Presented by: Seyed Morteza Moghimi
Date: Friday, June 20, 2025
Time: 10:00 am
Place: Zoom - see below.
Abstract: Energy efficiency is a critical tool for reducing greenhouse gas emissions, particularly in the residential building sector. This research presents a Machine Learning-based energy optimization framework for Connected Smart Green Townhouses (CSGTs) in Burnaby, BC, focusing on cost, emissions, and occupant comfort. A series of five integrated studies address energy demand prediction, load optimization, and real-time energy management using hybrid models. A deep LSTM-CNN model demonstrated high accuracy (MAPE < 5%, R² > 0.85) across different bedroom types in grid-connected mode. In island mode, incorporating Electric Vehicles (EVs) and V2G functionality improved resilience, achieving MAPE of 4.43% and RMSE of 3.49 kWh. An adaptive load management system was developed using occupancy and pricing data, with MOPSO enabling automatic transitions between modes and up to 12% efficiency improvement. An occupant-centric framework leveraging IoT data reduced energy loads by 13%, peak loads by 11%, and emissions by up to 24%, while improving thermal comfort by 19%. These results validate the model's effectiveness in real-world smart buildings. The proposed solutions are scalable and adaptable for future urban housing. Ongoing work explores advanced topics like federated learning, edge AI, and grid-interactive controls to further enhance energy system intelligence.
Date & Time: Friday, June 20, 2025 | 10:00 AM (Eastern Time)
Join Zoom Meeting:
https://uvic.zoom.us/j/82916028226?pwd=krrYO8tLqAXAUagD9J4Himog20MOnt.1
Meeting ID: 829 1602 8226
Passcode: 885192 Dial-in Options: Note: Please log in to Zoom via SSO and your UVic Netlink ID |
Skip to this page's content
