[Skip to content]

.

Loading Rate Effects on Tensile Properties and Fracture Toughness of Steel

Dr C S Wiesner, TWI, Cambridge, UK and Mr H MacGillivray [1], Imperial College, London, UK

Presented at 1999 TAGSI Seminar - 'Fracture, Plastic Flow and Structural Integrity' (dedicated to Sir Alan Cottrell in the year of his Eightieth Birthday), Held at TWI, Cambridge, UK, 29 April 1999

[1]Note: Mr MacGillivray is chairman of the ESIS TC5 committee, which is concerned with dynamic mechanical testing.

Abstract

An overview of the effects of elevated ('dynamic') loading rates on the tensile and initiation toughness properties of steels and their effect on structural behaviour is given. Issues affecting engineering critical assessments are also considered. Dynamic loading rates affect both the material resistance and the structural response of engineering components and it is the combination of these two influences which determines the structural behaviour.

Because of inherent complexities of testing at high loading rates, the development of suitable test techniques for measuring both tensile and fracture toughness properties has been a topic of marked investment in the last 40 years or so, but increasing consensus has now been achieved leading to the development of test standards. Generally, the effect of increasing loading rate is to increase strength (positive strain rate dependence) and the strain rate sensitivity increases with temperature, but there are exceptions when dynamic strain ageing effects intervene. The effect of loading rate on the fracture toughness of ferritic steels is dependent on overall material behaviour: for temperatures below the brittle to ductile transition region, toughness decreases with increasing loading rate. At upper shelf temperatures, the ductile initiation toughness and tearing resistance generally increases with loading rate, but exceptions exist which are outlined in the paper. In or near the transition region, increasing loading rates can cause a shift from fully ductile behaviour at static rates to brittle behaviour at high rates of loading. The brittle to ductile transition temperature of ferritic steels increases with increasing loading rates and methods have been proposed to predict this shift.

1. Introduction

It is well known that an increase in the rate of loading affects material properties in steel. This is recognised in national and international test standards for tensile and fracture toughness testing by prescribing a maximum rate of loading beyond which the test standard is no longer valid. For steels, the main effect of increasing loading rate is an increase of the yield and tensile strength which generally leads to a reduction of cleavage fracture toughness. This is implicitly recognised in the widely used Charpy test where a notched bar is impact tested in three point bending at an initial displacement rate of 5.5m/s, so as to include the detrimental effect of loading rate in the toughness determination.

For more accurate, fracture mechanics-based, structural integrity evaluations, Charpy results are not sufficient so fracture toughness input and fracture toughness tests using pre-cracked specimens are generally employed. When characterising the fracture toughness of an engineering component, the rate of loading to which the structure may be subjected should be taken into account, and the fracture toughness test should be carried out at the appropriate rate, so as to simulate the structural configuration of interest. Mechanistically, the reduction in cleavage toughness generally observed with increasing loading rate is due to the increase in yield strength associated with high loading rates, which elevates the crack tip stresses such that critical conditions in the crack tip region are reached at lower levels of remote load than under static conditions.

There are many examples where structural behaviour is affected by dynamic loading rates. For example, brittle fracture events in steel frame buildings during recent earthquakes in Japan and North America are likely to have been influenced by dynamic loading rate effects caused by the rapid ground movement during the earthquake event. Postulated high loading rate earthquake events are also worst-case accident scenarios during the design of nuclear pipework and storage tanks in certain locations. Other design scenarios causing high loading rates are accidental drops of heavy lids on nuclear storage containers, and impacts in land, sea and air transport.

Typical examples of loading rates are given in Table 1, which was produced from information given in Ref 1 to 3 . These should be taken as estimates only since the exact values will depend strongly on local geometry, loading configuration and flaw dimensions. Loading rates in Table 1 are expressed as strain rate, Symbol.1., and the rate of increase of the linear elastic stress intensity factor, Symbol.2.. The latter parameter is a more appropriate measure for loading rates in fracture mechanics specimens and flawed structures as it includes the relevant geometries and flaw dimensions.

Table 1 Typical loading rates in some engineering components

Application Symbol.1., s -1 Symbol.2., MPa √m 0.5s -1
Storage tanks, pressure vessels 10 -6 to 10 -4 10 -2 to 1
Bridges, cranes earthmoving 10 -2 to 0.1 10 to 10 3
Earthquake loading
Marine collisions
0.1 to 10 100 to 10 4
Land transport, aircraft undercarriage 10 to 1000 10 3 to 10 6
Explosion, ballistics 10 4 to 10 6 plus 10 7 to 10 10 plus

edot.gif

Symbol.1.

kdot.gif

Symbol.2.

The present paper provides an overview on test techniques for determining tensile and fracture toughness properties under dynamic conditions and examples of effects of dynamic rates on tensile properties and fracture toughness, as well as structural assessment considerations. The emphasis is on ferritic steels with some consideration given to stainless steels.

2. Experimental Methods

2.1 Introductory Remarks

Dynamic test methods and standardisation of such methods are developing rapidly. This overview paper cannot cover every aspect in detail, and it is inevitable that there will be some omissions.

Dynamic tests have been used and standardised to characterise qualitatively material properties for many years. Such tests include Charpy and Pellini drop-weight tests. Although valuable for comparative purposes, they offer little detailed insight into failure mechanisms and quantifiable properties. This has led to instrumented versions of these tests, and to new test methods being developed over the past 20 years or so. A number of test techniques, initially developed for research purposes, are now being refined and compared to establish if they can meet the strict requirements of standardisation. Repeatability, accuracy and ease of use are vital if such methods are to become accepted. There is co-operation within Europe and with the USA to ensure that any methods are truly international. Seminars arranged by ESIS TC5 at Mol, Belgium in April 1994 [4] , and by ASTM in Seattle, USA in May 1999, showed that there remains great interest in developing and using dynamic test methods.

Dynamic testing is usually expensive. Special testing machines may have to be used. Furthermore, it is normally necessary to instrument the test machine and/or the specimen in some way. High speed recording equipment will be needed, and the analysis of the test date may be complex. Skilled and experienced personnel will be needed to perform the tests and evaluate the results correctly. Extra safety precautions need to be taken. For all these reasons, it is increasingly realised in Europe and the USA that such tests should be performed and analysed using, as far as possible, agreed international standard methods. This paper will consider dynamic tensile testing and fracture testing in both linear-elastic and elastic-plastic regions.

There are three general regions at which dynamic tests can be performed.

  1. Medium rates, when inertial effects are negligible or can be controlled and quasi-static analyses remain applicable.

  2. High rates where inertia dominates, and special measurement and analyses techniques must be applied.

  3. Very high rates where stress wave loading is dominant.

This paper is limited to the first two regions.

2.2 Fracture Toughness Test Methods and Standards

The British standard BS 6729:1987 was the first standard for fracture testing at dynamic loading rates. It is based on static procedures with allowances made for dynamic condition. Round-robin testing carried out during the development of the standard is reported in Ref.1. The standard gives procedures for determining K Ic and CTOD values at Symbol.2. extending from the quasi-static region at 2.5MPam 0.5s -1 to dynamic loading at 3000MPam 0.5s -1, corresponding to machine crosshead rates of about 0.02 to 100mm/s. It is to be replaced in due course by BS 7448:Part 3. For Symbol.2. exceeding 3000MPam 0.5s -1, an appendix exists in BS 6729 which gives guidance as to how to carry out tests. The main issue with such high loading rates is that conventional instrumentation is no longer appropriate to record the actual conditions experienced by the specimen. For example, Fig.1 (from Ref 30 ) shows load-time plots recorded during static and dynamic fracture tests using both a conventional load cell and a calibrated strain gauge attached directly to the specimen. It can be seen ( Fig.1a) that both give identical results under static conditions, but only the strain gauge is responsive enough to record load data at high loading rates ( Fig.1b), whilst the conventional load cell data are dominated by inertial effects.

spcswapr99f1a.gif

Fig.1a. Load-time plots under static loading [30]

spcswapr99f1b.gif

Fig.1b. Load-time plots under dynamic loading [30]

For even higher rates, the impact response curve concept has been developed in Germany (e.g. Ref 5 ) and the calibrated crack tip strain gauge method has been developed in the UK [6] and the USA [7] . The overall principle of the latter is essentially to turn the specimen into its own load cell. An instrumented 100mm thick BxB SENB dynamic fracture toughness specimen and a typical load time record obtained from an instrumented 12mm thick SENB specimen are shown in Fig.2

spcswapr99f2a.gif

 Fig.2a. Instrumentation of 100mm thick high rate fracture mechanics specimens

spcswapr99f2b.gif

Fig.2b. Load-time record for fracture toughness test

There is considerable interest in obtaining fracture mechanics information from testing small specimens, and a number of methods are in use or have been proposed (e.g. [8] ) to employ precracked Charpy specimens for fracture toughness tests. ESIS TC5 has produced a draft document available from the second author which defines the range of tests available and the types of test machines which are suitable.

To characterise fully ductile behaviour, fracture toughness values at the initiation of tearing or tearing resistance curves are required. There are no accepted standards yet to carry out such tests under dynamic loading conditions but a number of test methods have been suggested [9-13] and reviewed in 1989 by MacGillivray and Turner [14] . One method, due to Chipperfield, is illustrated in Fig.3. Shoulders are machined into the sides of an impact specimen which cause the specimen to 'slip' between the impact anvils at a given amount of load line displacement, associated with a given amount of tearing. Various shoulders can be machined to give various amounts of crack growth to establish a dynamic R-curve.

spcswapr99f3.gif

Fig.3. Chipperfield impact test set-up


2.3 Dynamic Tensile Tearing

Dynamic tensile testing methods are standardised in a ESIS TC5 document which builds on the tensile test method included in BS 6729:1987, and incorporates the wide experience of the ESIS TC5 members, particularly MPA Stuttgart [15] . The strain rates are divided into three ranges; from quasti-static 10 -3 up to 1s -1, 1 to 100s -1, and 100 to 1000s -1, with increasing complexity. At the lower rates, a load cell and low-mass extensometer can be used to measure force and strain, respectively. At intermediate rates, the specimen must be instrumented with a calibrated strain gauge on an enlarged dynamometer section (see Fig.4a), which remains elastic during the test; and strain is measured by strain gauges or rapid response non-contact methods. A typical load-time trace from a dynamic tensile test using a specimen configuration as in Fig.4a is shown in Fig.4b. For tests at the highest strain rates where equilibrium conditions may not be achieved, special considerations are necessary, see for example Ref.16 .

spcswapr99f4a.gif

Fig.4a. Tensile specimen for high rate tests

spcswapr99f4b.gif

Fig.4b. Load-time record for tensile test


A ESIS TC5 round-robin programme has been performed by seven laboratories to compare the proposed standard with reasonable success reported [17] . The ESIS test protocol [15] also specifies procedures for determining stress/strain curves.

3. Trends in Test Results

3.1 Tensile Properties

It has been long known (e.g. Ref.18 , cited in Ref.16 ) that increases in loading rate tend to increase the yield and tensile strengths of steels. An example of typical results for a 20 Mn Mo Ni 55 (similar to A533B) pressure vessel steel is shown in Fig.5, where yield and tensile strength results obtained from round-robin testing according to the proposed ESIS protocol are plotted against strain rate.


spcswapr99f5.gif

Fig.5. ESIS Round-Robin - Strain rate effect on strength


A further data collection for plain carbon steel (reproduced from Ref.16 ) is shown in Fig.6, showing the increase in yield strength with respect to the static value, plotted against strain rate. This plot reveals three regions. Up to a strain rate of about 1s -1 only a small to moderate increase in yield strength of about 70MPa for this steel type is obtained. The strain rate sensitivity increases beyond 1s -1, up to a value of about 10 3s -1. (Data in Fig.5 were up to this strain rate values only.) Beyond 10 3s -1 the sensitivity to strain rate becomes extreme.


spcswapr99f6.gif

Fig.6. Increase in yield strength due to strain rate [16]

The effect of temperature on strain rate dependence of yield and tensile strength [18] is shown in Fig.7 for a mild steel (figures adopted from Ref.16 ). The room temperature behaviour ( Fig.7a) is similar to that shown in Fig.5 and 6. At 600°C ( Fig.7b), there is an increased strain rate dependence compared to room temperature, especially at lower strain rates. At a temperature of 200°C ( Fig.7c), the steel investigated displays negative strain rate sensitivity of the tensile strength up to strain rates of about 10s -1, due to strain rate dependent microstructural processes (dynamic strain ageing). Hence, for the purpose of establishing constitutive equations, care must be taken when extrapolating or even interpolating data obtained at temperatures or strain rates not representative of the situation under consideration.


spcswapr99f7a.gif

Fig.7a. True stress versus strain rate at RT [18]

spcswapr99f7b.gif

Fig.7b. True stress versus strain rate at 600°C [18]

spcswapr99f7c.gif

Fig.7c. True stress versus strain rate at 200°C [18]

An example of the effect of increasing strain rate on stress-strain curves is shown in Fig.8 for a 20 Mn Mo Ni 55 pressure vessel steel. Increasing strain rates result in higher stress/strain curves, and an increased susceptibility to the formation of upper yield strength behaviour.

spcswapr99f8.gif

Fig.8. ESIS Round-Robin - Dynamic stress-strain curves

3.2 Fracture Toughness Behaviour

3.2.1 Brittle Fracture

Early investigations (e.g. Ref.19, 20 ) focused on the effect of strain rate on the plane strain (brittle) fracture toughness K Ic, noticing especially that an important effect of loading rate is the shifting of the fracture toughness transition curve to higher temperatures, and that this shift is dependent on the strength of the steel, higher shifts being generally obtained for lower strength steels. Figure 9 depicts this trend for ABS-C and 18 Ni marageing steels (yield strength approximately 270MPa and 1700MPa, respectively). The ABS-C steel ( Fig.9a) exhibits marked loading rate sensitivity (the 40 ksi √in ( ≈ 44MPa √m) transition temperature is shifted by about 170°F ( ≈ 100°C)), whilst there is no shift for the very high strength steel ( Fig.9b).

spcswapr99f9a.gif

Fig.9a. Effect of temperature and strain rate on crack toughness of ABS-C steel [19]

spcswapr99f9b.gif

Fig.9b. Effect of temperature and strain rate on crack toughness of 18Ni (250) maraging steel [19]


At a given temperature, the effect of loading rate (expressed as the rate of increase of the applied stress intensity factor, kdot, in the elastic loading region of the specimen) is generally to reduce the measured K Ic values. Examples are shown in Fig.10 for a BS11 rail steel ( Fig.10a) and a A533B pressure vessel steels ( Fig.10b and 10c) [6,21,22] . Figure 10c also shows that the loading rate sensitivity of A533B steel fracture toughness increases with increasing temperature. However, not all steels exhibit negative fracture toughness loading rate sensitivity. Figure 11 shows data for alloy steels showing positive loading rate sensitivity over a range of temperatures ( Fig.11a) or a change in loading rate sensitivity of fracture toughness with increasing kdot at some temperatures ( Fig.11b). Whilst the reasons have not been investigated in detail, it is likely that strain rate dependent microstructural process such a dynamic strain ageing are affecting the test results.


spcswapr99f10a.gif

Fig.10a. Effect of loading rate on fracture toughness of Rail steel [6]

spcswapr99f10b.gif

Fig.10b. Dynamic K Ic data for A533B [21]

spcswapr99f10c.gif

Fig.10c. Loading rate and temperature dependence of K Ic for A533B [22]

spcswapr99f11a.gif

Fig.11a. K Id data for 0.4% C Ni Cr Mo steel [22]

spcswapr99f11b.gif

Fig.11b. K Id data for 0.4% C Ni Cr Mo V steel [22]

However, for the majority of steels, the K Ic dependence can be summarised as shown in Fig.12: K Ic decreases with loading rate and the sensitivity to loading rate increases with increasing temperature. At very high loading rates an increase in K Ic may be observed (indicated by the dotted line in Fig.12), attributed to adiabatic heating of the specimen during the test.

spcswapr99f12.gif

Fig.12. Schematic K Ic vs Symbol.2. plot for low to medium strength steels


3.2.2 Ductile Fracture and Tearing Behaviour

Ductile tearing resistance of steels is strongly affected by the tensile strength and the strain hardening behaviour of the material. As tensile properties tend to increase with loading rate, it is generally observed that increased loading rates lead to increased ductile fracture toughness. Examples for a BS 4360 Grade 50D C-Mn structural steel are shown in Fig.13 where increasing ('dynamic') loading rates lead to increased resistance (R) curves. This behaviour is shown schematically in Fig.14, the effect on R-curves being illustrated in Fig.14a, whilst the effect on the toughness at initiation, or at a given amount, of ductile tearing is shown in Fig.14b.


spcswapr99f13a.gif

Fig.13a. Effect of loading rates on R-curves [14]

spcswapr99f13b.gif

Fig.13b. Dynamic and static J-R curves for C-Mn steel (BS4360) [13]

spcswapr99f14a.gif

Fig.14a. Rated effect on R-curves for high work hardening

spcswapr99f14b.gif

Fig.14b. Rate effect on ductile fracture toughness

Fully ductile behaviour does not always lead to an increase in toughness with increasing loading rate. Recent work by the International Piping Integrity Research Group (IPIRG) [23] has demonstrated that whether or not an increase in ductile toughness is observed depends on steel type. For stainless steel piping material, an increase in dynamic ductile toughness with respect to static values was determined, similar to the data shown in Fig.13-14 . However, for ferritic carbon steel pipe (typically A106 Grade B or A333 Grade 6), a decrease in ductile initiation toughness (or toughness at a given amount of tearing) was frequently observed. These steels exhibited low yield to tensile strength ratios and the toughness behaviour of the two steel types investigated is shown in Fig.15 which shows the ratio of dynamic to static toughness versus yield to tensile strength (Y/T) ratio. It can be seen that carbon piping steels exhibiting Y/T ratios of less than 0.5, dynamic ductile toughness values can be smaller than those observed at static rates. Whether this trend holds for different ferritic steel types with similar Y/T properties remains, at present, unknown. However, whilst the use of static R-curves is conservative when assessing dynamic scenarios in stainless steel components, the data shown in Fig.15 demonstrate the need to verify this assumption for ferritic steels which may exhibit negative loading rate dependency of ductile fracture toughness.


spcswapr99f15.gif

Fig.15. Ductile toughness versus Y/T ratio [23]


3.2.3 Fracture Behaviour in the Transition Region

The influence of loading rate on fracture toughness in the transition region of ferritic steels is complicated by the fact that there exists the possibility of a loading rate-induced change in fracture mode. Specimens behaving in a fully ductile fashion at static loading rate may exhibit completely brittle behaviour when subjected to impact loading rates. For instance Fig.16a shows static and dynamic fracture toughness test results at -23°C of an A333 steel. At static and intermediate loading rates, fully ductile behaviour is observed and fracture toughness values in excess of 10,000N/mm 3/2 ( ≈ 310MPa √m) are observed. However, at impact loading rates, a change in fracture mode occurs in some specimens and these exhibit markedly decreased fracture toughness (minimum value less than 2000N/mm 3/2 ( ≈ 60MPa √m). Mechanistically, this effect can be ascribed to the increase in yield strength with loading rate. In the crack tip region this means that higher crack opening stresses can be achieved before plastic flow occurs. If the loading rate-induced yield strength elevation is sufficient, the opening stresses can reach critical cleavage stress magnitude and hence induce a change in fracture mode from ductile to brittle.


spcswapr99f16a.gif

Fig.16a. K Id versus Symbol.2. for A333 steel

A further example of this behaviour is shown in Fig.16b which shows data of a deliberately temper embrittled C-Mn submerged arc weld metal at three temperatures. At -25°C the expected decrease in brittle fracture toughness with increasing loading rate is obtained (compare Fig.10). At 50°C, the fully ductile toughness increases with loading rate (as expected, compare Fig.14b) until at the highest loading rate a change in fracture mode is induced which causes a marked reduction in fracture toughness. The 0°C data show behaviour lying between these two boundaries. The effects are summarised schematically in Fig.16c.


spcswapr99f16b.gif

Fig.16b. Fracture toughness versus loading rate

spcswapr99f16c.gif

Fig.16c. Rate effect in transition region

3.2.4 Overview

The overall effect of loading rate on toughness can be illustrated with reference to the typical fracture toughness transition curves shown in Fig.17 for static and dynamic loading rates. For brittle behaviour (at temperature T 1 in Fig.17) increasing loading rates lead to a reduction in fracture toughness (compare data in Fig.10 and schematic in Fig.12, but beware material-specific exceptions as shown in Fig.11). For ductile behaviour (at temperature T 3 in Fig.17), the initiation toughness increases for most steels with increasing loading rate (see Fig.13, 14, but note exceptions shown in Fig.15). In the transition region (at temperature T 2 in Fig.17), increasing loading rate ( Symbol.2. 2 > Symbol.2. 1) can lead to an increase in toughness if ductile behaviour prevails, but a loading-rate induced change in fracture mode is possible, leading to marked reduction of toughness at a given temperature (compare data and schematic in Fig.16).


spcswapr99f17.gif

Fig.17. General effect of loading rate on transition


4. Structural Behaviour and Engineering Critical Assessment Considerations

4.1 Structural Behaviour

So far, the paper has focussed on the effect of loading rate on material properties. For structural integrity assessment purposes, the effect on the structural response has also to be considered and whilst this paper is not intended to provide detailed comments regarding this aspect, it is felt necessary to mention the need to consider these effects for completeness. Finite element stress analyses can quantify the effect of inertia and stress waves on the structural stress state due to dynamic loading. Results can be validated using experimental methods (such as strain gauge techniques). The overall approach for both uncracked components and flaw-containing structures is outlined in Fig.18.

spcswapr99f18.gif

Fig.18. Effect of elevated loading rates on structural behaviour


4.2 Engineering Critical Assessment Considerations

Once the actual stress state and applied loading rates for the component or structure to be assessed have been determined, appropriate input parameters (tensile and toughness properties) can be determined and the assessment can be carried out using assessment procedures such as BS 7910:1999 or R6. The importance for consideration of loading rate effects is recognised in the text of BS 7910 which states in Clause 7 that 'fracture toughness tests should take account of (...) rate of loading (...) experienced in service'.

The importance of this issue has recently been demonstrated by Lorentzon and Eriksson [24] who measured fracture toughness as a function of loading rate for existing bridge steels, and compared the results with both the current limit of one of the relevant static fracture toughness test standards and the possible applied loading rates experienced by a bridge steel girder containing a flaw. The results are summarised in Fig.19 which shows the typical decrease of brittle toughness with loading rate. (Note that a clear reduction of fracture toughness even below the current limit of the test standard is evident). Figure 19 also includes calculated typical applied Symbol.2. values of bridge steel girders of freespan, L, containing a through-thickness crack of length, a, subject to train loadings travelling at a speed, v. It can be seen that the resulting Symbol.2. values can reach 300MPa √m/s which by far exceeds the limit of the static test standard. Hence, a fracture assessment of a bridge should be based on fracture toughness values determined at the appropriate loading rate.


spcswapr99f19.gif

Fig.19. Toughness results and loading rates in short span bridges [24]

It should be noted that toughness requirements in some fabrication codes for bridge structures do include consideration of dynamic loading, and whilst the above comments apply generically, the loading rate effects on significance of potential fabrication flaws in bridges have been taken into account in the derivation of toughness stipulations (e.g. Ref.25 ).

4.3 Prediction of Loading Rate effects on Material Properties

Frequently, it is impossible or impractical to carry out tensile or fracture toughness tests on steel used in existing structures. Therefore, efforts have been underway since the 1970s to predict the effect of loading rate on tensile and toughness properties without having to carry out additional tests. Based on an extensive test programme on steels of varying strength, Rolfe and Barson [26] developed an empirical predictive relationship for structural steel to calculate the influence of the time period and the temperature of a dynamic event on the yield stress:

spcswapr1999e1.gif
[1]

σ yd is the yield stress in MPa at the temperature T (in °K) and the time period t (in seconds) of the event; σ ys is the static, room temperature yield stress in MPa. This formula would give a rise in yield strength of 65MPa for a isothermal test time of 40ms (corresponding to a strain rate of about 10s -1 for a 50mm gauge length steel specimen at room temperature) and of 140MPa for a isothermal test time of 0.4ms (corresponding to strain rate of about 1000s -1 for conditions as above).

Barson [27] also developed relations predicting the shift in fracture toughness transition curve with increasing loading rate, as shown in Fig.20. It can be seen that the predicted shifts decrease with increasing yield strength (compare data in Fig.9). The relationship shown in Fig.20 has been enhanced more recently by Ainsworth and Neale [28] to include a term depending on the actual value of the fracture toughness transition temperature as experimental evidence has shown that loading rate-induced shifts in transition temperature are greater for steels exhibiting lower absolute transition temperature values.


spcswapr99f20.gif

Fig.20. Predicted fracture toughness transition shift [27]


An additional predictive framework has been developed by Wallin [29] . The predicted fracture toughness transition temperature shift, Δ T o (in °K) depends on loading rate, Symbol.2. (in MPa √m/s), static fracture toughness transition temperature, T o stat (in °K for a loading rate, of Symbol.2. = 1MPa √m/s) and the static yield strength, σ ys (in MPa), as follows:

spcswapr1999e4.gif
[2a]

with:

spcswapr1999e5.gif
[2b]

For a ratio of dynamic to static loading rate of 10 4, the predicted shifts are as shown in Fig.21. It can be seen that the shifts are a function of yield strength (with lower predicted shifts for higher yield strength, as in Fig.20), but also depend strongly on absolute value of static fracture toughness transition temperature, with markedly lower predicted shifts for higher transition temperatures.


spcswapr99f21.gif

Fig.21. Predictions of fracture toughness transition temperature shifts for a ratio of dynamic to static loading rate of 10,000 according to the method proposed by Wallin [29]

Equations [1] and [2] permit predictions of material property changes to be made when material data cannot be generated. This enables preliminary engineering assessments to be made to study the sensitivity of the integrity of a particular component with respect to the effect of dynamic loading rates.

5. Summary and Concluding Remarks

An overview of the effects of high loading rates on the tensile and initiation toughness properties of steels and their effect on structural behaviour is given. Issues affecting engineering critical assessments are also considered. Dynamic loading rates affect both the material resistance and the structural response of engineering components and it is the combination of these two influences which determines the structural behaviour of uncracked components. For structures or components containing crack-like flaws, it is the effect of loading rate on the yield strength which affects crack tip stress fields and through this the initiation fracture toughness at elevated loading rates. Again, it is the combination of the structural response to high loading rates and the effect on dynamic fracture toughness which determines structural behaviour of flawed components.

Because of the inherent complexities of investigating material behaviour under high loading rates, the development of suitable test techniques for measuring both strength/fracture toughness has been a topic of marked investment in the last 40 years or so, but increasing consensus has now been achieved leading to the development of test standards.

There is an important effect of loading rate on tensile properties, especially at very high rates. Generally, the effect of increasing loading rate is to increase strength (positive strain rate dependence), but microstructural influences (such as dynamic strain ageing) can cause negative strain rate dependence). Strain rate sensitivity increases with temperature.

The effect of loading rate on the initiation fracture toughness of ferritic steels is dependent on overall material behaviour. For temperatures below the brittle to ductile transition region (i.e. for cleavage toughness or overall brittle behaviour), toughness decreases with increasing loading rate. At upper shelf temperatures, the ductile initiation toughness and tearing resistance generally increases, but this depends on strain hardening behaviour as decreases in ductile toughness with increasing loading rates have been observed for certain ferritic steels with low yield to tensile strength ratios. In or near the transition region, increasing loading rates can cause a shift from fully ductile behaviour at static rates to brittle behaviour at high rates of loading. The brittle to ductile transition temperature of ferritic steels increases with increasing loading rates and methods have been proposed to predict this shift.

6. References

1 P.R. Christopher et al (IIW Commission X, UK Briefing Group on Dynamic Testing), 'Some Proposals for Dynamic Toughness Measurement', Proc. Conf 'Dynamic Fracture Toughness, The Welding Institute, 1977.
2 M.G. Dawes, 'The Relevance of Deformation Rate to Brittle Fracture Initiation', The Welding Institute Research Bulletin, 1976, pp.153-156.
3   WES 2805:1997, 'Method of Assessment for Flaws in Fusion Welded Joints with Respect to Brittle Fracture and Fatigue Crack Growth', Japan Welding Engineering Society, 1997.
4   ESIS TC5 publication of Mol Seminar: 'Evaluating material properties by dynamic testing', ESIS Publication No. 20, EMAP, 1996.
5 M. Böhme and J.F. Kalthoff, 'On the quantification of Dynamic Effects in Impact Loading and the Practical Application for K Id Determination', Journal de Physique, Colloque C5, Vol.46, August 1985, pp.C5-213 + C5-218.
6 H.J. MacGillivray, 'Project to Develop a Standard Method of Fracture Toughness Testing at Very High Loading Rates', Prepared by the European Group on Fracture Working Party on High Rate Conventional Fracture Testing, DTI Ref. RTP2/155/78, 1990.
7 J. Duffy and C.F. Shih, 'Dynamic Fracture Toughness Measurement Methods for Brittle and Ductile Materials', Proc Conf ICF 7, Vol.1, pp.633-642, Pergamon Press, 1984.
8   Proposed standard method for instrumented impact testing of sub-size Charpy V-notch specimens of steel, draft 7, ESIS, TC5, June 1997.
9 H.J. MacGillivray, V. Grabulov and E.R. Akum, 'Development of the D.C. Potential Drop Method for Dynamic J-R Curve Testing of Charpy Specimens, International', International Symposium on Welding and Testing of Metal Structures in Civil Engineering, Belgrade, 1987.
10 R. Salzbrenner and T.B. Crenshaw, Multiple Specimen J-Integral Testing at Intermediate Rates, Experimental Mechanics, September 1990, pp.217-223/
11 M. Satoh, T. Funada, Y. Urabe and K. Hojo, 'Measurement of Rapid-Loading Fracture Toughness J Id', Factors That Affect the Precision of Mechanical Tests, ASTM STP 1025, R. Papirno and H.C. Weiss, Eds, American Society for Testing and Materials, Philadelphia, 1989, pp.63-76.
12 J.A. Joyce and E.M. Hackett, 'An Advanced Procedure for J-R Curve testing Using a Drop Tower', Nonlinear Fracture Mechanics: Vol.1 - Time-Dependent Fracture, ASTM STP 995, A. Saxena, J.D. Landes and J.L. Bassani, Eds, American Society for Testing and Materials, Philadelphia, 1989, pp.298-317.
13 C.S. Wiesner, 'Dynamic J-R Curve Determination', Oberflächen Werkstoffe, Vol.34, No.12, 1993, pp.12-16.
14 H.J. MacGillivray and C.E. Turner, 'A comparison of Dynamic R-Curve Methods', Fourth International Conference on the Mechanical Properties of Materials at High Rates of Strain, Oxford, 1989.
15   Proposed standard method for dynamic testing, Draft 4, ESIS TC5, March 1997.
16   Metals Handbook, 9 th Edition, Vol.8 Mechanical Testing.
17 H.J. MacGillivary and A. Klenk, 'High rate tensile testing - progress towards a standard method', Euromat Conference 1994, Hungary.
18 M.J. Manjoine, 'Influence of Rate of Strain and Temperature on Yield Stresses of Mild Steel', J. Appl. Mech, Vol.II, 1944, pp.1-211 to 1-218.
19 A.K. Shoemaker and S.T. Rolfe, 'Static and Dynamic Low-Temperature K Ic Behaviour of Steels', Transactions of the ASME, 1969, pp.512-518.
20 A.K. Shoemaker and S.T. Rolfe, 'The Static and Dynamic Low-Temperature Crack-Toughness Performance of Seven Structural Steels', Engineering Fracture Mechanics, Vol.2, Pergamon Press, UK, 1971, pp.319-339.
21 D.R. Ireland, 'Critical Review of Instrumented Impact Testing', Proc. Conf. Dynamic Fracture Toughness, Paper 5, The Welding Institute, 1977.
22 A.M. Priest, 'Influence of Strain Rate and Temperature on the Fracture and Tensile Properties of Several Metallic Material', Proc. Conf. Dynamic Fracture Toughness, Paper 10, The Welding Institute, 1977.
23 D.L. Rudland, C. Marscall and G.M. Wilowski, 'Effect of Dynamic Loading Corresponding to Seismic Rates on the Fracture Toughness of Nuclear Piping Steels,' ASME PVP - Vol.350, ASME 1997, pp.89-96.
24 M. Lorentzon and K. Eriksson, 'Influence of Intermediate Loading Rates and Temperature on the Fracture Toughness of Ordinary Carbon-Manganese Structural Steels', Fatigue and Fracture of Engineering Materials and Structure, 1998, Vol.21, pp.805-817.
25 J.M Barson, 'Development of the ASSHTO Fracture Toughness Requirements', Engg. Frac. Mechs., Vol.7, 1975, pp.605-618.
26 S.T. Rolfe and J.M. Barson, 'Fracture and Fatigue Control in Structures, Prentice-Hall, New Jersey, 1977, p.86.
27 J.M. Barson, 'Effect of Temperature and Rate of Loading on the Fracture Behaviour of Various Steels', Proc. Conf. Dynamic Fracture Toughness, Paper 31, The Welding Institute, 1977.
28 R.A. Ainsworth and B.K. Neale, 'Effect of Strain Rate and Temperature on Fracture Toughness', Nuclear Electric Technology Division Report TD/SEB/MEM/4021, June 1992.
29 K. Wallin, 'Effect of Strain Rate on the Fracture Toughness Reference Temperature, T o, for Ferritic Steels', Proc. Conf. 'Recent Advances in Fracture' (R K Mahidhara, Ed), pp.171-181, The Minerals, Metal and Materials Society, 1997.
30 R.L. Jones and P.C. Davies, 'Experimental characterisation of dynamic tensile and fracture toughness properties', Fat. Frac. Engng. Mat. Struct., Vol.12, (1989), 423-437.