24 Residual stresses in composite materials DHD allows the resolution of residual stresses through even very thick components.In this method,a hole with a narrow diameter is drilled through a part that has residual stresses.The diameter of the hole is then measured carefully (typically using an air gage)as a function of depth and angular position inside the hole.The residual stresses are then released by co-axially removing a core of larger diameter from around the hole.The release of residual stress causes the shape of the reference hole to change.The diameter of the hole is then re-measured at the same angular positions and depths as in the original measurements and using the same equipment.Changes in the shape of the hole are then related to the residual stresses that were present before the hole was drilled(Reid.2009). The main assumption of this technique is that the introduction of the reference hole has little effect on the residual stress state,and that cutting the core allows residual stresses around the hole to thoroughly relax (Mirzaee-Sisan,2007). Another assumption is that the core is comprised of many independent lengths. Therefore,a thick part can be approximated as a set of stacked layers unconnected by through-thickness shear stresses (Bateman et al.,2005).The alterations in the core length can be considered as a measure of through-thickness residual stresses as the outer core depth increases(Procter and Beaney,1987).The DHD technique has been utilized occasionally in cases of welds in metal parts(Bouchard et al., 2005;Brown et al.,2006;George and Smith,2005;Mirzaee-Sisan,2007),railway track(Stefanescu et al.,2003)and rolls in steel mills(Kingston and Smith,2005). This method has also been used to measure residual stresses in a laminated carbon-fiber composite (Bateman et al.,2005).In this case,the removal of the core around the hole could not be performed using EDM,and as an alternative,a diamond encrusted hole saw was used.The analysis technique was extended to account for orthotropic material to allow the change in hole shape to be related to the original residual stresses. The basic assumption of the DHD method is that the material at each depth can be treated as a continuum.When this technique is applied to a laminated composite, averages of the stresses within the fiber and matrix are measured.As a result,this method cannot be directly used to measure the micro-scale residual stresses.The method is only able to resolve meso-scale residual stresses,which in unloaded uni-directional laminates are non-existent (Reid,2009).Drilling a small diameter hole results in the cutting of fibers on the hole wall.Consequently,the micro-scale residual stresses in the fibers are released and the material near the ends of the fibers responds in an elastic way.By carefully monitoring this phenomenon,it is possible to modify this method to measure the longitudinal micro-scale residual stresses in uni-directional GFRP(Reid,2009). Attempting to modify the technique poses at least three major difficulties(Reid, 2009),first that the mechanical methods ofcreating the hole create cutting stresses on the hole wall.These occur at the same place where the breaks in the fibers occur,so the elastic responses from cutting stresses and from the release ofmicro- scale residual stresses are coincident.These two effects cannot clearly be Woodhead Publishing Limited,2014
24 Residual stresses in composite materials © Woodhead Publishing Limited, 2014 DHD allows the resolution of residual stresses through even very thick components. In this method, a hole with a narrow diameter is drilled through a part that has residual stresses. The diameter of the hole is then measured carefully (typically using an air gage) as a function of depth and angular position inside the hole. The residual stresses are then released by co- axially removing a core of larger diameter from around the hole. The release of residual stress causes the shape of the reference hole to change. The diameter of the hole is then re- measured at the same angular positions and depths as in the original measurements and using the same equipment. Changes in the shape of the hole are then related to the residual stresses that were present before the hole was drilled (Reid, 2009). The main assumption of this technique is that the introduction of the reference hole has little effect on the residual stress state, and that cutting the core allows residual stresses around the hole to thoroughly relax (Mirzaee-Sisan, 2007). Another assumption is that the core is comprised of many independent lengths. Therefore, a thick part can be approximated as a set of stacked layers unconnected by through- thickness shear stresses (Bateman et al ., 2005). The alterations in the core length can be considered as a measure of through- thickness residual stresses as the outer core depth increases (Procter and Beaney, 1987). The DHD technique has been utilized occasionally in cases of welds in metal parts (Bouchard et al ., 2005; Brown et al ., 2006; George and Smith, 2005; Mirzaee-Sisan, 2007), railway track (Stefanescu et al ., 2003) and rolls in steel mills (Kingston and Smith, 2005). This method has also been used to measure residual stresses in a laminated carbon- fi ber composite (Bateman et al ., 2005). In this case, the removal of the core around the hole could not be performed using EDM, and as an alternative, a diamond encrusted hole saw was used. The analysis technique was extended to account for orthotropic material to allow the change in hole shape to be related to the original residual stresses. The basic assumption of the DHD method is that the material at each depth can be treated as a continuum. When this technique is applied to a laminated composite, averages of the stresses within the fi ber and matrix are measured. As a result, this method cannot be directly used to measure the micro- scale residual stresses. The method is only able to resolve meso- scale residual stresses, which in unloaded uni- directional laminates are non- existent (Reid, 2009). Drilling a small diameter hole results in the cutting of fi bers on the hole wall. Consequently, the micro- scale residual stresses in the fi bers are released and the material near the ends of the fi bers responds in an elastic way. By carefully monitoring this phenomenon, it is possible to modify this method to measure the longitudinal micro- scale residual stresses in uni- directional GFRP (Reid, 2009). Attempting to modify the technique poses at least three major diffi culties (Reid, 2009), fi rst that the mechanical methods of creating the hole create cutting stresses on the hole wall. These occur at the same place where the breaks in the fi bers occur, so the elastic responses from cutting stresses and from the release of microscale residual stresses are coincident. These two effects cannot clearly be
Destructive techniques in the measurement of residual stresses 25 distinguished from each other,making an accurate measurement difficult.Second, it is not easy to measure the profile of the hole wall at scales smaller than that of the diameter of the glass fiber,due to problems with measurement resolution. Third,the material that tends to move radially inwards in response to the fiber being cut cannot be removed in the process of making the hole,and consequently the profile of the hole wall cannot be monitored properly.The DHD technique cannot be used to measure the micro-scale residual stresses present in the fiber direction of uni-directional GFRP directly,and every modification of the technique leads to additional problems(Reid,2009). Initial development of the DHD method was carried out by Zhandanov and Gonchar (1978),Beaney (1978)and Jesensky and Vargova(1981).Zhadanov and Gonchar(1978)used the DHD method to measure residual stresses in steel welds. Beaney's(1978)methodology was later improved by Procter and Beaney (1978) with the introduction of non-contacting capacitance gages to measure the hole diameter.Jesensky and Vargova(1981)measured residual stresses in steel welds using strain gages attached to the sides of the hole to measure the strain relaxation following trepanning. Bateman et al.(2005)described an extension to the DHD method for the evaluation ofresidual stresses in thick section composite laminated plates.Leggatt et al.(1996)described the development of a DHD method,based on earlier techniques,for measuring the through-thickness distribution of residual stresses. Some researchers have made improvements to the DHD method,by gun-drilling a hole of 3 mm nominal diameter and measuring the change in diameter of the hole using an air probe.Trepanning the core was carried out using an electro- discharge machining(EDM)operation(Bonner and Smith,1996;George et al., 2000,2002;George and Smith,2000). 2.5 The ring-core method The ring-core method (Fig.2.4)follows a procedure similar to the hole-drilling technique.Instead of residual stresses being released by drilling a hole and monitoring the elastic response of the neighboring material,the ring-core method releases stress by cutting an annular groove into the surface of a part containing residual stress.A strain gage rosette is used to measure the elastic response on the end of the core within the groove.An incremental increase in the depth of the groove makes it possible to determine the stress variation through the thickness (Keil,1992).The through-thickness stresses can be determined by monitoring the change in core length with increasing groove depth(Wern et al., 1997,Werm,1997). Compared with the hole-drilling method,the ring-core method offers a number of benefits.Since the strains are more fully relaxed,the measured response is considerably larger.There is no stress concentration effect in the annulus and as a result,this method can measure residual stresses up to the yield stress of the Woodhead Publishing Limited,2014
Destructive techniques in the measurement of residual stresses 25 © Woodhead Publishing Limited, 2014 distinguished from each other, making an accurate measurement diffi cult. Second, it is not easy to measure the profi le of the hole wall at scales smaller than that of the diameter of the glass fi ber, due to problems with measurement resolution. Third, the material that tends to move radially inwards in response to the fi ber being cut cannot be removed in the process of making the hole, and consequently the profi le of the hole wall cannot be monitored properly. The DHD technique cannot be used to measure the micro- scale residual stresses present in the fi ber direction of uni- directional GFRP directly, and every modifi cation of the technique leads to additional problems (Reid, 2009). Initial development of the DHD method was carried out by Zhandanov and Gonchar (1978), Beaney (1978) and Jesensky and Vargova (1981). Zhadanov and Gonchar (1978) used the DHD method to measure residual stresses in steel welds. Beaney’s (1978) methodology was later improved by Procter and Beaney (1978) with the introduction of non- contacting capacitance gages to measure the hole diameter. Jesensky and Vargova (1981) measured residual stresses in steel welds using strain gages attached to the sides of the hole to measure the strain relaxation following trepanning. Bateman et al . (2005) described an extension to the DHD method for the evaluation of residual stresses in thick section composite laminated plates. Leggatt et al . (1996) described the development of a DHD method, based on earlier techniques, for measuring the through- thickness distribution of residual stresses. Some researchers have made improvements to the DHD method, by gun- drilling a hole of 3 mm nominal diameter and measuring the change in diameter of the hole using an air probe. Trepanning the core was carried out using an electrodischarge machining (EDM) operation (Bonner and Smith, 1996; George et al ., 2000, 2002; George and Smith, 2000). 2.5The ring- core method The ring- core method ( Fig. 2.4 ) follows a procedure similar to the hole- drilling technique. Instead of residual stresses being released by drilling a hole and monitoring the elastic response of the neighboring material, the ring- core method releases stress by cutting an annular groove into the surface of a part containing residual stress. A strain gage rosette is used to measure the elastic response on the end of the core within the groove. An incremental increase in the depth of the groove makes it possible to determine the stress variation through the thickness (Keil, 1992). The through- thickness stresses can be determined by monitoring the change in core length with increasing groove depth (Wern et al ., 1997; Wern, 1997). Compared with the hole- drilling method, the ring- core method offers a number of benefi ts. Since the strains are more fully relaxed, the measured response is considerably larger. There is no stress concentration effect in the annulus and as a result, this method can measure residual stresses up to the yield stress of the
26 Residual stresses in composite materials c(sc) 5 b(sb) 95 a(ea) Strain gage dz 02 dz o dz 2.4 Principle of the ring-core method for determination of plane residual stress(Keil,1992).Notation:dz,amount of step-by-step milling of annular groove;z,depth of groove;s,strain in direction a; o,and o,stress acting in principal directions 1 and 2 respectively. material.The relaxation of strain that occurs on the end of the separated core is uniform,so this method is less sensitive to errors in the strain gage rosette positioning (Reid,2009).As the ring-core method ruptures the fibers while the groove is being formed,it is possible to monitor the elastic response of the neighboring material in order to measure the magnitude of micro-scale residual stresses(Reid,2009). Despite these benefits,the ring-core method has not been employed extensively. This could be because it was not widely known until 1988(Keil,1992).It could Woodhead Publishing Limited,2014
26 Residual stresses in composite materials © Woodhead Publishing Limited, 2014 material. The relaxation of strain that occurs on the end of the separated core is uniform, so this method is less sensitive to errors in the strain gage rosette positioning (Reid, 2009). As the ring- core method ruptures the fi bers while the groove is being formed, it is possible to monitor the elastic response of the neighboring material in order to measure the magnitude of micro- scale residual stresses (Reid, 2009). Despite these benefi ts, the ring- core method has not been employed extensively. This could be because it was not widely known until 1988 (Keil, 1992). It could 2.4 Principle of the ring- core method for determination of plane residual stress (Keil, 1992). Notation: dz, amount of step-by-step milling of annular groove; z, depth of groove; εa, strain in direction a; σ1 and σ2, stress acting in principal directions 1 and 2 respectively
Destructive techniques in the measurement of residual stresses 27 also be due to common use of the standard hole-drilling technique for measuring residual stresses in FRPs.It is not practical to use EDM,since composite materials are either non-conducting or are poor conductors ofelectricity,and this complicates the conditions for composites(Reid,2009).This method has been applied to large cast steel parts and forgings (Keil,1992),forged aluminum parts (Witt et al., 1983),welds in stainless steel (Roy et al.,2005)and hot rolled laminates of stainless and carbon steel (Schroder et al.,1995).The ring-core method can also be used in regions with high stress gradients,such as laser welds(Ren and Li, 2007)and ultrasonic spot welds (Li et al.,2007).Lu (1996)devoted a part of his book to this method. 2.6 The cutting method The cutting method is based on similar principles to those in the hole-drilling method.Once more,stress is relaxed by the removal of material.This time a notch is removed from a specimen,resulting in the creation of a free edge.A Moire interferometry grating can be applied to the specimen to record the resulting strain field (Filiou et al.,1992:Lee et al.,1989),which is then calculated and related to the residual stresses using finite element analysis(Myers,2004). This method is not without its faults and needs modifications to be totally accepted in the experimental community.Niu (1999)posed many questions regarding the application of the grating.Since the grating was applied after an edge of the composite had been trimmed,some of the residual stresses were released prior to data recording and a complex strain field was created under the surface.The resulting measurements thus contained data based on a stress field different from the original residual stress(Myers,2004). Lee et al.(1989)studied the residual strain distribution in a thick composite ring experimentally using interferometric techniques.The specimen was a segment of a graphite/epoxy composite cylinder,with the fibers orientated in the hoop and axial directions.Casari et al.(2006)presented a method for the characterization of residual stresses in thick filament wound tubes.A second technique derived to employ the cutting method was proposed by Sunderland et al.(1995).Their successive grooving technique involved cutting a groove through the thickness at successive depths.Strain gages were placed opposite the groove and recorded the changing strain field.The residual stress was then calculated from the strain by use of a numerical 2D model for each layer(Myers, 2004)- 2.7 The contour method The contour method is used to measure 2D residual stresses.In this method,a part which contains residual stresses is cut through by a planar surface.This results in the release of the residual stresses across the plane and consequently the new Woodhead Publishing Limited,2014
Destructive techniques in the measurement of residual stresses 27 © Woodhead Publishing Limited, 2014 also be due to common use of the standard hole- drilling technique for measuring residual stresses in FRPs. It is not practical to use EDM, since composite materials are either non- conducting or are poor conductors of electricity, and this complicates the conditions for composites (Reid, 2009). This method has been applied to large cast steel parts and forgings (Keil, 1992), forged aluminum parts (Witt et al ., 1983), welds in stainless steel (Roy et al ., 2005) and hot rolled laminates of stainless and carbon steel (Schröder et al ., 1995). The ring- core method can also be used in regions with high stress gradients, such as laser welds (Ren and Li, 2007) and ultrasonic spot welds (Li et al ., 2007). Lu (1996) devoted a part of his book to this method. 2.6 The cutting method The cutting method is based on similar principles to those in the hole- drilling method. Once more, stress is relaxed by the removal of material. This time a notch is removed from a specimen, resulting in the creation of a free edge. A Moiré interferometry grating can be applied to the specimen to record the resulting strain fi eld (Filiou et al ., 1992; Lee et al ., 1989), which is then calculated and related to the residual stresses using fi nite element analysis (Myers, 2004). This method is not without its faults and needs modifi cations to be totally accepted in the experimental community. Niu (1999) posed many questions regarding the application of the grating. Since the grating was applied after an edge of the composite had been trimmed, some of the residual stresses were released prior to data recording and a complex strain fi eld was created under the surface. The resulting measurements thus contained data based on a stress fi eld different from the original residual stress (Myers, 2004). Lee et al . (1989) studied the residual strain distribution in a thick composite ring experimentally using interferometric techniques. The specimen was a segment of a graphite/epoxy composite cylinder, with the fi bers orientated in the hoop and axial directions. Casari et al . (2006) presented a method for the characterization of residual stresses in thick fi lament wound tubes. A second technique derived to employ the cutting method was proposed by Sunderland et al . (1995). Their successive grooving technique involved cutting a groove through the thickness at successive depths. Strain gages were placed opposite the groove and recorded the changing strain fi eld. The residual stress was then calculated from the strain by use of a numerical 2D model for each layer (Myers, 2004). 2.7 The contour method The contour method is used to measure 2D residual stresses. In this method, a part which contains residual stresses is cut through by a planar surface. This results in the release of the residual stresses across the plane and consequently the new
28 Residual stresses in composite materials surface undergoes an out-of-plane deformation(Reid,2009).These out-of-plane deformations are measured and the original residual stresses across the cut are determined using the finite element method.In order to do this.displacement boundary conditions (equal to the negative of the measured deflections)are imposed on the cut surface.The method has proved to be efficient in mapping complicated residual stress fields,such as in railway track(Kelleher,2003)and welds (Prime et al.,2006;Zhang et al.,2003),as well as those caused by hypersonic impact (Martineau et al.,2004). This procedure is mainly ideal for measuring existing longitudinal residual stresses in uni-directional GFRP materials.The heterogeneous structure of GFRP is easy to model using the finite element method,and with accurate measurement of out-of-plane displacements in the vicinity of the fibers using methods such as laser probe scanning (Prime et al,2004),the fiber residual stresses can be determined easily. The main requirement of the contour method is that the planar section must be cut through the material under stress with great care.Mechanical methods tend to create cutting stresses and trim down out-of-plane deflections while they are being observed,so electric discharge wire machining(EDWM)is commonly used as a suitable alternative.The advantage of this method is that it only removes material at the tip of the cut and no significant cutting stresses are created.Unfortunately, EDWM cannot be utilized to create the cut in GFRP materials,because neither constituent is conducting.The contour method therefore cannot be applied to the measurement of longitudinal micro-scale residual stresses in GFRP(Reid,2009) Prime(2001)presented a powerful new method for residual stress measurement, concluding that the contour method was more powerful than other relaxation methods because it could determine an arbitrary cross-sectional area map of residual stress,yet at the same time simpler,because the stresses could be determined directly from the data without an inversion technique.He verified this method with a numerical simulation,and then experimentally validated it,using a steel beam with a known residual stress profile.Zhang et al.(2003)performed contour measurements on a MIG 2024-T351 aluminum alloy welded plate.They compared their results with the results from neutron and synchrotron X-ray and observed a favorable agreement between them. Prime et al.(2006)joined plates of aluminum alloys 7050-T7451 and 2024- T351 in a butt joint by friction stir welding (FSW).In their work,a 54mm long test specimen was removed from the parent plate,and cross-sectional maps of residual stresses were measured using neutron diffraction and the contour method. Martineau et al.(2004)impacted a thick plate of high-strength low-alloy (HSLA- 100)steel with tungsten carbide spheres travelling at velocities ranging from 0.8 to 2.5 km/s.Good agreement was shown between the numerical simulation of the impact event and the experimental data.Prime et al.(2004)described noncontact scanning using a confocal laser probe to measure surface contours for applications in residual stress measurement Woodhead Publishing Limited,2014
28 Residual stresses in composite materials © Woodhead Publishing Limited, 2014 surface undergoes an out- of-plane deformation (Reid, 2009). These out- of-plane deformations are measured and the original residual stresses across the cut are determined using the fi nite element method. In order to do this, displacement boundary conditions (equal to the negative of the measured defl ections) are imposed on the cut surface. The method has proved to be effi cient in mapping complicated residual stress fi elds, such as in railway track (Kelleher, 2003) and welds (Prime et al ., 2006; Zhang et al ., 2003), as well as those caused by hypersonic impact (Martineau et al ., 2004). This procedure is mainly ideal for measuring existing longitudinal residual stresses in uni- directional GFRP materials. The heterogeneous structure of GFRP is easy to model using the fi nite element method, and with accurate measurement of out- of-plane displacements in the vicinity of the fi bers using methods such as laser probe scanning (Prime et al ., 2004), the fi ber residual stresses can be determined easily. The main requirement of the contour method is that the planar section must be cut through the material under stress with great care. Mechanical methods tend to create cutting stresses and trim down out- of-plane defl ections while they are being observed, so electric discharge wire machining (EDWM) is commonly used as a suitable alternative. The advantage of this method is that it only removes material at the tip of the cut and no signifi cant cutting stresses are created. Unfortunately, EDWM cannot be utilized to create the cut in GFRP materials, because neither constituent is conducting. The contour method therefore cannot be applied to the measurement of longitudinal micro- scale residual stresses in GFRP (Reid, 2009). Prime (2001) presented a powerful new method for residual stress measurement, concluding that the contour method was more powerful than other relaxation methods because it could determine an arbitrary cross- sectional area map of residual stress, yet at the same time simpler, because the stresses could be determined directly from the data without an inversion technique. He verifi ed this method with a numerical simulation, and then experimentally validated it, using a steel beam with a known residual stress profi le. Zhang et al . (2003) performed contour measurements on a MIG 2024-T351 aluminum alloy welded plate. They compared their results with the results from neutron and synchrotron X-ray and observed a favorable agreement between them. Prime et al . (2006) joined plates of aluminum alloys 7050-T7451 and 2024- T351 in a butt joint by friction stir welding (FSW). In their work, a 54 mm long test specimen was removed from the parent plate, and cross- sectional maps of residual stresses were measured using neutron diffraction and the contour method. Martineau et al . (2004) impacted a thick plate of high- strength low- alloy (HSLA- 100) steel with tungsten carbide spheres travelling at velocities ranging from 0.8 to 2.5 km/s. Good agreement was shown between the numerical simulation of the impact event and the experimental data. Prime et al . (2004) described noncontact scanning using a confocal laser probe to measure surface contours for applications in residual stress measurement