Academy of Interdisciplinary Studies Division of Environment and Sustainability 234 Building Retrofit Benefit Project Supervisor: FUNG Jimmy Chi Hung / ENVR Co-Supervisor: LAU Alexis Kai Hon / ENVR Student: SUI Chang / MATH-CS Course: UROP 1000, Summer This project evaluates the energy consumption of buildings at the Hong Kong University of Science and Technology (HKUST) and, more broadly, across Hong Kong. For each HKUST building, annual energy consumption was estimated by combining gross floor area (GFA), a published energy-use intensity (EUI) benchmark, and end-use adjustment factors. This methodology was scaled to the city level using GIS data to classify buildings by type, height, and footprint, enabling the application of appropriate EUI values. Comparisons with actual utility data revealed significant discrepancies, largely attributed to inaccuracies in input data, particularly estimated floor counts and building heights, which affected GFA calculations. The discussion highlights how such data limitations distort results and identifies potential strategies to address these issues. Building Retrofit Benefit Project Supervisor: FUNG Jimmy Chi Hung / ENVR Co-Supervisor: LAU Alexis Kai Hon / ENVR Student: ZHANG Xiaoan / MATH-CS Course: UROP 1100, Summer Accurate building energy consumption prediction is critical for identifying cost-effective retrofit measures and tracking decarbonization targets. This project develops a reproducible GIS-based workflow to convert open 2D building footprints into detailed EnergyPlus models without manual drafting. Floor-plate polygons were enhanced with geometric and operational attributes in ArcGIS, exported as shapefiles, and automatically transformed into IDF input files using a custom Python script. The script assigned constructions, loads, schedules, and ideal-loads HVAC systems. Hourly simulations, driven by typical meteorological year data, produced baseline energy-use intensities and enabled scenario analysis. The methodology, demonstrated on the HKUST campus, reduced model-authoring effort by over 90% compared to conventional manual EnergyPlus workflows for Hong Kong. Estimation of Wet Deposition Chemical Components in Northern Hemisphere by Using Deep-Learning Technique Supervisor: FUNG Jimmy Chi Hung / ENVR Student: ZHANG Yuyang / DSCT Course: UROP 1100, Fall This study employed a Back-Propagation Neural Network (BPNN) to establish the relationship between 25 environmental parameters, including PM2.5, and Aerosol Optical Depth (AOD). Data from 2018–2021, focusing on the East Asia and Pacific region, were utilized. The methodology involved data matching and a 10-fold cross-validation approach to enhance model reliability. Results demonstrated a strong correlation between PM2.5 and AOD, emphasizing the challenges of model tuning and the importance of optimizing neuron configurations in hidden layers. This research highlights the potential of BPNNs in AOD prediction and suggests further exploration of neural network architectures for environmental applications.
RkJQdWJsaXNoZXIy NDk5Njg=