THEFLEXIBLETEXTILEMESH
Manufacture of Curved Perforated Cladding Panels
YING YI TAN1 and TAT LIN LEE2 1,2Singapore University of Technology and Design [email protected] [email protected]
Abstract. This paper presents a new approach to manufacture lightweight perforated panels using textile reinforced composites (TRCs) for curved building designs. It explores the design variation of a graded mesh as a knitted textile formwork created by CNC knitting technology that can be edge-shaped by bendable elements and sprayed with polymer resin to form the composite panel.
Keywords. Textile-reinforced composites; Knitted textiles; Perforated Panels.
1. Introduction This paper presents a preliminary study towards a new method to fabricate lightweight curved panels made of textile-reinforced composites. We tap on the inherent stretch-ability of knitted textiles and knit customisability enabled by CNC-knitting technology (Shima Seiki, 2017). The idea is to use the textile both as a flexible mould and an internal reinforcement of a composite panel. The desired curved geometries can be obtained by using bendable strips along the textile edges and subsequent process of spraying polymer resin matrix to create the cladding panel. In addition, by customising the knit patterns, we foresee the possibility of fabricating porous curved panels similar to ones made from sheet metals. In practice, perforated curved cladding panels are conceived by discretising the global geometry, e.g. a building envelope, into multiple cladding panels of manageable dimensions for fabrication. For doubly curved perforated panels, fabricators use a two-step manufacturing process of shaping and cutting an initially flat metal sheet. They use forming techniques to shape sheet metal to the target geometry, such as single or multi-point forming (Lee & Kim, 2013). Subsequently, the solid surface undergoes laser-cutting or stamping to subtract material based on a designed perforation pattern.
T. Fukuda, W. Huang, P. Janssen, K. Crolla, S. Alhadidi (eds.), Learning, Adapting and Prototyping, Proceedings of the 23rd International Conference of the Association for Computer-Aided Architectural Design Research in Asia (CAADRIA) 2018, Volume 2, 349-358. © 2018 and published by the Association for Computer-Aided Architectural Design Research in Asia (CAADRIA) in Hong Kong. 350 Y.Y. TAN AND T.L. LEE
Figure 1. Curved Perforated Metal panels of Dongdaemun Design Plaza (left) (Catherine, 2016) shaped by multi-point stretch forming (right, credit: SteelLife) (Lee & Kim, 2013).
Although this produces lightweight perforated panels of curved geometries with good precision, these methods require heavy expensive machinery along with skilled manpower to operate them. These shaping techniques also risk compromising the aesthetic finish quality of the sheet metal, as they cause indentations upon the metal’s surface (Lee & Kim, 2013; Vazquez & Coleman, 2017).
1.1. PROPOSED FABRICATION In response, we propose an alternative approach to fabricate curved perforated panels with a good degree of flexibility for customization. Our research explores the fabrication of a textile reinforced composite (TRC) panel - a textile reinforcement layer made from high-performance yarns (eg. glass/carbon fibre etc.) and encased in a polymer matrix material which enables both to function in unison. TRCs are lightweight, feature mechanical properties comparable to certain metals, and are corrosion resistant, making them a viable substitute material for façade cladding panels (Nguyen et al., 2013). Our research starts with a knitted mesh textile formwork that simplifies the shaping and cutting steps. The textile formwork performs as a flexible mould and reinforcement layer to fabricate the panels based on the following workflow: (a) Design and CNC-knitting of the textile formwork; (b) Edge-Shaping of textile formwork using bendable strips; (c) Spraying of polymer resin matrix. THE FLEXIBLE TEXTILE MESH 351
Figure 2. Proposed Fabrication Workflow.
For this scenario, the choice of knitted textiles stems from their innate stretch-ability in comparison to woven, non-woven and braided counterparts (Padaki, 2006). Their characteristic intermeshed loop structure enables the formation of deep curvatures while maintaining a continuous surface with minimal wrinkling. We adopt CNC knitting technology which involves line-by-line knitting instructions input via a design software that describes the movement of the needles and yarn carriers to designate the position and type of stitches (Shima Seiki, 2017). Thus, this provides the opportunity for designed perforations to be created directly within the formwork itself without undergoing subtractive processes. In addition, the technology allows one to knit the sleeves around the edges to insert bendable strips during the shaping process.
Figure 3. Designed Perforations of a graded knitted mesh pattern.
Next, we shape the knitted formwork by inserting and flexing bendable aluminium strips along the boundary sleeves. The strips function as shaping devices to directly control the global geometry of the knitted textile while keeping the textile in tension and avoid unwanted deformations. The edge members also facilitate the alignment of adjacent edge conditions with neighbouring panels to achieve a visually continuous façade envelope. Once shaped, we spray laminating polyester resin as a matrix material onto the textile formwork using a pneumatic 352 Y.Y. TAN AND T.L. LEE air-gun. The formwork solidifies to form a TRC panel and this can be further polished to enhance its surface finish. As a starting point, this paper documents an initial study of the shaping and spraying processes of knitted textiles into TRC perforated panels. It also seeks to address the issue of geometric deformations in the form of wrinkling by varying stitch densities through differentiated mesh patterns. We hypothesize that this regulates the flow of forces within the textile and mitigates the occurrence of wrinkles on the surface. We investigate this through the design and fabrication of several singly and doubly-curved small-scale prototypes using graded mesh knits to create perforated panels.
2. Literature Review 2.1. EDGE SHAPING TEXTILE FORMWORKS Textile formworks have been used since the 1890s primarily as a stay-in-place flexible mould for concrete casting of building components to full-sized buildings. Textile shaping methods of draping and pneumatic systems have been researched extensively to provide viable alternatives to conventional rigid timber formwork which is laborious to construct for curved geometries (Veenendaal et al., 2011). Our research focuses on a different strategy which entails edge-shaping of a textile formwork. An emerging method that employs this is the Textile-hybrid system - a reciprocal structural system composed of bending-active elastically deformed rods connected to a tension-active textile membrane (Ahlquist et al., 2013). The interaction between both structural types enables the membrane to resolve the residual forces induced from the bent glass-fibre reinforced polymer rods, while the rods stabilise the textile in place. These components create a self-structuring system that has been developed for installations and used indirectly to create small-scale non-planar composites (Ahlquist et al., 2013; Sharmin & Ahlquist, 2016; Tamke et al., 2015). The system offers a good degree of direct control of the textile’s geometry with a minimal amount of shaping material while expanding the achievable geometric vocabulary of curved surfaces (Lienhard et al., 2013).
Figure 4. Semi-Torodial Installation (left) (Ahlquist & Menges, 2013) and Post-forming Composite Morphologies (right) (Ahlquist et al., 2014). THE FLEXIBLE TEXTILE MESH 353
2.2. KNITTING FOR GEOMETRIC SHAPING Manipulating elastic knitted textiles along its edges to accommodate specific curved geometries imposes a strain on the textile, which in turn self-organises to maintain a continuous surface. Such might result in the phenomenon of wrinkling arising from compressive forces within its surface (Amirbayat & Hearle, 1989; Zhang et al., 2007). This causes the textile to buckle into localised doubly curved regions leading to geometric inaccuracies. It also diminishes the load-bearing performance of the textile and its resultant composite (Leong et al., 1999). In textile hybrid systems, the act of customising a knitted membrane to regulate the flow of forces could be considered as a potential solution to mitigate the occurrence of wrinkling. Differentiation of the textile comes in form of designed knit patterns and stitch densities, providing regions of graded elasticity to better accommodate to the area within the bounding rods (Ahlquist & Menges, 2013).
3. Research Scope & Methodology In this paper, we investigate the shaping and casting processes of fabricating TRC perforated panels using a knitted mesh formwork, which not only offers porosity but also potentially reduces wrinkling. We edge-shape knitted textiles with different stitch densities in the form of varied mesh sizes. This aims to achieve the following: (a) to document the geometric vocabulary achievable for singly and doubly curved surfaces without the appearance of wrinkling and (b) to understand the geometric parameters where perceivable wrinkling occurs. We also design graded mesh patterns for several doubly curved surfaces based on the hypothesis that stitch density directly influences the elasticity of the textile which can minimise surface wrinkling. Subsequently, this paper outlines the procedures involved in spraying of polymer resin on the shaped textile surface.
3.1. SHAPING STUDIES Our research selects three mesh patterns of varying perforation sizes with a fixed panel size of 200 x 200mm (see Fig. 5 for specifics). Each panel consists of tubular sleeves along its perimeter edges to allow the edge-shaping members to be inserted. The panels are knitted using the Shima Seiki flat-bed weft-knitting machine, out of white cotton yarn material to provide visualization of wrinkles. Additionally, these knits are ironed before shaping to eliminate initial creases.
Figure 5. Different Mesh sizes used and U/S-curves as Edge members. 354 Y.Y. TAN AND T.L. LEE
We create singly and doubly curved geometries based on a vocabulary of interconnected curved edges. These curved edges are generated from the manipulation of one or two control points along the curve and translated along the Z-axis to form U- and S-curves respectively (see Fig. 5 right) . The curves are restrained to a maximum length of 200mm which adhere to the textiles’ dimensions. The curved edges are realized in physical space by laser-cutting MDF boards as a substitute for the bendable aluminium strips. For singly curved surfaces, we insert these edge curves along the more stretch-resistant wale-wise (longitudinal) top and bottom sleeves of the knitted textile meshes and mount these onto the aluminium rails (Fig. 6, left). The rails are adjusted to tension the textiles until wrinkling and sagging are minimised. In addition, straight edge members are inserted into the course-wise (transverse) left and right sleeves to maintain tension along the sides. For doubly curved surfaces, thin flexible aluminium strips replace the straight edges and are flexed to create the U- and S-curves. Once the planar textile is shaped, we identify and document regions of wrinkled deformations. The height of the wrinkles is measured from the adjacent flat region of the textile.
Figure 6. Edge shaping and Measuring the extent of wrinkling .
3.1.1. Shaping Results For singly curved geometries, the results show that the finer the mesh size, the more aptitude the textile has for reducing surface wrinkling and this makes it viable for a larger range of curved geometries. This could allude to a sparser stitch density having an increased capacity to self-organise and reduce the extent of wrinkling of its surface. Such is shown below is a matrix of the possible geometric vocabularies tested for each surface: THE FLEXIBLE TEXTILE MESH 355
Figure 7. Matrix of Achievable Geometric Vocabulary (Singly Curved Surfaces).
Wrinkling occurs mainly along the wale direction and these manifest as straight channels that run from one edge to the next, (see Fig. 6, left). This is more intense for geometries which have significant differences in curvatures or peak heights between both parallel curved edges (curvature difference: 1.98; peak height difference: >125mm). It is also noted that the number and height of wrinkle lanes are dependent on the distance between the peaks and the valleys of both edges. Furthermore, anticlastic geometry is significant for edge curves with none to small height difference (ie. < 25mm). This manifests as a negative curvature on the top or flattening along the sides of the textile’s surface. For doubly curved surfaces, it is noted that buckling still occurs in straight lanes, but also at certain end-corners where one edge meets another perpendicularly and manifest as S-shaped bends. These end-corners are at the perpendicular interface between two positively curved edges with an angle difference of >20 degrees (see Fig. 8), resulting from an accumulation of textile. This buckling behaviour intensifies as the angle difference increases. Alternatively, wrinkling is mitigated in other corner scenarios where a positively curved edge meets a negatively curved edge, or where two negatively curved edges intersect. Thus, altering the geometry is a possible solution to overcome wrinkling as this stretches out the textile in the course-wise direction.
Figure 8. Doubly Curved Surface with problematic edge corner wrinkling. 356 Y.Y. TAN AND T.L. LEE
3.1.2. Design of Graded Mesh Formworks In this case, the design of a graded mesh pattern is dependent on where wrinkling is predicted to occur and its corresponding intensity: areas where we anticipate larger wrinkles are created with the more elastic finer mesh and vice versa. It is also observed that these mesh patterns need to be continuous and span from one edge to another, preferably from wale-wise top to bottom edges based on how wrinkling patterns were recorded.
Figure 9. Graded Mesh Shaping & Spraying.
From our prototypes, it is essential to consider the entire textile surface as a continuous entity when designing the flow of forces within the knitted textile. Assigning localised mesh patterns based on previously observed wrinkling regions might lead to a transfer of compression forces and results in different extents of wrinkling emerging in other regions of the textile. Therefore, this design is a non-trivial process that requires a more precise understanding of the compressive stresses within the textile through mechanical simulations to enable more exact mesh designs. Such an application will be investigated in depth in future studies.
3.2. SPRAYING STUDIES After shaping, we spray the curved textile panels with polyester laminating resin mixed with hardener catalyst. This step uses a pneumatic air-gun to coat the resin directly onto the textile’s surface with a low intensity of spray and the resin is uniformly distributed using an aluminium spring roller. The formwork is left to cure at room temperature for approximately 12 hours until the surface is no longer tacky. We repeat this process to progressively build up the thickness of the perforated TRC panel until it gains stiffness to the degree that it does not bend under its self-weight.
3.2.1. Spraying Results From the spraying attempts, the textile needs to be sprayed with an initial thin layer of resin first to solidify the geometry before spraying more layers. This step helps to prevent excessive resin from accumulating on the underside of the surface, which might lead to undesirable bumps of resin solidifying or clogging of THE FLEXIBLE TEXTILE MESH 357 the mesh perforations. Additionally, orienting the frame vertically is preferable to ensure that both sides can be sprayed evenly with resin. A total of three sprayed layers are needed to form a stiff TRC panel which does not deform under its own self-weight. The panel solidifies to a depth of 1.5mm and can be increased by 0.25mm in thickness for subsequent spraying attempts. It is also stiff enough to withstand drilling for attaching mechanical fixtures. The initial surface finish is of a mildly rough texture, which can be polished by wet-sanding of the surface to achieve a smooth surface texture.
Figure 10. Perforated TRC Panels.
4. Conclusion and Future Work This paper presents an alternative method to fabricate perforated screens using edge-shaping and spray casting of a textile mesh formwork. It is also a prelude study of the textile’s behaviour during edge shaping of the formwork and gives an initial insight into how stitch density can be designed to mitigate geometric inaccuracies incurred by wrinkling. Furthermore, it documents the spraying experiments to maintain a uniformly smooth surface while ensuring porosity of the mesh. Nonetheless, we recognise that certain challenges are needed to be addressed, particularly dealing with the textile formworks at larger scales. These will be implemented in future research, which includes upscaling the fabrication process and documenting the scalability of material behaviour of the graded textile formworks. Moreover, our research also seeks to develop a more precise method of designing graded mesh patterns through the realization of a computational framework. This could be done through mechanical simulations (eg. finite element methods) to determine the regions of compression forces when the textile is shaped along its edges. Thus, this informs the placement and sizing of the knitted mesh patterns to negate these geometric deformations. Acknowledgements: This research is supported by the Computational Framework for Functional Textiles project funded by SUTD Digital Manufacturing and Design (DManD) Centre, supported by the Singapore 358 Y.Y. TAN AND T.L. LEE
National Research Foundation.
References “Shima Seiki CNC Knitting” : 2017. Available from