humanizing digital reality

Humanizing Digital RealityDesign Modelling Symposium Paris 2017

Editors : Klaas De Rycke, Christoph Gengnagel, Olivier Baverel, Jane Burry,Caitlin Mueller, Minh Man Nguyen, Philippe Rahm, Mette Ramsgaard Thomsen.

Ice Formwork for Ultra-High PerformanceConcrete: Simulation of Ice Melting

Deformations

Vasily Sitnikov(&)

KTH Royal Institute of Technology, 100 44, Stockholm, [email protected]

Abstract. This research project asks how can we efficiently solve structuralproblems with much smaller masses of concrete while maintaining the samelevel of cement consumption, through the use of Ultra-High PerformanceConcrete (UHPC), degradable ice formwork and through an active involvementof digital simulation. A survey undertaken in the recent advancements ofconcrete-related technologies has revealed an opportunity to propose a newconcept of casting. The new method of fabrication of UHPC elements employsice as the main material of the temporary formwork construction, involvesautomation of the fabrication process and solves the problems of waste materialand manual labor at the stage of demolding. This casting method is beingdeveloped for cases, where unique and customized concrete elements of com-plex geometry are needed, e.g. topology-optimized structural elements. Toimpose a desirable geometry onto ice, the concept considers the use of 2.5DCNC milling operations. To minimize machining time, milling is thought to becombined with controllable melting of ice that allows to achieve high-qualitysurface finish on a large scale in a very simple way. The main hypothesis of theresearch is, that the control over melting could be possible through digitalsimulation, that is the melting deformations can be pre-calculated if parametersof the material system are known.

Keywords: Ultra-high performance concrete � Formwork � Simulation

Context—Fabrication of Complex Geometry in Concrete

A Strategy for Efficient Use of Ultra-High Performance Concretein Architectural Constructions

The contemporary industry of prefabricated structural concrete elements aims atminimizing its environmental and economic cost. For this purpose, the industryemploys a strategy of reducing the share of cement in concrete, decreasing it down tothe minimum of approx. 10% (Fig. 1a). This approach results in a relatively weakconcrete, leading to thicker profiles of floor slabs, beams and columns. The increasingmass of these concrete elements negatively affects the total embedded energy caused bytransportation and handling, increasing the final CO2 footprint. The current researchassesses alternative strategies of improving the overall performance of the industry.

© Springer Nature Singapore Pte Ltd. 2018K. De Rycke et al., Humanizing Digital Reality,https://doi.org/10.1007/978-981-10-6611-5_34

Counter-intuitively, a possible solution is to boost mechanical properties of concrete byincreasing the content of cement. Higher material performance allows us to solvestructural problems with much smaller masses of concrete, maintaining the same levelof cement consumption per structure. This strategy is based on the use of Ultra-HighPerformance Concrete (UHPC), where cement constitutes approximately 40% of thebulk mass (Fig. 1b). While the chemical composition of UHPC is almost identical withregular concrete, the mechanical properties are simply incomparable. For instance,some types of UHPC gain compressive strength of 100 MPa in twenty-four hours ofhardening (Fig. 1c), while regular concrete grades can barely reach half of this strengthin a month time (Sitnikov and Sitnikov 2017).

The radically improved material properties of UHPC entail new concepts in thefield of structural design. For instance, Phillipe Block claims that the new material cansave up to 70% of the mass of concrete constructions, if structural geometry is rede-fined according to the improved material properties (López et al. 2014). However,contemporary computational methods of structural optimization generate complexdoubly-curved geometry which is very uneconomical to produce using existingindustrial methods. In many cases, the only fabrication method is based on CNC-milledEPS formwork. Considering the high amount of non-degradable waste caused by thistype of production, and the shortage of opportunities to reuse these molds, the speci-ficity of design needed for UHPC concrete arguably needs another system for moldproduction.

Fig. 1. Comparison of mix design of C30 concrete grade and UHPC

396 V. Sitnikov

A Strategy for Minimizing Formwork-Related Material Waste

Driven by the intention to eliminate the material waste in the process of production ofconcrete structures, several researchers have proposed constructing formwork out offully reusable materials, adopting the format of “closed loop recycling”. Following thistrace of thought, Mainka et al. (2016) and Oesterle et al. (2012) have developedstrategies for producing concrete constructions using formwork made from industrialwax. Considering that wax can be CNC-milled and can be fully recycled at a relativelylow cost, the casting technology was assumed to have a straightforward implementa-tion in the building industry. However, wax is an extremely sensitive material withvariable softness that reacts dynamically to changes in temperature. This lack of sta-bility makes it difficult to use for on-site concrete casting.

But the employment of an easily melting, reusable material for construction of atemporary formwork is evidently a good choice for sustainable and economic fabri-cation strategy of bespoke concrete elements. To use such a process actively involves anew technological parameter—temperature—that so far has been used rather passivelyby the concrete industry. In these circumstances, a viable way to reduce risks offabrication is to conduct concrete casting procedures in a controlled environment, thatis prefabrication of concrete elements in an equipped factory, followed by trans-portation of this elements on construction site. Historically, the form of prefabricationhas proved its efficiency in countries where climate conditions does not allow in situcasting during the winter season. However, when talking about 70% lighter elements ofUHPC, the expenses on transportation are getting substantially reduced, offering newcases for efficient implementation of prefabrication.

Arriving to the Concept of Ice-Formwork

Control over temperature throughout the casting process enables even broader range ofmaterials that can constitute a temporary container for concrete of a specific geometry.In this case even ice can be thought of as a main material for the formwork, providingnot only a wasteless, but an ultimately clean process of fabrication. Given that theenergy consumed by the refrigeration system can come from renewable sources, theconcept has a potential to be fully sustainable.

To anticipate results of the implementation of this proposal in practice, the currentresearch seeks to look into all the effective changes that fundamentally reorganizedesign and fabrication workflows, and reconfigure the economy of the process, startingwith the re-reading of industrial refrigerator as a new kind of kiln.

Research in the Material Science of Concrete

The concept of the ice formwork for concrete builds upon the assumption that, apartfrom the increased mechanical properties, the chemical design of UHPC entails otherqualities, which distinguish it from conventional concrete grades. Therefore, theresearch attempts to exploit the side effects caused by the extremely low water-cementratio possessed by UHPC, which offers drastically different ways of processing it.

Ice Formwork for Ultra-High Performance Concrete … 397

First of all, the almost complete absence of unbounded water in UHPC mixesprovides possibilities for hardening in sub-freezing temperatures. This aspect wasinvestigated and proved with a series of material tests (Fig. 2a) during the first stage ofthe current research. With an addition of a very basic antifreeze admixture it waspossible to conduct concrete setting and hardening in the temperature of −18 °C.

Second of all, experiments carried out in this research show that the temperatureincrease as UHPC hardens can be kept below 0 °C. The material studies identified thatthe amount of heat released due to the exothermal reaction of cement hydration isdirectly dependent on the initial temperature of concrete mix (Fig. 2b).

Above mentioned means that, if UHPC mix of a negative temperature is poured in aformwork made of ice, it will harden without melting the contact surface. This fact isillustrated in the Fig. 3, where very minor details that are present on ice surface offormwork remain on the surface of the negative concrete cast. These findings constitutethe initial argumentation of the concept of ice formwork.

The Ice Machining Techniques—Imposing Geometry onto Ice

The Basic Types of Automated Methods of Processing Ice Shape

From the design perspective, the most important question is how to impose a desirablegeometry onto ice. Assessment of this question has revealed a range of alreadyestablished computer numerical controlled methods for processing ice.

The most established method of ice processing is the three-axe CNC-milling.Machines, designed specifically for ice milling, has been developed by many com-panies and are now available on the market in different sizes. Yet, they have no

Fig. 2. Kinetics of sub-freezing hydration of UHPC (a left, strength tests results of reference andexamined samples; b right, graphs illustrating kinetics of hydration)

398 V. Sitnikov

practical implementation, and are used only for advertisement and entertainment. Thistype of machining allows to make 2.5D reliefs of several decimeters in depth, that isreliefs without undercuts.

Next, abrasive wire-sawing has been used by Søndergaard and Feringa (2017) toproduce large-scale elements of ice of a ruled-surface geometry for a temporarypavilion in Kiruna, Sweden. The wire saw actuated by an industrial robot exemplifies apossibility of machining ice on an architectural scale, producing elements that aredirectly implemented in architectural structures. However, this concept is limited bymerely ruled geometry and therefore can’t fully respond to the demand ofstructurally-optimized geometry.

Finally, several concepts for ice 3D printing have been proposed by differentauthors. Perhaps, the most well-established has been developed at McGill Center forIntelligent Machines. The PhD dissertation of Barnett (2012) delivers a robot-assistedrapid prototyping system for producing ice object of almost any geometry by incre-mental deposition and crystallization of water. While reaching the tolerance of 0.5 mm,the method requires parallel deposition of a supporting material (shortening methylester) that should be removed manually after the printing is finished. It is hard toovercome these obstacles while scaling the fabrication up for the intended architecturalapplication. However, an alternative strategy has been proposed by Peter Novikov fromAsmbld, the architectural technology consultancy. A project of an Ice 3D Printer offersa method of aggregating ice through direct deposition of overcooled purified water,utilizing the effect of immediate crystallization. While being a very promising tool, it isstill on a very early stage of development.

Due to the objectives of this research it was considered that CNC-milling is the bestfitting method in terms of speed, precision, scale range and formal freedom.

Fig. 3. Accuracy of the local geometry transfer and the surface finish quality. Left ice formwork,right negative cast in concrete


Arriving to the Concept of Simulation-Based Hybrid Method of ProcessingIce

Considering that the ice formwork could be most efficient in cases when producingunique structurally optimized concrete elements, and therefore to be for single use only,it is highly desirable to minimized as much as possible the time-consuming process ofmilling. This intention has led to a concept of a simulation-based method of hybridfabrication, where minimal and elementary milling interventions in a volume of aregular ice block are allowed to develop into a more complex geometry through thefollowing process of melting. It should be mentioned that the intention to linkcomputer-generated geometry of optimized structures to the natural phenomenon ofmelting deformations is driven both by the necessity to minimize the effort duringfabrication and to provide a natural materialization strategy, that does not solely rely onthe transitory machining equipment, but aspires to involve new perpetual physicalprocesses in the culture of fabrication.

Conducted laboratory experiments (Figs. 4, 5 and 6) helped to identify conditionsthat reduce complexity and provide homogenous, incremental melting that is quite

Fig. 4. Melting test #1—a negative cast of the surface developed by projecting a heated waterjet onto the surface of ice. Full documentation of the fabrication process: goo.gl/FPIMW8

400 V. Sitnikov

suitable for geometrical simulation without excessive challenges. The chosen conditionis illustrated with the Fig. 6 and consist of a static interaction between ice and salinewater solution of equal negative temperature.

Underlying Physical Principles of Ice Melting in Saline Water Solution

The physical phenomenon that underlies the above-mentioned formation processhappens on the molecular level and depends on quite a few factors. When a block of iceis submerged in a saline water of the same negative temperature, the randomly walkingmolecules of liquid start to collide with the static molecules of ice, held by its crys-talline lattice. At every collision of a salt molecule with ice, one of the water moleculesleave the crystalline lattice, changing its state from solid to liquid. Therefore, theinterface between ice and water is constantly retreating. Important to notice that themolecules located on the flat areas are open only from 180°, and those located on edgesof the block are open from 270°. That is free-swimming molecules will collide with theedges more frequently than with molecules located on the flat sides, resulting in dis-proportional deformation of the initial volume and development of doubly-curved

Fig. 5. Melting probe #2—a negative cast of the surface molten by static contact with warmwater


surface. Depending only on the concentration of salt and on the level of constanttemperature, the same deformations will progress faster or slower.

Even though this physical condition of melting largely avoids ambiguity of bulkflow caused by thermal convection and turbulence, it still involves aspects of diffusionand flux due to the dilution of saline solution with fresh water that separates from ice.In fact, in the process of constantly decreasing concentration of salt, the self-containedsystem of saline solution, ice and constant negative temperature can reach a state ofequilibrium, so that deformation to the interface would no longer occur (Kim andYethiraj 2008). Therefore, it is necessary to consider changes in concentration that areconditioned by the diffusion coefficient of salt in water at a given temperature, and bythe flux caused by difference in local densities.

A Preliminary Framework for Iterative Digital Simulation of MeltingDeformations

In the recent decades, there has been proposed a number of strategies for computersimulations of ice melting. It was oceanology that has first raised the mathematical

Fig. 6. Melting probes #3—a negative casts of the geometry of the ice surface created throughsubmerging ice in a saline solution of sub-zero temperature. When the geometry of ice wasimprinted in concrete it was digitally recovered through 3D scan of the negative concrete castsfor further analysis

402 V. Sitnikov

problem of the moving boundary (Stefan problem) for numerical simulation of sea icemelting. On the other hand, molecular simulations used in chemistry allow precisemodelling of almost any molecular reaction, including water phase change. However,the only discipline that has been interested in melting of life-sized lump of ice is theindustry of visual animation. Since the scientific credibility is not the objective of thisdiscipline, the epistemological value of melting animations is quite small. Nevertheless,contemporary methods of animation encompass aspects of thermal transfer and fluiddynamics on a quite advanced level, and can contribute to the intention of this research.

In terms of geometrical modelling, animations either use level-set or particle-basedalgorithms. The implementation of particle-based method is advisable when theintention is to capture the deformations caused by fluid dynamics, such as deformationsof water flowing down the surface of ice (Iwasaki et al. 2010). However, in cases, whenmelting is caused not by mechanical, but rather chemical factors, thelevel-set algorithms seem to be a better choice (Wojtan et al. 2007).

The intended simulation, therefore, is based on the representation of polygon meshwith marching cubes algorithm (Lorensen and Cline 1987). The initial surface condi-tion that separates liquid and solid phases (that is the initial geometry of ice) is used toassign the initial decimal values between 0 and 1 to corners that belong to ice or salinesolution correspondingly. These values will be recalculated at every iteration accordingto the diffusion level of salt particles at a given negative temperature. The flux causedby the difference in local densities should also be taken into account. It is obvious, thatthe total sum of all the corners’ values in this closed system should be constantthroughout the simulated period. At every iteration, the polygon mesh that representsthe interface will be transformed according to the new distribution of values, usinglinear interpolation to separate vertices with values <0.5 (ice) from those of >0.5 (salinesolution). Due to the fact that changes of corner values are conditioned by changes inthe neighboring corners, in principal it can be computed using cellular automataalgorithm.

To evaluate the accuracy of simulation and estimate its tolerance, the additionallaboratory tests will be conducted. By comparing the spatial data calculated throughsimulation of melting with the data gained through 3D-scanning of actual materialsamples, the level of correspondence between digital model and physical process canbe identified. A simple way to scan deformations of ice melting is to imprint itsgeometry in concrete, that is to make a negative cast. The feasibility of this method hasbeen already tested. As a well-functioning methodological tool, it can either serve toimprove and calibrate the simulation.

Conclusion

The current stage of the research does not allow to draw an overall conclusion, since alot of aspects lack certified documentation and need to be further examined. However,the crucial aspects have been validated with principal empirical tests. It is, therefore,possible to claim that concrete casting in ice, as well as fabrication of ice formworkthrough melting deformations, is viable and potentially efficient (Figs. 7, 8 and 9).


Fig. 7. Melting test #3—the ice formwork

Fig. 8. Melting test #1—the process of producing ice formwork (full documentation: goo.gl/FPIMW8)

404 V. Sitnikov

The research targets on fabrication of a full-scale structural element out of UHPC.The process of its fabrication is thought to be the major experiment of this research andthe main epistemological source for evaluation of the concept, revealing all the hiddenaspects that can’t be anticipated beforehand. However, the studies that are separatelyundertaken in material science of concrete and physics of phase transfer can, to acertain extent, ensure a positive result and reduce risks related to the increased scale oftechnological operations.

Acknowledgements. This project is a part of the Innochain Research Training Network. It hasreceived funding from the European Union’s Horizon 2020 research and innovation programmeunder the Marie Sklodowska-Curie Grant Agreement No 642877.

Fig. 9. Melting test #1—the ready formwork (full documentation: goo.gl/FPIMW8)


References

Barnett, E.: The Design of an Integrated System for Rapid Prototyping with Ice. McGillUniversity, Montreal (2012)

Iwasaki, K., Uchida, H., Dobashi, Y., Nishita, Y.: Fast particle-based visual simulation of icemelting. J. Compil. Pac. Graph. 29(7), 2215–2223 (2010)

Kim, J.S., Yethiraj, A.: The effect of salt on the melting of ice: a molecular dynamics simulationstudy. J. Chem. Phys. 129(12), 124504 (2008)

López, D.L., Veenendaal, D., Akbarzadeh, M., Block, P.: Prototype of an ultra-thin, concretevaulted floor system. In: Proceedings of the IASS-SLTE Symposium, Brasilia (2014)

Lorensen, W.E., Cline, H.E.: Marching cubes: a high resolution 3D surface constructionalgorithm. In: Proceedings of the Annual Conference of the ACM (1987)

Mainka, J., Kloft, H., Baron, S., Hoffmeister, H.-W., Dröder, K.: Non-Waste-Wachsschalungen:Neuartige Präyisionsschalungen aus Recycelbaren Inustriewachsen. Beton- und Stahlbeton-bau 111(12), 784–793 (2016)

Oesterle, S., Vansteenkiste, A., Mirjan, A.: Zero waste free-form formwork. In: Proceedings ofSecond International Conference on Flexible Formwork, pp. 258–267. Bath (2012)

Sitnikov, I., Sitnikov, V.: Production of thin-walled unreinforced elements out of UHPC. Concr.Plant Int. 17(2), 46–50 (2017)

Søndergaard, A., Feringa J.: Scaling architectural robotics: construction of the Kirk Kapitalheadquarters. In: Proceedings of Fabricate 2017, pp. 264–271. UCL Press, London (2017)

Wojtan, C., Carlson, M., Mucha, P.J., Turk G.: Animating corrosion and erosion. In: Ebert, D.,Mérillou, S. (eds.) Eurographics Workshop on Natural Phenomena 2017, Prague (2007)

406 V. Sitnikov

humanizing digital reality

Documents