A5038517524- Cellular Mechanics and Interactions
- Particle accelerators and beam dynamics
- Composite Material Mechanics
- Force Microscopy Techniques and Applications
- Advanced Mathematical Modeling in Engineering
- Elasticity and Material Modeling
- Particle Accelerators and Free-Electron Lasers
- Superconducting Materials and Applications
- Gyrotron and Vacuum Electronics Research
- Probabilistic and Robust Engineering Design
- Distributed and Parallel Computing Systems
- Magnetic confinement fusion research
- Microstructure and mechanical properties
- Coagulation, Bradykinin, Polyphosphates, and Angioedema
- Heat Transfer Mechanisms
- Metal and Thin Film Mechanics
- Machine Learning in Materials Science
- Real-time simulation and control systems
- Plasma and Flow Control in Aerodynamics
- Scientific Computing and Data Management
- Advanced Neuroimaging Techniques and Applications
- Non-Destructive Testing Techniques
- Adhesion, Friction, and Surface Interactions
- Hydrogels: synthesis, properties, applications
- Advanced Fluorescence Microscopy Techniques
Rensselaer Polytechnic Institute
2020-2025
Institute of Cancer Research
2024
Los Alamos National Laboratory
1993-2023
Imperial College London
2023
Abstract Genetic variation at the 19q13.3 KLK locus is linked with prostate cancer susceptibility in men. The non-synonymous KLK3 single nucleotide polymorphism (SNP), rs17632542 (c.536 T > C; Ile163Thr-substitution PSA) associated reduced risk, however, functional relevance unknown. Here, we identify that SNP variant-induced change PSA biochemical activity mediates pathogenesis. ‘Thr’ variant leads to small subcutaneous tumours, supporting risk. However, also displays higher metastatic...
Abstract This work presents a novel sensitivity approach that quantifies to regimes of model’s state variables rather than constitutive model parameters. Physical Regime Sensitivity (PRS) determines which independent have the biggest influence on an experiment or application. PRS analysis is demonstrated strength used in simulation copper Taylor cylinder. In series simulations, was perturbed sequentially local plastic strain, strain rate, temperature and pressure, then prediction cylinder...
Stochastic athermal networks composed of fibers that deform axially and in bending strain stiffen much faster than thermal axial elements, such as elastomers. Here we investigate the physical origin stiffening network materials. To this end, use models stochastic subjected to uniaxial deformation identify emergence two subnetworks, stress path subnetwork (SPSN) support (BSSN), which carry most energies, respectively. The BSSN controls lateral contraction modulates organization SPSN during...
model-traits is a C++ library for setting up scientific models and computational analysis.It provides minimal API applying boundary conditions (or other attributes) to the geometry of model.model-traits can either be used directly as library, or generate input files an existing analysis code.The design optimized make adding new output file formats easy maintainable without patching core library.
We develop a new neural network architecture that strictly enforces constitutive constraints such as polyconvexity, frame-indifference, zero strain energy with deformations, and the symmetry of stress material stiffness. Additionally, we show for this network, accuracy is significantly improved by using Sobolev minimization strategy includes derivative terms. Using our minimization, obtain NMSE 0.15% energy, 0.815% averaged across components stress, 5.4% This machine learned model was...
This paper presents efforts to improve the hierarchical parallelism of a two scale simulation code. Two methods GPU parallel performance were developed and compared. The first used NVIDIA Multi-Process Service second moved entire sub-problem loop into single kernel using Kokkos PackedView data structure. Both approaches improved with method providing greatest improvements.
This article presents MuMFiM, an open source application for multiscale modeling of fibrous materials on massively parallel computers. MuMFiM uses two scales to represent such as biological network (extracellular matrix, connective tissue, etc.). It is designed make use multiple levels parallelism, including distributed parallelism the macro and microscales well GPU accelerated data-parallelism microscale. Scaling results microscale show that solving problems concurrently can lead a 1000x...
The growing scale of high-performance computing systems increasingly enables scientists to develop more complex applications as in situ workflows composed coupled simulation and analysis codes. It is therefore important that workflow programming runtime middleware support the composition execution these intuitively efficiently. scientific community has put significant effort into purpose-built codes have been optimized for specialized use cases. However, development involving coupling...
Abstract The next generation of aero engines feature a more compact design with shorter distance between the fan rotor and bypass outlet guide vane (OGV) assembly. As such, potential forced response due to rotor-stator interactions has become significant concern regards high cycle fatigue component lifespan. It been shown that deploying non-axisymmetric OGV configuration is able reduce first engine order (1EO) forcing on fan’s blades. This paper explores effects aerodynamic vanes themselves....
An experimental demonstration confirming the beam dynamics of a ramped-gradient drift-tube linac (RGDTL) was performed at Los Alamos National Laboratory. The RGDTL designed to meet requirements maximum acceleration, minimum longitudinal and transverse emittance growth, acceptable wall power loss. At low energies, transverse-magnet focusing is weak RF defocusing must be minimized. As energy increases, stronger can tolerated electric field gradient increase. A detailed comparison theory...
This article presents the displacement field produced by a point force acting on an athermal random fiber network (the Green function for network). The problem is defined within limits of linear elasticity, and obtained numerically nonaffine networks characterized various parameter sets. classical solution applies at distances from larger than threshold which independent parameters in range studied. At smaller distances, nonlocal nature interactions modifies solution.