Tensile testing of materials at high rates of strain

A tension version of the split Hopkinson bar or Kolsky apparatus is developed for conducting tests in tension at high rates of strain up to 10³ s^?1. A number of aluminum, titanium, and steel alloys tested in tension show increasing degrees of rate sensitivity above 10 to 10² s^?1. Tests on 6061-T651 and 7075-T6 aluminum show measurable strain-rate sensitivity in tension at the highest strain rates, although similar tests in compression in the literature show essentially no strain-rate sensitivity. Details of the apparatus and instrumentation and guidelines for its use are presented.     

Techniques for measuring stress-strain relations at high strain rates

The qualitative dependence of the mechanical behavior of some materials on strain rate is now well known. But the quantitative relation between stress, strain and strain rate has been established for only a few materials and for only a limited range. This relation, the so-called constitutive equation, must be known before plasticity or plastic-wave-propagation theory can be used to predict the stress or strain distribution in parts subjected to impact stresses above the yield strength. In this paper, a brief review of some of the experimental techniques for measuring the stress, strain, strain-rate relationship is given, and some of the difficulties and shortcomings pointed out. Ordinary creep or tensile tests can be used at plastic-strain rates from 10^?8 to about 10^?1/sec. Special quasi-static tests, in which the stress- and strain-measuring devices as well as the specimen geometry and support have been optimized, are capable of giving accurate results to strain rates of about 10²/sec. At higher strain rates, it is shown that wave-propagation effects must be included in the design and analysis of the experiments. Special testing machines for measuring stress, strain and strain-rate relationships in compression, tension and shear at strain rates up to 10⁵/sec are described, and some of the results presented. With this type of testing machine, the analysis of the data requires certain assumptions whose validity depends upon proper design of the equipment. A critical evaluation of the accuracy of these types of tests is presented.     

Characterization of Anisotropic Polymeric Foam Under Static and Dynamic Loading

An orthotropic polymeric foam with transverse isotropy (Divinycell H250) used in composite sandwich structures was characterized at various strain rates. Uniaxial experiments were conducted along principal material axes as well as along off-axis directions under tension, compression, and shear to determine engineering constants, such as Young??s and shear moduli. Uniaxial strain experiments were conducted to determine mathematical stiffness constants, i. e., C _ij . An optimum specimen aspect ratio for these tests was selected by means of finite element analysis. Quasi-static and intermediate strain rate tests were conducted in a servo-hydraulic testing machine. High strain rate tests were conducted using a split Hopkinson Pressure Bar system built for the purpose using polymeric (polycarbonate) bars. The polycarbonate material has an impedance that is closer to that of foam than metals and results in lower noise to signal ratios and longer loading pulses. It was determined by analysis and verified experimentally that the loading pulses applied, propagated along the polycarbonate rods at nearly constant phase velocity with very low attenuation and dispersion. Material properties of the foam were obtained at three strain rates, quasi-static (10^?4 s^?1), intermediate (1 s^?1), and high (10³ s^?1) strain rates. A simple model proposed for the Young??s modulus of the foam was in very good agreement with the present and published experimental results.     

Mechanics of the rate-dependent elastic–plastic deformation of glassy polymers from low to high strain rates

A combined experimental and analytical investigation has been performed to understand the mechanical behavior of two amorphous polymers—polycarbonate and poly(methyl methacrylate)—at strain rates ranging from 10⁻⁴ to 10⁴ s⁻¹. This range in strain rates was achieved in uniaxial tension and compression tests using a dynamic mechanical analyzer (DMA), a servo-hydraulic testing machine, and an aluminum split-Hopkinson pressure bar. DMA tension tests were used to characterize the viscoelastic behavior of these materials, with focus on the rate-dependent shift of material transition temperatures. Uniaxial compression tests on the servo-hydraulic machine (10⁻⁴ to 1 s⁻¹) and the split-Hopkinson pressure bar (10³ to 10⁴ s⁻¹) were used to characterize the rate-dependent yield and post-yield behavior. Both materials were observed to exhibit increased rate sensitivity of yield under the same strain rate/temperature conditions as the β-transition of the viscoelastic behavior. A physically based constitutive model for large strain deformation of thermoplastics was then extended to encompass high-rate conditions. The model accounts for the contributions of different molecular motions which become operational and important in different frequency regimes. The new features enable the model to not only capture the transition in the yield behavior, but also accurately predict the post-yield, large strain behavior over a wide range of temperatures and strain rates.     

A Combined Theoretical/Experimental Approach for Reducing Ringing Artifacts in Low Dynamic Testing with Servo-hydraulic Load Frames

One of the most challenging tasks facing computer-aided engineering (CAE) analysis is the acquisition of accurate tensile test data that spans quasi-static to low dynamic (10^?5/s?≤ $ \overset{.}{\varepsilon } $ ≤5?×?10²/s) strain rates ( $ \overset{.}{\varepsilon } $ ). Critical to the accuracy of data acquired over the low dynamic range is the reduction of ringing artifacts in flow data. Ringing artifacts, which are a consequence of the inertial response of the load frame, are spurious oscillations that can obscure the desired material response (i.e. load vs. time or load vs. displacement) from which flow data are derived. These oscillations tend to grow with increasing strain rate and peak at the high end of the low dynamic range on servo-hydraulic tensile test frames. Common practices for addressing ringing are data filtering, which is often problematic since filtering introduces distortion in smoothed material data, or trial-and-error design of test specimen geometries. This renders techniques for reducing ringing based upon the mechanics of the load frame and optimization of tensile specimen geometry quite attractive. In the present paper, relationships between load, stress wave propagation, and specimen geometries are addressed, to both quantify ringing and to develop specimen designs that will reduce ringing. A combined theoretical/experimental approach for tensile specimen design was developed for reducing ringing in flow data over the low dynamic range of strain rates (10^?5/s≤ $ \overset{.}{\varepsilon } $ ≤5?×?10²/s). The single camera digital image correlation (DIC) method was used to measure the displacement fields and strain rates with specimens resulting from the combined theoretical/experimental approach. While the approach was developed on a specific commercial load frame with a TRIP steel subject to a two-step quenching and partitioning heat treatment (Q&P980), it is readily adaptable to other servo-hydraulic load frames and metallic alloys. The developed approach results in a 90 % reduction in ringing artifact (with no filtering) in a tensile flow curve for Q&P980 at $ \overset{.}{\varepsilon}\kern-4pt $ = 5?×?10²/s. Results from split Hopkinson bar tests of Q&P980 were performed at $ \overset{.}{\varepsilon } $ = 500/s and compare favorably with the test data generated by the developed testing approach. Since the Q&P980 steel represents a new generation of advanced high strength steels, we also evaluated its strain rate sensitivity over the low dynamic range.     

Novel Microtensile Method for Monotonic and Cyclic Testing of Freestanding Copper Thin Films

This paper presents the results of new microtensile tests conducted to investigate the mechanical properties of submicron-thick freestanding copper films. The method, used in this study, allows the observation of materials response under uniaxial tensile loads with measurements of stress at strain rates up to 5.5 × 10⁻⁴/s. It also facilitates tension–tension fatigue experiments under a variety of mean stress conditions at cyclic loading frequencies to 20 Hz. The sample processes involve fabrication of a supporting frame with springs and alignment beams all made of electroplated nickel. Electroplating took place on top of a previously deposited sample rather than creating a structure by subtractive fabrication. Tensile sample loading is applied using a piezoelectric actuator. Load was measured using a capacitance gap sensor with a novel mechanical coupling to the sample. Tension–tension fatigue experiments were carried out with feedback to give load control. Fatigue tests were conducted on sputter-deposited 500 and 900 nm copper films with grain sizes ∼50 nm. Fatigue life reached 10⁵ cycles at low mean load, which decreased with an increase in the mean load. The results indicate decreasing plasticity with increasing mean load.     

Determination of stress-strain characteristics at very high strain rates

A modified version of the Kolsky thin-wafer technique is described. The method permits one to obtain the dynamic plastic properties of materials at strain rates as high as 10⁵ sec^?1. Data obtained from compression tests on high-purity aluminum are presented for strain rates ranging from 4000 to 120,000 sec^?1 at room temperature. Specimen-size effects and the effect of lateral inertia are taken into account in analyzing the data. The results plotted as stress vs. strain rate at constant strains (5 to 20 percent) show that, at the highest strain rates, the stress rises very rapidly with strain rate suggesting that a limiting strain rate is being reached. At the lower strain rates (10³ to 10⁴ sec^?1), the stress is linearly proportional to the strain rate indicating that the material is deforming in a viscous manner.     

Impulsive Tension of Hexane and Glycerol under Shock‐Wave Loading

The strength of hexane and glycerol was measured under impulsive tension produced by interaction of a triangular compression pulse with a free surface. The tests were performed for strain rates of 10⁴–10⁵ sec^–1. It is established that the strength of hexane is equal to 14 MPa and does not depend on the strain rate, whereas the strength of glycerol increases from 57 to 142 MPa with an order of magnitude increase in the strain rate. The possibility of using the model of homogeneous nucleation to interpret the data obtained is discussed.     

Relationship between Local Strain Jumps and Temperature Bursts Due to the Portevin-Le Chatelier Effect in an Al-Mg Alloy

Local strain and temperature of an AA5754-O aluminum alloy sheet have been full-field measured during monotonous tensile tests carried out at room temperature. Sharp strain increases and temperature bursts which are locally generated by the Portevin-Le Chatelier phenomenon have been measured at the same point for two strain rates: V2?=?1.9?×?10^?3?s^?1 and V10?=?9.7?×?10^?3?s^?1. A relationship, which is based on the underlying physical mechanisms, has been established between the strain and the temperature and experimentally verified for the highest strain rate V10. The discrepancy between the theoretical and experimental results for the lowest strain rate V2 suggests that the localized plastic deformations do not follow an adiabatic transformation. Such a set-up seems to offer a direct and experimental method to check the adiabatic character of localized plastic deformations.     

High-velocity deformation properties of polyurethane foams

This paper presents the results of dynamic uniaxial-stress tests performed on polymer-foam material. A water-blown ester polyurethane foam designated as rigid and a castor-oil-base polyurethane foam designated as semirigid were tested in tension and compression at rates of loading from 10^?3 in./in./sec to 10³ in./in./sec at room temperature. A gas-operated medium-strain-rate machine was used for rates of loading from 10^?3 to about 10² in./in./sec. Tests at higher rates were performed on a split Hopkinson-bar device. Highspeed photographic techniques were used to study dynamic fracture.     

Effects of strain rate and moisture content on the behaviour of sand under one-dimensional compression

The influence of strain rate and moisture content on the behaviour of a quartz sand was assessed using high-pressure quasi-static (10^?3 s^?1) and high-strain rate (10³ s^?1) experiments under uniaxial strain. Quasi-static compression to axial stresses of 800 MPa was carried out alongside split Hopkinson pressure bar (SHPB) experiments to 400 MPa, where in each case lateral deformation of the specimen was prevented using a steel test box or ring, and lateral stresses were recorded. A significant increase in constrained modulus was observed between strain rates of 10^?3s^?1 and 10³s^?1, however a consistently lower Poisson’s ratio in the dynamic tests minimised changes in bulk modulus. The reduction in Poissons ratio suggests that the stiffening of the sand in the SHPB tests is due to additional inertial confinement rather than an inherent strain-rate dependence. In the quasi-static tests the specimens behaved less stiffly with increasing moisture content, while in the dynamic tests the addition of water had little effect on the overall stiffness, causing the quasi-static and dynamic series to diverge with increasing moisture content.     

Large Strain Shear Compression Test of Sheet Metal Specimens

A hybrid experimental-computational procedure to establish accurate true stress-plastic strain curve of sheet metal specimen covering the large plastic strain region using shear compression test data is described. A new shear compression jig assembly with a machined gage slot inclined at 35° to the horizontal plane of the assembly is designed and fabricated. The novel design of the shear compression jig assembly fulfills the requirement to maintain a uniform volume of yielded material with characteristic maximum plastic strain level across the gage region of the Shear Compression Metal Sheet (SCMS) specimen. The approach relies on a one-to-one correlation between measured global load–displacement response of the shear compression jig assembly with SCMS specimen to the local stress-plastic strain behavior of the material. Such correlations have been demonstrated using finite element (FE) simulation of the shear compression test. Coefficients of the proposed correlations and their dependency on relative plastic modulus were determined. The procedure has been established for materials with relative plastic modulus in the range 5?×?10^?4?<?(E _p /E)?<?0.01. It can be readily extended to materials with relative plastic modulus values beyond the range considered in this study. Nonlinear characteristic hardening of the material could be established through piecewise linear consideration of the measured load–displacement curve. Validity of the procedure is established by close comparison of measured and FE-predicted load–displacement curve when the provisional hardening curve is employed as input material data in the simulation. The procedure has successfully been demonstrated in establishing the true stress-plastic strain curve of a demonstrator 0.0627C steel SCMS specimen to a plastic strain level of 49.2 pct.     

Load-Inversion Device for the High Strain Rate Tensile Testing of Sheet Materials with Hopkinson Pressure Bars

A high strain rate tensile testing technique for sheet materials is presented which makes use of a split Hopkinson pressure bar system in conjunction with a load inversion device. With compressive loads applied to its boundaries, the load inversion device introduces tension into a sheet specimen. Two output bars are used to minimize the effect of bending waves on the output force measurement. A Digital Image Correlation (DIC) algorithm is used to determine the strain history in the specimen gage section based on high speed video imaging. Detailed finite element analysis of the experimental set-up is performed to validate the design of the load inversion device. It is shown that under the assumption of perfect alignment and slip-free attachment of the specimen, the measured stress–strain curve is free from spurious oscillations at a strain rate of 1,000 s^?1. Validation experiments are carried out using tensile specimens extracted from 1.4 thick TRIP780 steel sheets. The experimental results for uniaxial tension at strain rates ranging from 200 s^?1 to 1,000 s^?1 confirm the oscillation-free numerical results in an approximate manner. Dynamic tension experiments are also performed on notched specimens to illustrate the validity of the proposed experimental technique for characterizing the effect of strain rate on the onset of ductile fracture in sheet materials.     

The effect of strain rate and heat developed during deformation on the stress-strain curve of plastics

Polymethylmethacrylate, cellulose acetate butyrate, polypropylene and nylon 6–6 have been characterized in compression at various strain rates from 10^?4 s^?1 to 10³ s^?1 at room temperature. A medium strain-rate machine and a split-Hopkinson-bar apparatus are used in conducting the experiments. The temperature rise developed during deformation is also measured by using a thermocouple. All four materials tested definitely show a viscous effect at the beginning of the deformation and a plastic flow follows thereafter. Test results also indicate that the temperature rise developed during deformation cannot be neglected in determining the dynamic response of those materials investigated in this study.     

Pressure-shear waves in 6061-T6 aluminum and alpha-titanium

Thepressure-shear plate impact technique is used to study material behavior at high rates of deformation. In this technique, plastic waves of combined pressure and shear stresses are produced by impact of parallel plates skewed relative to their direction of approach. Commercially pure alpha-titanium and 6061-T6 aluminum are tested under a variety of pressure and shear tractions by using different combinations of impact velocities and angles of inclination. A laser interferometer system is used to monitor simultaneously the normal and transverse components of motion of a point at the rear surface of the target plate. The experimental results are compared with numerical solutions based on an elastic/viscoplastic model of the material. Both isotropic and kinematic strain hardening models are used in the computations. The results indicate that unlike the normal velocity profiles, the transverse velocity profiles are sensitive to the dynamic plastic response and, thus, can be used to study material behavior at high strain rates. For the materials tested the results suggest that the flow stress required for plastic straining increases markedly with increasing strain rate at strain rates above 10⁴s^?1. Hydrostatic pressure of the order that exists in the tests (up to 2 GPa) does not affect the plastic flow in 6061-T6 aluminum and appears to have at most a minor effect on the deformation of the titanium.     

Development of an apparatus for fatigue testing in vacuum

The development and use of a small, light-weight, remotely controlled fatigue apparatus are described. The particular application was for fatigue tests of 7075-T6 aluminum alloy in vacuums of 8.6×10^?8 to 4×10^?9 Torr. The fatigue life in vacuum was 1.8 times greater than the fatigue life in air based on the upper 95 percent confidence limit for tests in air and the lower 95 percent confidence limit for tests in vacuum.     

Influence of Portevin-Le Chatelier Effect on Shear Strain Path Reversal in an Al-Mg Alloy at Room and High Temperatures

The main objective of this study is to characterize the mechanical behaviour of an Al-Mg alloy in conditions close to those encountered during sheet forming processes, i.e. with strain path changes and at strain rates and temperatures in the range 1.2×10^?3–1.2×10^?1 s^?1 and 25–200°C, respectively. The onset of jerky flow and the interaction of dynamic strain ageing with the work-hardening are investigated during reversed-loading in specific simple shear tests, which consist of loading up to various shear strain values followed by reloading in the opposite direction, combined with direct observations of the sample surface using a digital image correlation technique. Both strain path changes and temperature are clearly shown to influence the occurrence and onset of the Portevin-Le Chatelier (PLC) effect. Moreover, the Bauschinger effect observed in the material response shows that the PLC effect has a major influence on the kinematic contribution to work-hardening as well as its stagnation during the reloading stage, which could open up interesting lines of research to improve theoretical plasticity models for this family of aluminium alloys.     

Torsional split Hopkinson bar tests at strain rates above 104s−1

The torsional split Hopkinson bar is used for testing materials at strain rates above 10⁴s⁻¹. This strain rate, which is an order of magnitude higher than is typical with this technique, is obtained by using very short specimens. Strain rates of 6.4×10⁴s⁻¹ have been achieved with specimens having a gage length of 0.1524 mm. Results from tests on 1100 aluminum show an increase in rate sensitivity as the strain rate increases.     

High strain-rate testing: Tension and compression

Details of the split Hopkinson pressure-bar method for obtaining complete stress-strain curves at strain rates on the order of 1000 sec^?1 in either tension or compression are presented. In compression, a gage for measuring radial strain and, therefore, Poisson's ratio is also described. Some typical results are presented for aluminum, and various factors pertaining to the accuracy of the results are discussed.     

The dependence of acoustic emission on strain rate in 7075-T6 aluminum

The dependence of RMS acoustic-emission output on strain rate was determined during tensile tests on 7075-T6 aluminum over a range of strain rates from 0.015 min^?1 to 0.230 min^?1. A special normalization technique was developed to eliminate the inherent differences in the acoustic emission from supposedly identical specimens tested under identical conditions. It was found that the RMS acoustic-emission output increases approximately linearly with strain rate, and that the slope of the linear dependence decreased with increased plastic strain. The effect of pinned dislocation breakaway on the acoustic-emission output vs. plastic-strain relationship is the subject of a preliminary theory.