Towards Self‑Sufficient High‑Rises Performance Optimisation using Artificial Intelligence

Abstract

Population growth and urbanisation trends bring many consequences related to the increase in global energy consumption and CO<inf>2</inf> emissions and decrease in arable land per person. Alternative design proposals for sustainable living are on the agenda of researchers and professionals to respond to the needs of the 21st century for a sustainable future. Since the early examples in the 19th century the high-rises have been one of the inevitable buildings of metropolises to provide extra floor space in compact cities. Based on the facts of the 21st century high-rise buildings should fulfil more than provide extra floor space in the limited urban plot. This research suggests “self-sufficient high-rise buildings” that can generate and efficiently consume vital resources in addition to dense habitation for sustainable living. Optimisation of highrise buildings has been the focus of researchers because of significant performance enhancement mainly in energy consumption and generation. However optimisation of self-sufficient high-rise buildings requires the integration of multiple performance aspects related to the vital resources of human beings (e.g. energy food and water) and consideration of large numbers of design parameters related to these multiple performance aspects. Hence the complexity of self-sufficient high-rise buildings is more challenging than optimising regular high-rises that have not been addressed in the literature. The purpose of this dissertation is to present a framework for performance optimisation of self-sufficient high-rise buildings using artificial intelligence focusing on the conceptual phase of the design process. Chapter 1 is the introduction to the dissertation. The necessity and the definition of the self-sufficient high-rise buildings are explained after presenting recent proposals of scholars and professionals related to sustainable living alternatives. Additionally the complexity level of self-sufficiency which consists of four categories as scale period parameters and performance is described by indicating the focus in the overall chart. Until now high-rise buildings have been optimised to improve the energy performance that reflects self-sufficiency only in energy consumption. The contribution of this study which focuses on optimising high-rise buildings for multiple resources (e.g. energy food and water) to decrease their environmental impact is described. The research method consists of four main steps: literature review tool development and pilot study computational method development and case study. After presenting the research problem questions aim objectives and output of the dissertation the research method explains the abovementioned steps. Finally the chapter is concluded by discussing the social and scientific relevance of the research. Chapter 2 presents the literature review on optimising form-finding parameters in performative computational architecture that entails form generation performance evaluation and optimisation. A systematic review is conducted based on multiple databases to elaborate the trends for investigating well-performing design alternatives using optimisation algorithms in the architectural design domain. Therefore the review focuses on studies involving form-finding parameters. One hundred studies are systematically reviewed focusing on swarm and evolutionary optimisation algorithms frequently used in architectural design. The chapter concludes by presenting the gaps and needs considered while developing the optimisation tool and computational framework focusing on form-finding parameters performance evaluation and optimisation applications. Chapter 3 presents the development of the optimisation tool called Optimus and the pilot study to test the efficiency of the multi-zone optimisation approach in high-rises. Part A of Chapter 3 presents the Optimus tool which considers a self-adaptive ensemble evolutionary algorithm that can cope with large numbers of design parameters. Tests 1 and 2 are presented to indicate the relevance of the developed tool based on 30-dimensional Congress on Evolutionary Computation 2005 benchmark problems and a 70-dimensional design problem. Part B explains Test 3 to utilise the efficiency of the multi-zone optimisation approach. The main idea of this method is to divide the building into several subdivisions (zones) to be considered different optimisation problems. The pilot high-rise model considers one of the most used façade parameters reported in Chapter 2 (overhang length) and glazing type for two conflicting daylight metrics predicted by the basic version of artificial neural network models and optimised by the initial version of Optimus tool. Chapter 4 presents the multi-zone optimisation (MUZO) methodology that entails the parametric high-rise model machine learning for surrogate models computational optimisation and decision-making. Part A of this chapter presents the entire methodology and two design scenarios indicated as Tests 4 and 5 to demonstrate the relevance of the MUZO. Both scenarios focusing on quad-grid and diagrid façade designs integrate frequently used form-finding parameters for building shape and façade design reported in Chapter 2. Additionally Part A conducts the machine learning results using the parametric high-rise models to cope with the computationally expensive simulation time while assessing the performance of the entire building. Afterwards Part B presents the optimisation problems and results of both design scenarios using the predictive models developed in Part A and the released version of the Optimus tool presented in Chapter 3. Since the study focuses on optimising the entire design of the high-rise scenarios are considered 260 and 220 design parameters respectively for quad-grid and diagrid scenarios. Consequently Part B presents the relevance of the MUZO methodology by comparing the results with the regular high-rise scenarios which use the same design parameters in the entire building. Chapter 5 investigates utilising the MUZO methodology and Optimus tool to optimise the Europoint complex in Rotterdam the Netherlands for self-sufficiency in terms of energy consumption and food production. The sufficiency in food production is demonstrated for lettuce crops grown in vertical farms. Building-integrated photovoltaic panels are used in several building parts regarding sufficiency in energy. The optimisation problem which involves 117 decision variables related to the façade design and the thermal properties of the glazing addresses the self-sufficiency at the building scale in detail. Moreover another optimisation problem reports the potentials at the neighbourhood scale using the same self-sufficiency aspects and design parameters. Among 13 algorithms used to optimise both problems the Optimus tool presented the most favourable self-sufficiency performance. Chapter 6 concludes the dissertation by summarising the contribution of the research addressing the answers to research questions presenting the limitations of the research and highlighting future recommendations. After completing the development of the optimisation tool and conducting preliminary results of the pilot high-rise model the research results are conducted in weeks instead of years during the development of the MUZO methodology and case study. Thanks to artificial intelligence decision-makers can utilise the proposed computational framework for optimising self-sufficient high-rise buildings. In this way consequences of the decisions on performance aspects of self-sufficiency become possible for such a complex design task with high awareness of the alternatives in search space within a reasonable timeframe. © 2024 Elsevier B.V. All rights reserved.