Thesis title: Multi-objective optimization for sustainable building design. From schematic design phases to retrofit strategies optimization using genetic algorithms
The intrinsic complexity of the built environment is increasing, due to the intricate interplay of various factors related to environmental, social and economic objectives. Such complexity poses significant challenges in reaching the demanding performances required to buildings today but, at the same time, offers opportunities to innovate the building design process. This doctoral thesis explores the integration of building performance optimization (BPO) methodologies within different design stages, coupling parametric modeling, dynamic simulations and optimization algorithms. The study elaborates on how multi-objective BPO can address complexity in the design process of sustainable and low-energy buildings from the schematic design phase of a new construction up to the retrofit of an existing building. In so doing, it focuses on multi-family residential buildings, as they are a priority in energy efficiency measures at the national and international levels and play a key role in addressing current issues, such as climate change and fuel poverty.
The research takes its moves from the concept of the built environment as a complex system, emphasizing the necessity of advanced digital tools to deal with such complexity. Indeed, design decision-making phases are characterized by numerous possible combinations of passive strategies which can improve building performance without using extra energy. Such design solutions influence the often conflicting objectives that a sustainable building aims to achieve, towards high conditions of comfort and functionality, without the consequence of excessive energy consumption and emissions.
First, a systematic literature review, carried out in parallel with an online survey distributed to architectural firms, highlights, on the one hand, the great interest in BPO in current research and, on the other, the limited diffusion of both simulation and optimization tools in architectural practice.
Once misalignments between research and professional practice have been analyzed, the thesis then proposes specific methodologies, rooted in the results of the review, for the use of multi-objective optimization both for new constructions and for existing buildings, demonstrating them by applications on relevant case studies. The optimization carried out during the schematic phase of the project (thus for the case of a new construction) takes into consideration the geometric variables of the building to minimize energy consumption and maximize daylight accessibility. The proposed methodology integrates multi-disciplinary simulations and optimization, in a parametric modeling environment. Results emphasize the critical role of design choices during the early stages, and thus, the necessity of a meticulous optimization of geometric variables. Moreover, in order to analyze the optimization during an advanced and highly detailed phase of the project, the case of a retrofit of an existing building is considered, and a methodology is proposed, testing an advanced algorithm, the active-archive Non-dominated Sorting Genetic Algorithm-II (aNSGA-II). Architecturally-compatible passive retrofit strategies on the building envelope are analyzed and the optimization process is carried out to find an energy- and cost- efficient solution. Furthermore, active strategies and renewable energy are also considered, reflecting on their role in the decarbonization of the building stock and in making inhabitants less subjected to energy price fluctuations.
The thesis concludes by illustrating the answers to the research questions raised in the introduction, emphasizing multi-objective BPO's potential to address and inform design decisions from early phases, aligning with the current necessity of collaborative and multi-disciplinary design process. Therefore, the thesis demonstrates how BPO can be of support in different phases of a building design, showing the significant variables to be analyzed based on the level of detail of the design phase considered and proposing specific methodologies.
The impact of the research here presented extends beyond academia, influencing professionals, software developers, and policymakers in promoting sustainable and low-energy buildings towards the mitigation of climate change and fuel poverty.