The Computational Study of Sand Dune Architecture - Pre-simulation for Sand Forming with Wind-Directing Robots –
Haruna Okawa Faculty of Policy management, Keio University, Fujisawa, Japan
[email protected] 5322, Endo Fujisawa-shi, Kanagawa-ken, Japan
Kensuke Hotta The graduate school of Media and governance, Keio university, Japan
[email protected] 5322, Endo Fujisawa-shi, Kanagawa-ken, Japan
Yasushi Ikeda The graduate school of Media and governance, Keio university, Japan
[email protected] 5322, Endo Fujisawa-shi, Kanagawa-ken, Japan
The Computational Study of Sand Dune Architecture - Pre-simulation for Sand Forming with Wind-Directing Robots – This paper examines the relation between computer fluid dynamics (CFD) and mass physics simulation in a desert -like space with obstacles. By using this information, a shape and its behaviour of obstacles; called wind-directing robots, are designed. Those shape as well as behaviours make sand dune much controllable in order to realize large scale additive construction method. By utilizing local materials and natural energy, in here sand and wind, as the main sources of power to form materials, the final outcome of this research enables us to increase the efficiency in future construction as well as create architectural ecosystems that consumes less energy. Keywords: word; additive manufacturing; construction automation; material computation; natural energy; CFD simulation Subject classification codes: Parallels in Advanced Fabrication: Large Scale Additive Manufacturing.
1.
Introduction Many machines are used in the construction industry and there are various ways
of processing materials. Table 1 shows how different approaches are used for moving and shaping materials in terms of construction automation. It is clear that each machine uses two mechanical systems, one for moving materials and the other for moving the machine itself.
Shovel Car
Vacuum Excavator
Sand Cutter Blower
Concrete Pump Truck
Excavator
Baler
Wheel TractorScraper
Subtraction
Scrape
Vacuum
Nab
N/A
Drill /screw
Reap
Push
Addition
Release
Impound
Blow
Extrude
N/A
Roll
Leave
Material Processed
55 m3/h
2040m3/h
2300 t/h
87 m3/h
2880m3/h
55 t/h
88.75 m3/h
Energy Consumption
522 kW
74 kW
260 kW
60 kW
597 kW
56 kW
373kW
Photo
Table 1. The comparison among representative onsite construction machines, that is used for forming particle -like materials. Below diagram (Figure 1-1) shows the efficiency of the precedent examples. In the case there is no limitation in time, calcareous cave can be the one way of creating form with the least energy, vice versa energy usage, a shovel car would be the best. Yet, these cannot serve a proper function in the context of architecture since the cost is extremely high. Therefore, a compromise has to be found.
Figure1-1. The comparison between energy consumption and speed of Onsite Construction Machines in table 1 From emergent architectural theories such as 3d printing or additive manufacturing, the trend can be seen time and accuracy. In contrast, this research intends to focus on energy consumption while sacrificing time and accuracy. Thus, this project uses airflow to move sand. In order to reduce energy consumption even further, it uses natural winds to move the sand instead of transporting it.
2.
Aim
2.1 Aim of this whole research project As the evaluation factors have already set, this project aims to maximize the efficiency between time and energy consumption. Thus, it investigates possibilities of utilizing natural energy in architectural sand forming. If natural energy can be used as a source of power to move materials, then it follows that comprehensive construction processes become more energy efficient. The final goal of this project is to develop methodology to intentionally form mountain of sand. In order to achieve this, the following process have to be completed. Figure 2-1 shows the future vision of how this methodology will be used in the real environment. Process 1
Accurate Simulation for Sand Forming and Robot’s Outer Shape
Process 2
Robot’s Design including Mechanics, Inner Circuit, and Behaviour
Process 3
Establishment of Sand Solidification Methodology with Adhesive Materials
Process 4
Consideration to Finishing for the Formed Mountain
Process 5
Physical Experiment of Entire System
Table 2. Entire research process.
Figure 2-1. The image of implementation for whole research; the aroud 3m height machine called wind-directing robots moves around on the desert automatically. These forms the higher dune mountain than natural one, that will be utilize for construction.
2.2 Aim of this paper This paper deals with the initial simulation part of the entire project, which consists of mass physics simulation. The aim of this paper is to establish simulation methodology for in situ additive manufacturing of sand using natural wind as the source of power. If it could accumulate sand vertically, the technology can be applied to create three dimensional forms. Thus, this paper uses the height as the evaluation index along with time and energy consumption. It compares the time and energy usage spent to create sand forms with a certain height.
3.
Statement of the Art From the perspective of applications of additive manufacturing technology to
architecture, there have been precedent studies that were done on exploring sand solidification on deserts. Larsson (2007) introduced scheme to use bacteria as adhesive materials and create walls to prevent desertification. Debnarova (2014) proposed to use lens to collect sunlight and heat up the sand to cause crystallization. In the former case, the drill goes under the sand to inject bacteria. As is evident from the above figures, drilling into the sand is not energy-efficient way of processing the material. In the latter case, the surface of the sand was solidified first and when the wind blew, the nonsolidified sand underneath is blown away and the left space finally becomes usable. The shape of the solidified sand that appears from this method depends on the existing dune shape meaning few freedoms for designing is left. The other strategy is to have a highly pressured inflatable dome as a support structure to print regolith on (Kestelier et al. 2014). This approach is also not efficient as the machine has to transport materials from the ground to the target area. Therefore, significance of this research is the use of natural winds as an agent that carries the material instead of the machines.
On the other hand, from particle simulation method, there are many precedents. In the realm of geography, numerical simulations of sand dunes have been examined using the Werner-model-based approaches (Barchan and and Hugenholtz 2012). However, this model doesn’t consider collisions between individual particles. Yet, the computing power for mass physics simulation is not enough in current computing speed in 2016.
Advantage
Werner Model Simulation
CFD Simulation
Particle system
Mass physics Simulation
Fast computing time
Relatively fast
Relatively fast
Can manipulate mass and friction, collision
Disadvantage
Cannot simulate with
Cannot predict
Can detect
Heavy and time
obstacles
sedimentation
collision but
consuming for compute
of sand
abstract
Table 3. Advantage and disadvantage of the particle simulation methodology
4.
Problem Statement and Hypothesis
4.1 Problems that examined in this paper •
Research the relationship among 3 data; velocity and pressure from CFD, the height of particles in the experiment space from the mass physics simulation.
•
As single machine, simulate vertical / horizontal flow (both velocity and pressure) around vehicles in order to guess where is the area that sands are excavated and be precipitated.
•
As multiple machine, through layout between the two vehicles to navigate wind by intent (parameters: location, angle)
4.2 Problems that could not be achieved to examine in this paper •
More number and more complex layout of machines.
5.
•
Dynamic behaviour of the obstacles: called wind-direct machine
•
The scale-ability of this machine(s).
Methodology/ Data Analysis In this research, the divergence between the physics simulation and CFD
simulation was examined first. Then, CFD simulation is used to understand the approximate behaviour of the wind. The simulation process consists of four steps. The first step is for form finding of the robot. The others are for creating preferred shapes of the sand. Reducing the deviation from the physical behaviour is one of the concerns.
5.1 Validity of Substituting Physics Simulation with CFD Simulation: In this section, comparisons between physics simulation and CFD simulation are examined. The physics simulation with mass is done by using ‘MassFX’ function in 3ds Max 2016 (Autodesk). Compared to so-called particle system in various sort of software, this function has following features: 1) mass and density, 2) collision detection, 3) parametric friction, 4) parametric wind force including turbulence, and 5) parametric precision by controlling the number of iterations. The estimated result of this experiment is the correlation between the distribution of sedimentation of sands and wind velocity and pressure showed in Flow Design CFD analysis. If this hypothesis is verified, it follows that the CFD simulation is available for estimating sedimentation. First attempt is to compare below three candidates (figrure5-1). The assumption is the pattern of Velocity and/or Pressure somehow correspond to mass physical simulation by using image processing. A lot of way of overlaying are generated (figure 4), but finally the appropriate result did not come out in any way. The sample result is below
Figure 5-1. Left: the velocity in the field with rect-obstacles by CFD Middle: the pressure in a space with rect-obstacle by CFD Right: The Mass physics simulation and particles’ height (red)
Figure 5-2. Left: 4 step patterning for both velocity and pressure, Second Left; differentiate patterning in XY directions, Right three: Various way of multiples calculation in RGB values. Second attempt is to see the vector directions of each particle in CFD, both in horizontal and vertical as 2D, then finally 3D directions are also examined. Simply the moving direction of particles are set synthetic vectors made by gravity and each motion vectors. (Hensel, Menges, and Weinstock 2010) From this hint, the diagram (figure 5-3) shows that the end of vector towards ground is correspond with mass simulated pattern in both
X, Y, direction. Those are tested certain amount of times.
Figure 5-3. Horizontal vector direction by 3D CFD
6.
Conclusion
6.1 Conclusion of this paper Firstly, through the analysis of present construction machines, evaluation criterion for future construction automation are set below; accuracy, energy consumption, and time. Secondary, in order to confirm the validity of the proposed case study, simulation workflow was established. Finally, the information from CFD simulation substituted the mass physics simulation in certain probability, though it need to be carefully operated.
6.2 Future Work The relationship between physical model and mass physical simulation, has not yet tested well, though there are obvious similarity (Figure6-2). Especially there are issues; units, scale factor and parameters’ balanced point between 3 models are ambiguous. Those need to be precisely examine.
Figure 6-1. Comparison between 3 models (expected image) Expected Shaping Process: Unlike conventional construction machine, robots command to transport the material. In this project the sand is blown away by natural wind power and the robots only direct where the sand is deposited. Figure 5 shows how the shaping process takes place.
Figure 6-2. Sectional diagram of expected shaping process The future development includes, but not limited to, investigation on the size of the particles and patterns for adhesive material as well as full-scale experiment of the robots. It would be great if the result of this simulation supported the research hypothesis that in situ forming with utilization of natural energy satisfies, but the tradeoff between time and energy consumption at higher level.
Figure 6-3. Mechanical diagram of wind-direction robot, which is under develop
References: Barchan, Thomas E. , and Hugenholtz, Chris H. 2012. "A new tool for modeling dune field evolution based on an accessible, GUI version of the Werner dune model.” Geomorphology 138 415–419 Debnarova, Adriana. 2014. “Sandwright.” Accessed June 4. https://www.behance.net/gallery/20326693/Sandwright Hensel, Michael., Menges, Achim., and Weinstock, Michael. 2010. Emergent technologies and design. Oxon : Rutledge. Kestelier, De Xavier., Dini,Enrico., Cesaretti, Giovanni., Colla, Valentina., and Pambaguian, Laurent. 2014. "THE DESIGN OF A LUNAR OUTPOST: 3D PRINTING REGOLITH AS A CONSTRUCTION TECHNIQUE FOR ENVIRONMENTAL SHIELDING ON THE MOON." In Fabricate : negotiating design & making, edited by Gramazio, Fabio., Kohler, Matthias., and Langenberg, Silke, 198-205. Zurich : gta Verlag. Larsson, Magnus. “DUNE (AA THESIS 07-08).” Accessed June 4. http://www.magnuslarsson.com/architecture/dune.asp Poon, Wingchi. “Mesquite Flat Sand Dunes” Accessed July 5. https://upload.wikimedia.org/wikipedia/commons/1/15/Rolling_Mesquite_Flat_ Sand_Dunes.JPG
