2022 Computer Science and Engineering

The pre-recorded talks and posters on this page showcase the work of students who received NC Space Grant research funding during 2021-2022. The menu at right provides links to pre-recorded talks and posters by other funded students on additional topics.

Josh Kokatnur

Career Internship Award at HyperX Software – Summer 2021
North Carolina State University
Undergraduate Student (Sophomore), Computer Science

Automation of Cloud Computing Systems for Structural Analysis

What resources can we use to process extremely large finite element models most efficiently? Collier Aerospace’s HyperX software utilizes both finite element models and finite element analysis results to perform structural analysis. With increasingly complex models, those components can scale to terabytes of data and the analysis could take weeks on an average computer. Because Collier Aerospace had already been experimenting with Amazon Web Services, I opted to use their tools and services. As for designing the process, I had three main decisions: the programming language, the method for file transfer, and the method for controlling the cloud computing instance. I chose PowerShell because it is a simple, yet powerful scripting language with important native functionality. Specifically, PowerShell is installed by default on all Windows machines and has native .net functionality (important for interacting with HyperX). For file transfer, I chose AWS S3 cloud storage, as it can store unlimited data and is already configured to work with remote compute instances. Similarly, for remote control, I opted for AWS Systems Manager, which allows you to run PowerShell code on the instance. The script initially ran through the PowerShell console, which is a basic command prompt application. The user had limited control of the process, but this version of the script was able to transfer necessary files to the instance, run analysis, and retrieve result files. Next, to improve upon the script’s functionality, I implemented a GUI (also through PowerShell). Through various iterations, the GUI ended up with controls for file/project selection and instance selection and also provided the live console output of the commands that are run on each instance. In our tests, we were able to process a model about five times faster on AWS than on a powerful laptop.

Mentor: James Ainsworth, Collier Research Corporation – HyperX Software

Shaun Murtha

NASA Internship Award at Langley Research Center, NASA Academy – Summer 2021
University of North Carolina Wilmington
Recent Graduate, Computer Science

2021 NASA Academy Summer Program Communications Sub Team

Wildland fires are growing in size, intensity, frequency, and destruction. During the 60 years between 1960 and 2020, the three years with the most acreage burned were 2017, 2019, and 2020. This heightened fire activity is straining existing firefighting resources and budgets across the nation and worldwide. For this reason, the 2021 NASA Academy Team was tasked with applying NASA technology to the challenges faced by on-the-ground wildland firefighters. The initial phases of the project involved identifying key problems to be solved, and the later stages involved splitting into sub-teams to tackle multiple areas of need. The approach was three-fold: Data Acquisition, Communications, and Equipment Development.The Communications Team focused on improving data latency for fire mapping and advancing communications between civilians and incident dispatch. The team developed two main concepts to address these areas: First, a software package implemented via an Uncrewed Aerial System (UAS), and second, a fully functioning mobile app.

Mentor: Elizabeth Ward, NASA Langley Research Center

Kyle Schultz

2021-22 NC Space Grant MSI STEM Bridge Scholar
Fayetteville State University
Undergraduate Student (Senior), Mathematics

Improving the Design of Heat Exchangers with Artificial Intelligence

This project concerns development of AI-Designed Heat Exchangers through the use of machine learning, and predictive neural networks. With advancements in electrical aviation reaching new heights, a need for multifunctional lightweight thermal management structures to reduce payload mass, and enable new aero efficiencies is becoming more prevalent. Design considerations include reduction in time to design by 10x over current industry standard, and ability to select multifunctional designs to reduce the weight of the system by 15%. To satisfy this demand, current efforts involve development of AI based natural language data analysis to comb existing taxa to curate data. The curated data is then divided up into non-overlapping subsets: the training set and the test set. The predictive neural network is trained with the training set with a loss function as the measure of how well the neural network has learned to predict the outputs from the inputs. With decreasing loss on the training set, the network predictive ability is also measured on the hold-out test set. The decreasing loss in both training and test data suggests that our choice of neural network and curated dataset has started to produce trained models. Given the complexity of the problem, the focus of our work will be in design and testing other neural network architectures while also finding other data sources. With the sheer volume of data ever-increasing the use of AI, and Machine Learning techniques to analyze data becomes ever more prevalent, and can significantly reduce design times, and improve industrial productivity.

Faculty Advisor: Sambit Bhattacharya, Fayetteville State University

Patrick Wilson

NASA Internship Award at Langley Research Center, NASA Academy – Summer 2021
Duke University
Undergraduate Student (Senior), Mechanical Engineering

Aerial Monitoring of Wildfires

Wildland fires are growing in size, intensity, frequency, and destruction. During the 60 years between 1960 and 2020, the three years with the most acreage burned were 2017, 2019, and 2020. This heightened fire activity is straining existing firefighting resources and budgets across the nation and worldwide. For this reason, the 2021 NASA Academy Team was tasked with applying NASA technology to the challenges faced by on-the-ground wildland firefighters. The initial phases of the project involved identifying key problems to be solved, and the later stages involved splitting into sub-teams to tackle multiple areas of need. The approach was three-fold: Data Acquisition, Communications, and Equipment Development. The Communications Team focused on improving data latency for fire mapping and advancing communications between civilians and incident dispatch. The team developed two main concepts to address these areas: First, a software package implemented via an Uncrewed Aerial System (UAS), and second, a fully functioning mobile app.

Mentor: Elizabeth Ward, NASA Langley Research Center