Modular construction allows for a faster, safer, better controlled, and more productive construction process, yielding quality results with low risk and controlled costs. However, despite the potential advantages of this methodology, its adoption has remained slow due to the reasonably high degree of standardisation and repetition that projects require, inexorably clashing with the unique building designs created to meet the clients’ needs. The present article proposes performing a modularisation process after the building design is complete, reaping most benefits of modular construction while preserving the unique vision and design of the building. This objective is achieved by implementing a semi-supervised methodology reliant on the clustering of individual rooms and subsequent user validation of the obtained clusters to identify base modules representative of each cluster. The proposed methodology is applied in a case study of an existing apartment complex, in which the modularisation process was previously performed manually—thus serving as a baseline. The acquired results display a 99.6% reduction in the modularisation process’ duration, while maintaining a 96.4% Normalised Mutual Information Score and a 93.3% Adjusted Mutual Information Score, justifying the continuous development and assessment of the methodology in future works.

Self-Reconfigurable Robots have shown great versatility and promise for building dynamic and self-adapting structures of modular robots. Unfortunately, the structural requirements for building structurally-sound robotic structures of modular robots at the architectural scale, where structural performance and stability of the assembly are crucial for success, have yet to be properly addressed.

This thesis addresses these requirements and proposes a structurally-driven control strategy for a self-reconfigurable robotic system based on the structural analysis and performance of not only the final target configuration of a robotic assembly but also of the intermediate transitional configurations achieved during self-reconfiguration. To formulate a structurally feasible target shape, a topology optimization is used to evolve the target shape based specific boundary and loading conditions and to maximize structural stiffness. Thereafter, the control strategy drives the modules’ decision-making process using three fitness criteria for action selection: modules’ convergence towards the given target configuration, stability of the overall assembly, and the structural performance of the assembly. While the proposed control strategy succeeds in filtering out unstable and structurally unsafe configurations, corrective measures fail in completely dealing with a declining structural performance. Nevertheless, this thesis exposes some of the difficulties in using local decision-making to solve global structural issues and extends the state-of-art on structurally-aware self-reconfigurable robots.

Algorithmic Design (AD) is a programming-based approach to design where, instead of creating a model of the intended design, the designer creates a program that generates the model of the intended design. This approach provides many opportunities for innovation and improvement in the design process. However, despite having been extensively explored with geometry-based Computer-Aided Design (CAD) tools over the past years, AD has only recently started to be explored with the Building Information Modelling (BIM) methodology brought by BIM tools. BIM brings substancial gains to the design activity and, as a result, has started to become mandatory in architectural practices all over the world. By combining algorithmic processes with the BIM methodology, a new approach to design emerges, one that we designate Algorithmic-based Building Information Modelling (A-BIM).

In this thesis, we define, explore and evaluate A-BIM in the context of architectural design. Through a case study, we compare A-BIM to two other design approaches, namely an algorithmic approach to geometry-based CAD and a manual BIM approach, and we show that A-BIM can offer great benefits to architectural practices.

Apesar da construção modular permitir aumentar significativamente a produtividade na construção, a sua adoção permanece lenta devido ao elevado grau de repetição e normalização necessário para a sua implementação, tipicamente pouco implementa‑do pelos projetistas e requisitado pelos clientes. Para abordar este problema, o pre‑sente artigo propõe uma metodologia de modularização a ser implementada após a fase de design estar concluída, aproveitando as vantagens da construção modular enquanto mantem a visão e o design original do edifício. Esta metodologia é basea‑da numa metodologia de semi‑supervised clustering para agrupar divisões individuais em grupos conforme a sua semelhança, identificando, posteriormente, um módulo base para representar cada cluster. A metodologia proposta é aplicada num caso de estudo no qual foi realizado previamente um processo de modularização manual, servindo como ponto de referência e comparação. Os resultados mostram uma re‑dução de 99,6% do tempo necessário para executar o processo de modularização, justificando o desenvolvimento continuado desta metodologia em trabalhos futuros.

The lack of interoperability and the diversity of required documentation in the development of architectural projects often results in inefficient design processes. Integrated design approaches such as Building Information Modeling seek to tackle this problem, but still require strenuous and time-consuming manual work when it comes to design exploration and the implementation of design changes. Algorithmic design approaches facilitate this process by supporting quick change propagation and exploration of design variations, as well as automating the production of the required documentation. This paper presents an integrated algorithmic design workflow, encompassing all design stages, from conceptual design to fabrication. The workflow is tested throughout the design, analysis, visualization, and fabrication of a classroom renovation project, resulting in a more fluid and efficient design process.

Os processos de Design Algorítmico têm um enorme potencial, pouco explorado pela grande maioria dos gabinetes de arquitetura de pequena dimensão: na produtividade, na redução de custo/tempo ou na liberdade experimental. Para esse fim, é necessário combinar os processos de projeto arquitetónico BIM com os de raiz algorítmica.

Este artigo apresenta dois casos práticos resultantes dessa combinação. O primeiro centrou-se no controlo da incidência solar na fachada duma habitação isolada, tendo em vista o seu conforto luminíco. O segundo focou-se na integração de equipamento em unidades de quartos de hotel, como método de remodelação e reutilização de dois edifícios centenários. A partir da descrição destes dois casos práticos, são discutidos e analisados os obstáculos encontrados ao longo do processo, as vantagens e desvantagens do mesmo, aslições aprendidas, e a implementação e avaliação de soluções de protótipo pela equipa de arquitetos, de modo a fomentar a sua aplicação futura na prática arquitetónica.

Algorithmic design processes have enormous potential for architecture. Even though some large design offices have already incorporated such processes in their workflow, so far, these have not been seriously considered by the large majority of traditional small-scale studios. Nevertheless, as the integration of algorithmic techniques inside architectural studios does not require mastering programming skills, but rather taking advantage of a collaborative design process, small design studios are therefore able of using such strategies within their workflow. This paper discusses a series of challenges presented by one of these studios, where we had to integrate algorithmic design processes with the studio’s traditional workflow.

Algorithmic-based Building Information Modelling (ABIM) allows the development of BIM models through algorithms. In a collaborative environment, A-BIM requires management strategies to deal with concurrent development of architectural projects. However, despite there being several tools that support this type of collaborative work, they are not appropriate for A-BIM because: (1) they track changes in the generated model instead of the code where the changes originate from, and (2) they are vendor-specific while A-BIM models might be generated for different BIM applications. In this paper, we discuss the use of Version Control (VC) for project management and concurrent development of A-BIM projects. We evaluate VC for A-BIM through a series of scenarios in the context of a case study.

O Projeto Algorítmico (PA) é uma abordagem de design baseada em programação em que, ao invés de ser criado o modelo 3D do design pretendido, é criado um programa que gera o modelo 3D do design pretendido. Esta abordagem permite trazer grandes vantagens a todo o processo de design e de construção, incluindo, por exemplo, a produção de geometrias complexas e a automatização de tarefas e processos repetitivos que seriam muito morosos na modelação manual. Graças à sua natureza parametrizável, esta abordagem permite também a fácil e rápida exploração de alternativas ao modelo inicial. Apesar destas vantagens, o PA tem sido pouco explorado em combinação com a metodologia Building Information Modelling (BIM). Ao combinar processos algorítmicos com a metodologia BIM surge uma nova abordagem de design baseada no desenvolvimento de programas que geram modelos BIM do design pretendido. A esta abordagem designamos Algorithmic-based Building Information Modelling (A-BIM).

Neste paper, definimos, exploramos e avaliamos o A-BIM num contexto arquitetónico. A partir de um caso de estudo, comparamos o A-BIM a uma abordagem BIM manual, e demonstramos que o A-BIM pode oferecer grandes vantagens à prática arquitetónica.

Generative Design (GD) is a valuable asset for architecture because it provides opportunities for innovation and improvement in the design process. Despite its availability for Computer-Aided Design (CAD), there are few applications of GD within the Building Information Modelling (BIM) paradigm, and those that exist suffer from portability issues. A portable program is one that will not only work in the application it was originally written for, but also in others with equivalent results. This paper proposes a solution that explores portable GD in the context of BIM. We also propose a set of guidelines for a programming methodology for GD, adapted to the BIM paradigm. In the end, we evaluate our solution using a practical example.