Textile Logics: Designing Structural Membranes

August 22-24 , 2011 the last and final workshop of the Digital Crafting Workshops series was held at the Design School in Kolding Denmark.   Workshop 5, titled  Textile Logics:  How to Brace,  was run by Mette Ramsgard Thompsen Head of CITA in Copenhaen, with invited guest Sean Alquist from the Institute for Computational Design at the University of  Stuttgart.The workshop focussed on the relationship between digital tools and the fabrication of variegated textile tension structures.  Textile designers Vibeke Riisberg , textile engineer Joy Boutrup, and textile manufacturing technical designer Helene Jensen  from the Design School at Kolding helped guide 3 teams of architects, engineers, artists and designers in creating three versions of tension structures from different types of variegated textiles.On Day 1, Teams designed the external shape of the tension structures as well as the textile itself  on laptops.    Vibeke and Helene worked with teams to produce  knitting samples and tested techniques.  The picture above is the the external shape and textile my team designed, in this image it is fully tensioned.  In this case the textile acts like a net.  Stresses are moving primarily through the thicker parts of the textile.On Day 2 teams began to process their textile patterns for the knitting machine.  The picture below shows the fabric our team designed directly from the industral knitting machine.  You can see that the fabric has a lot of bubbles that do not carry so much stress.


The workshop provided a unique opportunity to fabricate using tools from the textile and commercial apparel industry. Here are some pictures of the tools used to fabricate our tension structure.  The sewing machine below, sewed the loose ends of the knit together with a polyester and nylon thread, and sliced off the leftover part of the seam as you stitched.   Very fast,  scary at first.

Sewing on an Edger Machine

In the picture below is the industrial knitting machine that made our textile, which was a double knit structure.  The machine was started with plain wool to get it going and then the yarn was changed to the polyester yarn.

flat Knitting Machine

Day 3 was a seminar day, where there were invited guests, and teams discussed what was learned during the entire process  that moved from digital information to the fabricated tension structure.  You can read more about the Digital Crafting 5 Workshop here:  http://www.digitalcrafting.dk/wp-content/uploads/2012/08/DC-DigitalCrafting_Web_S.pdf

Cloth Emerging from Knitting Machine

Python Powered Textiles

Matters of Tension and Compression within Geometries

Woven Half Sphere_Computational Textile_Boundary Stretched


Issues of tension and compression seems to be thematically recurring in the process of developing these printed textiles works.  Above  is an  ABS 3d printed model created with the plain weave script from the PointCrowd RhinoPython workshop run by Ari Kardassis and Masoud Akbarzadeh in January 2011.  Many of the tension and compression issues stem from the phenomena of what I call getting gauge.   It is that trial and error part of getting the right tension on a knit textile for example to meet the size requirement for a sweater or blanket etc.   Geometrically for this diameter pipe or  yarn, this is about the maximum distance the outer boundary can stretch from this specific original sphere before the piped surface starts to twist along the spine of the curve and warp the circular section of the pipe.   This behavior makes for zero sectional thickness in the 3d digital model, and cannot be printed.   This first experiment had a very simple version of piping that is computationally much lighter.  This pipe script works, up to a limit.  In this version of this model, the piping or yarn does not intersect at the weave crossings on this model, but the geometry of the sphere holds the separate woven strands together.




Compression of Single Link Textile


The single link scripted model above was printed as a flat, rectangular model, as you can see with no boundary return along the open ends of the strand, permitting this model a way to slide diagonally  in the xy plane to have different states of density.   Each of the strands is also free to rotate in the z direction allowing for some curvature parallel to the direction of the strand.

Knit Half_DetailKnit Half Sphere_Detail


This is a model of interloping or knitting and presented the most issues with the pipe or yarn torquing into nothing,  it required a new pipe script to get it to print at all.   The section on this model is true and circular.   The script is computationally much more demanding in terms of what it needs to track, but you get good results in the section of the yarn.   It is posted here,  thanks to Skylar Tibbets for help with solving the twisted pipe.  This was made with some intersections in the yarn, but geometrically will hold together just fine if the yarn does not self intersect.  The “ApplyCurvePipe” command in Rhino yields beautiful pipe results,  especially for the knit curves and  I have an update from Pascal Golay at McNeel.  You can use   “ExtractPipedCurve” to get a nice meshed curve from CurvePipe object that can be exported to an stl file.

Knit Half Sphere_InteriorKnit Half Sphere_Interior


Having made these physical models,  I think the most useful model is the script, that can be used repeatedly to look at many different kinds of things both physically and digitally.   These textile samples are neither surfaces nor bones  and can reside somewhere in between, because you can push the script to be either/or if you so wish.    These qualities have meant that it has been possible to look at these models as outer perimeters or skin and as  structure.  This kind of model does not require one kind of outer boundary or shape but is made up of many interrelationships of parts that can then be formed for multiple purposes.

Experiments to continue…

Thank you to Chris Dewart and Seth Hall at RPL for use of the ABS printer.