Being able to predict cooling load, the rate of heat removed from the indoor space, is crucial to achieve thermal comfort and high energy performance of central air-conditioning (AC) systems in buildings. In this talk, we first demonstrate feature engineering from time series data using Pandas package. With the extracted features, we develop our ML model (Random Forest Regressor) using Scikit-learn for cooling load prediction. The tradeoffs between model performance and business need are discussed by comparing two ML models with different features. Finally, how our clients benefit from the above application will be illustrated.
Speaker: Mr. Benny Tam / Hong Kong / ATAL Building Services Engineering Ltd Language: English Date and Time : October 8, 2021 / 14:00-14:30 (UTC+8)
Speaker Introduction
Benny Tam is an R&D Engineer at ATAL Building Services Engineering Ltd, working on HVAC (Heat, Ventilation, and Air-Conditioning) Optimization, Fault Detection & Diagnosis (FDD), and Data Pipeline Design. Benny graduated from The Chinese University of Hong Kong (CUHK) and obtained BSc and MPhil degree in Physics. His MPhil Thesis is related to statistical inference from stochastic process on complex networks. He has rich experience in programming such as Python, C++ and MATLAB. Currently, he heavily uses Pandas and Scikit-learn for building ML models. He also works with the backend team on data pipeline design using AWS cloud services.