A Comprehensive Optimization Approach for Modular Facades: The Case of PULSE Sunshading

Ioannis Chatzikonstantinou; Michela Turrin; Cemre Cubukcuoglu; Ayca Kirimtat; Sevil Sariyildiz

Ioannis Chatzikonstantinou Department of Architectural Engineering + Technology, Faculty of Architecture and the Built Environment, Delft University of Technology https://orcid.org/0000-0002-8282-928X
Michela Turrin Department of Architectural Engineering + Technology, Faculty of Architecture and the Built Environment, Delft University of Technology
Cemre Cubukcuoglu Department of Architectural Engineering + Technology, Faculty of Architecture and the Built Environment, Delft University of Technology
Ayca Kirimtat Faculty of Informatics and Management, Center for Basic and Applied Research, University of Hradec Kralove
Sevil Sariyildiz Department of Architectural Engineering + Technology, Faculty of Architecture and the Built Environment, Delft University of Technology

Keywords: computational intelligence-based decision support, Modular Façade, Optimization Approach

Abstract

It is well established in literature that computational intelligence-based decision support methodologies are able to efficiently support design decisions in complexity-defined environments. In this paper, we follow a case-based research methodology to develop and propose a comprehensive decision support system for treating complex design problems pertaining to large-scale building facades. The overall approach is based on the well-known combination of geometry generation via parametric models, performance assessments via numeric calculations and/or simulations, optimization via search algorithms. The specific challenge of this case focuses on the high-dimensionality of the decision space, which makes the computational process difficult to manage. The proposed method tackles this challenge in the parametric model and in the performance assessment. Specifically, the proposed method is initialized with parametric design techniques that enable managing the excessive amount of decision variables associated with complex facade design projects. A parallel computation scheme and simulation parameter estimation study enables carrying out the complex computations necessary to establish design performance. A multi-objective evolutionary algorithm is then used to identify non-dominated solutions to the facade design problem. The performance of the proposed method is shown and validated in an application to a real- world design case study. The case study is a facade design problem for a new building in the university campus of Delft University of Technology. The facade features a striking large glass area protected by an innovative shading elements composed by hundreds of double-curved elements, each one different from the others. Its design includes inputs from research groups and design professionals collaborating in an extensive research effort, which parametric models and performance assessments had to incorporate. The proposed decision making approach has as a result a range of promising facade configurations in accordance to multi-objective optimization premise that are effective in negotiating indoor daylight conditions, while being economically feasible to the extent allowed by the design space. The paper closes by discussing potential limitations of the proposed method, as well as extensions to deal with the latter.