DCU Expo 2023 Final Year Projects

40 42. Techno-Economic Analysis of Implementing an Air Source Heat Pump in University Campuses The purpose of this project was to conduct a techno-economic review of air-source heat pumps (ASHP) for use on a university campus in Ireland. The relevant literature on ASHP was reviewed. From this review energy modelling took place based on the Irish climate. Simulations were then run to find the energy consumption of a heat pump with three different COP values. The results of these simulations were compared. The energy consumption of the heat pumps was also compared with that of a gas boiler. The results of these simulations found that for higher COP values the energy consumed was less than for lower COP values. It was found that the ambient temperature had an effect on the energy consumption. It was also found that the ASHP consumed less energy than the gas boiler. Class Mechatronic Engineering Project Area 3-DModelling, Energy Conservation, Simulation, EnergyModelling Project Technology Excel/VB, EnergyPlus, DesignBuilder Student Name(s) Fiona O’Shea Email fiona.oshea8@mail.dcu.ie Supervisor Dr Reihaneh Aghamolaei 43. Locker Locker is a door lock authentication system that aims to utilise face and speech recognition. Locker allows users to train face and speech recognition models which can be used to authenticate users. Using the Locker system, users can add and control multiple devices connected to the Locker network. Using a phone camera andmicrophone, users can then train the necessary face and speech models used for user authentication. Images of the user are used to train face recognition models, while speech patterns are analysed to uniquely verify the identity of the speaker. Registered devices running the Locker system then use these recognition models to authenticate users’ faces and speech accordingly. Class Computer Applications Project Area Artificial Intelligence, Biometrics, Computer Vision, Internet of Things, RaspberryPi, Speech Recognition, Web Application Project Technology CSS, Docker, HTML5, JavaScript, Nodejs, Python, REST, SQLite, React.js, Machine Learning Student Name(s) Joseph Libasora  |  The Ky Lien Email joseph.libasora2@mail.dcu.ie   |  the.lien3@mail.dcu.ie Supervisor Dr David Sinclair 44. Explainable AI in Pathology: Concept-Based Explainability for Mitotic Figure Detection inWhole Slide Images This project investigates the utility of human-interpretable concepts in understanding the predictions of an object detectionmodel to detect mitotic figures in whole slide images of various tissues. An automated approach to generating potential concepts is used, eliminating the need to annotate visual examples of these concepts. The aimof this project is to explore the potential for this automated approach to provide meaningful insights into the predictions for a visually complex task that typically requires an expert. Class Data Science Project Area Artificial Intelligence, Computer Vision, Explainable AI Project Technology Python, Machine Learning Student Name(s) AdamTegart Email adam.tegart2@mail.dcu.ie Supervisor Dr AlessandraMileo