Occupancy significantly influences building energy use; however, large-scale occupancy estimation, particularly in residential buildings, remains a challenge. While surveys and sensor data, particularly from remotely readable electricity meters, have been used for non-intrusive occupancy detection, water meter data have so far only been applied at daily resolution.This study introduces, validates, and applies a novel algorithm to infer hourly occupancy using data from remotely readable water meters. A new day definition, spanning noon to noon, is proposed to reflect typical daily rhythms better. Daily occupancy is inferred using water usage events and average consumption, akin to established methods. For days classified as occupied, hourly occupancy is estimated by associating water use with occupancy, followed by smoothing. Nighttime occupancy is inferred by linking the last evening and first morning occupancy.Validation was performed using one year of manually labelled data from ten households for daily occupancy (MCC = 0.982), and 147 days of unseen ground-truth data from nine households for hourly occupancy (MCC range: 0.267 to 0.795; building-weighted MCC = 0.594). Short absences were often misclassified as occupied, while longer absences (>3h) were reliably detected. Comparison with a state-of-the-art electricity-based algorithm showed superior performance of the proposed water-based method.Finally, the algorithm was applied to one year of data from 2,690 households, yielding plausible and interpretable occupancy patterns at both daily and hourly scales. These results demonstrate the method’s robustness and scalability, offering a promising approach for large-scale, non-intrusive occupancy monitoring.

Energy performance certificates (EPCs) have gained increasing attention in recent years in policies and research. However, little attention has been paid to the consistency of EPC input data over time and the ability to assess EPC quality by comparing it with prior or subsequent EPCs. This study analysed the EPC input data from 52,000 Danish single-family houses with two or more EPC reports. In total, about 105,000 EPC reports were investigated. Statistical analysis of the building geometry and thermal building envelope revealed significant changes. The results show that the building envelope area varies nearly independently of the reported floor area. Furthermore, the U-values of the opaque envelope increased by 45

The increasing demand for energy-efficient buildings has driven the development of innovative renovation concepts, such as the I-DIFFER system, which integrates Double Skin Facades (DSF) with Diffuse Ceiling Ventilation (DCV). This system has demonstrated comparable or improved energy performance compared to traditional renovation approaches while ensuring a good indoor environment. The efficient implementation of I-DIFFER requires a configuration for each specific building, ideally supported by Building Performance Simulation (BPS) tools. Despite their potential, BPS tools face challenges in accurately modeling DSF and DCV due to limitations in simulating specific physical processes occurring within these systems. Furthermore, the lack of studies validating DCV models in BPS tools with empirical data raises questions about their reliability for configuring and optimizing combined systems such as I-DIFFER. This study aims to experimentally validate two BPS models, evaluating their respective capabilities in assessing the performance of the I-DIFFER system in two operational modes: cooling mode, where airflow in the DSF is driven by natural forces and mechanically driven in the DCV plenum, and heating mode, where airflow in both the DSF and DCV is mechanically driven. The experimental data is obtained from a full-scale experimental set-up, and the simulation results show that both models are sufficiently accurate for predicting thermal conditions in the room with the I-DIFFER system. However, both models show weaknesses in predicting peak-hour thermal performance and face challenges in accurately simulating heat transfer between the DCV plenum and the room, potentially limiting the development of optimal control strategies.

The discrepancy between theoretical and actual heat demand-commonly referred to as the energy performance gap (EPG)-is well-documented across national contexts. However, less attention has been given to how this gap varies within energy efficiency classes and between household types. This study investigates the heterogeneity of the EPG in Danish single-family homes rated A2010 and A2015, using administrative data from 8,930 households. The dataset links Energy Performance Certificates (EPCs) with actual annual heat use data, covering both space heat and domestic hot water for the period 2019 to 2021. Results show modest variation in actual heat use across income groups, diverging substantially from the theoretical demand. For homes rated A2010, actual use ranges from 62.6 kWh/m2/year among low-income households to 64.6 kWh/m2/year among high-income households, which resembles an EPG of 36.6% and 41.1% respectively. Similar patterns are observed in A2015 homes. The results also showed modest variations across the variables, number of adults and children in the household, whereas the use patterns across years varied slightly. These findings highlight the role of household composition, occupant behavior, and socio-demographic factors in shaping actual energy use-even in highly energy efficient buildings. The study underlines that energy efficiency alone does not ensure low energy use and that energy policies and performance metrics should more explicitly account for household-level variation.

The operational building data presented in this paper has been collected from six office rooms located in an office building (research and educational purposes) located on the main campus of Aalborg University in Denmark. The dataset consists of measurements of occupancy, indoor environmental quality, room-level and system-level heating, ventilation and lighting operation at a 5-minute resolution. The indoor environmental quality and building system data were collected from the building management system. The occupancy level in each monitored room is established from the computer vision-based analysis of wall-mounted camera footage of each office. The number of people present in the room is estimated using the YOLOv5s image recognition algorithm. The present dataset can be used for occupancy analysis, indoor environmental quality investigations, machine learning, and model predictive control.

Buildings can deliver short-term thermal energy storage by utilising the thermal capacity of the building construction and/or by activating the water tanks included in the heating/cooling installation. The flexibility potential of demand management using decentralized thermal energy storage has been quantified in many theoretical modelling studies, and it is considered an essential technology for an affordable energy transition. We have investigated the drivers and barriers to the adoption of demand management in buildings in district heating and cooling systems via a Strengths, Weaknesses, Opportunities and Threats (SWOT) analysis and presented 17 elements that shape the current and future application of this concept. The results indicate that the application of the DR concept has left the theoretical studies and moved towards real-life applications. Yet, there is a lack of feasible business models and regulatory frameworks supporting the large-scale application of the concept. Utilities and their customers do not fully understand the benefits of the DR concept; therefore they are reluctant to adopt it outside of the research projects where the test environment is fully controlled and with limited impact and timeline. Therefore, the regulatory framework must be adjusted to allow DHC operators to develop new business models and DR tariffs that will incentivise the customers to deliver flexibility to the system without compromising their comfort and everyday practices and increasing energy poverty.

The ongoing digitalisation of the district heating sector, particularly the installation of smart heat meters (SHMs), is generating data with unprecedented extent and temporal resolution. This data offers potential insights into heat energy use at a large scale, supporting policymakers and district heating utility companies in transforming the building sector. Clustering is crucial for representing this wealth of data human -understandable groups, necessitating consideration of seasonality. Advancing current research in clustering SHM data, this work applies an established co -clustering approach, FunLBM, considering seasonal variation without fixed season definitions. Furthermore, to enhance the understanding of differentiating factors between clusters, the possibility to understand cluster memberships based on 26 building characteristics was analysed using classification and variable selection methods. Applying FunLBM on a large-scale hourly dataset from single-family houses revealed six well -separated energy use clusters each distributed over six -temporal clusters, which are correlated with the exterior temperature, yet not following fixed seasons. Variable selection and classification showed that building characteristics describing the building with a high level of detail are insufficient to explain cluster membership (Matthew's correlation coefficient (MCC) approximate to 0 . 3 ). By merging the energy use clusters based on profile and magnitude similarities, classification performance significantly improved (MCC approximate to 0 . 5 ). In both cases, simple and readily available building characteristics yield similar insights to detailed ones, emphasising their cost-effectiveness and practicality.

A dataset containing measurement data for six office rooms in Aalborg Denmark.All the measurements have been resampled to 5 minute resolutionThe measurements consists of: BMS data for the rooms Occupancy for the rooms (from cameras) BMS data for the AHU supplying the rooms BMS data for the Heating system supplying the rooms Changes from v1Corrections have been made to the following measurement points in the period from 2023-02-27 00.00 until 2023-7-13 at 08.00 as a miscoupling on the pressure difference sensor on the fans was found to measure a too high pressure difference. The airflows and pressure differences are therefore lower in this version. Ventilation:Fan__air_flow__exhaust Ventilation:Fan__air_flow__supply Ventilation:Fan__pressure_difference__exhaust Ventilation:Fan__pressure_difference__supply

This study explores two modeling approaches for occupancy detection at room level for residential buildings in Denmark. The aim is to assess the performance and generalizability of occupant detection models using XGBoost method trained on a rich dataset comprising indoor environmental quality (IEQ) variables and occupancy ground truth. A global approach and a room-specific approach are considered. After a thorough feature selection and importance analysis process, the occupancy detection models (ODMs) are trained and tested in a nested cross-validation schema. The time of the day, indoor CO2 concentration, and feature transforms related to short-term IEQ dynamics were found to be the most important features of the ODMs. Both the global and room-specific models show good occupancy prediction performance, especially for bedrooms. When tested for generalizability with an unseen dataset from a different residential building, the ODMs maintain very good performance for the bedroom but not for the office room. This discrepancy could be explained by significant differences in occupancy and ventilation patterns, and large air infiltration from adjacent rooms. Although currently limited in terms of generalizability, XGBoost-based ODMs using IEQ data have the potential to provide robust and scalable occupancy detection for occupant-centric control and occupant-aware building performance assessments. The IEQ dataset with occupancy ground truth collected for this study is made available in open access.

This paper develops a computationally fast algorithm to disaggregate the total energy use recorded by smart heat meters into space heating (SH) and domestic hot water (DHW). The algorithm trains a regression model on SH-only hours and predicts SH for all other hours with potential DHW use. Data from smart water meters were used to identify hours with only SH. Assessing 13 regression models, an untuned random forest was identified as the best-performing regression model with an acceptable computational cost (median: 7.4s per building). Furthermore, using one year of data from over 2400 single-family homes, it was shown that the best result (median CVRMSE: 0.13) can be obtained using only smart heat meters based regressors, increasing the applicability of the developed method. Furthermore, two tests were implemented, and it was successfully demonstrated that they identify situations where the regressors are distributed differently for hours with SH only and hours with SH and DHW, to identify possible training bias; thus, buildings where the disaggregation is potentially unreliable. Validation against data from three single-family houses with known SH and DHW use confirmed the good performance of the method and that the performance can be estimated without ground truth data via nested cross-validation.

Comparing, validating and assessing the accuracy of dynamic models is crucial for multiple applications in the field of energy, buildings and indoor environment modelling. Various comparison metrics and key performance indicators have been developed or adopted from other research fields, with popular guidelines recommending some for building energy models.This article gives an overview of the metrics used by this research community based on a large-scale review of scientific publications over the last 40 years. Furthermore, main trends are discussed, gaps are revealed, and future research activities are suggested. These research activities have been started by a multi-institutional group of researchers and should greatly benefit the entire building energy simulations community.

The data presented were sourced from 34,884 commercial smart heat meters and 10,765 commercial smart water meters, spanning a timeframe of up to 5 years (2018–2022). All data primarily originated from single-family houses in Aalborg Municipality, Denmark. Furthermore, comprehensive building characteristics were collected for each building, where available, from the Danish Building and Dwelling Register (BBR) and Energy Performance Certificate (EPC) input data. This effort yielded an extensive pool of up to 86 distinct characteristics per building. All smart meter data were processed employing a well-established methodology, resulting in equidistant hourly data without any erroneous or missing values. The building characteristics derived from the EPCs were additionally filtered using rule sets to improve the data quality. This dataset holds substantial value for researchers involved in the domains of the built environment, district heating, and water sectors.

Buildings can deliver short-term thermal energy storage by utilising the thermal capacity of the building construction and/or by activating the water tanks included in the heating/cooling installation. The flexibility potential of demand management using decentralised thermal energy storage has been quantified in many theoretical modelling studies, and it is considered an essential technology for an affordable energy transition. However, adoption in practical applications is very limited to date. There is a lack of feasible business models and regulatory frameworks supporting the large-scale application of the concept. Utilities and their customers do not understand the concept and therefore are reluctant to adopt it. We have investigated the drivers and barriers to the adoption of demand management in buildings in district heating and cooling systems via a SWOT analysis and presented the work of IEA EBC Annex 84 “Demand management of buildings in thermal networks” as a project aiming to deliver knowledge to overcome the weakness and threats identified in SWOT analysis.

Recent research has demonstrated the fundamental potential of smart heat meter (SHM) data. However, it has also been shown that the usability of the data is reduced because SHM energy measurements are commonly rounded down (truncated) to kilowatt-hour values. This study therefore investigates, for the first time, the error introduced by truncation using a high-resolution dataset. Furthermore, a method is developed to reduce the loss of information in the truncated data by combining smoothing with a ruleset and scaling approach (SMPS). SMPS is shown to increase the pointwise accuracy and correlation of the truncated data with the full-resolution data.

This technical report describes the evaluation process of various machine learning algorithms' performance used for supervised binary classification for occupant detection, using a dataset from a residential building in the North of Denmark.

With the aim to offer an alternative renovation concept for schools and thus contribute to a better indoor environmental quality and reduced energy consumption, a system combining diffuse ceiling and double-skin facade with an existing exhaust ventilation (I-DIFFER) was proposed in a previous work by the authors. The initial analyses of the novel system were promising, showing a potential 11 % reduction in primary energy consumption compared to a traditional renovation where the facade was insulated, windows were replaced and a balanced ventilation system was installed. Consequently, this work further investigates the performance of I-DIFFER under different boundary conditions using BPS. The influence of orientation, thermal mass and reflectance of the existing classroom facade, future climate change, extreme weather conditions, and varying occupant densities are studied. The results show that I-DIFFER can compete with the traditional renovation approach for all investigated orientations but north and leads to superior results for southern orientations (SE, S, SW). It was found that a high thermal mass facade with low reflectance is favourable for the classroom facade. For forecasted future climate conditions and climates with mild winters and mild to hot summers, I-DIFFER showed superior results compared to the traditional renovation. An equal performance was seen for a varying occupant density. With this study, I-DIFFER can be confirmed as a competitive alternative to a traditional renovation and thus contributes to improving not only the energy efficiency but also the IEQ of schools.

The correct evaluation of the performance of models used in the field of energy, building and indoor environment modelling is crucial to correctly assess the reliability of the results and the suitability of the model for the purpose. This technical report is supplementary material to the work of Johra et al., 2023, who conducted an extensive literature review of 259 papers to provide an overview of the evaluation metrics used by the energy, building and indoor environment research community. The information gathered from the 259 reviewed papers is compiled in the spreadsheet attached to that technical report. This technical report provides an overview of all the time series comparison metrics found for building energy and indoor environment modelling validation, using a consistent notation and naming convention and any alternative names for the respective metric. Such an overview should provide valuable guidance to both practitioners and researchers within the energy, buildings and indoor environment community. Furthermore, the use of a consistent naming and equation notation should reduce the possibility of misunderstanding, which has been highlighted in Johra et al., 2023. In addition to this overview, Section 3 discusses possible limitations and pitfalls when evaluating models within the field of energy, building and indoor environment modelling, which provide additional support.

With the aim to increase today's low retrofitting rate and thus decrease the currently rising energy consumption of the building sector, this work proposes a novel renovation concept for schools. Therefore, two components, double skin facade (DSF) and diffuse ceiling ventilation (DCV), are combined with an exhaust ventilation system, providing an alternative to a traditional renovation where windows are replaced, the external wall is insulated and a balanced ventilation system is installed. Further, based on a Danish case study classroom, different configurations of the novel system, with varying glazing properties and DSF cavity thicknesses are compared against a range of traditional renovations through building performance simulation using IDA-ICE. The results indicate that the new system can achieve up to 11% lower total primary energy consumption as the best traditional renovation while achieving an equal indoor environmental quality and therefore offering designers and engineers a competitive alternative for school renovation projects. Additionally, employing a sensitivity analysis, it could be demonstrated for the primary energy consumption and the global thermal comfort that the glazing (number of panes and physical properties) has a superior influence on the novel system compared to the cavity thickness.

The now widespread use of smart heat meters for buildings connected to district heating networks generates data at an unknown extent and temporal resolution. This data encompasses information that enables new data-driven approaches in the building sector. Real-life data of sufficient size and quality are necessary to facilitate the development of such methods, as subsequent analyses typically require a complete equidistant dataset without missing or erroneous values. Thus, this work presents three years (2018-01-03 till 2020-12-31) of screened, interpolated, and imputed data from 3,021 commercial smart heat meters installed in Danish residential buildings. The screening aimed to detect data from not used meters, resolve issues caused by the data storage process and identify erroneous values. Linear interpolation was used to obtain equidistant data. After the screening, 0.3% of the data were missing, which were imputed using a weighted moving average based on a systematic comparison of nine different imputation methods. The original and processed data are published together with the code for data processing ( https://doi.org/10.5281/zenodo.6563114 ).