Unsupervised learning of energy signatures to identify the heating system and building type using smart meter data

Westermann, Paul; Deb, Chirag; Schlueter, Arno; Evins, Ralph

doi:10.1016/j.apenergy.2020.114715

Cited by 67 publications

(28 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 1. Additional information for the Italian building typology matrix (MDc: mass distribution class, according to Reference [37], Table 1; SHCc: specific heat capacity class, according to Reference [34], Table A. 14). Subscripts "rf", "ew" and "gf" indicate roof, external walls and ground floor elements, respectively.…”

Section: Climate-related Input Datamentioning

confidence: 99%

“…Noussan and Nastasi [12] analyze data from the regional registry of heating plants in Lombardy, Italy, while Mancini and Nastasi [13] adopt a survey approach to collect information about building features, technical services and appliances. Westermann et al [14] develop a method to automatically infer building type and heating system type from the energy signature, expressed as the electricity consumption for heating and/or cooling detected by smart meters as function of outside temperature recordings.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic Approach to Evaluate the Effect of Reducing District Heating Temperature on Indoor Thermal Comfort

et al. 2020

View full text Add to dashboard Cite

To reduce energy consumption for space heating, a coordinated action on energy supply, building fabric and occupant behavior is required to realize sustainable improvements. A reduction in district heating supply temperature is an interesting option to allow the incorporation of renewable energy sources and reduce distribution losses, but its impact on the final users must be considered. This aspect is especially critical as most European countries feature an old building stock, with poor insulation and heating systems designed for high-temperature operation. In this study, a complete methodology is devised to evaluate the effect of district heating temperature reduction on the end users by modeling all the stages of the system, from the primary heat exchanger to the indoor environment. A dynamic energy performance engine, based on EN ISO 52016-1:2017 standard and completed with a heat exchanger model, is implemented, and its outputs are used to calculate thermal comfort indicators throughout the heating season. As a practical application, the method is used to evaluate different scenarios resulting from the reduction of primary supply temperature of a second-generation district heating network in Northern Italy. Several building typologies dating back to different periods are considered, in the conservative assumption of radiator heating. The results of the simulations show that the most severe discomfort situations are experienced in buildings built before 1990, but in recent buildings the amount of discomfort occurrences can be high because of the poor output of radiators when working at very low temperatures. Among the possible measures that could help the transition, actions on the primary side, on the installed power and on the building fabric are considered. The investigation method requires a limited amount of input data and is applicable to different scales, from the individual building to entire urban areas lined up for renovation.

show abstract

Section: Climate-related Input Datamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Dynamic Approach to Evaluate the Effect of Reducing District Heating Temperature on Indoor Thermal Comfort

et al. 2020

View full text Add to dashboard Cite

show abstract

“…They are based on energy interval data (dependent variable) and weather data (independent variables), together with other independent variables that may be derived from contextual information. External air temperature is the most important independent variable, used for weather normalization of energy consumption (Masuda and Claridge, 2014;Lin and Claridge, 2015;Westermann et al, 2020). Additionally, rather than using energy data directly, we can transform them to derive the average power, called energy signature (ISO, 2013), over the amount of operating hours in the time interval considered.…”

Section: Regression Models In Operational Phase Analysismentioning

confidence: 99%

Linking Design and Operation Phase Energy Performance Analysis Through Regression-Based Approaches

2020

View full text Add to dashboard Cite

The reduction of energy usage and environmental impact of the built environment and construction industry is crucial for sustainability on a global scale. We are working towards an increased commitment towards resource efficiency in the built environment and to the growth of innovative businesses following circular economy principles. The conceptualization of change is a relevant part of energy and sustainability transitions research, which is aimed at enabling radical shifts compatible with societal functions. In this framework, building performance has to be considered in a whole life cycle perspective because buildings are long-term assets. In a life cycle perspective, both operational and embodied energy and carbon emissions have to be considered for appropriate comparability and decision-making. The application of sustainability assessments of products and practices in the built environment is itself a critical and debatable issue. For this reason, the way energy consumption data are measured, processed, and reported has to be progressively standardized in order to enable transparency and consistency of methods at multiple scales (from single buildings up to building stock) and levels of analysis (from individual components up to systems), ideally complementing ongoing research initiatives that use open science principles in energy research. In this paper, we analyse the topic of linking design and operation phase’s energy performance analysis through regression-based approaches in buildings, highlighting the hierarchical nature of building energy modelling data. The goal of this research is to review the current state of the art of in order to orient future efforts towards integrated data analysis workflows, from design to operation. In this sense, we show how data analysis techniques can be used to evaluate the impact of both technical and human factors. Finally, we indicate how approximated physical interpretation of regression models can help in developing data-driven models that could enhance the possibility of learning from feedback and reconstructing building stock data at multiple levels.

show abstract

“…Unsupervised ML methods are applied in the analysis of performance and control of buildings [45]. Cluster analysis is suitable for data preprocessing and often combined with supervised learning [46,47,48]. Reinforcement learning is a promising area for control, but its practical application is still limited [49,50,51].…”

Section: Introductionmentioning

confidence: 99%

Classification of retrofit measures for residential buildings according to the global cost

2021

View full text Add to dashboard Cite

Data-driven black-box surrogate models are widely used in research related to buildings energy efficiency. They are based on machine learning techniques, learn from available data, and act as a replacement for or an addition to complex and computationally intensive knowledge-based models. Surrogate models can predict energy demand, indoor air temperature, or occupants behavior; explore search space in optimization problems; learn control rules; etc. This paper analyzes surrogate models that classify building retrofit measures directly according to the global cost. In addition, they quantify the importance of each variable for the classification process. The models are based on random forest classifiers, which are fast and powerful ensemble learners. They can be applied to effectively reduce search spaces when optimizing energy renovation measures or to rapidly identify projects that deserve financial support. This approach is applied to two residential buildings and three scenarios of price development. The training process uses a small share of retrofit options assessed with standard calculations of the heating and cooling demands, as well as the global cost. The results show very high classification performance, even when the models are trained with small and imbalanced training sets. The obtained recall, precision, and F-score values are mostly above 95%, except for extremely small training sets.

show abstract

Unsupervised learning of energy signatures to identify the heating system and building type using smart meter data

Cited by 67 publications

References 29 publications

Dynamic Approach to Evaluate the Effect of Reducing District Heating Temperature on Indoor Thermal Comfort

Dynamic Approach to Evaluate the Effect of Reducing District Heating Temperature on Indoor Thermal Comfort

Linking Design and Operation Phase Energy Performance Analysis Through Regression-Based Approaches

Classification of retrofit measures for residential buildings according to the global cost

Contact Info

Product

Resources

About