Curved Folding



Among the challenges of scaling the hairline thickness of paper to thicker plastic materials approaching 2-3mm, is the necessity to create localized fault lines whilst maintaining material surface integrity. For instance, in paper the option to either compress the fibers of a fold line with a bonefolder or perforate with a rotary cutter or lasercutter works very well. However, the technique does not translate to thicker materials. With compression of plastics requiring mechanized pressures beyond availability of the school and perforation holes shearing apart as the upper material fibers fail to tension during folding, a different design technique was required for folding thick materials. (Image here of perforated study cracking under fold)

We imagined how else digital milling might mass produce the fold lines of the design to scale besides perforating and cutting, turning toward milling for cutting grooves in partial thickness of the final material. With a steady hand, armed with dozens of router bits, we began testing the analogue equivalent hand router making test grooves in the final 3mm polypropylene and subsequently checking for material inconsistencies in the groove, and ability to be folded by hand. Variations included “V” and “U” bits, differing bit widths and routing depths resulting in variations in plastic smearing across the groove and fold ability.

Curved Folding Rocking Chair Crease Mill Groove Details

We learned slower tooling speeds build up heat at the bit which smear the grooves creating inconsistent depths. Additionally continuing grooves to intersect one another or to the edge of the sheet resulted in material fracturing upon folding. Speeding up tooling and offsetting the termination of grooves a few mm proved successful. (Figure XYZ – Milling Depth Studies)


Standard sheet stock sizes such as A4, A3, and A2 do not pose an immediate problem in early concept studies and design. In fact embracing a bit of imprecision of scapular exactitude at these sizes by zooming in photocopiers and fitting to extents in laser cutting only seems to expedite the iterative process. However a model of 1:1 scale has very real material sheet size constraints to be addressed. In our particular design requiring a roughly 2.6m x 2.6m continuous plastic sheet, the largest domestically available polypropylene sheet was 1.2 m x 3m.

As a result either the design needed to accommodate the sheet dimensions by retaining halves to max extents on the sheets that would fasten together, or the sheet needed to accommodate the design by becoming a larger continuous plastic sheet. We tried both.

Curved Folding Rocking Chair Research Sheet Fastening Details

To fasten together smaller sheet sizes we explored a series of mechanical interlacing zippers which fastened halves together be means of tabs and slots. Variations included tabs scaled to estimated loading at that point in the design, square and arrow type tabs, as well as partial weaving. ( Figure XYZ – Fastening Studies) This method failed structurally under load for two main reasons. Primarily, the tabs under tension failed while the compression surface buckled and the design abruptly splayed apart. Secondly, the zippers did not maintain the design surface curvature across the zipper fastening.  Rather than conform to a singular surface curvature, the shared edges remained pronounced under the stress of each half creating a centerline ridge in the study models.

Curved Folding Rocking Chair Research Sheet Fusing Details

Results were more promising when augmenting sheets to larger sizes through thermal fusing (plastic welding). Using a flat tipped soldering iron, we set out exploring variants in joining 3mm plastic sheets each butt and overlay seams. (Figure XYZ – Plastic Fusion Studies). Successive attempts proved a combination of butt joining and followup backer overlay seaming provided sufficient continuity and rigidity across the seam line. In addition, this method allowed a the finish face of the chair design to remain uninterrupted, shiny and glossy the sea line reduced to a mere hairline in the prototype.


An interesting branch of our material research involves exploring material solutions governed by different shipping methods. When acquiring sheets of roughly 3m x 3m delivery could either be in continuous roll form delivered in a shipping tube or stacked in discrete sheets on a pallet. We worked with Forbo and their London supplier DeBruyn to test vinyl roll flooring for structural resiliency under load, finding the material to slump over time resulting in deformation of the design under self load, and collapse under loading. (Figure XYZ –  Forbo roll flooring studies)

Curved Folding Rocking Chair Research Material Studies Forbo Linoleum Roll Flooring

However, discrete sheet stock of polypropylene plastics of equal thickness retained the rocking chair form under self load of the material and subsequent performance loading tests. The internal rigidity and “memory” of the sheets to spring back to flat plays a vital role in the rocking chair design to maintain tension in the concealed seat back tabs which lock the behavior of the folding into a stable form.


Unique to a rocking chair design is the necessity to perform under continuous change of loading condition on the frame in addition to when static at rest. To simulate and design for motion, center of gravity, balance, and safety, we devised a number of test methods to assist and design mechanisms during development. Before any dynamic test began a simple static loading at rest was performed using resting barbells.(Figure XYZ – Static Loading Performance)

Curved Folding Rocking Chair Research Loading Performance Studies

Dynamic Motion tests employing a bean filled keyboard wrist rest as a scale human model were conducted to simulate a low center of gravity through successive rocking motions. (Image here Beanbag model rocking frames) Initial rocking tests developed smooth back and forth motion, which proceeded to refine motion by including design features. Features including back springs, front and rear motion stoppers, as well as angle of maximum and minimum rocking position were developed iteratively through trial and error. The best performer proceeding. (Figure XYZ – individual features evolution)

Curved Folding Rocking Chair Research Feature Evolution


Attempts were made to breed design traits by assigning fitness criteria to individual fold lines and cross breeding curvatures without explicit performance testing of the feature. Performance metrics were governed by a predetermined set of variants, and highest values were compared, and curvatures averaged , then the entire design re evaluated. The results were less than encouraging, and the process abandoned for a more focused development of individual features in the design approaching one problem after the next until each was resolved elegantly considering the features already resolved. (Figure XYZ –  Evolution of design template concept to proto)

Curved Folding Rocking Chair Research Crease Pattern Evolution


In creating the rocking chair prototype we employed a number of low-tech techniques. The innovation was in the sequencing and embracing speed over precision when possible to rapidly realize, evaluate, and fabricate successive iterations toward a 1:1 scale rocking chair prototype. Often hand tracing, using available hand tools, and residual spaces at the AA to fabricate the design.

01. projection and tracing halves (image here tracing wall template)

02. Backlit tracing valleys on reverse side ( image here tracing against window)

03. Routing grooves each side ( image here hand routing on floor)

04. Jigsaw cutting off residual edges (image here jigsaw cutting on a table)

05. Heat fusing centerline of two adjoining sheets (image here soldering hundreds of holes)

06. Folding with 4 and then 13 into stable shape (image here of chinos yard 4 folding and time out lot DRL class folding)

Curved Folding Rocking Chair Research Prototyping Analogue Methods

Curved Folding Rocking Chair Research Prototyping Analogue Methods



Credits & Special Thanks

AADRL Team: Henry David Louth (USA), Guillermo Oliver (Mexico), Julian Lin (China), Yang Du Messi (China)

Course Tutor: Shajay Bhooshan – Zaha Hadid Architects

Curved Folding Primer: Gregory Epps – RoboFold

Consultation: Rasti Bartek  – Buro Happold Engineers



document icon

Curved Folding Behaviour

AADRL Book 2012


>>References & Further Reading

Curved Folding Research Papers:

>>David Huffman, “Curvature and Creases: A Primer on Paper“. 1976.

>>Erik Demaine, “Reconstructing David Huffman’s Legacy in Curved-Crease Folding” 2010.

>>Erik Demaine  et all, “Curved Crease Origami” xxxx.

>>Martin Kilian et all, “Developable Surfaces with Curved Creases” 2008.

>>Dmitry Fuchs et all. “More on Paperfolding” 1999.

>>SD Guest et all. “Inextensional Wrapping of Flat Membranes” 1992.


Curved Folding Practicioners:

Gregory Epps Douwe JacobRan AmitaiPolly VerityRichard SweeneyYoshinobu MorimotoTokujin YoshiokasDavid Huffman,