients should a disruptive event occur. This would ensure that the hospital has a proactive planning strategy to keep the patient flow running, even in case of the worst disruptions. Recommending a plan started with identifying the different types of disruptions MD Anderson has historically experienced, along with the types of natural calamities that Houston has been impacted by. The impact of disruption was classified into levels, so that each level would have its own planning strategy. The flow of patients was modeled as a queuing system, where patients arrive and receive their surgeries, with the arrival and surgery times defined by distributions obtained using historical data. The simulation of patient flow was realized using Arena, a popular software to simulate discrete events. Disruptions were then progressively introduced while simultaneously deploying additional ORs to check queue behavior. Multiple metrics were introduced to quantify the reduction of patient backlog, the metrics being the basis for recommending action plans, based on the severity of the disruption. The assumptions and limitations of this analysis are discussed thoroughly in the concluding sections of this report, along with discussing some of the possible actions that could be taken in refining the model in the future.
Active interfaces in nanofluidics : from nonequilibrium statistical mechanical foundations to fluid-structure interactions
(2026-05) Martínez Cordeiro, Juan Pablo; Aluru, Narayana Rao; Thomas J.R. Hughes; Donald Siegel; Li Shi
Fluid-structure interactions (FSIs) at the nanoscale are essential for understanding and mimicking ultraefficient biological nanofluidic systems. However, while historically there have been developments in Molecular Dynamics (MD) that allow simulations of nonequilibrium systems, most MD tools are designed for equilibrium and their applicability in nonequilibrium cases in nontrivial. Partly due to this limitation, the coupling of external surface-driven perturbations to the internal dynamics of water have not been studied in detail. This dissertation establishes a generalized framework for performing nonequilibrium MD simulations with existing and widely available tools and uses these developments to perform simulations studying FSIs at the nanoscale to develop a fundamental understanding of nonequilibrium interfacial nanofluidics.
Evaluating the simulation-to-real transfer of point cloud completion for autonomous surface inspections
(2026-05) Diaz, Jorge Alejandro, M.S. in Engineering; Pryor, Mitchell Wayne; Anderson, Robert Blake
Robotic systems are increasingly deployed for inspection and surveying tasks in hazardous environments, such as corrosion detection in refineries and radiation monitoring in nuclear facilities. Despite advances in autonomy, many systems still require human supervision to ensure safety and task completeness, limiting their ability to reduce human exposure to dangerous environments. In cluttered and unstructured settings, the robot's sensors are subject to limited fields of view and occlusions, resulting in partial observability impacting path planning and task execution. These limitations lead to incomplete inspection coverage and increased collision risk due to unknown scene geometry. This thesis investigates the integration of learning-based point cloud completion (PCC) into robotic perception and planning pipelines to infer missing geometric structure from partial observations. To address the challenge of acquiring real-world training data, we develop a synthetic dataset generation pipeline using computer-aided design (CAD) models of assets. Results indicate effective simulation-to-real transfer on real sensor observations, supporting the use of synthetic data for real-world deployment. We also present ongoing integration of PCC as a coarse geometric prior within an inspection pipeline that combines active perception and coverage planning in a behavior tree framework, enabling modular, hardware-agnostic deployment. These efforts demonstrate that PCC can help enhance autonomy and reliability in robotic systems operating with partial observability.
The influence of step width on individual muscle contributions to balance control during walking
(2026-05) Emerson, Helen Elizabeth; Neptune, Richard R.; Vistamehr, Arian
The failure to maintain dynamic balance may lead to falls, which can cause significant long-term physical injuries, limit mobility, and lead to a decreased overall quality of life. Older adults and those with neuromuscular deficits often have difficulty maintaining dynamic balance and walk with wider step widths. Wider step widths increase the mediolateral moment arm between the body center of mass and the foot center of pressure and have been associated with a higher time rate of change of angular momentum (Ḣ), which indicates poor dynamic balance control. Previous simulation work has identified the primary muscle contributors to frontal-plane Ḣ, but it remains unclear how those contributions change with altered step width. Therefore, the purpose of this study was to analyze how step width influences individual muscle contributions to frontal plane Ḣ using musculoskeletal modeling and simulation. A cohort of healthy young adults completed walking trials on an instrumented treadmill at four step widths: Narrow, Self-Selected, Wide and Extra Wide. Experimental body segment kinematic, ground reaction force and EMG data were used as simulation inputs and individual muscle contributions to Ḣ were quantified and compared across step widths. The results revealed the overall muscle contributions to net Ḣ changed little across step widths. As expected, individual muscle contributions to the vertical GRF component of Ḣ increased with step width, due to the increased mediolateral moment arm. However, this effect was opposed primarily by the contribution of gluteus medius to the medial GRF component, and by a decrease in contributions of the gastrocnemius and soleus to the opposing lateral GRF component. The opposing effects of step width on the vertical and mediolateral components of Ḣ resulted in the observed lack of change in contributions to net Ḣ. These results support the premise that the hip abductors and ankle plantarflexors are critical for frontal plane balance control and thus should be a central consideration in developing rehabilitation strategies for older adults and those with neuromuscular deficits who walk with wider steps.
Building patient-specific silicone phantoms from explanted human tricuspid valves
(2026-05) Abrahams, Luke S.; Rausch, Manuel Karl; Fuhg, Jan
This work presents a fabrication and testing platform for patient-specific silicone phantoms of the tricuspid valve. The tricuspid valve exhibits complex geometry and significant variability, making controlled and repeatable experimental study difficult when using biological specimens alone. Starting from computationally generated geometries, a workflow was developed to construct deformable silicone valves with integrated chordae and well-defined mounting conditions. The valves were evaluated in a pulsatile flow loop designed to reproduce physiological right-heart pressure loading. The fabricated valves exhibited stable and repeatable closure behavior, with full leaflet coaptation and no visible central regurgitation observed during testing. The measured pressure waveform showed physiologically relevant characteristics with minimal cycle-to-cycle variability, indicating consistent loading conditions. The fabrication approach was further extended to create a full silicone heart phantom using a sacrificial molding technique, demonstrating the ability to reproduce more complex anatomical geometries. These results demonstrate that synthetic valve models can be constructed with controlled geometry and consistent material properties while capturing key aspects of valve function. This platform enables systematic investigation of valve mechanics under well-defined conditions and supports future model validation and data-driven approaches.

Introduction

Discover

Communities in TSW

Recent Submissions