Influence of pre-tension on ballistic impact performance of multi-layer Kevlar 49 woven fabrics for gas turbine engine containment systems

Lulu LIU,Zhenhua ZHAO,Wei CHEN,Gang LUO

Jiangsu Province Key Laboratory of Aerospace Power System,College of Energy and Power Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China

1.Introduction

High-strength woven fabrics are widely used in structural systems where high energy absorption is required.Due to their high strength per weight ratio and good ability to resist high-speed impacts,they are more weight-efficient than metal materials.One typical application is the ‘softwall” containment design in a gas turbine engine.As far as safety concerned,the containment structure in a propulsion engine is required to contain a fan blade in the rare event of blade loss during engine operation according to Federal Aviation Regulations.1For the‘softwall” containment structure,a metal case is designed to be penetrated by a releasing blade and captured by outer wrap aramid fabrics,where most of the kinetic energy of the releasing blade is consumed.2Currently,the only woven fabric widely used in engine containment systems is Kevlar 49 developed by DuPont.To develop a lightweight and reliable‘softwall” containment structure,the performance of wovenfabrics subjected to a blade-out load needs to be evaluated,which is directly related to the ‘softwall” containment case designing.

Designing a containment structure for a fan blade from an origin concept to a final certification is a really expensive and time-consuming process.Though there have been successful applications on CF6-80 and GE90 engines,there are much less data available in public and very limited experience on test/analysis correlations for fan containment systems made of fabrics.As a conventional containment test costs millions of dollars,it always follows several fan blade containment rig tests and ballistic tests on subscale and full-scale fan case models.These early impact tests can also be used to demonstrate the effects of many factors on the ballistic performance of fabrics and support design and optimization of containment systems.

Efforts to develop light-weight containment systems have been started since 1970s.Weaver3reported evaluation and development of Kevlar fabric used to contain blades released from gas turbine rotors;ballistic impact evaluations,laboratory tests,spin pit tests,and engine tests were conducted,which proved that Kevlar fabric had better impact resistance capacity than that of stainless steel.Stotler et al.4,5conducted containment tests on containment systems wrapped with dry Kevlar,and developed a relationship between the required thickness and the released blade energy.In 1990s,Pepin Associates Inc.6and Pepin7developed a Kevlar/titanium containment ring design to contain turbine engine rotor failures,which was successfully tested on a full-scale T-53 engine.Salvino and DeLucia8presented experimental results to determine design guidelines for turbine rotor fragment containment rings made of dry fabric.Le9evaluated lightweight materials for turbine engine rotor failure protection,and found that dry Kevlar performed better than Kevlar impregnated with phenolic resin.

Since 2004,under the Federal Aviation Administration’s Airworthiness Assurance Center of Excellence and with support from the Aircraft Catastrophic Failure Prevention Program,testing procedures and computational models were developed for designing and evaluating ballistic fabric turbine engine containment structures.Rajan,10Sharda,11and Bansal12et al.conducted tensile tests on Kevlar and Zylon fabrics to help understand the strength,stiffness,and energy absorption ability of fabrics.Pereira,13,14Simmons,15and Gomuc16et al.developed material models for fabrics in LSDYNA which could be used in fan blade out analysis of turbofan engines.In the second stage of this project,Rajan et al.17conducted additional tensile tests,friction tests,and static ring tests to develop a material model.They found that the damage evolution modes,the direction of yarns,the friction equations,and the contact between fabrics could all influence the numerical results.Revilock and Pereira18investigated ballistic impact test results,and found that the energy absorption per areal density was in a linear relationships with the projectile contact area instead of the projectile shape.Simmons and Shockey19and Vintilescu20modified the containment modeling program and material model to improve its accuracy and stability.In the third-stage research,Rajan et al.21,22carried out highstrain rate tensile tests and single- fiber pull-out tests to develop a material model.A micromechanical model was also developed with the real geometries of fill yarn and warp yarn.Naik23and Stahlecker24et al.introduced experimental and simulation works in developing reliable containment modeling methodologies for fan blade out containment analysis in this program.

Besides the containment application,Kevlar fabrics have been widely investigated in the ballistic defense area.It is well known that many parameters play a role in the ballistic resistance of a soft armor system,including fiber- fiber friction,fiber-projectile friction,fabric weave structure, fiber/yarn mechanical properties,projectile impact velocity,projectile impact obliquity,projectile nose geometry,and projectile material properties.25,26Under impact,the ballistic performances and energy absorption mechanisms of fabrics can vary significantly with the impact velocity.Shim et al.27showed that the energy absorption by fabrics during low-velocity impact was obviously different from that in high-velocity impact.A much larger region of creasing and stretching was produced in low-velocity conditions.Tan and Ching28observed pyramidal shape deformation in high-speed photographs of ballistic impact tests with different impact velocities.Cheeseman and Bogetti26described several mechanisms responsible for perforation of a fabric,and found that projectiles broke a few yarns to form a small opening and then wedged their ways through the opening by pushing aside the remaining yarns.When testing fabric armor systems for ballistic impact,the size of the specimen and the clamping methods can be important.Shockey et al.29performed experiments to study the effects of boundary conditions on absorbed energy,and found that when targets were gripped on two edges,more energy was absorbed than that on four edges.Cepuset al.30discovered that fixed boundaries promoted a more rapid development of strain energy and an increase in strain localization,with relatively small deformation prior to complete energy absorption or failure.Zeng et al.31found that targets that were clamped on two edges could absorb more impact energy than those with all four sides clamped as well.Friction plays an important role on the impact performance of fabrics as well.The work of Lee et al.32showed that increasing the friction between a projectile and a fabric and yarns themselves hindered the mobility of the yarns and required the projectile to break more yarns,which resulted in greater energy absorption.Duan et al.33modeled fabrics with 3D elements in yarn levels,and results indicated that friction dramatically affected the local fabric structure in the impact region by hindering the lateral mobility of principal yarns.Rao et al.34modeled the impact of a rigid sphere onto a plain-weave Kevlar,and results indicated that the ballistic performance depended upon the friction,elastic modulus,and strength of yarns.While friction improves the ballistic performance by maintaining the integrity of the weave pattern,yarn properties have significant influences on the friction.

Designing a containment system involves determining the type of fabric,the number of fabric layers,fabric width required,etc.In these designing factors,the number of fabric layers is the most important,which determines the weight of the containment system.The complexity of analysis increases with the number of fabric layers due to the inclusion of new energy dissipation mechanisms and increased interactions between different plies of the fabric.Cunniff25impacted panels of Kevlar,Spectra,and Nylon,and found that the energy absorbed by spaced single plies was greater than that absorbed by layered systems.Lim et al.35impacted panels of Twaron,and found that the absorbed energy for layered systems was more than that of spaced systems for certain impact velocities and projectile shapes on the contrary.Figucia36conducted bal-listic impact tests on Kevlar fabrics with different numbers of layers,and concluded that the fabric areal density and the energy absorption were in linear relationships.Sharda et al.11found that although the peak load seemed to scale proportionally to the number of layers,dynamic responses were highly nonlinear due to the progressive mechanism of failure.Pereira and Revilock14showed that the normalized energy absorbed was relatively insensitive to the number of layers.Stotler and Coppa4conducted ballistic impact tests on subscale metalfaced/Kevlar designs,and generated a quadric correlation between the Kevlar thickness and the absorption of kinetic energy.However,ballistic impact results obtained by Pereira and Revilock13,37showed that the normalized energy absorbed was relatively insensitive to the number of layers.In addition,Lim et al.35conducted ballistic impact tests on both single-and double-layer fabrics,and found that fabrics displayed similar failure mechanisms,while the increase of energy absorption did not double when the ply number increased from one to two.Obviously,current literature hasn’t reached an agreement on the effect of the number of fabric layers.

Besides,for a wrapped containment system particularly,pre-tension should be implemented in fabrics in the wrapping process so as to make sure that fabrics are smoothly rolled around a metal ring without being loose.For a single-ply fabric system,pre-tension may give rise to initial deformation of the fabric,but for a multi-layer fabric system,interactions between fabric layers should also be taken into account besides initial deformation.38There is very limited work on ballistic impact of fabric systems warped with different pre-tension.Revilock and Pereira18performed two sets of tests,one without tension and one with a tension of 25 N,on a Zylon fabric system,and results revealed no significant difference.As only two sets of tests were conducted,more tests need to be conducted to investigate the influence of pre-tension.

In this paper,ballistic impact tests on a wrapped multilayer fabric system were carried out with a flat blade projectile for replicating the impact response during a fan blade out event.The influences of the number of Kevlar layers and pre-tension were investigated and discussed.Test results were used to analyze failure modes and energy absorption of multi-ply Kevlar fabric.The work in this paper can provide guidance for designing light-weight multi-layer fabrics containment systems.

2.Ballistic impact tests

2.1.Testing specimen

In this test,a 1490 Denier Kevlar 49 plain weave fabric with an areal density of 227.5 g/m2was used.The main properties of this fabric are listed in Table 1.Though a Kevlar fabric is capable of absorbing enormous energy through tensile and stretching,it lacks the ability to maintain its shape and provide sufficient stiffness required as casing.Thus a Kevlar fabric is always wrapped around a thin metal inner casing in a softwall containment casing.Considering the large amount of design factors of the inner casing and the complexity of the interaction between the Kevlar fabric and the inner metal casing,the influence of the inner casing was not taken into account in this paper.Only the dynamic behavior of the Kevlar fabric was investigated.

Table 1 Parameters of Kevlar 49 fabric.

Since boundary conditions can greatly affect the ballistic resistance of fabric targets,ring-shaped specimens were designed to simulate the boundary conditions of fabrics used in a lightweight fan blade containment system.The target design used Naik’s23test method for reference,but with a smaller size.A continuous Kevlar fabric of a 200 mm width was wrapped around a back-up steel ring with an outer diameter of 600 mm,a thickness of 25 mm,and a height of 200 mm,as shown in Fig.1.There was a 200 mm circumferential opening in the back-up ring,and thus the fabric at this region was stretched to a flat status due to the fabric tension.The head of the Kevlar fabric was glued to the back-up ring opposite to the opening to avoid slipping,indicated as glue site 1#in Fig.1(b).The Kevlar fabric was cut through after forming certain plies,and its end was glued to the neighboring layer of fabric,indicated as glue site 2#in Fig.1(b).The glue chosen for attaching the fabric on the metal ring is 502#glue,which is a singlecomponent instant-cure adhesive based on α-ETHYL CYANOACRYLATE.Six stiffener panels were welded to the ring circumferentially to provide radial supporting stiffness.The back-up steel ring was fixed to the ground by steel fixtures.

As Naik’s tests mainly focused on the energy absorbing capability comparison on different types of multi-layer fabric,and did not involve the effect of pre-tension exerted on fabrics.In this study,the influence of pre-tension on the ballistic performance is discussed in particular.To ensure the Kevlar fabric being continuously wrapped around the back-up ring under constant tension,a tension control winding system was designed.A schematic of the tension control winding system is shown in Fig.2.Constant tension control was achieved by feedback loop control through a tension controller,a pressure transducer,a magnetic particle brake,and a hall switch.The head of the fabric was attached on the metal ring firstly through 502#glue.The glue site was pressed under certain pressure for 10 min and placed for another 24 h to guarantee fully cured.Then the metal ring was clamped on the tension control winding system as shown in Fig.2.In the fabric winding process,the tension controller converts the pressure from the transducer to tension,and compares it with the set tension value.Meanwhile,the winding speed read from the hall switch is transmitted to the tension controller.If the fabric tension is higher than the set value,the magnetic particle brake receives signals from the tension controller to speed up unreeling.Otherwise,it slows down unreeling until the tension in the fabric reaches the set value.After certain plies were wrapped on the metal ring,502#glue was used at glue site 2#to attach the outmost layer on the inner layer,and the tension force was maintained in the meanwhile.Gluing the end of the fabric used the same method as gluing the head.The completed specimen was cut off from the redundant fabric strips after 24 h curing to ensure the pre-tension force not changed.

Fig.1 Testing specimen with multi-layer fabrics being wrapped around back-up ring.

Fig.2 Photo and schematic of tension control winding system.

Projectiles of different geometries perforating single-ply fabric were investigated by Tan and Khoo,39and test results showed a quantitative distinction on the amount of energy absorbed by the fabric with these projectiles.For containment issues of turbofan engines,a projectile similar to a blade fragment is generally specified for energy absorption evaluation of containment systems.12–16The projectile employed in the tests is a flat titanium alloy blade with a rounded front edge,which has a similar geometric shape to Naik’s23projectile but is made in a smaller size and of a different material.The primary geometric size of the blade projectile is shown in Fig.3.It has a length of 120 mm,a width of 45 mm,and a thickness of 6 mm,with a mass of 138.52 g.

Fig.3 Geometry of projectile specimen.

Fig.4 Schematic of testing device—a single-stage gas gun.

Fig.5 Movement of projectile and inclined target ring.

2.2.Testing device

A gas gun with a 65 mm caliber used in ballistic tests can accelerate the projectile to a maximum velocity of 200 m/s,as shown in Fig.4.The projectile is supported by foam in an aluminum sabot.The sabot and the projectile are pushed and accelerated by compressive air in the gun barrel.The projectile launch velocity is determined by the air pressure.The sabot is captured at the end of the gun barrel by a sabot separator,and the projectile moves forward and impacts a target.The fragments,spallation from the target or projectile with a residual velocity,can be arrested by a fragment collector behind.

The projectile impact velocity is obtained by a velocimeter system installed between the gun muzzle and the target.When the projectile passes through the velocimeter system,laser beams are obstructed successively,generating a transformation of the voltage trigger signal.With the time interval of two trigger signals and the distance of the velocimeter,the velocity can be determined.High-speed videos with an average frame rate of 5000 frame/s were taken for all tests.These photos can provide an observation of target deformation,the projectile trajectory,and an estimation of the residual velocity of the projectile during tests.The back-up ring wrapped with the fabric is located at an inclination angle of 15°along the projectile trajectory direction to allow the projectile to pass over the front edge of the metal ring and impact the center of the Kevlar fabric,as shown in Fig.5.

3.Testing results and analysis

In the current paper,21 ballistic impact tests were conducted to characterize the ballistic performance of woven fabrics and identify the influences of the number of layers and fabric tension.Factors that influence the impact performance of warped fabrics are more complicated,such as the blade impact angle,boundary conditions,and so on,which are not discussed in this paper.

3.1.Testing results

This paper mainly focuses on the investigation of damage modes and primary design factors for the containment system.Testing results are presented in Table 2.The tests were divided into two series:various numbers of layers including 2,4,6,and 8 layers under 25 N tension and different fabric tensions as 25,50,and 75 N with 4-layer fabric.Three different tension values used for testing(25,50,and 75 N)were achieved by the tension control winding system,and the tension action was maintained and not relaxed after the wrapping process.The residual velocity of the projectile was estimated from high-speed photographs.In an impact analysis or containment design,the ballistic performance is normally measured by energy absorption of a specified material.Energy absorption of fabric is defined as the consumption of the kinetic energy of the projectile.For cases in which projectiles rebound,it is considered as the total kinetic energy of the projectile that is absorbed by fabric targets.According to the initial velocity and the residual velocity,the energy absorption of a fabric can be calculated as

in which m denotes the projectile mass,Videnotes the initial velocity of the projectile,and Vrdenotes the residual velocity of the projectile.

3.2.Influence of impact velocity

Fig.6 illustrates the typical perforation process of the 4-layer fabric with 25 N tension impacted by the projectile with a velocity of 103.17 m/s(Test 14)taken by a high-speed camera.At the moment when the projectile contacts the innermost Kevlar fabric,longitudinal stress wave generates in primary yarns that directly contact the projectile,and propagates away from the impact point at the sound speed.The transverse motion of the primary yarns is induced.Meanwhile,the stress wave that has developed in the primary yarns propagates tothe adjacent yarns and layers through crossovers,driving more neighboring yarns move transversely with the projectile.

Table 2 Results of Kevlar 49 fabric impact tests.

Fig.6 High-speed photographs of perforation process of 4-layer Kevlar fabric,Test 14.

The impact process of a fabric mainly involves the generation of a pyramid-shaped deformation perpendicular to the fabric plane in the impact region during the perforation or rebound process.Under impact,the kinetic energy of the blade is mainly absorbed by the tension action of yarns in the pyramid deformation.The primary yarns provide the main resistance action to stop the blade penetrating while the secondary yarns in the pyramid area also contribute to part of energy absorption by deformation.If the initial energy of the blade is totally absorbed during the pyramid deformation generation process before the yarns fracture,the blade rebounds after its velocity decreases to zero.Otherwise,when the strain of the yarns in the pyramid-shaped deformation area reaches the fracture strain,these yarns rupture instantly,usually at the impact point.The blade perforates the fabric by breaking the primary yarns and forming a penetrating opening,which dissipates great kinetic energy of the projectile for a highstrength fabric with a large elongation.The blade pushes through the opening and causes fibers being pulled out from the fabric.

The damage observation suggests that there are two kinds of impact damage for fabrics,that is,global deformation and local damage,as shown in Figs.7 and 8,respectively,which were taken from Test 14 for the perforation conditions of the 4-layer fabric.From the appearance in Fig.7,global deformation mainly involves stretching of yarns in the impact region and fabric wrinkle from both sides to the impact zone along the full circumferential direction.This is because the longitudinal wave propagates at a higher velocity than that of pyramid-shaped deformation,and thus yarns far away from the impact region are also subjected to extensive deformation.

Fig.7 Global damages of 4-layer Kevlar 49 fabric,Test 14.

Local damage is mainly characterized by yarn fracture,yarn pull-out,and yarn unraveling.A yarn fracture and pullout effect can be observed from the side where the blade exits,as shown in Fig.8(a).Yarns which contact the projectile and are involved in the impact process directly are known as primary yarns,while yarns which are involved in the deformation process through the interaction action with the primary yarns are known as secondary yarns.In this study,it is deserved to be noticed that major damages occur on the primary yarns.The main damage areas present as a narrow strip area along the height direction and a long wide strip area along the circumferential direction,as shown in Fig.8(b).The blade escapes from the fabric by pushing through the approximately rectangular opening formed by yarns fracture successively.As the constraint condition is special for the containment casing system,the yarns in the height direction have a free boundary.Under impact,these yarns will be extracted from the weaving architecture and cause yarn unravelling in the narrow strip area.

In our testing method,the back-up ring wrapped with a fabric is located at an inclination angle of 15°along the projectile trajectory direction to allow the projectile to pass over the front edge of the metal ring and impact the center of the Kevlar fabric.This special design makes the projectile impacts the Kevlar fabric in an oblique angle of 15°,as shown in Fig.9.The upper edge of the blade projectile contacts the fabric first.As the fabric is free in the width direction and the head of the blade projectile has round edges,the projectile could slide upward.In addition,the stretch of the yarns above the projectile is more severe than those below the projectile.Both factors contribute to the phenomenon that the blade is closer to one of the free edges of the fabric.

Fig.8 Local damages of 4-layer Kevlar 49 fabric,Test 14.

Fig.9 Contact process of blade projectile with fabric.

3.3.Influence of number of layers

Tests were conducted for fabric systems with different layer numbers,and typical results are presented in Fig.10 respectively.2-and 4-layer fabric systems were all perforated during several tests.Similar impact velocities were selected,as shown in Figs.10(a)and(b),respectively.Unperforated conditions were selected for the 6-and 8-layer fabric systems,as shown in Figs.10(c)and(d),respectively.It is obvious that,similar to the analysis in Section 3.2,global wrinkle and stretch,as well as local damage are observed in all cases.Major damages occur on the primary yarns,which are in a ribbon pattern.For perforated cases,there are wide penetrating openings in the bulge deformation area where the blade impacts and extrudes directly.Along the blade exit direction,fractured yarns in the opening were pulled out from the free edges.For unperforated fabrics systems with superior layers,the damage pattern was similar,while the opening was narrower than the blade shape,but not enlarged by the blade penetration.Though not perforated,the fabric systems were severely damaged with few yarns corresponding to the impact point fractured.

Fig.10 Damages of fabrics with different numbers of layers with same pre-tension of 25 N.

The blade rebound process taken by the high-speed camera in Test 7 is illustrated in Fig.11.It is shown that there are two distinct stages during the impact process:the pyramid deformation generation stage and the deformation recovery stage.Before time 1.0 ms,pyramid deformation developed with the blade’s continuous extrusion.The blade kinetic energy was totally absorbed by yarns tension action.After the blade velocity decreased to zero,the blade was accelerated in the reverse direction under the elastic deformation recovery of the fabric.As the blade kinetic energy was transformed to various damage and deformation modes,not limited to the elastic deformation,the rebound velocity was much lower than the impact velocity,and time consuming in the rebound process lasted about 3 times that of the penetrating process.From the damage appearances shown in Fig.10(c)and(d),the stretching deformation at the impact point and wrinkle in the circumferential direction are not as evident as those with lesser layers and lower velocity.

The impact point at the top of the pyramid deformation area has the largest displacement.The displacement at the moment before the blade perforates the fabric or rebounds is taken as the largest displacement during the impact process,which can be obtained from high-speed photographs.The relationship between the maximum displacement along the blade impact direction and the number of fabric layers is shown in Fig.12.Fig.13 presents the photographs taken by the highspeed camera at the moment of the maximum displacement.The maximum displacement increases with the number of fabric layers and the blade initial velocity.As the failure strain almost remains unchanged for fabrics,the increasing displacement is mainly due to special boundary conditions.With an increase of the layer number,more yarns are involved in the impact event.The slide of the yarns contributes to extra displacement,which also makes a fabric system with more layers have an improved ballistic performance and energy absorption ability.

Fig.12 Relationship between maximum displacement and number of fabric layers.

3.4.Energy absorption characteristic

Energy absorption modes are related to various damage modes,such as yarn tension,yarn fracture,yarns interaction,yarn sliding,and yarn pull-out.Fig.14 shows the variation of energy absorption with the impact energy.For conditions when the blade rebounds and the fabric is not perforated,the kinetic energy of the blade is considered as being absorbed totally by the fabric.The energy absorbed by the fabric increases with the impact energy.For a complete perforation situation,when the impact velocity is close to the ballistic limit,energy absorption increases with the impact energy and then decreases instead after exceeding the scope of the ballistic limit velocity,as shown for the 2-layer fabric with 25 N pre-tension.

Fig.11 Rebound process taken by high-speed photograph,Test 7.

Fig.13 Maximum displacements of fabric systems with different numbers of layers.

Fig.14 Variation of energy absorption by fabrics with impact energy.

Fig.15 Specific energy absorption with specific impact energy.

Fig.16 Relationship between number of fabric layers and energy absorption.

One of the parameters evaluating the energy absorption ability is specific energy absorption,i.e.,the energy absorbed by a fabric per unit area.For an aero-engine fan casing,this parameter is extremely important for the weight-reduction requirement.Fig.15 shows the variation of specific energy absorption with the specific impact energy.The specific energy absorption of the 2-layer fabric ranges from 9.39 to 11.69 kJ/(g/cm2)in the perforation cases while that of the 4-layer fabric ranges from 7.53 to 9.47 kJ/(g/cm2)in the perforation cases.For 6 layers in the rebound cases,the specific energy absorption is lowest,only in the range of 4.5–6.58 kJ/(g/cm2).The potential of the 6-layer system was not fully realized because no perforation result was obtained.The specific energy absorption of the 8-layer fabric ranges from 5.8 to 8.31 kJ/(g/cm2)in the rebound cases,and reaches 9.67–9.87 kJ/(g/cm2)in the perforation cases.Generally,the 2-layer system achieves the maximum specific energy absorption,which means that it could dissipate the impact energy more effectively.It is noteworthy that the specific energy absorption decreases with the specific impact energy for the 2-layer system,which indicates that the energy absorption capability decreases when it is highly above the ballistic limit.

Fig.17 Local damage of each layer for 4-layer fabric system for a perforation case,Test 13.

Fig.16 shows the energy absorbed by fabrics with different numbers of layers under a constant tension of 25 N.For the same ply number,perforation and rebound results are plotted with different symbols.As rebound conditions underestimate the energy absorption abilities of fabrics,only perforation conditions were analyzed.It is apparent that a fabric system absorbs more energy as the number of fabric layers increases,nearly in a linear relation.With the number of layers increasing,more yarns need to be broken for the blade to pass through all fabric layers,leading to an increased amount of energy absorbed by the fabric.The only exception is the 6-layer case,in which no perforation result was obtained.Thus the current maximum energy absorption is plotted with an arrow upward,which means that the energy absorption of the 6-layer fabric could be much higher.

For a multi-ply fabric system,the innermost layer that contacts the bladefirstly is defined as the first layer,and the other layers are numbered consecutively from the innermost to the outermost.Fig.17 shows the local damage of each layer for the 4-layer fabric system after impact in Test 13.The opening in the first layer is much larger than those in the back layers.The tension and pull-out of failed yarns show a gradually rising trend from the first layer to the fourth layer.It can be inferred that yarns mainly have shear failure in the first ply and are stretched to failure in the subsequent plies when the impact velocity is beyond the ballistic limit of the fabric.

4.Influence of pre-tension

In this section,the influence of pre-tension on the ballistic performance of fabric systems was especially discussed.The specific energy absorbed by a fabric under various pre-tension is shown in Fig.18.The specific energy absorption is the energy absorption per areal density.Perforation and rebound cases are presented in different colors.The specific energy absorbed by a fabric tends to increase slightly at first and then decreases significantly with fabric pre-tension.The specific energy absorption of fabrics with 25 N pre-tension ranging from 7.53 to 9.47 kJ/(g/cm2)is in a medium level,while the specific energy absorption of fabrics with 50 N pre-tension could reach the maximum value of 10.8 kJ/(g/cm2).However,for fabrics with higher pre-tension(75 N),the specific energy absorption is lowest,ranging from 6.67 to 7.3 kJ/(g/cm2).It indicates that there exists an optimal value of pre-tension,and increasing pre-tension above the critical value cannot improve the ballistic performance of containment fabric systems,which should be considered in the design and fabrication of fabric containment systems.

Fig.18 Specific energy absorption under different pre-tension values.

For pre-tension implemented on the fabrics in the tests,the maximum average stress in the single yarn of a fabric is calculated as 5.77 MPa(when the pre-tension is 75 N).In this stage,yarns are undergoing a stretch from crimp,with the strain level below 0.01%.Since the cross-sections and deformation of yarns in the fabric indicate no significant change below a strain level of 1%,22energy variation caused by the initial deformation due to the tension as well as the stress stiffening effect of Kevlar fabrics can be ignored.The tendency of fabric energy absorption under different pre-tension may be attributed to fabric layer-layer friction and fabric-metal friction.As the contact force grows with pre-tension,inter-layer friction and metal-fabric friction increase with the contact force.17The increasing friction hinders the slippage between fabric layers and that between the fabric and the metal ring,thus decreasing the amount of energy dissipated by the fiction action.A combination of layer-layer friction and fabric-metal friction results in a reduction of energy absorption.The fabric local damage under different pre-tension in Fig.19 can also prove the friction assumption indirectly.The local damage shows that the abrasion between yarns is more obvious as pre-tension grows and the yarn pull-out area decreases.The penetrating hole caused by the blade tends to be larger with increasing pretension,and the damage in the contact area between the fabric and the metal ring increases as well.

Fig.20 presents the maximum displacement before perforation for the 4-layer fabric system.The maximum deformation is 41.13,49.11,and 54.92 mm for pre-tension values of 25,50,and 75 N,respectively.It is found that with a similar velocity,the maximum displacement increases with the pretension exerted on the wrapped fabric system.Thus the energy absorbed by the pyramid deformation increases with the pre-tension.A combination of the friction action and the fabric deformation results in firstly an increase and then a decrease of the energy absorption.As the pre-tension increases from 25 N to 50 N,the energy absorption increase caused by the fabric deformation exceeds the deduction caused by the weakened friction action.Thus the energy absorption increases slightly.The influence of the friction overrides the influence of the deformation when the pretension increases to 75 N,and thus the energy absorption decreases instead.

Fig.19 Fabric local damage under different pre-tension.

Fig.20 Maximum displacement for 4-layer fabric system with different pre-tension.

5.Conclusions

In this study,ballistic impact tests on multi-layer fabric systems with various layer numbers and pre-tension values were implemented to investigate their influences on the ballistic performance and provide guidance for containment design.Conclusions based on the tests can be obtained as follows:

(1)Under a blade projectile impact,the impact process of fabrics mainly involves the generation of a pyramid shaped deformation perpendicular to the fabric plane in the impact region during the perforation or rebound process.There are two kinds of impact damage for fabrics:global deformation and local damage.Global deformation mainly involves stretching of yarns in the impact region and fabric wrinkle from both sides to the impact zone along the full circumferential direction.Local damage is characterized by yarn fracture,yarn pull-out,and yarn unraveling.

(2)For fabric systems with different numbers of layers,global wrinkle and stretch deformation,as well as local damage,were observed in all cases.Two distinct stages,the pyramid deformation generation stage and the deformation recovery stage,werefound in blade rebound situations. The maximum displacement increases with the number of fabric layers and the blade initial velocity.A fabric system with more layers has improved ballistic performance and energy absorption ability.The specific energy absorption reaches the maximum value approximately at the ballistic limit velocity.

(3)The energy absorption increases firstly and then decreases with the pre-tension implemented on fabrics.As the inter-layer friction and metal-fabric friction increase with the pre-tension,the slippage between fabric layers and that between the fabric and the metal ring are hindered,causing a decrease of the energy dissipated by the fiction action.On the contrary,the maximum displacement and the energy absorbed by the pyramid deformation increase with the pre-tension,causing an increase of the energy absorption by the deformation.A combination of the friction action and the fabric deformation results in firstly an increase and then a decrease of the energy absorption in the range of tested tension forces.

Acknowledgements

This study was co-supported by the National Natural Science Foundation of China(No.51575262),the China Postdoctoral Science Foundation(No.2015M571754),and the Aeronautical Science Foundation of China(No.2015ZB52008).

1.FAR33.Airworthiness standards:Aircraft engines.Washington,D.C.:Federal Aviation Administration;1984.

2.Mcmillan A.Material development for fan blade containment ring.J Phys Conf Ser 2008;105(1):12–22.

3.Weaver AT.Lightweight engine containment.In:Witmer EA,editor.Workshop on an assessment of technology for turbojet engine rotor failures;1977 Mar 29–31;Cambridge,USA.Washington,D.C.:NASA;1977.

4.Stotler CL,Coppa AP.Containment of composite fan blades.Washington D.C.:NASA;1979.Report No.:NASA-CR-159544,R79AEG197.

5.Stotler CL.Development of advanced lightweight system containment—Final report.Washington,D.C.:NASA;1981.Report No.:NASA-CR-165212,R81AEG208.

6.Pepin Associates Inc.Fiber reinforced structures for small turbine engine fragment containment—Phase II.Washington,D.C.:Of fice of Aviation Research,Federal Aviation Administration,U.S.Department of Transportation;1995.Report No.:DOT/FAA/AR-99/54.

7.Pepin J.Fiber-reinforced structures for turbine engine rotor fragment containment.Washington,D.C.:Of fice of Aviation Research,Federal Aviation Administration,U.S.Department of Transportation;1999.Report No.:DOT/FAA/AR-99/6.

8.Salvino JT,DeLucia RA,Russo T.Experimental guidelines for the design of turbine rotor fragment containment rings.Washington,D.C.:U.S.Department of Transportation;1988.Report No.:DOT/FAA/CT-88/21.

9.Le DD.Evaluation of lightweight material concepts for aircraft turbine engine rotor failure protection.Washington,D.C.:U.S.Department of Transportation;1997.Report No.:DOT/FAA/AR-96/110.

10.Rajan SD,Mobasher B,Sharda J,Yanna V,Deenadaylu C,Lau D,et al.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems.Part 1:Static tests and modeling.Washington,D.C.:U.S.Department of Transportation;2004.Report No.:DOT/FAA/AR-04/40,P1.

11.Sharda J,Deenadayalu C,Mobasher B,Rajan SD.Modeling of multilayer composite fabrics for gas turbine.J Aerosp Eng 2006;19(1):38–45.

12.Bansal S,Mobasher B,Rajan SD.Development of fabric constitutive behavior for use in modeling engine fan blade-out events.J Aerosp Eng 2009;22(3):249–59.

13.Pereira JM,Revilock DM.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems.Part 2:Ballistic impact testing.Washington,D.C.:U.S.Department of Transportation;2004.Report No.:DOT/FAA/AR-04/40,P2.

14.Pereira JM,Revilock DM.Ballistic impact response of Kevlar 49 and Zylon under condition representing jet engine fan containment.J Aerosp Eng 2009;22(3):240–8.

15.Simmons J,Erlich D,Shockey D.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems.Part 3:Model development and simulation of experiments.Washington,D.C.:U.S.Department of Transportation;2004.Report No.:DOT/FAA/AR-04/40,P3.

16.Gomuc R.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems,Part 4:Model simulation for ballistic tests,engine fan blade-out,and generic engine.Washington,D.C.:U.S.Department of Transportation;2004.Report No.:DOT/FAA/AR-04/40,P4.

17.Rajan SD,Mobasher B,Sankaran S.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems,Phase II Part 1:Fabric material tests and modeling.Washington,D.C.:U.S.Department of Transportation;2009.Report No.:DOT/FAA/AR-08/37,P1.

18.Revilock DM,Pereira JM.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems,Phase II.Part 2:Ballistic impact testing.Washington,D.C.:U.S.Department of Transportation;2009.Report No.:DOT/FAA/AR-08/37,P2.

19.Simons J,Shockey D.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems,Phase II,Update to Part 3:Material model development and simulation experiments.Washington,D.C.:U.S.Department of Transportation;2006.Report No.:DOT/FAA/AR-10/22.

20.Vintilescu IV.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems,Phase II,Part 4:Model simulation for ballistic tests,engine fan blade-out,and generic engine.Washington,D.C.:U.S.Department of Transportation;2008.Report No.:DOT/FAA/AR-08/37,P4.

21.Rajan SD,Mobasher B,Stahlecker Z,Bansal S,Zhu D,Morea M,et al.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems,Phase III,Part I:Material model and numerical simulations.Washington,D.C.:U.S.Department of Transportation;2010.Report No.:DOT/FAA/AR-10/23,P1.

22.Rajan SD,Mobasher B,Stahlecker Z,Bansal S,Zhu D,Morea M,et al.Explicit finite element modeling of multilayer composite fabric for gas turbine engine containment systems,Phase III,Part 2:Arizona State University fabric material tests.Washington,D.C.:U.S.Department of Transportation;2010.Report No.:DOT/FAA/AR-10/23,P2.

23.Naik D,Sankaran S,Mobasher B,Rajan SD,Pereira JM.Development of reliable modeling methodologies for fan blade out containment analysis.Part I:Experimental studies.Int J Impact Eng 2009;36(1):1–11.

24.Stahlecker Z,Mobasher B,Rajan SD,Pereira JM.Development of reliable modeling methodologies for engine fan blade out containment analysis.Part II:Finite element analysis.Int J Impact Eng 2009;36(3):447–59.

25.Cunniff PM.An analysis of the system effects in woven fabrics under ballistic impact.Natick:U.S.Army Natick RD&E Center;1992.Report No.:AD-A273 891.

26.Cheeseman BA,Bogetti TA.Ballistic impact into fabric and compliant composite laminates.Compos Struct 2003;61(1):161–73.

27.Shim VPW,Tan VBC,Tay TE.Modeling deformation and damage characteristics of woven fabric under small projectile impact.Int J Impact Eng 1995;16(4):585–605.

28.Tan VBC,Ching TW.Computational simulation of fabric armour subjected to ballistic impacts.Int J Impact Eng 2006;32(11):1737–51.

29.Shockey DA,Erlich DC,Simons JW.Improved barriers to turbine engine fragments—Interim report III.Washington,D.C.:U.S.Department of Transportation;1999.Report No.:DOT/FAA/AR-99/8.

30.CepusE,Shahkarami A,Vaziri R,Poursartip A.Effect of boundary conditions on the ballistic response of textile structures.In:Vautrin A,editors.ICCM12 conference:12th international conference on composite materials;1999 Jul 5–9;Paris,French;1999.

31.Zeng XS,Shim VPW,Tan VBC.Influence of boundary conditions on the ballistic performance of high-strength fabric targets.Int J Impact Eng 2005;32(1–4):631–42.

32.Lee BL,Walsh TF,Won ST,Patts HM,Song JW,Mayer AH.Penetration failure mechanisms of armor-gradefiber composites under impact.J Compos Mater 2001;35(18):1605–33.

33.Duan Y,Keefe M,Bogetti TA,Cheeseman BA,Powers B.A numerical investigation of the influence of friction on energy absorption by a high-strength fabric subjected to ballistic impact.Int J Impact Eng 2006;32(8):1299–312.

34.Rao MP,Duan Y,Keefe M,Powers BM,Bogetti TA.Modeling the effects of yarn material properties and friction on the ballistic impact of a plain-weave fabric.Compos Struct 2009;89(4):556–66.

35.Lim CT,Tan VBC,Cheong CH.Perforation of high-strength double-ply fabric system by varying shaped projectiles.Int J Impact Eng 2002;27(6):577–91.

36.Figucia F.Energy absorption of Kevlar fabric under ballistic impact.Natick:Army Natick Research&Development Command;1980.Report No.:AD-A090390.

37.Revilock DM,Pereira JM.Explicit finite element analysis modeling of multilayer composite fabric for gas turbine engines containment systems,Phase II,Part 2:Ballistic impact testing.Washington,D.C.:U.S.Department of Transportation;2009.Report No.:DOT/FAA/AR-08/37,P2.

38.Tabiei A,Nilakantan G.Ballistic impact of dry woven fabric composites:A review.Appl Mech Rev 2008;61(1):10801.

39.Tan VBC,Khoo KJL.Perforation of flexible laminates by projectiles of different geometry.Int J Impact Eng 2005;31(7):793–810.