Indoor occupancy prediction can play a vital role in the energy-efficient operation of building engineering systems and maintaining satisfactory indoor climate conditions at the lowest possible energy use by operating these systems on the basis of occupancy data. Many methods have been proposed to predict occupancy in residential buildings according to different data types, e.g., digital cameras, motion sensors, and indoor climate sensors. Among these proposed methods, those with indoor climate data as input have received significant interest due to their less intrusive and cost-effective approach. This paper proposes a deep learning method called CNN-XGBoost to predict occupancy using indoor climate data and compares the performance of the proposed method with a range of supervised and unsupervised machine learning algorithms plus artificial neural network algorithms. The comparison is performed using mean absolute error, confusion matrix, and F₁ score. Indoor climate data used in this work are CO₂, relative humidity, and temperature measured by sensors for 13 days in December 2021. We used inexpensive sensors in different rooms of a residential building with a balanced mechanical ventilation system located in northwest Copenhagen, Denmark. The proposed algorithm consists of two parts: a convolutional neural network that learns the features of the input data and a scalable end-to-end tree-boosting classifier. The result indicates that CNN-XGBoost outperforms other algorithms in predicting occupancy levels in all rooms of the test building. In this experiment, we achieved the highest accuracy in occupancy detection using inexpensive indoor climate sensors in a mechanically ventilated residential building with minimum privacy invasion.

Accurate long-term forecasts of aggregate energy load profiles are crucial for effective energy system planning at regional and national scales. This study aims to validate PROFet, a flexible load profile modeling tool. PROFet forecasts weather-dependent heating and electrical load profiles at an hourly resolution for both residential and non-residential buildings connected to district heating systems. Given that the tool’s accuracy for residential buildings has been demonstrated in previous studies, further validation is needed for non-residential buildings. To achieve this, our study evaluates PROFet’s extrapolation performance using an out-of-sample test dataset from Trondheim municipality. The results show that PROFet accurately forecasts the hourly heating load during the space-heating season but is less accurate during periods when domestic hot water needs dominate. The tool exhibits nearly identical performance across the three cases, except for efficient kindergartens, where the estimation accuracy is lower.

Behind-the-meter data were collected in three residential buildings in Trondheim, Norway, heated with direct electric panels and a wood stove, as part of the SusWoodStoves research project, which aimed to investigate the influence of wood stove operation on electric power consumption. The dataset is characterized by high resolution in time, space (that is, room level), and the number of recorded physical quantities, allowing for a wide range of investigations related to indoor climate, energy use, and building–grid interaction. Specifically, the dataset includes indoor and outdoor environmental variables such as air temperature, CO₂, TVOC, radon concentrations, and particulate matter; global solar irradiation has been recorded from a nearby meteorological station; electric power is measured for each heat emitter and on the main meter; and the surface temperature of the wood stove is recorded to monitor its operation. Measurements were sampled every 5 min, with periods ranging from two to six months.