Furnace and Process Data Analysis
Student Researchers: Phil Sizek, Melissa Griffith, Jai Kedia
Faculty Advisors: Jennifer Bowen and Drew Pasteur (Math)
The AMRE team was given Furnace, Batch, and Load Data logged by the SCADA system for Fives North American Combustion’s Furnace 6F43, located at Scot Forge, Spring Grove, Illinois. The aim of the project was to determine which factors of operation of Furnace 6F43 were significant in determining the energy usage of the furnace (measured by the total gas flow of the furnace in BTUs). The ultimate goal was for Fives to make recommendations to Scot Forge, based on the AMRE team’s findings, regarding Scot Forge’s operation procedures in an attempt to reduce Scot Forge’s fuel consumption
Idealized Cornering Footprint
Student Researchers: Michael Sokolich, Dagmawi Zegeye
Advised by Denise Byrnes (CS) and Pamela Pierce (Math)
The goal of this project was to develop software to outline a corning tire footprint for Goodyear Tire and Rubber Company. The cornering footprint of a tire is the image of the portion of the tire touching the ground as the tire goes through a turn. The initial goal of the project was to use a curve-fitting technique to model the cornering footprint of a tire given 8 input measurements and to output points that are on the curve to a text file. As time project progressed, more features were added to the program. Some of the requested features were viewing only the left or right edge of the footprint, viewing a designated “zone” of the footprint, multiple curve fitting techniques, and a menu to use the features.
Predictive Analysis of Tire Quality Metrics for Manufacturing Equipment
Student Researchers: Joshua Houtz, Stuart Young, Brian Maddock
Faculty Advisors: Drew Pasteur and Robert Wooster (Math)
The purpose of this project was to develop a predictive model of tire production failure for the Goodyear Tire and Rubber Company. Goodyear can use this model to alert machine operators of impending failure, who can then maintenance the machine accordingly. We examined a variety of tire measurements in a sample data set of tire production.Using R, MATLAB, and Excel, we explored several statistical and predictive analysis techniques.
Our techniques did not identify a consistent trend within the data set. This implies that scrap tire production is either random, or that it is determined by a different step of the process.
Identifying manufacturing factors that cause faulty displays
Student Researchers: Nisa Usman and Joseph Smith
Faculty Advisors: Sofia Visa (CS) and Robert Wooster (Math)
The purpose of this project was to isolate production factors that cause faulty displays. Machine learning analysis techniques including decision trees found combinations of productions factors that can be used to improve yield. Multinomial logistic regression confirmed these general trends and explored the effect of late shift production times on yield. Regions of high and low yield were compared to replicate desirable production conditions and to avoid undesirable conditions. Data analysis scripts were written in Scilab and R to provide Kent Displays with access to these analysis techniques in the future.
Undergraduate Research in Knot Theory
Student Researchers: Brian Foley and Kiera Dobbs
Faculty Advisors: Jennifer Bowen and John Ramsay (Math)
Our team decided to explore new approaches to problems in knot theory through the use of computer programs. We looked at problems such as: finding the Alexander Polynomial of a knot, the unknotting problem, and knot equivalence. We developed a program that automatically calculates the Alexander Polynomial of any given knot or link, a recursive process that is difficult to compute by hand. We also defined new methods for untangling complex knots and implemented them into a program that untangles knots automatically. Lastly, we revised the paper on Klein link hole placement from last summer, and wrote up our results from this summer in three other papers. We plan on submitting our work for publication speaking at several conferences this fall.
Finding promoters in tomato fruit genome
Student Researchers: Laith Sersain and Carlos Gonzalez Mendoza
Faculty Advisor: Sofia Visa
The purpose of this project was to develop a frequency- and clustering-based approach for finding promoters in the Solanum Lycopersicum genome. The students worked in collaboration with researchers at the Ohio Agricultural Research and Development Center to identify promoter elements in the genome of the tomato fruits. The new methodology was implemented in programs (C++) and scripts (Python, MATLAB), and used to identify several likely promoters.
Design Thinking for Language and Assistive Technology
Student Researchers: Steven Schott, Arpan Roy, Shannon McKnight
Faculty Advisors: Simon Gray (CS) and Diane Uber (Spanish)
The AMRE team was hired to serve as consultants to PRC and aid in the process of redesigning their communication software. PRC has been making assistive technology for over half a century and communication software for the past 30 years. Their initial client base was patients with Cerebral Palsy, but is now more predominantly patients with Autism Spectrum Disorder, which has resulted in a high learning curve when clients are first given the devices. This was the problem the team was asked to address. The AMRE team goals for the final product design were to make it intuitive, efficient, flexible, and if possible, able to aid in the language acquisition process. The following report contains a functional specification for the system the AMRE team designed. Other general recommendations the AMRE team has for PRC include:
- Having two separate systems: one for cognitively impaired populations, and one for motor impair populations;
- Developing a parent or caregiver website that allows for easy modification of the device and to encourage parent involvement;
- Combining core vocabulary with a personalized vocabulary to better encourage communication;
- And to install a suite of various interactive games for the purpose of teaching social skills and language to the clients.
If these recommendations and the AACtiveNet system are implemented into a device, the goals the AMRE team set out to accomplish will be met. Communication using these devices will be quicker, the learning curve will lowered, and the software will be a better contender in the 21st century technological market.
Automated Software Validation
Student Researchers: Doug Code and Pratistha Bhandari
Faculty Advisor: Denise Byrnes (Computer Science)
The PRC/Saltillo testing team worked to build a centralized framework for code review, static analysis, and unit testing for Linux and Windows operating systems. The resulting framework consists of Gerrit code review, SonarQube analysis, and Jenkins continuous integration. This framework shifts all code analysis to the server, which eliminates the need to run static analysis on developer machines. Gerrit lets developers participate in code reviews from any internet-connected location, and solves the issue of scheduling conflicts between developers. In addition, SonarQube tracks how full-project metrics have changed over time and between software versions. The Jenkins server works to continuously integrate changes into the product through regular software builds, and immediately reports build success or failure to developers. Through the use of these three components, the amount of incorrect code submitted is minimized, debugging is reduced, and code quality is improved.
Insurance Policy Geography Variable Analysis
Student Researchers: Meredith Schervish, Priyanka Datta, Varunavi Newar
Faculty Advisors: John Ramsay and Pam Pierce (Math)
The purpose of this project was to use a mapping software called ArcGIS to map and analyze the territorial segmentation of insurance rate factors for one of Progressive’s competitors. These factors are multipliers to a base rate for the calculation of someone’s auto insurance rate and are based on location. The competitor has increased segmentation of territories for their factors, which Progressive will investigate further using maps created for them in this project.
Cost Analysis
Student Researchers: Jeremy Lenz, Daniel Miller, Karlena Luz
Faculty Advisor: Jim Burnell (Economics)
The purpose of this project was to determine the costs that were associated with different parts of the Cleveland United Way 2-1-1. To do this we broke the costs into different major categories. We found three categories: call handlers costs, database costs, and business costs. By looking at the 2-1-1 expense report and working closely with management we were able to ascribe different costs to each area, allowing the 2-1-1 to better price their services for potential customers. We also did a productivity analysis using regression to help improve efficiency.
Year-round Schooling: Literature Review
Student Researchers: Alyssa Hullings, Sheamus Dalton, Abigail Frank
Faculty Advisors: Alison Schmidt and Megan Werely (Education)
Through a review of significant, scholarly literature and a qualitative study of the Wooster City School District and its various stakeholders and constituents, our AMRE team created a comprehensive review of year-round schooling, which includes (a) the history of both the traditional school calendar and the year-round schooling model, (b) a definition of year-round schooling, (c) definitions of the primary models of year-round schooling, (d) the fiscal impact of a year-round schooling calendar, (e) a review of the “summer slide” phenomenon and overall student academic achievement, and (f) the implications of year-round schooling on students, teachers, staff, and the community as a whole. Each of these topics were investigated with a focus on their context relevant to the Wooster City School District. In conclusion, recommendations for further research were included in the report.