ISE Senior Design Symposium Award Winners

Project Abstract
Republic Finance is a consumer lending corporation which specializes in personal loans for consumers with FICO credit scores between 500-700. They wanted to minimize profit loss from charge-offs when lending and expand their online lending presence. Charge-offs occur when an individual does not pay back a loan, which results in Republic Finance taking the burden of the cost left on the loan as a total loss. Traditionally, consumer loan risk analysis uses FICO credit scores, but these models fail for online and younger consumer demographics. Bank transactional data is different from the traditionally used credit data, providing a competitive advantage to Republic Finance’s risk modeling operations. We implemented a new risk model utilizing bank transactional data that reduced losses that would occur through charge-offs. The model also allowed more efficient expansion in digital channels and growth in target consumer demographics without reliance on opening new physical locations. We present a sampling-based machine learning ensemble model that produces higher accuracy in prediction classification than the existing Republic Finance credit modeling suite does on the provided bank transactional dataset. Our modeling techniques have a projected impact on savings of over $2,100,000 annually for our client in charge off losses and a net increase of over $1,300,000 annually in revenue.

Project Abstract

The FutureHAUS project is an interdisciplinary research initiative pioneering the development of sustainable, affordable housing solutions that can be mass-produced through prefabricated manufacturing. This unique approach offers a faster, more efficient alternative to traditional construction since these modular homes are constructed off-site in sections and assembled on location. As the U.S. housing market faces a shortage of over 3.7 million units with rising costs pricing 2.4 million potential buyers out of the market, FutureHAUS aims to partner with industry leaders to combat these challenges through the scalable production of its energy-efficient, modular home designs. FutureHAUS will present a new concept home at the 2026 Solar Decathlon, demonstrating its potential for affordable, large-scale manufacturing. To attract potential sponsors and partners, FutureHAUS must develop a solid business plan that clearly demonstrates how their new design can be produced successfully. Following a phase-gate methodology, we researched current industry practices and spoke to industry professionals to outline the necessary business plans for the scalable production of the concept home. The team created a bill of materials for the concept home, outlining the costs of each modular cartridge, and proposed a potential factory process layout. Our proposed prefabricated manufacturing approach has the potential to reduce labor costs by approximately $111,690 per unit, while also saving roughly $20,000 in administrative costs. The team also selected energy-efficient materials for the production process, which are expected to save future homeowners around $3,750 over time by lowering utility expenses.

Project Abstract
Roanoke County General Services manages critical infrastructure and services, including waste collection via fleet vehicles, providing solid waste and recycling collection services for citizens of the county. Frequent breakdowns, tows, and downtimes of the collection vehicles are contributing to high repair expenses. These failures stem from driver neglect, wear and tear, untimely maintenance, and truck overpacking. Additionally, the current solid waste and recycling routes are outdated and inefficient. The accumulation of these issues is costing the county almost $800,000, annually. Our team initiated the project by joining truck drivers for ride-alongs and interviewing key maintenance personnel. We employed a variety of tools for our approach, including Python, Microsoft Excel, differential equations, mathematical modeling, data analysis and management, and routing techniques. Our initial recommendations included installation of hopper cameras and multiple grease guns for driver use. Historical and live data from their FASTER Fleet Management maintenance record database and Trash View active route tracking software were used to identify future areas of improvement. For our deliverables, we finalized a preventative maintenance schedule, an official recycling route, reconfigured zones for bandit truck routes, and maximum can count thresholds drivers must adhere to prevent overpacking. A year after implementation, General Services is expected to experience a 12-hour reduction in downtime, a $1,936 reduction in towing, and a $39,922 reduction in repair costs. Our solutions are also projected to have $36,742 annual savings in material and labor costs, as well as improved service for the residents and increased driver morale through the year.

Project Abstract

Moog produces over 10,000 different types of slip rings for the space and defense industries. Moog’s overall customer demand has doubled since 2019, and the company has not been able to meet this demand because they do not have a complete understanding of their process capabilities. We were asked to focus on improving throughput, on time shipments, scrap and rework for specific slip rings considered of high importance based on volume and profit margins. We created a total of five value stream maps for five different slip rings through Visio. We used production clocking data, insight from production leaders and consulted with engineers to make our maps an accurate reflection of the manufacturing floor. A value stream map is an illustration of all the steps in a production process specifically in relation to material flow. Upon the completion of each value stream map, we presented the map to management along with the recommendations to improve the manufacturing process. Our recommendations successfully improved throughput by 156%, on-time shipments by 61%, scrap by 15%, and rework by 6%. Our recommendations mainly focused on addressing bottlenecks and meeting production goals. It is expected that these recommendations will help guide Moog to make focused improvements on the production line.