7m ago
1.27 MB
5 Pages

AUTEX Research Journal, Vol. 12, No2, June 2012 AUTEXOPTIMISATION OF THE WARP YARN TENSION ON A WARP KNITTING MACHINEVivienne Pohlen, Andreas Schnabel, Florian Neumann, Thomas GriesInstitut für Textiltechnik der RWTH Aachen University, GermanyOtto-Blumenthal-Str. 1, D-52074 Aachen, Phone: 49 (0)241 80 23462, Fax: 49 (0)241 80 22422,E-Mail: [email protected], gations (calculations) based on a warp yarn tension analysis on a warp knitting machine with multiaxialweft yarn insertion allow prospective reduced yarn tension differences in technical warp knits. From this a futureopportunity is provided to substitute the subjective warp let-off adjustment by a model of tension control. Theoutcome of this is a higher reproducibility with associated increasing process reliability and rising productquality.Key words:Multiaxial fabric, non-crimp fabric (NCF), warp knitting machine, warp yarn tension, yarn tension control.Initial situationwarp knitted glass or carbon fibre layers. These are made onwarp knitting machines with multiaxial weft insertion. Theseenable the production of NCFs for different applications byadjusting several parameters.Increasing political pressure is demanding that industry nowalways produces environmentally sound products. Theautomobile sector is developing more powerful, lighter electricvehicles (e.g. Megacity Vehicle, MCV). The MCV of BMW incooperation with SGL Automotive Carbon Fibers is based upona composite construction of steel, aluminium, and CFRP(carbon fibre reinforced plastic) of 40% [1]. The aviation andaerospace industry is working on lighter aircraft which enablelower operating costs and result in reduced fuel and CO2emissions. The new Airbus A350 will consist of approximately55% CFRP. Materials such as CFRP or GFRP (glass fibrereinforced plastics) provide light weight with low density (2030% lighter than aluminium), and high stiffness while providingstrength and durability [2]. High-tech fabrics can significantlycontribute to weight reduction. These benefits involve highcosts: expensive raw materials (especially carbon fibre) andthe high cost of manual production. Automated systems forcost-effective mass production would ensure a consistentproduct quality. This goal and a more homogeneous NCFprocessing (non-crimp fabric, warp knitted multiaxial layers)are dealt with in this present work. It describes the approach,compensating the fluctuating warp tension on a warp knittingmachine with multiaxial weft insertion. This is importantbecause different tensions have different effects on the NCFquality, among others, poorer resin permeability, and lowermaximum tensile strength initiated, for instance, by strongerundulations.A detailed study regarding various adjustable parameters onsuch a machine has not occurred in previous works. A basisfor this is provided by warp tension studies on conventionalwarp knitting machines. Research projects at the ITM of theTU Dresden showed maximum tensile forces with peakamplitudes of more than 50 cN per warp yarn. The maximaoccurred at the point in time when the guide needle does theoverlap in the knitting process [4] [5]. J. Hausding conductedresearch with the Malimo 14022 P2-2S bonding machine chinenfabrik GmbH, Chemnitz) to produce a biaxialstitch-bonded fibre layer regarding an extended warp knittingprocess. He studied, amongst others, the warp yarn tension,but only with respect to their maxima averaged without regardto the tension curve during one-stitch formation [6]. The resultsshow that these values are not comparable with those in thiswork. In terms of technical progress, it is necessary toincorporate here other parameters or variables and the exactwarp yarn tension curve during the one-stitch formation cycle.Currently, the operator adjusts the already stored positions viathe CNC machine panel. Parameters such as the mass perunit area of glass layers or the warp yarn tension can be addedsubjectively and vary depending on experience. The operatorfeels the warp yarn tension whilst the machine is running.Since this varies within a stitch formation, an accurateassessment is not possible, mainly because the operatorperceives the stress peaks. Thus, the human intervention inthe repetition of positions leads to non-identical products.In the future, a control system may provide for almost constantwarp yarn tension, resulting in a reproducibility of the productsto ensure consistent quality.State of technology for the production of NCFsExperimental procedure: measurement of theyarn tensionTextile products are technically versatile. For example, in theA350 the fuselage structure and the engine fixtures are madeof CFRP. The fuselage is exposed to lateral or vertical gustsand the undercarriage manoeuvres with a constantly varyingload. Therefore, multiaxial fabric is used, which allows variablereinforcement effects [3].In this study, glass fibre rovings (600 tex) were warp-knittedwith polyester fibres (175 dtex) on a Copcentra MAX 3 CNC(LIBA Maschinenfabrik GmbH, Naila). The two active weftcarriage systems laid the roving into the transport chain at anangle of 45 .The manufacturing and processing of NCF currently requiresa high level of manual operational action. Generally NCFs areTo measure the warp yarn tension, in total six piezoelectricyarn tension sensors (TS 44/A100, BTSR International S.p.A.,http://www.autexrj.com/No2-2012/0006 12.pdf29

AUTEX Research Journal, Vol. 12, No2, June 2012 AUTEX2. Main shaft angle 20 : Falling yarn tension – The compoundneedles rise gradually and release more yarn. Yarn is deliveredcontinuously.Olgiate Olona, Italy; Figure 1) were mounted along the machinewidth, above and below the yarn tensioning device. The sameyarn ran through two sensors. For the purpose of mapping theextremely fluctuating tension amplitude, it was necessary todetermine the exact time of its origin during the warp knitting.Two additional LVDT-displacement sensors (WA / 200 ofHottinger Baldwin Messtechnik GmbH, Darmstadt) measuredthe movement of the guide bar.3. Main shaft angle 160 : Re-falling yarn tension – Overlap ofguide needles takes place and expands slightly more than theyarn delivered.4. Main shaft angle 300 : Extreme rise of yarn tension – Thecompound needles sink, begin to pull down the yarns andeach yarn is laid down on the sinkers.Warp yarn5. Main shaft angle 345 : Tension drops abruptly – The stillsinking compound needles pull the yarns. Meanwhile, thefabric take-up pulls the warp knitted glass layers, which forcesthe stretched yarns to fall off the sinkers. This briefly causesloose yarn tension.6. Main shaft angle 350 : Immediate re-rise of tension – Thecompound needles finish sinking and tighten the yarns, butwill not exceed the previous amplitude due to generated yarnstorage on the sinkers which were released by dropping.Depending on the machine parameters set, differentcharacteristic tension curves result. On the one hand, theamplitudes increase or decrease, on the other hand a numberof intermediate or no amplitudes are visible. The amplitudelevel and the different runs depend on the intensity and type ofparameters selected.Yarn guide12 mm15 mmOffset switchFastenerFigure 1. Piezoelectric yarn tension sensor with fastener.Influence of parameters on the warp tensionThe following parameters with a foreseeable major influenceon the tension were realised in the trials: pattern, stitch length[mm], glass layer mass per unit area [g / m] and warp let-off[mm / rack1] (amount of delivered yarn per main shaft revolutionor stitch course, see Table 1). The material composition of thewarp yarns and the running-in layers were kept constantthroughout the entire experiment.As mentioned already, the greatest yarn tension is initiated bythe sinking movement of the compound needles (cast-off –starting underlap). This depends on the set position.Alternating machine speed has little effect on the yarn tension,which has a linear relationship to the machine speedaccelerated yarn delivery. By contrast, pattern, stitch length,mass per unit area and warp let-off have distinctive impacts.With increasing underlap (pillar to satin), the average yarntension increases (Fig. 3). Throughout the underlap, the guideneedle movement takes place at large distances, such assatin, with higher acceleration than at short distances. Thisdifference is not considered in the yarn supply, since the warpbeam drive does not respond within a stitch course. The yarntension increases with smaller stitch length (2 mm) by trend.The reason is that the yarn supply is much lower for the entirewarp knitting process as compared to a large stitch length(6 mm). For the moment of different stitch length formation, for6 mm stitch length a larger amount of yarn is available than for2 mm. An increase of the mass per unit area of the in-runningglass layers causes an increase in average yarn tension dueto the increased friction between the warp yarns and the thickerglass layer.Table 1. Experimental parametersParameterPattern (determinesunderlap length)Stitch length, mmMachine speed, U / minMass per unit area, g / m²Warp let-off, mm / racksubjectively setLevel 1Level 2Level 135250-1200“Max”(loose)“Medium”Evaluation and processing of the measurementseriesAll the sensors’ measured values were recorded usingLabVIEW software (National Instruments Germany GmbH,Munich). Analogue or digital filters were not used in order notto distort the useful signal. The series of measurements of allthe sensors were processed using Fourier analysis. Thisprocess enabled the elimination of high-frequencydisturbances such as machine vibrations. Thus the actualuseful signal could be better evaluated.A separate parameter in the experiments is the warp let-offvalue. For each new position, the value was set subjectively bythe operator. Subsequently, the individual parameters are tobe regarded as tendencies (see outliers in Fig. 3). The positionchange from tight (min) to loose (max) set shows significantdifferences in tension. The tension minima for loose settingsreach almost 0 cN, which may lead to process problems: Theyarns cannot be guided accurately enough around the warpknitting elements. The same applies to tight settings, whichlead to machine shutdown triggered by the yarn tensioningdevice and stress the warp yarn. Various warp let-off valuesare clearly perceivable in the textile appearance and lead todifferent displacement strengths of the glass layers. The linearyarn tension curve by trend (according to Fig. 3) assists in theFigure 2 shows the run of warp tension in dependence on themain shaft angle at a speed of 250 U / min. One stitch courseis formed during one main shaft revolution (360 ,corresponding to a duration of 0.24 sec per revolution). Strikingpoints for the different yarn tensions are:1. Main shaft angle 0 : Maximum yarn tension occurs (in thiscase about 11 cN) – The compound needles are in the lowestposition and pull more yarn as delivered.1Rack equals 480 main shaft revolutions or stitch courses.http://www.autexrj.com/No2-2012/0006 12.pdf30

AUTEX Research Journal, Vol. 12, No2, June 2012 AUTEXWarp yarn tensionTricot - 250 rpm - 6 mm - 600 g/m²Mean value originalMean value 6. orderGuide bar swingingGuide bar shogging53th period moving mean8201860 0 120 195 255 315 7345 5145Yarn tension [cN]1124261038261043Guide bar movement [V] respectively [cm]6164-12Approximatedneedle 600640680720Machine main shaft angle [ ]Figure 2. Warp yarn tension (Comparison of original vs. 6th Fourier series and moving mean considering guide bar movement).Average yarn tensionPillarPattern type - Stitch length - Mass per unit area - Warp let-off valueYarn tension regarding 6 mm stitch length [cN]600 g / m²TricotSatin1200 g / arn tension reagrding 2 mm stitch length [cN]Yarn tension regarding 2 mm stitch length [cN]Figure 3. Warp yarn tension (2 mm vs. 6 mm stitch length) regarding pattern, warp let-off and mass per unit area.Algorithm for calculating the required loop lengthin the NCF and the delivered quantity of yarn,depending on machine parametersyarn amount for the NCF, a model for determining the quantityof yarn per stitch course has been developed. The previouslyset values are based on the yarn feeding unit mm per rack.The warp let-off values are set subjectively, but do not alwayslead to optimum yarn tension. Therefore, knowledge of therequired thread length is necessary so that the extremes ofthe yarn tension can be diminished.The tension must be balanced by the warp let-off value. It isgreater, the longer the stitch length, the underlap and the greaterthe mass per unit area are. In order to deliver an optimizedBased on a previous model for loop length calculation in theconventional warp knitting process, the new model has beendeveloped by using elliptical geometrical shapes.development of a mathematical correlation for the requiredloop length for various parameters.http://www.autexrj.com/No2-2012/0006 12.pdf31

AUTEX Research Journal, Vol. 12, No2, June 2012 AUTEXThe elliptical model provides more accurate delivery values(average tension for all positions). Slight deviations with regardto the stitch length of 2 mm (see underlined values in Table 2)occur, where the calculation result is too low. The source ofthis is presumably the large displacement of the guide bar atsmall stitch length. Due to the high level of correlation with theadjusted values, the yarn Let-values serve as a reference point(Table 2).The conventional loop length calculation [7] adapted to NCFsis made up of individual components (lk, lu, ls, dg). The length iscalculated from the fractions as they are to be found in thefinished NCF. The formula for calculating the