Correlation between Vickers Hardness Number and Yield Stress of Cold-

Formed Products

Bülent YAVUZ, Middle East Technical University, Ankara, Turkey

(e-mail: byavuz@metu.edu.tr)
A. Erman TEKKAYA, Middle East Technical University, Ankara, Turkey
(e-mail: tekkaya@metu.edu.tr)

ABSTRACT

Hardness measurements provide an easy way of obtaining the increased yield

strength of the cold-formed products. Since these tests have minimum destruction on
workpieces and their performance is very simple and inexpensive. However, the
available models to convert the measured hardness into the corresponding flow stress
have an error of more than 20%. It can be noted that this is resulted from the
hardening behavior and inhomogeneous material properties especially at the
workpiece surface. This study aims to improve the accuracy of the available
correlations between Vickers hardness and yield strength of the material. The
suggested improved model decreases the conversion error to about 5%. The models
incorporate currently only the isotropic strain-hardening behavior of the work material.
The studies to incorporate anisotropic hardening and inhomogeneous hardening
distribution in the surface regions are still continuing.
Keywords: Vickers hardness, flow stress, cold forming.

1. INTRODUCTION

Among the metal forming processes, cold metal forming processes are the most
advantageous one due to low scrap, high production rates, and increased yield
strength of cold formed products after forming operation. In general, the customer of
these products needs the new increased yield strength distribution for design
calculations. However, the products are not appropriate for standard material testing
experiments such as simple tension, simple compression or simple torsion. Thus,
usually either simplified analytic computations or finite element models are used to
predict the new increased yield strength distribution. Alternatively, the local flow stress
is determined by means of hardness measurements, which are converted to the
respective flow stress at the point of measurement.

Hardness tests have for a long time been a standard method for material
characterization as they provide an easy, inexpensive, non-destructive, and objective
method of evaluating basic properties from small volume of materials. As well as
resistance to plastic deformation; stiffness, strength of thin coatings, residual stresses
near the surface, and the fracture toughness of the material are some basic properties
that can be measured by the hardness tests.

In the literature, Vickers hardness number (HV) has been the most popular in
investigation of the relationship between hardness and the flow stress of the material
because of two reasons. Firstly, its superior resolution as compared to spherical
indenters. And secondly, the Vickers indenter is self-similar, through which the
hardness is ideally independent of the indentation load and indentation depth.
Therefore, in this study, Vickers indentation will be the main concern.

2. PREVIOUS STUDIES

A basic review of the first results is covered by Tabor. [1,2]. Tabor has shown that the
mean contact pressure, Pm, (or hardness) can be related to the yield stress of the
material, Y, by an expression based on the theory of indentation of rigid perfectly-
plastic solid.

Indenter Force (kg)

HV =
Surface Area of the Imprint (mm 2 ) (1)
For non-strain hardening materials the Vickers hardness number, which is defined as
in the above formula, can be related to the constant yield stress, Y, by :

HV = 2.9 ⋅ Y to 3.0 ⋅ Y (2)

For strain hardening materials, Tabor suggests on empirical bases to use a similar
expression as in the case of non-strain hardening materials, however, as evaluated at
a representative plastic strain:

HV = 2.9 ⋅ Y (at an engineering plastic strain of 0.08) (3)

3.20
Ratio Vickers Hardness to Yield
Stress (in kg/mm2)

3.00

2.80
M ild S t e e l
A n n e a le d C o p p e r
2.60
0.00 0.10 0.20 0.30
E q u i v a l e n t P l a sti c S tr a i n ϕ

Figure 1: Experimentally determined Vickers Hardness Number and Yield Stress ratios for
various amounts of initial plastic strain, Tabor [1,2]. (yield stresses are evaluated for an
engineering representative strain of 0.08)

Figure 1 shows that the comparison of Tabors formula with his own experimental
results. In this plot, ϕ designates the equivalent true plastic strain, which the specimen
experienced before indentation. It can be observed that the agreement is rather good
for mild steel, whereas the annealed cupper a scatter of about %15 is present. It
should be emphasized that Tabor’s formula is given at various places with slightly
varying coefficients and representative strain values as compared with the values
given in Eqn.3.

4.5
Ratio Vickers Hardness to
Yield Stress (in kg/mm 2)

4.0

3.5

No prestrain (Dannenm ann et al.)

3.0
Prestrain 0.08 (Tabor)

2.5
0 0.4 0.8 1.2 1.6
Equivalent Plastic Strain ϕ

Figure 2: Ratio of Vickers Hardness number and yield stress by Dannenmann and Wilhelm. [3]

The relationship between Vickers hardness number and yield stress as related to
metal forming is investigated by Dannenmann, Wilhelm et al. [3]. The experimental
results can be seen in Figure 2. In their original work, no use of Tabor’s model as
explained above was made, so that an interpretation of their experimental results
utilizing Tabor’s model is also given in this figure. However, even this improved curve
indicates a conversion error of 20%.

In addition to these experimental analyses, analytical models regarding the relation

between hardness number and the yield stress are established for instance by Bishop
et al. [4], Hill et al. [5], Grunzweig et al. [6], Locket [7], Bay et al. [8]. However, these
works are giving insight to the physics of the process and they are restricted by the
assumptions such as non-hardening predefined plastic zone, rigid plastic material law,
etc.

Since 1990s, with incredible development in computers and programming, numerical

analyses based on the finite element method of the indentation process have been
performed. Although these studies gave quantitatively impacts on the understanding of
the indentation process, they are not focusing on workpieces which are cold formed
and, furthermore, they don’t attempt to construct an improved relation between the
hardness number and the yield stress.

As a consequence of the previous researches, it can be said that experimental studies

lack the separation of various factors affecting the hardness and flow stress
relationship, and, analytical and numerical studies lack either quantitative accuracy
and/or they don’t cover metal-forming issues properly.

3. ANALYSIS METHODS AND RESULTS

 A modification was performed in 1998 by Tekkaya [9,10] for the relationship between
Vickers hardness number and the yield strength of the material. Firstly, the Vickers
indentation test has been analyzed by finite element method with the following
assumptions and simplifications:
• The material behavior is of elasto-plastic type.
• Strain hardening is isotropic. (this is a critical assumption since after cold
working it is known that the material exhibits an anisotropic hardening
behavior.)
• The deformation during indentation is temperature and velocity independent.
• Friction between indenter and the workpiece can be neglected.
• Prior work hardening of the workpiece material is homogeneous in all
directions. (this is also a critical assumption since after any technical metal
forming process, the material properties are not homogeneous any more.)
• Three-dimensional Vickers indentation process can be replaced by a cone
indentation with an equivalent cone angle 140.6o. (the equivalent cone angle is
selected such that the displaced volume of the material for the same depth of
penetration is the same for the cone and Vickers pyramidal indenter having
apex angle of 136o.)

Figure 3: Used mesh in the analyses Figure 4: Normalized flow curves of

with 773 axisymmetrical elements. analyzed materials.
o
Semi-cone angle of indenter is 70.3 .

Figure 3 shows a typical finite element mesh used for the analyses, which has been
conducted with the commercial code MARC. Axisymmetrical linear four node elements
have been used in the computation and no re-meshing has been applied.

Simulations have been conducted for five different materials, for which the normalized
flow curves are given in Figure 4. The materials have been selected such that the
strain-hardening exponents (n) cover the wide range. The Vickers hardness for each
material has been computed in the annealed state and four strain-hardened states
corresponding to the equivalent plastic strain of 0.04, 0.10, 0.24, and 0.40. In this
range of plastic strain the assumed Ludwik type of flow curve representation (Y=K⋅ϕn)
is at least moderately accurate.

2400 2400
Materials:
C10, C15, C35, St38, CuZn40
2000 2000
ness HV in MPa
ess HV in MPa

Prediction Interval
1600 1600

99% Confidence
1200 99% Confidence 1200 Interval
Interval Prediction Interval
 Figure 5: Vickers hardness number v.s. Figure 6: Vickers hardness number v.s. the
initial yield stress of material. yield stress of the material computed at an
optimum offset equivalent plastic strain of
0.112.

Figure 5 and Figure 6 exhibit the Vickers hardness numbers (in MPa) versus the initial
yield stress and yield stress at an offset equivalent plastic strain of 0.112, respectively.
In the first plot, it can be concluded that the scatter is unacceptably large with the large
prediction interval. However, the second gives the better accuracy. Here, the offset
strain at which the flow stress is computed has been varied systematically until the
scatter in various hardness-flow stress pairs were minimized. The resulting value of
the equivalent plastic strain was 0.112. The slope of the regression line shown in the
figure corresponds to the coefficient in Tabor’s original formula as given in Eqn.3.
Hence, the modified Tabor’s equation can be given as:

HV = 2.475 ⋅ Y (at an equivalent plastic strain of 0.112) (4)

It can be shown that this new coefficient and new offset strain value result a correlation
between the Vickers hardness number and flow stress with an error less than 10% as
compared with the error of less than 20% for the Tabor’s original model. [10]

After the numerical analysis, the Vickers indentation test experiments of new five
different materials in various strain-hardened states has been done to check the
correlation obtained from the finite element simulations. Before the tests, these
materials have exposed to the upsetting test to obtain their flow curves up to the large
plastic strains. The list of these materials and the coefficients of Ludwik equation
representing their flow curves are given in Table 1.

16MnCr5 20MoCr4 AlMgSi05 AlMgSi1 Cf53

n 0.135 0.140 0.034 0.061 0.130

K (MPa) 912.40 906.60 381.80 481.10 1009.40

Y0 (MPa) 406 403 202 314 360

Table 1: List of the experimental materials and their material constants for Ludwik equation.

Cold direct-extrusion metal forming process has been chosen to get strain-harden the
 materials. Since the process is very suitable for this purpose and gives the exact
equivalent plastic strain at the center of the extruded product. Therefore, the hardness
measurements have been done around the center of the experimental specimens.
(See in Figure 7.)

Figure 7: Extruded specimens in various strain-harden states and the picture of the center
cross-section of the specimen.

The results of the Vickers hardness measurements (in kg/mm2) for the experimental
materials can be seen in Table 2.

16MnCr5 20MoCr4 AlMgSi05 AlMgSi1 Cf53

0.02 149.5 164.3 91.3 113.9 180.9
0.52 201.1 218.1 96.2 223.8
0.92 213.3 239.4 118.4 274.5
1.21 225.6 248.8 103.7 283.2
1.60 252.9 317.1

Table 2: Experimental results of HV for five different strain-harden states. (Vickers hardness
2
numbers in kg/mm )

300 30
25.9

Material: 20MoCr4 Classical Relation

Material: 20MoCr4
25 Tabor's Relation
250
Errors in Predicted HV in %

Proposed Relation
Vickers Hardness HV

17.3
17.3

200 20
16.3

15.5
15.3
14.8

14.7
16

14.6

15
150

10
100
Classical Relation
Tabor's Relation 5
2.2

1.7

-3.3

50
-1.3
-3.4

Proposed Relation
Experiment 0
0 0.02 0.52 0.92 1.21 1.60
0.02 0.52 0.92 1.21 1.60 -5
Equivalent Plastic Strain
Equivalent Plastic Strain

Figure 8: Comparison of Predicted and Measured Hardness Numbers for material 20MoCr4.
Modified relation (eqn.4) gives the best accuracy in the material 20MoCr4 for which the
hardness comparisons and error plots are given in the Figure 8. However, it is
observed that the errors for the other materials are about 5-10%. To increase the
accuracy and verify the proposed relation or to predict a new relation, some additional
works such as new simulations and experiments are required.

As additional analysis, the ten different materials, which are mentioned so far, have
been simulated with the same mesh in Figure 3 to obtain the hardness numbers in the
larger strain-hardened states. And all results are shown in Table 3.

Table 3: Finite element simulation results for Vickers hardness numbers (in MPa) of analyzed
materials.

In Figure 9., the plot of the hardness numbers versus corresponding yield stress
values (in MPa) can be seen.

3000
Vickers Hardness Number (MPa)

2500

2000
16MnCr5
20MoCr4
1500 AlMgSi05
AlMgSi1
C10
1000 C35
Cf53
Ck15
500 CuZn40
St38

0
0 200 400 600 800 1000 1200

Yield Stress Values (MPa)

Figure 9: Plot of Vickers hardness numbers versus yield stress values of simulated materials
(in MPa).
Among these results given in Table 3, the hardness values for non-hardened case
(this can be thought as the hardness measurement values at the head side of the
specimens.) have been analyzed separately. Since these data have a disagreements
with the results of the relations fitted for all hardening case. Therefore, a new relation
that is given in equation 5 is obtained. And the percentage error values between the
Vickers hardness values calculated from the equation 5 and the hardness
measurement values at the head side of the experimental materials can be seen in
Table 4.

HV = 3.04 ⋅ Y (at an offset strain of 0.03) (5)

HV (MPa) HV (MPa) Percentage Error

obtained by experiments obtained by equation 5 (%)

16MnCr5 1466.60 1573.74 7.31%

20MoCr4 1611.78 1582.27 -1.83%
AlMgSi05 895.65 887.54 -0.91%
AlMgSi1 1117.36 1119.54 0.20%
Cf53 1774.63 1800.58 1.46%

Table 4. Comparison of the modified relation (eqn.5) for the non-hardening case with the
experimental results.

Linear approximation analysis for the rest results obtained from the strain-hardened
cases have been done by using different offset strains. Unlike the modified Tabor
relation (eqn. 4), maximum regression is obtained with the offset strain of 0.120 and
the new relation is observed as following:

HV = 2.527 ⋅ Y ( at an offset strain of 0.120) (6)

To understand the effects of the strain-hardened states on the correlation, the

numerical results given in Table 3 have been divided into to two parts such as the
equivalent plastic strain in between 0.00 and 0.50, and, greater than 0.50. These two
parts were grouped separately and again linear approximation method was repeated.
The results are as followings:

HV = 2.528 ⋅ Y (at an offset strain 0.130) for 0.00 < ϕ < 0.50
(7)
HV = 2.520 ⋅ Y (at an offset strain 0.230) for ϕ ≥ 0.50

Another work has been done to analyze the influences of the strain-hardening
exponent (n) in the Ludwik type representation of the flow curves. In this work, three
relations were found for the given interval of n as followings:

HV = 2.50 ⋅ Y (at an offset strain 0.118) for 0.00 < n < 0.10
HV = 2.52 ⋅ Y (at an offset strain 0.116) for 0.10 ≤ n < 0.20 (8)
HV = 2.54 ⋅ Y (at an offset strain 0.135) for n ≥ 0.20

Same analysis has been performed for the material constant K in the Ludwik’s formula
(Y=K⋅ϕn) as following :
 HV = 2.51 ⋅ Y (at an offset strain 0.140) for K < 800 ⋅ MPa
(9)
HV = 2.51 ⋅ Y (at an offset strain 0.130) for K ≥ 800 ⋅ MPa

The comparisons between the predicted hardness numbers of material 20MoCr4 and
the numerical values of its hardness numbers obtained from the Vickers indentation
experiments are shown in Table 5.

Tabor’s Tekkaya’s Approach Approach Approach Approach

Experimental
Approach Approach 1 2 3 4
Results
(Eqn. 3) (Eqn. 4) (Eqn. 6) (Eqn. 7) (Eqn. 8) (Eqn. 9)

0.52 218.10 253.57 214.40 219.23 220.95 218.47 218.27

0.92 239.40 274.63 231.60 234.93 237.59 235.93 235.47

1.21 248.80 285.17 240.20 243.11 245.14 244.67 244.08

1.61 252.90 296.47 249.50 251.98 253.47 250.91 253.40

Percentage Errors with Experimental Results

0.52 16.26% -1.70% 0.51% 1.30% 0.17% 0.08%

0.92 14.72% -3.26% -1.86% -0.76% -1.45% -1.64%

1.21 14.62% -3.46% -2.28% -1.47% -1.66% -1.90%

1.61 17.23% -1.34% -0.36% 0.22% -0.79% 0.20%

Table 5. Comparison of the new modified relations with the experimental results of material
2
20MoCr4. (All hardness results are in kg/mm )

For practical applications, inhomogeneous and anisotropic material behavior should be

included to the model. And new experiments should be done to increase the reliability.

ACKNOWLEDGMENTS
 • This study was a AFP Research Project : AFP-98-03-02-03
• Research assistantship of Bülent Yavuz is supported by the METU Graduate
School of Natural and Applied Science.

REFERENCES
1. Tabor, D. (1947), “A Simple Theory of Static and Dynamic Hardness”, Proc.
Roy. Society Series A., 192, pp 247-274
2. Tabor, D. (1951), The Hardness and Strength of Metals, Oxford Clarendon
Press
3. Dannenmann, E. , Wilhelm, H. and Steck, E. (1968), “Uber den
Zusammenhang zwishen Eindringharte und Umformgrad bei
Kaltumformvorgangen”, Bander Bleche Rohre, pp 368-394
4. Bishop, R.F. , Hill, R. and Mott, N.F. (1945), “The Theory of Indentation and
Hardness Tests”, Phys. Soc., 57.
5. Hill, R. , Lee, E.H. and Tupper, S.J. (1947), “The Theory of Indentation of
Ductile Materials”, Proc. Roy. Soc. A., 188.
6. Grunzweig, J. , Longman, I.M. and Petch, N.J. (1954) “Calculation and
Measurements on Wedge-Indentation”, J. Mech. Phys. Solids, 2.
7. Lockett, F.J. (1963), “Indentation of Rigid-Plastic Material by a Conical
Indenter”, J. Mech. Phys. Solids, 11.
8. Bay, N. and Wanheim, T. (1974), “Axi-symmetric Upper Bound Theory Applied
to Indentation of Cones”, Int. J. Prod. Res., 12.
9. Tekkaya, A.E. (1998), “Hardness Measurements on Cold-Formed
Workpieces”, 31st Plenary Meeting of ICFG, Gothenburg, Germany.
10. Tekkaya, A.E. (1999), “Hardness Measurements on Cold-Formed
Workpieces”, 6thInternational Conference on Technology of Plasticity,
Nuremberg, Germany.