Engineering Commons^™

Voltage Scaled Low Power Dnn Accelerator Design On Reconfigurable Platform, Rourab Paul, Sreetama Sarkar, Suman Sau, Sanghamitra Roy, Koushik Chakraborty, Amlan Chakrabarti

Electrical and Computer Engineering Faculty Publications

The exponential emergence of Field-Programmable Gate Arrays (FPGAs) has accelerated research on hardware implementation of Deep Neural Networks (DNNs). Among all DNN processors, domain-specific architectures such as Google’s Tensor Processor Unit (TPU) have outperformed conventional GPUs (Graphics Processing Units) and CPUs (Central Processing Units). However, implementing low-power TPUs in reconfigurable hardware remains a challenge in this field. Voltage scaling, a popular approach for energy savings, can be challenging in FPGAs, as it may lead to timing failures if not implemented appropriately. This work presents an ultra-low-power FPGA implementation of a TPU for edge applications. We divide the systolic array of …

A Scalable Approach To Minimize Charging Costs For Electric Bus Fleets, Daniel Mortensen, Jacob Gunther

Electrical and Computer Engineering Faculty Publications

Incorporating battery electric buses into bus fleets faces three primary challenges: a BEB’s extended refuel time, the cost of charging, both by the consumer and the power provider, and large compute demands for planning methods. When BEBs charge, the additional demands on the grid may exceed hardware limitations, so power providers divide a consumer’s energy needs into separate meters even though doing so is expensive for both power providers and consumers. Prior work has developed a number of strategies for computing charge schedules for bus fleets; however, prior work has not worked to reduce costs by aggregating meters. Additionally, because …

Research In Robotics: Vex Robotics, Dylan Wagner, Darec Wagner, Lyric Yost, Alexander Redinger, Yue Fan, Mark Hamrick, Mufeng Shen, Lawrence Yan

SACAD: Scholarly Activities

Throughout the past year our group has been working toward the construction of a robot. Our goal with this robot was to compete in the competition laid out by VEX robotics. Our robot would be built using the VEX robotics pieces as well as some custom-made parts. We utilized the engineering design process to ensure that we were working as efficiently as possible, whilst also working towards making the greatest bot. The first semester had us split into two teams, with both going to Pittsburgh KS to compete against other universities from across the state. After this competition, we combined …

Successful Rocket Launch In Argonia, Seungyou Cho

SACAD: Scholarly Activities

The object of launching the PK-56 (The name of the kit) rocket is how to take off and land safely using the parachute deployment. To inspect the engineering process, I tried to get an L1 certification from Tripoli Rocketry Association who is judging whether the rocket was flying stable. This consisted of building the structure, deployment system, engine system, anddesigning process. The most important thing was to read the instructions thoroughly, and to build a safe rocket.

Since launching high-power rockets needs a comprehensive understanding of those processes, I tested with small rocket with a motor that impulses 1.26 N …

High-Fidelity Turbulence Modelling Of Double-Diffusive Mixing In Buoyancy Driven Flows, Sanaz Izadifar

Student Research Symposium

Double-diffusive convection is the flow caused by buoyancy gradient in a fluid, where two different scalars cause density gradients, and each scalar has significantly different diffusion rates.

Dynamics Of Rotating 2d Ellipse In Stratified Flow, Anton Kadomtsev

Student Research Symposium

• In nature, there exist innumerable flows with varying density, known as stratified flows.

• The density variation is caused by salt content or temperature variation.

• Submersibles in stratified flows approximated by spheroid.

• Fluid parcels perturbed from their neutral position gain potential energy, which becomes kinetic energy. Inherent oscillation is Brunt-Väisälä frequency.

• Stratification affects the flow structures and forces on bodies.

Wake Characteristics Of In-Phase And Out-Phase Array Of Oscillation Airfoils, Yi-Cheng Chen

Student Research Symposium

In order to solve the hydrodynamics of a school of fish using the open-source SEM solver Nek5000, an oscillating mesh is required to simulate fish oscillations and at the same time simplify the process of mesh partitioning and satisfy the limitation of nek5000 that the mesh must be quadrilateral.

Buckling Analysis Of A Novel High-Throughput Assembly For Accelerated Fatigue Test, Mehedi Hasan Mahfuz

Student Research Symposium

The novel high-throughput assembly was developed by Singh et al.[2024] previously to study creep, which we are using to study fatigue instead.

The main difference of testing creep and fatigue is, during fatigue test, the specimens are loaded with half-time compression and half-time tension where creep is only subjected to tensile load.

Machine Learning Based Control Of Digital Throttle For Impulse Design In Hybrid Rockets, Eric Larsen

Student Research Symposium

Research Goals

• Develop a Hybrid Rocket model to generate expected thrust profiles given a CD profile
• Create a ML method to generate expected Thrust profiles given an Injector CD Profile
• Develop a method to then generate CD profiles from a desired Thrust profile of a motor

Mechanical Property Variability In Lpbf Ss316l Due To Defects Formation And Process Interruption, Tasrif Ul Anwar

Student Research Symposium

This project aims to understand the correlation between the manufacturing process, defect formation, process interruptions, and the resulting properties of components made with laser powder bed fusion techniques.

Characterization Of Functionally Graded Regolith-Metal Powder Mixtures For Additive Manufacturing In Space, Ian Jones

Student Research Symposium

Additive manufacturing (AM), colloquially known as 3D Printing, is a manufacturing process in which a component is fabricated layer-by-layer. For metallic materials, the most common AM methods use powders as feedstock materials. To print in space, one abundantly available material is the loose powder covering the moon, i.e., lunar regolith. However, past research has shown that regolith does not have ideal properties for AM.

There are two main material properties that affect AM performance: (1) powder flowability, and (2) packing density. Powders must be able to flow over the print bed, otherwise the printer will jam. When powders naturally pack …

Dynamic Angle Of Repose Of Ss316l Powders Fabricated By Gas Atomization For Additive Manufacturing Applications, S. M. Mahid

Student Research Symposium

Metal additive manufacturing (AM) techniques, unlike conventional manufacturing methods like turning and milling, construct objects by incrementally adding layers of material. Various processes can be used to 3D print including laser or electron beams to melt or sinter metallic powder particles.

A Novel Camera Based Measurement Technique For High Temperature Applications, Prasenjit Dewanjee

Student Research Symposium

Digital Image Correlation (DIC) is an optical technique for camera-based measurements.

Uncertainty Quantification In Neural Networks For Cfd, Cody Grogan

Student Research Symposium

Neural Networks have surged to popularity in many fields because of their ability to model complex functions. However, for most scientists, this comes at the cost of interpretability.

The aim of this research is to compile the different methods to uncertainty quantification in Neural Networks and apply them to a CFD problem.

Trip Production Models For Persons With Disabilities, Pouyan Saeidian

Student Research Symposium

People with disabilities constitute nearly 13% of the US population, with their mobility being more restricted compared to other groups. To provide equitable access to transportation facilities, it's crucial to delve deeper into their travel patterns and mobility. Our research examines their mobility patterns by comparing daily trip metrics with those without disabilities and modeling and analyzes factors influencing their travel frequency to inform policy enhancements

An In-Depth Investigation Into Factors Influencing Pedestrian Crash Severity: Comparative Analysis Of Generalized Ordered Probit, Stacking Ensemble Model, And Tabnet, Amir Rafe

Student Research Symposium

• Pedestrian Vulnerability: In 2021, there were 7,388 pedestrian deaths in the U.S., with urban settings, intersections, and dark lighting conditions being significant contributing factors.

• Prediction Models in Research: Pedestrian safety research employs a range of statistical, machine learning, and deep learning methods to assess factors affecting crash severity. TabNet, a novel deep learning model, is being used for the first time in the analysis of pedestrian crash severity.

Simulating Low Gravity Particle Dynamics, Ryan Lewis

Student Research Symposium

Lunar and Martian dust can pose a serious risk to astronauts and equipment due to the abrasive nature. A better understanding of how dust particles behave in low gravity is crucial in minimizing risks for future NASA Lunar and Martian missions.

• Existing methods are too expensive for detailed studies.
• We are developing an experimental platform that offers an affordable and robust way to study low gravity particle- laden flow.
• In this novel approach, the total body force on a dust particle is a proxy for gravitational force and can be controlled by imparting a vertical coulomb force on a charged …

Seasonal Shifts And Insights: Updating The Great Salt Lake Halogen Study (2022/24), Himalaya Sapkota

Student Research Symposium

Halogens compounds, particularly chlorine- and bromine-related species, may alter the photochemical reactions influencing local and regional air quality – including particulate formation (PM_2.5) and ozone (O₃) cycling.

High-Throughput: An Approach For Simultaneous Creep Testing Of Nuclear Structural Alloys, Raushan Singh

Student Research Symposium

Creep is a time-dependent deformation at constant load and elevated temperature.

It is widely used in an application that involves high temperature under load for e.g., Nuclear reactors.

Creep tests are notoriously slow, taking weeks/months to complete one experiment.

• This makes High-Throughput techniques more desirable

Our Goals:

(1) Demonstrate novel High-Throughput methods to perform more creep tests more quickly

(2) Apply those methods to more quickly characterize 3D printed 316L (future work)

Fast Prediction Of The Trajectory And Coverage Of Aerially Delivered Fire-Retardants, Aditya Parik

Student Research Symposium

Motivation & Objectives

• The utilization of aircrafts in fighting wildfires has substantially increased over the last century, as incidences of forest fires increase all over the world.
• A constant effort of various fire control agencies throughout North America (such as USDA Forest Service) has been to improve the accuracy, efficiency, and overall effectiveness of aerial retardant delivery.
• In service of this goal, we develop a model for predicting the general trajectory a retardant plume takes when it is aerially delivered to a target.

Combined Radiation And Convection In Developing Flow In A Parallel Plate Channel With Real Gas Behavior, Kyle Davis Pulsipher

Theses and Dissertations

The laminar, thermally developing flow of gases in a two-dimensional parallel plate channel is studied here. Using computational fluid dynamics (CFD), the combined effects of convection and real gas radiation are explored. Gas mixtures composed of H2O, CO2, and N2 are considered. To account for radiation in the participating medium, the Spectral Line Weighted-sum-of-gray-gases model is used. The model formulation, verification, and validation are provided through published articles written by the primary author. The predictions made for the case of combined convection and radiation are compared to the solution for the classical pure convection case, known as the Graetz solution. …

Quantifying And Characterizing The Behavior Of Dissolved And Suspended Icp-Oes Detectible Elements In Utah Lake, Rachel Ann Valek

Theses and Dissertations

For Utah Lake in Utah, USA, we analyzed a uniquely comprehensive and longitudinal water quality dataset that extended over two summer sampling seasons. We measured 61 parameters using Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES), YSI ProDSS water sondes, and laboratory procedures to quantify and characterize dissolved and suspended behavior of trace elements in Utah Lake. We analyzed the data using principal component analysis (PCA) and multidimensional scaling (MDS) to understand the lake geochemistry and inform Utah Lake management decisions regarding nutrients and algal blooms. We found that 1) sediments rich in phosphorus (P) act as a source for bioavailable …

Strategies To Decrease Separation Rates Among Highly Skilled Manufacturing Employees To Gain A Competitive Advantage, Rebecca J. Mileham

Walden Dissertations and Doctoral Studies

Company leaders have a decreased competitive advantage when losing their highly skilled manufacturing employees due to separation from the company. Grounded in Herzberg’s two-factor theory, the purpose of this qualitative single-case study was to explore strategies aeronautical leaders used to decrease separation rates among highly skilled manufacturing employees to gain a competitive advantage. The participants comprised eight frontline managers in southern California with 2 years or more experience using strategies to decrease separation in their organization. Data were collected through face-to-face semistructured interviews and analyzed using thematic analysis, resulting in four themes: pay and promotion; benefits and feedback; treatment from …

A Multi-Directional Smart Hip Implant With Wireless Connectivity For Enhanced Health Monitoring, Leila Donyaparastlivari

Mechanical Engineering Theses

The thesis presents the culmination of the smart hip implant project, which signifies a significant stride in orthopedic technology. By integrating piezoelectric and triboelectric systems into conventional implants, this study demonstrates the feasibility and effectiveness of creating intelligent implants. Through a meticulous design process and rigorous testing, novel configurations were developed to optimize energy harvesting efficiency and enable precise detection of mechanical loads experienced during various physical activities. The application of Finite Element Analysis provided critical insights into the localization of contact points, informing the strategic placement of harvesters within the implant for enhanced performance. Utilizing a cam follower and …

Systems, Methods And Techniques For Torrefaction, Ezra Bar Ziv, Stas Zinchik, Shreyas Sunil Kolapkar

Michigan Tech Patents

A system for torrefaction of waste material comprising biogenic material and plastic material may comprise a material pre-processing system, a heating and compaction unit, a reactor system comprising a reaction portion and an extrusion portion, and a cutting unit adjacent an outlet of the reactor system. A method for operating a system for torrefaction of waste material comprising biogenic and plastic material may comprise processing the waste material to generate waste material having an aspect ratio between 0.8:1 and 1.2:1 and a largest dimension of less than 4 millimeters (mm); compressing and heating the pre-processed waste material in the heating …

Igniting Passion:​ A Detailed Journey Through Rocketry Course Activities, Krish M. Patel, Hannah Caycedo, Joshua Gage, Josi Maness, Kevin Park, Mufeng Shen

SACAD: Scholarly Activities

This course is a semester-long adventure in rocketry, led by Dr. Paul Adams. It covers everything about building and flying rockets, starting from the basics to more advanced rocketry. Students learn how to build rockets and use equipment used I payload systems like and altimeter and a GPS. We also learned about the importance of safety involved with building and launching rockets.

Improving Criticality Safety Benchmark Coverage By Developing A Benchmark Evaluation Of The Unm Agn-201m Reactor, Rowdy Davis

Nuclear Engineering ETDs

This research aimed to improve the benchmark coverage of the International Criticality Safety Benchmark Evaluation Project (ICSBEP) handbook and the International Reactor Physics Experiment Evaluation Project (IRPhEP) with respect to High-Assay, Low-Enriched Uranium Fuel (HALEU). This was accomplished by creating a benchmark of the University of New Mexico’s Aerojet General Nucleonics model 201-M reactor and submitting it to the ICSBEP. This research then aimed to quantify to what extent this benchmark improves coverage of the benchmark suites for HALEU and TRISO fuel applications. This investigation utilized WHISPER, TSUNAMI-IP, and TSURFER with 36 application cases developed in SCALE and 2,929 experiments …

04.08.2024 Orsp Connect, Liz Williamson

ORED Newsletter

Nature Masterclasses Overview, Mary Lea McMillan, Cost Transfer Policy, NSF Author and Allocation citing

Kitchen Ergonomics In Health And Healthcare: A Rapid Scoping Review, Haozhi Chen, Nicholas Anton, Richard J. Holden, Denny Yu

Purdue University Libraries Open Access Publishing Fund

Although traditional workplace settings such as hospitals and clinics are well-studied, non-clinical areas like home kitchens, which connect clinical healthcare with the care provided in daily living settings, have received less attention. During the early stages of the global pandemic, there has been an increase in home cooking activities due to widespread isolation and restrictions imposed on dining publicly. Ergonomics is at the intersection between work environments and human task performance, and kitchen ergonomics informs knowledge on effective, safe, and efficient cooking practices while maintaining the desired kitchen functionality. Numerous contributions across various disciplines have added knowledge to the kitchen …

Investigating Which Muscles Are Most Responsible For Tremor Through Both Experimental Data And Simulation, Daniel Benjamin Free

Theses and Dissertations

Tremor affects millions of people and many patients desire alternative treatment options to medication or neural surgery. Peripheral suppression techniques are gaining greater use, but are currently applied in a trial-and-error method. To optimize these techniques, the muscles most responsible for an individual patient's tremor need to be identified. In this dissertation, I explored two parallel paths that both could aid in identifying muscles responsible for tremor. The first method utilizies measured data and a technique (coherence) that quantifies the frequency dependent correlation between two signals. Using coherence to identify muscles contributing to tremor requires at least two parts: an …