Volume 4: Innovative Solutions for Energy Transitions: Part III

Data-Driven Outdoor and Indoor Temperature Prediction for Energy-Efficient Building Operation Yuzhen Peng, Arno Schlüter

Abstract

Outdoor and indoor temperature prediction of local buildings is important for optimal building operation and energy-demand management. This study collects data from a commercial building, covering outdoor and indoor climate, and variables of occupants and building system operation. Based on the selected data, two different data-driven methodologies using machine learning techniques are proposed to predict local outdoor and indoor temperatures at a high resolution. The proposed data-driven models with learning capabilities are based on k-nearest neighbor and artificial neural networks, showing good prediction performance for the case study building.

Keywords data-driven modeling, temperature prediction, machine learning, energy efficiency, smart buildings