Periodic Reporting for period 1 - ADDOPTML (ADDitively Manufactured OPTimized Structures by means of Machine Learning)
Période du rapport: 2021-05-01 au 2023-04-30
EU is a world leader in the development and manufacturing of products for the construction industry; an industry that generates an annual income of €562bn (9% in gross value) and 18M direct jobs for the European economy. Europe also aspires to become the first ever climate neutral continent by 2050 with zero net emissions and sustainable economic growth that is also decoupled from resource use. To achieve this, the recently published Green Deal for Europe1 sets to disrupt the construction sector. This is responsible for more than 40% of the overall energy consumption; it further results in an astonishing 25%-30% of all waste generated in the EU, hence being one of the heaviest and most voluminous waste streams generated. Within this setting, ‘Training’, ‘Innovation’, and ‘Energy Efficiency’ are three of the ‘Major Challenges’ for EU construction. To overcome these challenges and sustain Europe’s world-leading position, continuous technological advancements for more efficient and robust manufacturing for this high-value sector are crucial.
AM (3D printing), encompasses a range of material-processing technologies, which rapidly create physical parts directly from CAD models by sequentially depositing material in a layer-wise manner. AM components have already found numerous applications within the European aerospace and transport industry due to their well-known advantages, e.g. less machine, material, and labor costs, less manufacturing waste, and responsible raw material usage/management. Unfortunately, their application within the construction sector is long overdue; this is mainly due to the hugely interdisciplinary, complex, and inevitably fragmented design/analysis to construction workflow, which typically involves unique, custom-made structures of a significant scale and greatly deviates from the typical assembly line manufacturing paradigm. Thus, it becomes very difficult to disrupt one aspect of this workflow without affecting the rest, inevitably preventing the sector from keeping up with innovation. There is a clear societal need and a corresponding research challenge towards delivering innovative analysis, design, and construction workflows leveraging the unquestionable merits of additive manufacturing. The ADDOPTML overarching aim is to set a new manufacturing paradigm for the construction industry that will enable the rapid deployment of the resilient and adaptive structures of the future. Structures that can be rapidly tailored, manufactured and operate under diverse and dynamic environmental conditions. To achieve this, ADDOPTML sets to greatly advance the current state-of-the-art by employing a truly cross-disciplinary research methodology based on machine learning, topology optimization, structural engineering, and automated generative design. One construction type in which additive manufacturing may offer an important breakthrough is to address the huge challenges posed to humanitarian and development organizations by the escalating number of disasters, worldwide and particularly in developing countries.
Research objectives:
1) To derive a comprehensive library of data driven constitutive laws for structural materials
2) To develop a high-fidelity yet rapid data driven topology optimization environment for AM structural components
3) Fully automated design workflow for AM structures based on the generative design paradigm
4) To apply rapidly deployable steel and concrete structures custom fit for post-disaster sheltering solutions
5) To deliver design protocols for AM structural components to be deployed in space applications
AM (3D printing), encompasses a range of material-processing technologies, which rapidly create physical parts directly from CAD models by sequentially depositing material in a layer-wise manner. AM components have already found numerous applications within the European aerospace and transport industry due to their well-known advantages, e.g. less machine, material, and labor costs, less manufacturing waste, and responsible raw material usage/management. Unfortunately, their application within the construction sector is long overdue; this is mainly due to the hugely interdisciplinary, complex, and inevitably fragmented design/analysis to construction workflow, which typically involves unique, custom-made structures of a significant scale and greatly deviates from the typical assembly line manufacturing paradigm. Thus, it becomes very difficult to disrupt one aspect of this workflow without affecting the rest, inevitably preventing the sector from keeping up with innovation. There is a clear societal need and a corresponding research challenge towards delivering innovative analysis, design, and construction workflows leveraging the unquestionable merits of additive manufacturing. The ADDOPTML overarching aim is to set a new manufacturing paradigm for the construction industry that will enable the rapid deployment of the resilient and adaptive structures of the future. Structures that can be rapidly tailored, manufactured and operate under diverse and dynamic environmental conditions. To achieve this, ADDOPTML sets to greatly advance the current state-of-the-art by employing a truly cross-disciplinary research methodology based on machine learning, topology optimization, structural engineering, and automated generative design. One construction type in which additive manufacturing may offer an important breakthrough is to address the huge challenges posed to humanitarian and development organizations by the escalating number of disasters, worldwide and particularly in developing countries.
Research objectives:
1) To derive a comprehensive library of data driven constitutive laws for structural materials
2) To develop a high-fidelity yet rapid data driven topology optimization environment for AM structural components
3) Fully automated design workflow for AM structures based on the generative design paradigm
4) To apply rapidly deployable steel and concrete structures custom fit for post-disaster sheltering solutions
5) To deliver design protocols for AM structural components to be deployed in space applications
WP1 and WP2: A physics-rich topology optimization framework was developed, incorporating high-fidelity nonlinear response estimates. Machine learning techniques were integrated, based on data-driven models from WP2, calibrated through extensive material and component-level experiments. WP1 and WP2 combined to create a rapid yet precise digital platform for analyzing and optimizing large-scale structures.
WP3: The digital platform was seamlessly integrated into a unique digital design-based additive manufacturing workflow for large-scale structures. The central research question focused on developing manufacturing protocols for optimal structure topologies. This was achieved through a collaborative approach that delivered optimized designs, novel manufacturing protocols, 3D-printed model structures, and their experimental validation. The scope was to equip engineers with powerful tools for directly designing optimized, 3D-printable structures.
WP4: The goal was to develop a framework that leverages machine learning methods for prototyping using the tools developed in WPs 1-3. This framework, developed through collaboration among architects, structural engineers, 3D printing, and optimization specialists, facilitates intuitive yet guided prototyping from design drafting to 3D printing construction.
WPs5-7: Three novel demonstrators were created to benchmark scientific advancements, accelerating research impact. WP5 and WP6 focused on steel and concrete, respectively, addressing design targets, rapid manufacturability, deployment, and resilience. WP6 aimed to create safe living spaces that challenge stereotypical perceptions of gender and identities, while respecting cultural aspects. WP7 explored novel design paradigms for manufacturing optimal sub-components of space structures.
WP8 involves knowledge transfer, dissemination, and exploitation through international conferences, workshops, and student theses. WP9 is responsible for project management and coordination.
In terms of deliverables and milestones, the project has been successful. Multiple conference presentations and journal publications have been completed, enhancing open access to research outcomes. Additionally, a doctoral thesis has been concluded, with more underway, funded by ADDOPTML and expected to be completed this year.
WP3: The digital platform was seamlessly integrated into a unique digital design-based additive manufacturing workflow for large-scale structures. The central research question focused on developing manufacturing protocols for optimal structure topologies. This was achieved through a collaborative approach that delivered optimized designs, novel manufacturing protocols, 3D-printed model structures, and their experimental validation. The scope was to equip engineers with powerful tools for directly designing optimized, 3D-printable structures.
WP4: The goal was to develop a framework that leverages machine learning methods for prototyping using the tools developed in WPs 1-3. This framework, developed through collaboration among architects, structural engineers, 3D printing, and optimization specialists, facilitates intuitive yet guided prototyping from design drafting to 3D printing construction.
WPs5-7: Three novel demonstrators were created to benchmark scientific advancements, accelerating research impact. WP5 and WP6 focused on steel and concrete, respectively, addressing design targets, rapid manufacturability, deployment, and resilience. WP6 aimed to create safe living spaces that challenge stereotypical perceptions of gender and identities, while respecting cultural aspects. WP7 explored novel design paradigms for manufacturing optimal sub-components of space structures.
WP8 involves knowledge transfer, dissemination, and exploitation through international conferences, workshops, and student theses. WP9 is responsible for project management and coordination.
In terms of deliverables and milestones, the project has been successful. Multiple conference presentations and journal publications have been completed, enhancing open access to research outcomes. Additionally, a doctoral thesis has been concluded, with more underway, funded by ADDOPTML and expected to be completed this year.
The need for providing economical, sustainable, resilient and rapidly deployable structures means that construction technology suppliers are constantly looking for more efficient analysis and design experts and corresponding technologies. The fusion of Machine Learning, and engineering optimization with traditional engineering approaches resulting generative analysis and design workflows is expected to form the basis for Industry 4.0 in the construction sector. The ADDOPTML researchers will be prime candidates for developing these technologies, both in research organizations and in the industry due to the envisaged multidisciplinary and multisectoral training interactions. ADDOPTML will nurture many young researchers with a powerful technical background on developing advanced data driven constitutive modelling, generative design methods, and topology optimization methods, particularly focusing on additive manufacturing applications for the built environment.
Focal point of the ADDOPTML action is the delivery of methods and protocols for the Additive Manufacturing paradigm custom fit for the construction industry. This is a disruptive shift for an industry that largely relies on not automated, and complex design and construction workflows revolving around concrete and steel structures. Introducing additive manufacturing in the construction industry can only occur through the interactions of academics and engineers spanning the spectrum of civil and architectural design to robotics and additive manufacturing.
Focal point of the ADDOPTML action is the delivery of methods and protocols for the Additive Manufacturing paradigm custom fit for the construction industry. This is a disruptive shift for an industry that largely relies on not automated, and complex design and construction workflows revolving around concrete and steel structures. Introducing additive manufacturing in the construction industry can only occur through the interactions of academics and engineers spanning the spectrum of civil and architectural design to robotics and additive manufacturing.