Multi-Resolution in architectural design and robotic fabrication

Carnegie Mellon University

Masters of Advanced Architectural Design Thesis Research 2017-2018

Manuel Rodriguez Ladron de Guevara

Under the supervision of: Prof. Jeremy Ficca, Prof. Josh Bard, Prof. Daniel Cardoso, Prof. Mary Lou Arscott

Computational algorithm proficiency for the agent based design and robotic fabrication of the informed object through different resolutions. Research targets efficiently bespoke architecture responsive to divergent economies.

September 9, 2017 / Pittsburgh

Abstract   The masters’ thesis proposal aims to further develop the concept of Arxel, and its implementation in construction technology. Arxel is a digital architecture unit that may provide insight in the construction of digital architectures through the lens of genotype and phenotype but particularly through the lens of economy, fundamental in providing sustainable alternatives.

Advanced design and fabrication methods are immersed in the architectural shift and are aware about controlling different economies such as material, environment and functionality. However, architects are still far away from being computationally efficient. This is an imperative fact inherent to the system, therefore the target is not about eradicating inefficiencies but giving them a particular utility.

Unresolved gaps arise when design is materialized. This yields a potential loss of information and unexpected results that require an unforeseen post edition development. This common process can be rethought towards a higher level of efficiency, understanding the amount of information that is exclusively required during the whole operation of design and construction.

Targeting these misinformed pieces, the design strategy is reverted and a holistic process flow designed and implemented for additive manufacturing of multiresolution plastic surfaces suitable for use as architectural components is described. This holistic process flow begins with a geometrical input and includes topology optimization, finite element analysis verification, additive manufacturing and physical material testing. The whole procedure intends to implement bespoke design and fabrication in emanating markets that may bring new opportunities of appraising constructive requirements of an even more digital society.

Keywords

Arxel, resolution, additive manufacturing, multi-resolution, robotic fabrication, holistic process, agent based design, digital design, computational morphogenesis, feedback process.

Proposal and Objectives

Research focuses on the generation of a tool that introduces resolution at early design stage, which will inform the rest of the design by scrutinizing different types of data responding to some initial inputs. These inputs or demands react to the needs of the context, making the geometry as responsive as possible.

Among the different objectives are: development new software tool that help design according to strict necessities and testing whether or not it could be practical in real world. Development of a tool for the end of the arm robot with the ability to extrude mono-material /multi-material. Testing of materials such as plastic or/and resins. Fabrication and comparison of same object tested in different contexts.

Justification

Provide bespoke, efficient, optimized, digital design and fabrication alternatives in emergent markets may bring new opportunities of assessing the constructive requirements of an even more digital society. To research towards widening the additive manufacturing construction realm by generating larger scale construction solutions of a wider and more specific economy solution space may prove useful in approaching a more sustainable and efficient construction of digital architecture.

Methods

This section contains a brief description of the different stages of the Research. The holistic process flow will mainly have two major parts. Software design and testing, and Fabrication. I will focus most of the time during Fall Semester on the first half of the system, i.e. candidate part selection, inputs requirement, computational morphogenesis design, topology optimization for additive manufacturing and finite element analysis.  

The target of this phase is to generate a viable software that engenders the information needed for the fabrication part. Due to the closed cycle and continuous feedback between the design stage and fabrication, a decent amount of testing geometry through 3d printing will be necessary.

During Spring Term, I will address fabrication issues, such as material adjustments, robotic components, tooling design and development, workflow process review and material verification. This stage will require an important economic support, as the tool prototyping will be composed by some pneumatic/mechanic pressurized system, a canister for material containment and delivery, and a nozzle, which will have crucial determination in the final quality of fabrication. Material will be tested and the object is intended to be “furniture size’ which will require a decent amount of plastic / resin / concrete.

Design Structure

There are a decent array of variables in this stage and I would like to mention them all. It is possible that some of them are not necessary and on the other hand, there are missing ones too.

  1. Minimum unit, gen or arxel

The main element of the algorithm is the arxel or resolving power. This unit will house different layers of information, such as:

A.1 - Material quantity

A.2 - Material characteristics i.e. controlled anisotropy

A.3 - Thermal and visual properties

A.4 - Weight

A.5 - Size

A.6 - Shape and limit condition

The last one is crucial for the rest of the process. Depending on the limit condition the growing pattern will be one or another. Open to discussion.

  1.   Material  Research

This will inform the fabrication and the design process. PLA can be the option. Regarding the fabrication method, it could be either fused deposition, extruding the material with the robot arm, or just assembly of 3d printed pieces.

  1.   Material system and computation: Evaluation of structural performance

This is a crucial constituent of the generative computational framework. Recurring evaluation cycles that expose the system to embedded analysis tools. Analysis will play a critical role during the entire morphogenetic process, not only establishing and assessing fitness criteria related to structural capacity, but also in revealing the system’s material and geometric behavioral tendencies. In parallel to the growing factors the continual structural evaluation informs the development process or even directly interacts with the generation of the system’s morphology through processes of evolutionary structural optimization.

  1.  Material system and computation: Growth and evolutionary processes

Here is the difference between a top-down design and a bottom-up. The growth of individual instance and the evolution of the system across generations of populations of individual instances. Here we encounter two critical factors: on the one hand, the internal dataset or growth rules (see below), the genotype, and on the other hand the variable gestalt that results from the interaction of the genotype with the environment, the phenotype.  

Genotype variables:

D.1 - Shape: fractals, branching, walkers, etc.

D.2 - Density

D.3 - Limit condition

  1.  Missed information

These are the missing pieces of information that we inherently lose during the fabrication process. Strengthen this concept, this Research will depart from the 100% of the entity built (unnecessary and inefficient scenario), or which is the same, 0% multiresolution. The objective is to withdraw from this worst case scenario towards a percentage level of multiresolution that is established as optimum for a particular case.

To ensure a correct optimum stage finding, the geometrical design starts by selecting and positioning a number of arxels. These are mere data-cores that will have material information, such as type of material, quantity (from this data point, how long can you extrude material with a particular volume), quality and visual properties. The location of these arxels will inform the rest of the geometry. If the topology-optimized design does not meet the requirements of the finite element analysis, the design is stepped back in the loop to reconfigure the arxel units until it passes the FEM design verification step.  

  1.  Input parameter

To prototype the design of the geometry, some inputs will be demanded to activate the engine. These inputs talk about different economies:

F.1 - Material economy (available material)

F.2 - Economic budget

F.3 - Material characteristics (color, transparency, etc.)

F.4 – Surface quality (geometrical degree, smoothness vs roughness)

F.5 - Object or targeted general shape as “bounding box”

F.6 – Dimensional accuracy

F.7 – Loading conditions

F.8 – Residual stresses and strains

F.9 – Component size

F.10 – Quantity of arxels

  1.    Fabrication

This stage will lead the Research by the beginning of next semester. Some funding will be required to design, develop and prototype tools needed for the robotic arms in dFAB of CMU, as well as material purchase.

As a way to test the research, a decent amount of 3d printing will be needed during the whole process. This stage will be crucial as a proof of concepts. The main mechanism to fabricate will be the ABB 6640 and 4400 robots that live in dFAB at Carnegie Mellon University, with some testing in regular 3d printers. The aim of this stage is to create a refined tool that satisfies the scale and manufacturing characteristics and constrains envisaged during the design stage.

  1. Evaluation/Outcomes

This Thesis is a long and complex journey through multiple stages and scales. I don’t pretend to complete it at its full capacity but to achieve a decent level of termination that makes possible to actually build something. The intention of the Research is to figure out the potential impact it has in real practice, and eventually to question whether contemporary architecture responds to the speed in what the world is evolving, with all the components it entails.

As technology evolves and every year is more digital than the precedent one, the pretension of design and build by bespoke methods, as a tailored architecture is not far away. Immersed in the architectural shift towards digitalization and fabrication methods integrated in the design stage, we still design and build responding to some general patterns that follow a market framework, constrained by a standardized process. The democratization of the fabrication is closer and this Research intends to be a proof of it.

   J – Theoretical framework

The efficiency in design and construction has been pursued since the existence of trading and the beginning of democracy. Kicking off at pre-industrial age, this term is defined as the time before industrialization. In other words, it was life before machine manufacturing and mechanization.

When speaking of economies, most activity during pre-industrial age existed at the subsistence level, in which goods are produced for the consumption and survival of one’s family group. Scheme of production by the family for the family.

With the arrival of the Industrialization and the mass-production phenomena, small handmade labor cells disappeared, as well with the societal meaning of needing small personal factories. Everything shifted out of the rural areas to a focal distribution model, where all the supplies comes from big factories. This is known as Urbanization.

This fact made big companies to take over the bespoke design in architecture, obligating people to buy standard elements, creating millions of replicas over the world. This creates capitalism, where the same product is offered regardless the personal possibilities. The inaccessible and luxury items appear in the system alongside classes and world leaders. This triggers the excess and waste. And poverty, meaning the impossibility to access to certain quality, having to live under deplorable conditions, in a wide scope, including food and basic conditions. Within this theoretical framework, and if we focus on design and architecture, this is translated towards inhabitable houses, poor quality furniture and unhealthy conditions. With the growth of the excess and waste, rebellions started to appear, and society starts to get conscious about efficiency and sustainability.  And this happens at all levels, social, commercial, economic, etc.

Towards mid to late 20th century, there is a huge progression in technology and digital technologies start gaining importance in society. Immersed in the industrialization, companies are in a complete search for the efficiency in production, which is translated into economy. In design and construction, modularization, tessellation and standardization emerge as a way to economize the production and make profits out of it. In digital fields, such as software development, cinematic and videogames industry, the polygonization of the mesh is way more affordable than the smooth surface.

As a fact to close the gap between mass production and better design standards, movements like social housing, DIY, and corporations like IKEA appear in recent history. The idea is to bring a better quality or at least, better design to medium-to-low social classes, or the affordability to the inaccessible.

With the turn of the century, architecture starts shifting to the digital era, thanks to the fast evolution in technology. Summing it up, CAD-CAM and parametric tools, alongside with FEM analysis software make possible that the production of the object go back to the hands of the designer and gets integrated in the design process. Architecture is no longer visionary-based.

Inherent to the design and production processes, the inefficient as a fact is irreversible. As solutions, architects either override it assuming is part of the whole practice, or try to fix it appearing the post-production phase (which entails over-costs and it creates a paradoxical schema of combating inefficiency with more inefficiency). However, regardless the amount of fixed elements, there are always assumed gaps. This inefficiency is present in natural systems, with the difference that in some natural processes, inefficiency has a particular utility.

Imitating nature, the hypothesis is based on these inefficiencies, by means of controlling them by inserting even more. Assuming these errors as the set base for design and production, we can achieve higher levels of efficiency from having more inefficient gaps. Now these errors have a purpose. How much of these gaps are needed to meet the input demands of an object?

By reducing the amount of information an object has, the object is being adapted to the particularity, giving to each one what is needed, eliminating mass standardization processes, going back to the small handmade factories of the past, and welcoming the Small Digital Factory (SMD). SMD reduces cost, waste, inefficiency, excess, inequality and classes. In its broader sense, it is a critic against the system of excess and capitalism we live in.

   J - Framework Schedule

SOFTWARE DEVELOPMENT

A Workflow framework                3 WEEKS

B Bounding domain                2 WEEKS

C Inputs requirement                3 WEEKS

D Minimum unit ARXEL                2 WEEKS

E Morphogenetic algorithm        4 WEEKS

F Multi-resolution hotspots         3 WEEKS

G FEA modelling design setup        3 WEEKS                                

TOOL DEVELOPMENT

A Material container development        3 WEEKS

B End of arm Robot components        1 WEEK

C Nozzle development                3 WEEKS

D Material delivery testing                2 WEEKS

E Fabrication adjustments                3 WEEKS

F Fabrication                         4 WEEKS

H Material testing                 2 WEEKS


DISCUSSION AND EVALUATION OF OUTCOMES

A Presentation and dissertation         2 WEEKS

B Publication                        2 WEEKS

  K – General Schedule

CALENDARIO DE INVESTIGACIÓN

43e0fdaa9c057173d7e77d7d_rw_1200

   

Ll-budget

For tool development phases, to implement the knowledge and design acquired during Software development phase.

Physical utilities as 3d printer, material, components for the tooling part attached to the ABB 6640 and 4400 robots in dFAB at Carnegie Mellon, and software to make it feasible will be required to have a successful research.

Without the array of tools and software the Research will be highly undermined as it is a hands-on work tight to the constant manipulation of materials through tools and fabrication.

As for the software design stage, the purchase of licenses such as Karamba software, plugin for Grasshopper, will be required to properly integrate it to the holistic process. As parallel stage, a decent array of 3d printed prototypes will be necessary to mockup and analyze upon analog feedback.  Since there are some investigation further away from the expertise of architecture, some tokens for tutorials, external consultation, collaborations, etc will be desired.

The fabrication part will involve an important display of components and materials needed to prototype, mockup and fabricate the final tools that will allow the correct materialization of the design. Pieces such as pipes, fittings, pressurize options, testing materials, electronic devices, etc. will compound the set for the fabrication tools and work cell.dd00ffdb389a08f1e02836f3_rw_1200

a1ccf036275ea9a9fcbd94b7_rw_1200

References

Menges, A., Sheil, B., Glynn R., & Skavara M. (2017). FABRICATE UCLPress.

Yuval Noah Harari, Sapiens A Brief History of Humankind.

Rifkin, J. (2014). Zero Cost Marginal Society.

Oxman, N. (2010). Material-Based Design Computation,

Negroponte, N. (1975). The architecture machine. Computer-Aided Design, 7(3), 190-195.

Negroponte, N., & Abdala, M. (1995). Being Digital.

Gibson, L. (1985). The mechanical behaviour of cancellous bone. Journal of Biomechanics, 18(5), 317-328. 

Lynn, G. (1999). Animate Form.

Hensel, M., Menges A., & Weinstock, M. (2010). Emergent Technologies and Design.

Oxman R., & Oxman, R. (2014). Theories of the digital in Architecture.

Menges, A., & Ahlquist, S. Computational Design Thinking. AD Reader

Sheil, B. (2012). Manufacturing the Bespoke. AD Reader