Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Building Environment and Energy Engineering | en_US |
| dc.contributor.advisor | Fu, Xiao (BEEE) | en_US |
| dc.creator | Yan, Yiming | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/14110 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Electricity load prediction of HVAC system based on machine learning algorithm | en_US |
| dcterms.abstract | This study aims to develop a data-driven model based on the XGBoost machine learning algorithm to accurately predict the electricity load demand of Heating, Ventilation, and Air Conditioning (HVAC) systems. By integrating multi-dimensional features like weather, building physical characteristics, and operational patterns, the research systematically investigates the effect of various input parameters on the results of predictions and optimizes model hyperparameters to improve performance. The dataset, generated via EnergyPlus simulations, encompasses energy consumption, meteorological parameters, and building attributes of a large-scale office building. Three input feature scenarios (weather-only, weather + building features, and full-parameter features) were designed to validate the significant improvement in prediction accuracy through multi-dimensional inputs. | en_US |
| dcterms.abstract | Results demonstrate that combining weather and building features reduced the RMSE from 44,514.12 W to 22,839.71 W (a 48.7% decrease), improved R² to 0.9746, and optimized the MAPE to 7.55%. Feature importance analysis identified temperature (51.3%) and radiation (18.1%) as core drivers, while operational parameters (e.g., temperature setpoints) contributed minimally to short-term forecasting (R² increased by only 0.5%). | en_US |
| dcterms.abstract | The study further proposes prioritizing key feature parameters tailored to the energy consumption characteristics of different building types (e.g., data centers, factories). This research validates that XGBoost achieves efficient and high-precision predictions while reducing reliance on physical modeling, offering practical solutions for building energy management, smart control, and demand-side optimization. | en_US |
| dcterms.extent | viii, 36 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2025 | en_US |
| dcterms.educationalLevel | M.Eng. | en_US |
| dcterms.educationalLevel | All Master | en_US |
| dcterms.accessRights | restricted access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 8559.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 3.22 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/14110