A data-driven framework for energy-conscious design of building facade systems

B. Abediniangerabi,S.M. Shahandashti,A. Makhmalbaf

DOI: https://doi.org/10.1016/j.jobe.2020.101172

IF: 7.144

2020-05-01

Journal of Building Engineering

Abstract:A building façade system plays a critical role in building energy conservation since it has a large impact on the heat transfer between the outdoor and indoor environments. Considering complex interactions between building façade design parameters and building context is essential for evaluating the energy performance of building façade systems. Building energy simulations (BESs) can help designers to acquire necessary information about the energy performance of façade systems. However, many design firms do not have the capability of performance-based design due to the lack of trained personnel and technical resources, such as simulation tools and devices. Even with all the required resources, it is computationally expensive to analyze different façade design alternatives to compare their energy performance. The main objective of this paper is to develop a data-driven framework to extract hidden information and underlying structure from the thermal behavior of façade systems in a set of scenarios and provide recommendations for designers to select energy-efficient façade systems. In the proposed framework, clustering analysis was used to partition simulated data into different subsets with distinct patterns. Then, the association rule mining (ARM) technique was applied to each dataset to extract rules as recommendations for positive, negative, and neutral energy saving in favor of a selected façade panel. This framework provides simple recommendations on the energy performance of façade systems for different building types in different climate zones. The applicability of the proposed framework was tested on an innovative ultra-high-performance fiber-reinforced-concrete (UHP-FRC) façade panel. The extracted recommendations were validated using four different building models in different locations, which were not previously included in the building energy simulations. Results highlight the applicability of the proposed methodology in facilitating the energy performance analysis of different façade systems in different building contexts. This methodology can be used by designers as a decision support system to provide simple recommendations in the early stages of building envelope designs to obtain high-performance buildings.

construction & building technology,engineering, civil