COMPUTER AIDED EVALUATION OF THERMAL

FATIGUE CRACKS ON HOT-WORK TOOL STEEL

X. Wu and L. Xu

School of Materials Science and Engineering

Shanghai University

Shanghai 200072

P. R. China

Abstract

The Uddeholm heat-checking scale of thermal fatigue evaluation was ana-
lyzed by building on a system with image collecting and analyzing. The
results indicate that the Uddeholm heat-checking scale is not enough scien-
tific. It is very difficult to evaluate the degree of thermal fatigue accurately
by naked eyes. For avoiding the defect of the Uddeholm heat-checking scale,
the concept of a thermal fatigue damage factor was put forward, and a cor-
responding computational software was programmed. The thermal fatigue
behavior of 8407 and 4Cr5MoSiV1 steels were analyzed quantitatively by
the software, and the variational rule of thermal fatigue of the two steels
was unveiled. The computer-aided evaluation system of thermal fatigue was
actualized.

Keywords:

hot work steel, thermal fatigue, Uddeholm heat-checking scale, computer-
aided evaluation

INTRODUCTION

Surface cracking due to thermal fatigue is an important engineering prob-

lem since it is a major life limiting damage mechanism in the die casting
process. It is a hot spot in materials science, mathematics and physics to
raise the resistance to heat checking and clarify the mechanism for it [1, 2, 3].

The earliest test method of thermal fatigue, which now becomes a basic

experimental method, was developed by Z.F. Coffin [4] in 1950. Whereas,
this method needs large-scale equipment and the testing cost is too expensive.

781

782

6TH INTERNATIONAL TOOLING CONFERENCE

Moreover there is a great difference between the sample and actual hot
working dies in damaged condition using the test method. So the testing
result obtained by the Coffin method cannot represent real thermal fatigue
performance of tooling steel. Other testing methods of thermal fatigue use
plate sample and Uddeholm method [5], which are widely adopted, because
their experimental cost are very low and the stress distribution and damaged
form are very close to actual die-casting dies. The preparation of sample is
simple and the experimental condition is easy to control in the Uddeholm
method, but the evaluation of the degree of heat-checking is often affected
by assessment experience. So it is very difficult to compare the materials
which are close in properties because of the low testing precision. In addition,
some researchers designed their testing set and regulated their experimental
condition according to investigated objective, why the data on the thermal
fatigue obtained by different researchers can not compare to each other.
Up to now, there is not a test method or an evaluation standard that can be
widely acceptable. Because testing methods are different, the thermal fatigue
behaviors in the same condition present great difference. Therefore the work
of perfecting a method of thermal fatigue test and studying a computer-aided
evaluation of thermal fatigue cracking is necessary.

In this paper, the heat-checking scale of the Uddeholm was analyzed and

quantified in detail. A damage factor D for evaluating the degree of thermal
fatigue damage was put forward, a computer-aided evaluation method was
built and the question about the quantity of the Uddeholm thermal fatigue
testing method was resolved. The present work will is intended to be benefi-
cial to die casting industry to understand the mechanisms of thermal fatigue
in die casting dies and to further raise the die life.

TEST MATERIALS AND EXPERIMENTS

UDDEHOLM TESTING METHOD

The Uddeholm method [5] is based on high-frequency heating of the

surface of a small cylindrical solid sample. The unheated core material in
the sample then serves as constraint to thermal strains in the surface. By
carrying out cyclic heating and cooling of the surface between a top and a
bottom temperature thermal fatigue is obtained. The evaluation of the degree
of heat-checking on the surface of the sample is afterwards carried out by

Computer Aided Evaluation of Thermal Fatigue Cracks on Hot-Work Tool Steel

783

comparing with a heat-checking scale, which includes network as well as
leading cracks.

In the paper, thermal fatigue tests were held in a high frequency induction

furnace, which was contributed by Inductoheat Company and rebuilt in our
laboratory, possessing functions such as automatic controlling of heating,
cooling and recording the cycle number. According to the thermal fatigue
test standard of Uddeholm Company [5], which belonged to the category of
self-restricted thermal fatigue tests, the cycle was designed as follows:

Temperature range: room temperature ⇔ 700

◦

C ; heating time: 3.6 sec.;

holding at heat: 1 sec.; cooling time: 8 sec.; holding at heat: 1 sec.; cooling
medium: water; cycle number: 100∼3000.

Samples were treated with the above thermal cycles for 100, 200, 300,

400, 600, 800, 1000, 1200, 1600, 2000, 2500, 3000 times respectively, and
then dipped into dilute hydrochloric acid solution (10%) for 10∼15 min. to
eliminate the oxide skin. Then, the thermal fatigue cracks of samples were
observed in the stereomicroscope.

According to the Uddeholm heat-checking scale, there are A and B assess-

ments, in which A is the network and B is the leading cracks, each divided
into ten gradings separately. The finally thermal fatigue of a sample is to
add the two assessments. These two combined readings is the degree of heat
checking. The thermal fatigue is more serious with the grading increasing.
The excellence of Uddeholm heat-checking scale is very intuitionistic, par-
tial width and holistic scope of the cracks is considered. It’s disadvantage is
the difficulty to quantify, the evaluation of the degree of thermal fatigue is
often affected by the experience of the examiner. Especially, the difference
between A4, A5, A6 and A7, B4, B5, B6 and B7 in the heat-checking scale
is not very distinct, so it is very difficult to compare the materials which are
close in properties because of the low testing precision for beginner.

IMAGE ANALYZE OF THERMAL FATIGUE CRACKS

For evaluating the degree of thermal fatigue damage quantitatively, the

formation of surface crack of thermal fatigue was analyzed by the technol-
ogy of advanced image manipulation. The approach is that thermal fatigue
testing→scanning thermal fatigue cracks → analyzing image of the cracks
→

evaluating thermal fatigue grading automatically by a computer.

The computer-aided evaluation system of thermal fatigue was based on

the Uddeholm heat-checking scale, the area of thermal fatigue cracks and the

784

6TH INTERNATIONAL TOOLING CONFERENCE

average width of the main cracks that were calculated by a computer, which
is different to the conventional Uddeholm method. The final thermal fatigue
evaluation of a sample is to add A grading and B grading in Uddeholm
heat-checking scale. However, the area and the width of cracks can not
be added simply, because the physical unit of the two measured values are
different, otherwise the physics meaning will be confused. The degree of
thermal fatigue damage may be defined by the damage factor D which can
be expressed as D=A·W, where A is the area of thermal fatigue cracks and
W is the average width of the main cracks. The damage factor D increases
with the degree of thermal fatigue damage increasing.

TEST MATERIALS

4Cr5MoSiV1 steel was one of the H13 steel series smelted by electric

slag furnace in China. The contracted rate of forging area was about 84.8%.
After annealing and machining, the section size was 220×520 mm. 8407S
steel was superior H13 steel manufactured in Uddeholm, Sweden. The
contraction rate was about 89.6% and the section size 178×457 mm after
annealing and machining. The compositions of both steels are listed in
Table 1.

Samples were taken from the center of the cross section along the rolling

direction. The quenching temperature was 1025

◦

C and tempering temper-

atures were selected, investigating the effects of tempering temperature on
the fatigue behavior, in 5 levels as shown in Table 2. Every sample was tem-
pered for two times, each 2 hours. Sample size was determined according to
reference [5]. In order to eliminate the existence of grinding cracks, samples
were ground for several times to reach the mirror finish. Sign of samples
after different tempering processes were listed in Table 2.

Table 1.

Chemical composition of test steels,wt%

Brand

C

Si

Mn

P

S

Ni

Cr

Cu

Mo

W

Ti

V

4Cr5MoSiV1 0.42

0.98

0.30

0.018

0.005

0.07

4.93

0.06

1.40

0.02

0.01

0.87

8407S

0.40

1.02

0.41

0.009

0.0005

5.14

1.46

0.93

Computer Aided Evaluation of Thermal Fatigue Cracks on Hot-Work Tool Steel

785

Table 2.

Sign of samples after treatment by different tempering processes

Tempering temperature

Brand

550

◦

C

580

◦

C

610

◦

C

640

◦

C

700

◦

C

4Cr5MoSiV1

01

02

03

04

05

8407S

11

12

13

14

15

QUANTIFICATION OF THE UDDEHOLM
HEAT-CHECKING SCALE

The Uddeholm heat-checking scale was finally established by consider-

ing manifold factors, but it is only the same with naked eyes. To get a digital
standard, the Uddeholm heat-checking scale was analyzed by image manip-
ulation software. Figure 1 and 2 show the results. The percent of the area in
network cracks A and leading cracks B are shown. Comparing the grading,
we found that there was not a linear relation in evaluating network cracks
A3 to A8 and the leading cracks B3 to B8. This indicated that the evaluation
about thermal fatigue cracks was not very scientific and cannot be used to
accurate evaluating the grading of thermal fatigue cracks.

Comparing the average width of the leading cracks in B scale, an increase

rule by degrees come out. It indicated that the width of the leading crack in
the Uddeholm heat-checking scale B can be quantified, as shown in Fig. 3.
However, the estimating scathing degree only to consider the width of cracks
is not complete, the area of cracks should be considered yet. Such as Udde-
holm heat-checking scale similarly, the damage factor D of thermal fatigue
have both network cracks and leading cracks.

DAMAGE FACTOR D

From Fig. 2 and 3, orderliness could be found in the Uddeholm heat-

checking scale. Below the grading of A3 or B3, the area of cracks in the
scale presents similar linearity increase by degrees, and above the grading of
A6 and B6, the width of cracks in the scale presents a linearity relation with
the grading too. The degree of thermal fatigue damage may be defined by
the product of the area of cracks and the average width of the main cracks,
which can be expressed as D=A·W, where A is the area of thermal fatigue

786

6TH INTERNATIONAL TOOLING CONFERENCE

(a) A-1 4.32%

(b) B-1 1.91%

(c) A-2 7.24%

(d) B-2 9.29%

(e) A-3 17.52%

(f) B-3 12.54%

Figure 1.

Area of cracks in Uddeholm heat-checking scale(A-network cracks, B-leading

cracks).

cracks and W is the average width of the main cracks. The damage factor D
correspondingly increases with the grading of the Uddeholm heat-checking
scale. The results show that the change in damage factor D is very great
above the 8 to 10 grading, and that, the change in D from 3 to 7 grading
is small corresponding in the scale. The grading of quantitative standard
of thermal fatigue damage built up, based on the Uddeholm heat-checking
scale revised appropriately, as show in Table 3.

Computer Aided Evaluation of Thermal Fatigue Cracks on Hot-Work Tool Steel

787

(a) A-4 13.99%

(b) B-4 7.43%

(c) A-5 19.17 %

(d) B-5 10.48 %

(e) A-6 14.70%

(f) B-6 12.47%

Figure 1 (continued).

Table 3.

The grading of quantitative standard of thermal fatigue damage (1/1440inch )

The grading of thermal fatigue damage

1

2

3

4

5

6

7

8

9

10

Damage
factor D

0-15

16-35

36-50

51-70

71-90

91-

120

121-

180

181-

350

351-

1000

>1000

COMPARING THE THERMAL FATIGUE BEHAVIOR
OF 4CR5MOSIV1 AND 8407 STEELS

Figure 4 shows the D values of 4Cr5MoSiV1 and 8407 steels that were

treated at 1025

◦

C quenching and 610

◦

C tempering, which expressed the

788

6TH INTERNATIONAL TOOLING CONFERENCE

(g) A-7 16.37%

(h) B-7 12.76%

(i) A-8 14.69%

(j) B-8 4.75%

(k) A-9 15.46%

(l) B-9 15.30%

Figure 1 (continued).

relation between damage factor. It was shown in Fig. 4 that before 400 cycles,
D values for both steels were in very low level and there was no big difference
between the D values of the two steels before 1600 cycles. After 1600 cycles,
the D values increased rapidly showing the thermal fatigue cracks going into
the rapid propagation stage, the longitudinal cracks developed much further
and connected to each other to form the main crack. In this stage, the D
value of 4Cr5MoSiV1 steel was obviously higher than that of 8407s steel.

The effect of tempering temperature on thermal fatigue behavior of the

two steels was shown in Fig. 5, where one could find that thermal fatigue

Computer Aided Evaluation of Thermal Fatigue Cracks on Hot-Work Tool Steel

789

(m) A-10 38.35%

(n) B-10 41.54%

Figure 1 (continued).

Figure 2.

The relation between the area of cracks and the grading of the Uddeholm

heat-checking scale.

resistance of 8407 steel was obviously better than that of 4Cr5MoSiV1 steel
after tempering at 700

◦

C . From Fig. 5 one can still find that the thermal

fatigue damage factor got its lowest value after 610

◦

C tempering, obtained

the medium values after 580

◦

C and 640

◦

C tempering and the highest value

indicated the most serious damage after 700

◦

C . Thus it can be seen, the

thermal fatigue behavior of the two steels can be described accurately by
the thermal fatigue damage factor, and that it can not be carried out by the
Uddeholm heat-checking scale.

790

6TH INTERNATIONAL TOOLING CONFERENCE

Figure 3.

The relation between the average width of cracks and the grading of Uddeholm

heat-checking.

Figure 4.

The relation between D value and cycle number.

CONCLUSIONS

The relation between the area or width of cracks in Uddeholm heat-
checking scale and the grading in the scale was not a linear relation.
The results indicated that the Uddeholm heat-checking scale was not

Computer Aided Evaluation of Thermal Fatigue Cracks on Hot-Work Tool Steel

791

Figure 5.

The effect of tempering temperature on thermal fatigue behavior.

enough scientific. It is very difficult to evaluate the degree of thermal
fatigue accurately by the naked eyes.

The concept of a thermal fatigue damage factor and the grading of
quantificational standard of thermal fatigue damage were put forward,
and corresponding computational software was programmed. The
computer-aided evaluation system of thermal fatigue was actualized.

The thermal fatigue behavior of 8407 and 4Cr5MoSiV1 steels were
analyzed quantitatively by the software, and the variational rule of
thermal fatigue of the two steels was unveiled. The thermal fatigue
damage factor D value of the two steels were similar to each other
before 1600 cycles, but after 1600 cycles, the D value of 4Cr5MoSiV1
steel was obviously higher than that of 8407 steel.

ACKNOWLEDGMENTS

The authors acknowledge the financial support of the Development Foun-

dation Committee of Swedish Uddeholm Tooling AB, Sweden, Shanghai
Municipal Commission of Science and Technology, Shanghai Municipal
Commission of Education and Research.

792

6TH INTERNATIONAL TOOLING CONFERENCE

REFERENCES

[1] C. Burman, J. Bergstr¨om, Proc. Intern. Conf. on Tool Steel for Dies and Molds, Shanghai,

China, 14-16, April 1998, edited by Luoping Xu L. A. Norstrom( Shanghai University,
Shanghai, 1998)p.137

[2] C Rosbrook; R. Shivpuri, Conf.: 17th International Die Casting Congress and Exposi-

tion, Cleveland, Ohio, USA, 18-21 Oct. 1993, (North American Die Casting Association,
Rosemont, Illinois 60018, USA, 1993)p.181

[3] Lars-˚Ake Norstrom, Thermal Fatigue and Thermal Shork Behaviour of Some Hot work

Tool Steels. Proceedings of A Symposium Held in Sunne, Sweden,26-28 Sept. 1983.
(Uddeholm Swedish Institute for Metals Research 1982)p.177

[4] L. E. Coffin, R. P. Wesley, N. Y. Schenectady. Apparatus for Study of Effects of Cyclic

Thermal Stresses on Ductile Metals. J. Trans. ASME. 76 (1954):923

[5] L A Norstrom, M Svensson, N Ohrberg, Thermal-Fatigue behaviour of hot-work tool

steels, Met. Technology. 8(1981):376