Ellis Developments Limited

Nottinghamshire, United Kingdom

Embroidery for Engineering and Surgery

Presentation given at the Textile Institute World Conference in Manchester 2000


Julian G Ellis M.Phil, C.Text, FTI, MRSC, MAE


This paper first describes research and development work carried out by a collaborative team to investigate the potential of embroidery as a method of making textile preforms for structural composites and then demonstrates its use for surgical implant manufacture

Embroidery of composite textile preforms allows the accurate placement of fibres to produce the optimum fibre architecture of the finished reinforcement, leading to reduced manufacturing waste, lighter components (hence lower fuel consumption of vehicles), and consequent environmental benefits. In a major project, work was carried out to produce demonstration components in the form of a spacesaver wheel for a British luxury car manufacturer, a bearing housing for an aircraft component, and reinforcement for an automotive floor pan. The project included an underlying scientific component to examine the flow of resin through embroidered structures, and analytical approaches to the design of components. Follow-up work has examined the reinforcement of structures though the thickness to improve the interlaminar properties of finished components.

The programme of work on embroidery has extended into the development of a series of textile surgical implants. This is also described. In this second programme a number of devices, including a repair system for the minimally invasive treatment of potentially fatal abdominal aortic aneurysms, have been developed.

The manufacture of textile preforms for fibre reinforced composites remains a major difficulty in the industrialisation of composite manufacturing techniques based upon liquid moulding. The assembly by hand of reinforcement preforms is labour intensive and, for complex parts, tends to be very wasteful of raw materials. As an alternative to conventional fabric based technologies, two standard embroidery techniques, Cornely and Schiffli were assessed for suitability in laying down high modulus reinforcement fibres in near net shaped preforms.

The close control over fibre architecture offered by embroidery is potentially attractive for highly loaded structures, enabling fibres to be placed in the position and with orientations necessary to optimise strength and stiffness locally, while waste fibres. Labour costs are also reduced considerably.

As far as we are able to determine, the first work done using embroidery for structural composites was carried out in America by Xerkon Company who purchased a number of embroidery machines for producing structures in relation to the stealth bomber programme in the USA. Understandably, very little has been published of the work apart from a single patent (1). There has also been some work carried out on embroidery in Japan which led to the publication of a Japanese patent there (2). About the same time as our work programme commenced, a programme of work in a slightly different direction was started in Dresden, which work is ongoing (3).

Initial work on embroidery has been carried out for wound dressings by the ETH Institute in Switzerland (4). This was published at the 1999 Medical Textiles Conference in Bolton.

Ellis Developments Limited formed a consortium over a two year period which led to the setting up of a major Government supported research project under the LINK Structural Composites programme. This was 50% funded by the Department of Trade and Industry and the EPSRC. This work commenced in 1993 and consisted of a consortium of 10 companies and one University Department. The companies comprised two fibre manufacturers, two with different types of embroidery equipment, an embroidery machinery company, a textile development house, a resin manufacturer and two potential end users. The project had a value of some £700,000 (5).

The objective of the project was to investigate the use of embroidery techniques for the manufacture of reinforcement preforms for composites structures produced by liquid moulding.

Three technical demonstrators were developed: an automotive spacesaver wheel, demonstrating reduced weight over its steel equivalent; a generator drive end frame for an aircraft, demonstrating assembly time and waste reductions, plus weight saving over the conventional alloy part; and a patch reinforcement to strengthen the safety belt anchorage points on a prototype composite automotive floorpan.

The project assessed two different embroidery fibre placement techniques, and examined different embroidery substrate materials, fibre and yarn placement, Cornely and Schiffli and a selection of suitable materials. The characterisation of preforms and the design of components were studied in relation to the products together with an examination of the effect of fibre architecture on the structural performance.

2 Embroidery Techniques

Several machine embroidery techniques exist. The two studied were those considered to be the most appropriate for fibre reinforcement positioning of composites and that they are both able to lay fairly straight fibre configurations. Cornely embroidery uses a single needle head with a substrate material held in a pantograph, which is moved under computer control. One application of this technique in the garment industry is the tacking of heavy cords to substrates by coiling the cord in a wrapping yarn, then stitching the wrapping yarn to the substrate with a chain stitch. Schiffli embroidery uses rows of needle held on a horizontal rack, with a substrate material mounted on a vertically oriented pantograph. The primary yarn is passed through the thickness of the substrate and held in place by second interlocking yarn at the rear of the work.

In our studies embroidery technology using a multi-head lockstitch machine was not included. However, the stitching is similar to that produced by Schiffli embroidery machines, but does not have the same capacity for very high volume manufacture, as required by automotive manufacturers.

Both processes were used for manufacture of local reinforcement for conventional preforms (patches) and for manufacture of complete, near net-shape preforms. The project addressed the design, processing and performance of composite structures using these techniques covering the following major areas:

The Cornely and Schiffli embroidery processes were assessed using several reinforcement styles, stitching yarns and substrate materials. Glass, carbon and aramid fibres were tested as primary yarns, with polyester, glass and aramid as backing and stitching yarns. For the Cornely technique, reinforcement tows of linear density of between 600 and 2,400 tex of intermediate modulus carbon or E-glass were found to be suitable.

For the Schiffli technique, only aramid (Kevlar 29) primary stitching yarns were used successfully. Glass and carbon yarn broke easily due to the small turn radii required. Polymer fabric substrates performed well, glass fabric and pre-consolidated maps also perform adequately, although the latter tended to reduce needle latitude due to the abrasive nature of the fibres.

A systematic study of the effect of several of the Cornely embroidery parameters was made using Taguchi methods. Four parameters were studied, each at two levels and the results of permeability and mechanical property tests were used to optimise the embroidery process. As part of the study, a novel in-plane permeability test method was developed in order to devise rules for the rapid characterisation of materials and to provide an understanding of problems related to slow impregnation of areas during resin injection by the blocking of longitudinal flow channels by the stitching yarns. An evaluation of the effect of embroidery parameters on compaction into a mould showed the most significant parameters are for linear density and fibre diameter, with a higher linear density and larger filament diameter leading to a more compliant reinforcement. The linear density of the roving is important, since as in the case of permeability, this is accompanied by a reduction in the proportion of stitching yarn and a higher porosity for given reinforcement fraction. Thus, the stitching yarns play an important part in determining the achievable reinforcement content in a laminate. Larger filament diameters increased compliance, which may be due to the reduced filament count and lower internal friction for a given linear density.

Embroidered materials and control specimens from conventional fabrics were vacuum impregnated using unsaturated polyester resin to provide tensile test pieces. The result showed roving linear density to have the largest influence upon the textile modulus of quasi-unidirectional embroidered plaques with higher linear density yielding high modulus. Again, this was attributed to the lower proportion of non-reinforcing stitching yarns. For ultimate tensile strength, filament diameter appeared to be the most influential factor, with larger filaments yielding higher UTS, which reverses the usual trend. This effect was attributed to the superior processing characteristics of the preforms made from 17l fibres (6).

3. Design Analysis

The fibre orientation for the flat test specimens and the demonstrator components were designed with the aid of finite element analysis. The constantly changing fibre path, and therefore the orthotropic material properties, cannot be defined using conventional FE modelling techniques. In the present work, the structure was discretised by first aligning orthotropic elements with the local fibre direction, and then by consideration of the overall geometry.

Typically, this resulted in much finer meshes than will be required for isotropic analysis. Obtaining a satisfactory structure therefore became an iterative process:

Strength was estimated by a factor of safety on first ply failure indicated using the Tsai-Wu failure criteria. This required some re-meshing every time the fibre piles were altered.

4 Demonstration Components

4.1 Automotive spacesaver wheel.

Spacesaver wheels are fitted by an increasing number of manufacturers in order to reduce mass and cost as well as providing an increase luggage capacity space. The main body of a wheel for a Jaguar was designed based upon quasi-isotropic laminate produced in zero crimp carbon fibre fabric. A static design based on the bending fatigue requirement resulted in the hub of simple disc form and a weight saving of 56% compared with the steel version of the wheel. Cornely embroidered carbon patches were stitched using aramid yarn for the main body of the wheel, the patches being located on the two outer surfaces in the vicinity of the hub with the fibres aligned to inhibit crack propagation and to increase local stiffness of the highly stressed attachment points.

A design iteration was performed for which the regions of low stress in the hub were cut out to provide a spoke form, with local reinforcements surrounding the cut-outs and, as before, the bolt holes. The projected weight saving for this version in carbon was 63%.

4.2 Generator drive end frames (250 mm diameter).

The objective of this prototype study was to reduce the waste fabric generated during the preform manufacture and to reduce the preform assembly time while matching or improving the performance and reducing the weigh compared with the conventional aluminium alloy part. The first design iteration was developed using a quasi-isotropic laminate which met all the performance requirements with a 30% weight saving and a predicted factor safety of 2.2. A second iteration, in which the hub was designed with predominantly circumferential fibres with radially reinforced spokes and quasi-isotropic outer flange, was analysed with a predicted safety factor of 5.9.

The preforms for the second design iteration were embroidered using a Cornely machine with 1,200 tex glass tows on 200g/m2 plain-woven substrate. Manufacture involved producing multiple layers of fibres laid according to the FEA predictions. The near net- shaped potential for embroidery permitted the elimination of 55% waste fibre compare with conventional fabric, and the assembly time of fibre in the mould was very considerably reduced to about one eighth

The techniques were developed and are available for commercial take- up. This is inhibited by the fact that the design and manufacture of the parts must be technically co-ordinated, because the high structural efficiency obtained by this method is wasted if design and manufacture are not integrated. This approach does not suit all commercial organisations.

5. Through Thickness Stitching

Following the development of the drive end frame, it became clear that through thickness stitching would further simplify and speed up the manufacturing process.

Our initial interest in through stitching was developed through a study known as ZARC (Z-axis reinforcement of composites) carried out under a European Craft project in conjunction with Teritex of Nottinghamshire and the University of Nottingham. Many two-dimensional laminates have poor resistance to delamination, cracking under an impact loading because of their poor inter-laminar fracture toughness. As a consequence post-impact in-plane mechanical properties can be severely degraded, particularly compression strength and fatigue performance. Much of the work in stitched composites has concentrated on developing information to develop engineering criteria for use in future composites structures.

Much work on through stitching carried out by others has used simple lockstitch machinery. Our own studies have used conventional Barudan embroidery machines, which have the clear advantage of computerised control of stitching patterns. Designs and alterations to designs on simple grid stitching can be carried out extremely quickly using appropriate software, and the easily controllable variations of speed of stitching on an embroidery machine means that optimum stitching speeds can quickly be defined.

Heretofore, there has been a limitation in the size of the preform which can be completed using embroidery machinery, but the recent introduction of a very large embroidery bridges has resulted in a possibility of producing 2 metre square stitch panels without any need to rearrange the embroidery frames or the work within them. Larger structures can also be accommodated. There have been a number of studies of through stitching carried out, but the methodology of many of these have been severely criticised in a recent studies by Mouritz and Cox (7,8).

A considerable amount of data, for example tensile strength data, has been published for stitched laminates. As with other mechanisms, the findings appear contradictory, for example different authors have found that stitching (a) improves, (b) degrades or (c) does not alter significantly, the textile strength.

Although the cause of the variance has not been identified, it is suggested that some of the variations are caused by a failure of some workers to understand the complexities of textile structures, and do not tend to take into consideration some of the variables that can be designed into or out of textile structures. For example, a closely woven fully set fabric which has stitching yarn introduced through it will inevitably have distortions put in it by over-filling the fabric.

Conversely, if laminates with appropriate spacing between the fibre bundles is through-stitched properly, a highly packed structure will be produced. This will result in little damage to the in-plane properties and the through thickness properties reaching an optimum.

Mouritz and Cox similarly criticised the fact that relatively few authors have published details of fibre volume fraction changes due to stitching. They comment on the uncritical style in which data has been reported and the absence of observations of mechanisms during tests reduces discussions of mechanisms to speculation.

Unfortunately, research into textile composite structures has largely been engineering led rather than textile industry led. This is a failure by the textile industry to invest long term in the development of new textile products. The reason behind this is probably related to the long investment time before new developments in engineering structures can take to reach the market. Aerospace lead times are high and may well exceed 5 years.

I suggest that the manufacture of complex stitched structures using complex textile processes is probably best taken not from academic and engineering perspectives but through developing specific structures from the bottom-up using a small team of specialist engineers and textile scientists.

6. Surgical Implants

The manufacture of textile surgical implants often requires the production of low numbers of items. Modern textile manufacturing methods, however, are usually only cost effective if high volumes are produced.

It is, clearly desirable that textile implants should be able to be made cost effectively in small numbers on machinery that is mechanically or electronically controlled in such a way that each item produced to the same design will be virtually identical. It is also desirable that the design method used should be simple and quick, to minimise design cost. Low design costs also facilitate the cost effective production of bespoke implants, which may be desirable for unusual medical conditions or for use in patients where it is necessary to produce an implant of the exact size to fit that patient.

Although surgical implants have been made for many years using woven techniques, there are two difficulties, arising from the nature of woven fabric construction. The first of these is that the fibres, in general terms, lay at right angles to each other; the other is that the fabric frays if cut during surgical procedures. This can give rise to device failure.

Knitted fabrics, by their very nature, tend to be much more deformable and are, in many cases, totally unsuitable for use as load bearing textile structures within the body. The exception is the use of warp knitted structures with laid in warp and weft to provide dimensional stability in a quasi-woven construction. Again, there are difficulties related firstly to the fact that the setting up of machine is expensive, and secondly the very low volume required; textile implants are invariably small.

Ideally a textile implant should have its fibres placed in a position and direction which accord with the design requirements in order that they may carry out their function correctly, whether it be load-bearing or otherwise. Embroidery suits this perfectly. Because of the nature of an embroidered structure, the fabric can also be designed in such a way that even when cut as part of a surgical procedure, the finished structure will not fray or otherwise fall apart (9,10).

Although by definition, embroidery must be made upon a base cloth, the stitching can be carried out on a base cloth that is soluble or biodegradable so that after the implant is formed the base cloth can be dissolved away and only the embroidered stitched structure remains.

In a collaboration with a suture thread manufacturer, we have developed a number of specific textile structures using polyester suture threads embroidered onto a soluble base cloth. We have used this technology to manufacture for example, a patch for incisional hernia. In this condition internal tissues of the gut protrude through the wall of the abdomen. In such cases the internal tissues are only held within the body by a thin layer of tissue. The edges of the hernia are weak or friable, which makes repair using conventional sutures difficult. The provision of a patch assists the surgeon to obtain a good repair by enabling him to stitch the implant to a site remote from the weak edges. The patch can also provide a scaffold on which new tissues may grow.

The requirement specification for a patch includes that it shall have a high bursting strength, that it should be easily cut to size appropriate to the condition of the patient, without fraying or unroving from the cut edges. Although many woven structures have good properties as regards bursting strength, they will fray readily from cut edges. The use of a hot knife to cut thermoplastic fibres to prevent then fraying is undesirable because of the inconvenience of carrying this out in operating theatre conditions. Sharp edges (and possible toxic degradation products) can be left by thermal cutting. If the operative technique facilitates the assessment of the hernia dimensions, the dimensions of the implant can be passed to the implant designer and a customised implant made precisely to the required dimensions. Furthermore, should the operative procedure require amendment after the operation has commenced, the implant can still be cut to the correct shape, or a backup large patch provided. Flexibility is something that is highly desired by surgeons, in case pre-operative assumptions turn out to be incorrect.

In spinal surgery the replacement of intervertebral discs has previously been difficult. Replacement discs of silicone have often extruded from the intervertebral space by the movement of the spine. This is particularly so in the cervical spine; the neck is a part of the body that is only still when the patient is asleep. This problem has now been overcome by the use of an embroidered textile which can be wrapped around a shaped silicone block. The embroidered textile contains reinforced holes, which accept bone screws for fixation to the vertebra. The design of the textile mimics closely the shape and form of the natural ligaments of the cervical spine; extended in vitro trials have demonstrated that this implant is capable of withstanding some 10 million load and unload cycles (11).

Our major collaborative research programme, under the LINK scheme in the United Kingdom, partially supported by the Department of Health, has developed a graft stent for the repair of an abdominal aortic aneurysm graft stent. An aneurysm is a balloon-like swelling in the wall of an artery, and is the result of a weakening in the structure of the wall of that artery. The reason for this weakening may include genetic predisposition, trauma and atherosclerosis. The aneurysm gradually increases in size, usually over several years, and may eventually rupture often fatally. In the United Kingdom, approximately 10,000 people die annually as a result of ruptured or severely haemorrhaging abdominal aortic aneurysms. Conventional open repair using vascular textile grafts have been used for many years, and textiles used manufactured from polyester or expanded PTFE by weaving or knitting. They are available in a wide range of diameters and lengths and bifurcated designs are available for use in the abdominal aorta where it bifurcates into the iliac arteries.

The conventional repair system for an aneurysm is to bypass the diseased or distended section of an artery, to remove the aneurysmal sac from the circulation. The repair involves opening the abdomen, and accessing the aorta, opening the anterior wall of the aorta and clearing any debris from the region. A textile tube or graft is used to bridge the aneurysm. The top and bottom ends of the graft are sutured in place, and the remaining flaps are sewn together. The technique has a high level of trauma to the patient, and may not be suitable for many patients who have a low level of health or suffer from a number of conditions, which will exclude them from open repair. Therefore endovascular (keyhole) techniques offer significant potential benefits in the reduction of mortality and morbidity rates as well as having potential cost saving. The endovascular approach requires some method of fixing the graft to the vessel wall, which replaces the open suturing technique. It also requires that the graft can be compressed to fit inside a delivery catheter, and the catheter inside the blood vessel during introduction. Subsequently, the graft must expand to the final desired size. Usually a 7 mm-diameter catheter is used and a graft, which expands to some 34 mm in diameter, must be inserted through that catheter. This latter characteristic of expansion has been achieved by combining conventional vascular grafts with metal stents to give a hybrid product, a graft-stent.

There have been a number of difficulties with current designs of endovascular grafts. The design specification includes a requirement that the graft stent must not kink (and hence occlude blood flow) even in tortuous blood vessels. The device must also seal against the vessel wall to prevent blood leakage around the ends of the device, an endoleak. Also the device must provide a fixation system that prevents the device migrating. Our collaborative consortium have developed a design consisting of a tubular construction, with a polyester microfibre fabric forming a base, over which a series of Nitinol shape memory superelastic wire-form rings are attached by embroidery. A number of barbs are attached to the proximal ends of the device, forming an annulus. The use of Nitinol and its shape-memory condition means that the device can be fed down a relatively small catheter, and deploy into a lumen without any deleterious effect. The Nitinol provides a significant radial force that forms an efficient seal at the proximal and distal end of the device. This, at the same time, permits a certain degree of flexibility when sizing the device pre-operatively. The same rings of Nitinol, interspersed with the flexible fabric, provide a device that is highly kink resistant. The fixation barbs, again manufactured from Nitinol, withstand feeding through a catheter. Early indications are that a high level of fixation is achieved. The device, when emerging from the catheter, self expands upon contact with the blood and tissue (12,13,14).

The early results of our pre-clinical studies are encouraging, and we expect that on the completion of the long-term evaluation, the new graft stent will prove successful and ready to be introduced into clinical trials.


Link Mascet Collaborators: Dr C D Rudd, Dr N Warrior, Mr S P Gardner, Mr D Morris, University of Nottingham, Department of Mechanical Engineering, Ford Motor Company, Lucas Applied Technology Ltd, Hewetson-Leveaux Ltd, Cray Valley Total Ltd, PPG Industries UK Ltd, Tenax Fibres Ltd, Crescent Consultants Ltd, Jaguar Cars Ltd, Vidhani Brothers Ltd.

Site MedLINK Collaborators: Prof. B R Hopkinson, Department of Vascular Surgery, University of Nottingham, Pearsalls Implants, Taunton, Anson Medical Ltd, Didcot.

Department of Trade and Industry and Department of Health who provided financial support for this work


Julian Ellis will be delighted to hear from you. Telephone on +44 (0) 7976  425899

email: info@ellisdev.co.uk We are based at Far Close, Rolleston Road, Fiskerton, Southwell, Nottinghamshire, NG25 0UJ, United Kingdom

Note: although this is not a recent paper, the principles still apply and new equipment renders this technique more commercially viable than ever.