Thermal fatigue of cast iron brake disk materials -...

Journal of Mechanical Science and Technology 26 (6) (2012) 1719~1724

www.springerlink.com/content/1738-494x DOI 10.1007/s12206-012-0435-2

Thermal fatigue of cast iron brake disk materials†

Byeong-choon Goo1,* and Choong-hwan Lim2 1New Transportation Research Department, Korea Railroad Research Institute, Uiwang, 437-757, Korea

2R&D Institute, STX Metal Co. Ltd., Daegu, 704-946, Korea

(Manuscript Received June 24, 2010; Revised August 2, 2011; Accepted February 8, 2012)

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Abstract To develop cast-iron brake disks with high heat resistance to thermal shock loading, three candidate materials were developed. The

main components were Fe, C, Si, Mn, Ni, Cr, Mo, Cu and Al. The mechanical and thermal properties of the candidates were measured. Thermal fatigue tests were then carried out using equipment developed by the authors. The possible temperature range of the testing equipment was 20°C ~1500°C. Cylindrical solid specimens φ 20× 80 mm were heated by an induction coil and cooled in water. At an interval of 20~30 thermal cycles, the surfaces of the specimens were examined with a digital microscope to check for thermal cracks. To quantify the total length of the cracks, an image analyzing program capable of measuring the length of cracks on micrographs was devel-oped. It was found the fatigue lifetime of cast iron could be increased by regulating its composition and metallurgical structures.

Keywords: Brake disk; Cast iron, Railway; Thermal crack; Thermal fatigue ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 1. Introduction

Brake disks of rolling stock are exposed to heavy thermal and mechanical loadings during braking, and many thermal cracks in general occur on the surface of the brake disks. These thermal cracks may cause serious accidents, deteriora-tion of brake performance and an increase of maintenance costs due to the necessity for the frequent exchange of brake pads and disks. Therefore, engineers and researchers have tried to develop materials with high resistance to thermal load-ing. Under thermal shock loading such as that which occurs in brake disks, thermal cracks may occur after several thermal cycles. As a result, the crack initiation and propagation charac-teristics of the brake disk material significantly influence the lifetime of brake disks. Sakamoto and Hirakawa [1] developed forged steel disks mounted on wheels for Shinkansen trains. They used fracture toughness and stress intensity factors as parameters to select optimal candidate steels. They showed that thermal crack initiation was not influenced much by the characteristics of the material. Yamabe et al. [2] developed truck brake disks with high thermal strength. They used grey cast iron with nickel. Yamabe et al. also showed that cerium inoculation was effective to increase the number of graphite particles in the microstructure of brake material and that ther-mal fatigue strength was proportional to the number of graph-

ite particles. They used a pin-on-disk type wear tester. In case of rolling mill cylinders, turbine blades, nuclear reactor com-ponents, etc., thermal cracks are often related to high cycle thermal fatigue. Maillot et al. [3] studied thermal fatigue in AISI 304L and 316LN type stainless steel by using a thermal fatigue test device which used direct current for heating and water for cooling. They quantified surface crack lengths and density by image analysis. According to Fissolo et al. [4], thermal fatigue was more damaging than uniaxial isothermal fatigue. To simulate thermal crack networking Kamaya and Taheri [5] applied a body force method and Monte Carlo simulation. They found stress shielding effect played an im-portant role in the initiation of thermal cracks. Wang et al. [6] revealed addition of nano particles including rare earths to Cr-Mo-Cu alloys changed the graphite morphology and de-creased wear loss. Recently, biomimetic technology has been applied to the improvement of material properties. Tong at al. [7, 8] showed thermal fatigue resistance could be improved by making cast iron surface reticular or striated (biomimetic non-smooth surface).

The main objective of the present study is to develop disk materials with high heat resistance. From the results of pre-vious study [10], three new kinds of candidate materials were developed and thermal fatigue characteristics were evaluated by use of a thermal fatigue tester and image analysis program.

*Corresponding author. Tel.: +82 31 460 5243, Fax.: +82 31 460 5031 E-mail address: [email protected]

† Recommended by Editor Associate Chongdu Cho © KSME & Springer 2012

1720 B. Goo and C. Lim / Journal of Mechanical Science and Technology 26 (6) (2012) 1719~1724

2. Materials and thermal fatigue tests

2.1 Materials

To develop high heat-resistant castings for brake disks a va-riety of elements such as Cr, Mo, and Ni have been tried [9]. We, too, have tested cast iron with similar elements [10]. Ta-ble 1 shows three new candidate materials with different com-ponents and a conventional disk material (Conv.) in commer-cial service. The chemical compositions of the materials were analyzed using a spectrometer. Micrographs of the castings are shown in Fig. 1. Graphite particles in the conventional brake disks are shaped as flakes. For materials B, C and D, almost all the graphite particles are compact vermicular. Some spherical graphite particles are scattered in the matrix. To obtain vermicular graphite particles, Mg was added. To make castings, iron scraps were melted in an electric furnace at tem-peratures between 1500 ~ 1550°C. Sand molds were bound by furan resin. FeSi type inoculants were added. The solidified metals were kept in the sand molds until the temperature of the castings decreased to 350°C. At this temperature, the molds were broken and the castings were air-cooled. Me-chanical and thermal properties are shown in Table 2. The elongation, hardness and tensile strength values are the aver-age values of three specimen values. Brinell hardness and tensile strength increase from B to C to D, but the elongation rate decreases. The compressive strengths are much higher

than the tensile strengths, which is known to be characteristic of cast iron. Cast irons B, C and D with vermicular graphite have higher tensile strengths than Conv. with flake graphite, but their Brinell hardness is lower than hardness of Conv. In general, tensile strengths of steels are known to be propor-tional to Brinell hardness. But in case of cast iron, when the metallurgical structures of two types of cast iron are different, that relation is often not so [9]. Mechanical properties and metallurgical characteristics were presented in Ref. [11].

2.2 Thermal fatigue tests

Thermal fatigue test equipment is composed of a PC, a con-troller, an induction heater and a chiller. Tests were controlled and monitored on a personal computer [10, 11]. The possible temperature range of the equipment is 20~1500°C. Tempera-tures were measured by a K-type thermocouple inserted into the hole with a diameter of 1.5 mm. The tip of the thermocou-ple was just below the surface. The induction heater could quickly increase the temperature of specimens. Cylindrical specimens of 20 mm diameter were moved up and down by the air cylinder. They were heated by the induction heater coil in the upper position and cooled in water in the lower position. For the thermocouple to follow the temperature variation of the specimens, the specimens were heated slowly. It took 26.5 seconds for heating and 20 seconds for cooling. The cooling

Table 1. Chemical compositions (Wt. %). Type C Si Mn P S Cu Ni Cr Mo Mg

Conv. 3.13 2.05 0.71 0.041 0.018 0.3 - 0.8 - -

B 3.79 2.39 0.419 0.039 0.013 0.353 0.419 0.222 0.284 0.016

C 3.75 2.41 0.401 0.04 0.013 0.353 1.05 0.403 0.433 0.015

D 3.77 2.33 0.393 0.04 0.014 0.341 1.01 0.534 0.549 0.015

(a) Material Conv. (b) Material B

(c) Material C (d) Material D Fig. 1. Distribution of graphite particles (×100).

Table 2. Mechanical properties.

TypeTensile strength (N/mm2)

Tensile elongation

(%)

Compressive strength (N/mm2)

Hardness (HBW

10/3000)

Conv. 275.0 0.11 693.2 220

B 414.0 3.1 688.4 179

C 483.4 2.1 784.1 209

D 499.6 1.2 824.2 215

Fig. 2. Flowchart of the thermal fatigue test.

B. Goo and C. Lim / Journal of Mechanical Science and Technology 26 (6) (2012) 1719~1724 1721

water was kept at a constant temperature of 25°C by the chiller. After several thermal cycles, the surfaces of the speci-mens were examined with a digital microscope to check whether thermal cracks were generated. Micrographs were taken at six different positions on the surface of each specimen. Each micrograph covered 4.48 × 3.59 mm2. After a few ther-mal cycles thermal cracks were produced. To quantify the length of cracks, an image analyzing program that can measure the length of cracks from micrographs was developed. The basic principle of measurement of crack length was counting the pixels that the cracks on a micrograph occupied. Before this, the thickness the crack lines on the micrographs was made to be 1 pixel. Using this program, we were able to measure the length of all surface cracks on a microscopic photograph by counting the pixels occupied by the cracks and evaluate the propagation characteristics of cracks with respect to thermal cycles. Fig. 2 shows the flowchart of the thermal fatigue test.

3. Results and discussion

The characteristics of crack initiation and propagation may depend on the positions observed on the specimen surfaces, because the specimens are neither homogeneous nor heated and cooled uniformly. We examined six spots on the surface of each specimen and measured the total crack length at each spot. Each observed spot covers 4.48 × 3.59 mm2. Figs. 3 and 4 show the development of thermal cracks after 100 and 780 thermal cycles, respectively, in the specimens made from a disk in commercial service. Figs. 5-10 show the photographs for specimens B, C and D, respectively. On the photographs of 100 thermal cycles, corroded orange lines are seen. Follow-ing consecutive thermal cycles, cracks were generated along the corroded orange lines. It is found that in case of material D small branch cracks are sparser. Fig. 11 shows the image of cracks after 780 thermal cycles for material B. Figs. 12-17 show the total crack length measured on each observed spot for each specimen. Fig. 18 shows the total crack length on six observed spots. It is found that the candidate materials are

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 5. Cracks in material B after 100 thermal cycles.

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 6. Cracks in material B after 780 thermal cycles.

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 7. Cracks in material C after 100 thermal cycles.

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 8. Cracks in material C after 780 thermal cycles.

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 9. Cracks in material D after 100 thermal cycles.

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 10. Cracks in material D after 780 thermal cycles.

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 3. Cracks in material Conv. after 100 thermal cycles.

(a) Observed spot 1 (b) Observed spot 2 (c) Observed spot 3 Fig. 4. Cracks in material Conv. after 780 thermal cycles.


Fig. 15. Crack length vs. thermal cycles on observation spot 4.



Fig. 18. Total crack length vs. thermal cycles.

(a) Observed spot 1

(b) Observed spot 2 (c) Observed spot 3 Fig. 11. Crack images of material B after 780 thermal cycles.




B. Goo and C. Lim / Journal of Mechanical Science and Technology 26 (6) (2012) 1719~1724 1723

more resistant to thermal shock loading until 550 thermal cy-cles and that the initiation life of thermal fatigue of the conven-tional material is shorter than those of materials B, C and D.

The crack propagation characteristics are summarized in Table 3. The lifetimes of crack initiation of the developed materials increased by 93%, 185% and 71% compared to that of material Conv. In Table 3, the total crack length is the summation of crack lengths measured on six observed spots and the crack density is the total crack length divided by the total observed spot area. To examine the crack morphology in the direction of depth from the surface, we cut the specimens at the hole position. Fig. 19 shows crack morphology in the cross sections perpendicular to the axial direction after 780 thermal cycles. The micrographs were taken near the surfaces of the specimens before cutting. It was found the cracks initi-ated at the interfaces between the graphite particles and the matrix joined other cracks also generated at the interfaces. The graphite particles are softer than the surrounding cast iron, so

they act as void or notches. This kind of behavior was ob-served in flake, vermicular and nodular graphite cast iron [7]. In the view of fracture mechanics it can be said the flake cast iron has more acute cracks, because the flake type graphite is thinner and more acute than the vermicular.

4. Conclusions

We performed thermal shock fatigue tests for three kinds of candidate cast iron and a specimen made from a disk in com-mercial service. The main conclusions are as follows:

(1) The thermal fatigue test process used in this study was very effective.

(2) The mechanical properties of a casting can be controlled by varying the Ni, Co and Mo contents and the graphite parti-cle forms.

(3) The three candidate castings have compact vermicular graphite particles and higher resistance to thermal cracks than the disk material in use which includes flake graphite particles. Material C has the highest thermal resistance.

(4) Several dominant cracks propagate from graphite parti-cles to graphite particles.

(5) The three developed materials, which include more ob-tuse graphite particles than the conventional cast iron, have longer crack initiation lifetimes than the conventional material.

Acknowledgment

This study was supported by the Korea Research Council for Industrial Science in affiliation with the Ministry of Knowledge Economy. The authors are grateful for the support.

References

[1] H. Sakamoto and K. Hirakawa, Fracture analysis and material improvement of brake disks, JSME Interna-tional Journal, Series A, 48 (4) (2005) 458-464.

[2] J. Yamabe, M. Takagi, T. Matsui, T. Kimura and M. Sasaki, Development of disk brake rotors for trucks with high thermal fatigue strength, JSAE Review, 23 (2002) 105-112.

[3] V. Maillot, A. Fissolo, G. Degallaix and S. Degallaix, Thermal fatigue crack networks parameters and stability: an experimen-tal study, Int. J. Solids and Structures, 42 (2005) 759-769.

[4] A. Fissolo, S. Amiable, O. Ancelet, F. Mermaz, J. M. Stel-maszyk, A. Constantinescu, C. Robertson, L. Vincent, V. Maillot and F. Bouchet, Crack initiation under thermal fatigue: An overview of CEA experience. Part I: Thermal fatigue ap-pears to be more damaging than uniaxial isothermal fa-tigue, International Journal of Fatigue, 31 (3) (2009) 587-600.

[5] M. Kamaya and S. Taheri, A study on the evolution of crack networks under thermal fatigue loading, Nuclear Engineer-ing and Design, 238 (9) (2008) 2147-2154.

[6] Y. Wang, Z. Pan, Z. Wang, X. Sun and L. Wang, Sliding wear behavior of Cr–Mo–Cu alloy cast irons with and with-out nano-additives, Wear XXX (2011) in press.

[7] X. Tong, H. Zhou, L. Ren, Z. Zhang, W. Zhang and R. Cui,

Table 3. Thermal crack propagation characteristics.

Characteristics Conv. B C D Crack initiation, 40 µm

(cycle) 60 90 110 100

Crack initiation, 1 mm (cycle) 140 270 300 240

Crack length 10 mm (cycle) 240 330 450 330




Total crack length at 780 cycles (mm) 56.36 70.71 48.94 41.23

Crack density (mm/mm2) 3.52 4.42 3.59 2.58

(a) Material Conv. (b) Material B

(c) Material C (d) Material D Fig. 19. Crack propagation from the surface into the inside.


Effects of graphite shape on thermal fatigue resistance of cast iron with biomimetic non-smooth surface, International Journal of Fatigue, 31 (4) (2009) 668-677.

[8] X. Tong, H. Zhou, M. Liu and M. Dai, Effects of striated laser tracks on thermal fatigue resistance of cast iron samples with biomimetic non-smooth surface, Materials & Design, 32 (2) 796-802.

[9] T. Miyauchi and T. Tsujimura, The basic friction and wear characteristics of brake disks and linings in train brakes on the narrow-gage lines, QR of RTRI, 38 (2) (1997) 70-75.

[10] B. C. Goo and C. H. Lim, Thermal fatigue evaluation of cast iron discs for railway vehicles, Procedia Engineering, 2 (1) Fatigue 2010 (2010) 679-685.

[11] C. H. Lim and B. C. Goo, Development of compacted ver-micular graphite iron for railway brake discs, Met. Mater. Int., 17 (2) (2011) 199-205.

Byeong-choon Goo received his Ph.D from the University of Franche-Comte, France and has been a principal re-searcher at Korea Railroad Research Institute since 1996. His main research area includes brake system, brake squeal noise, structural analysis, wheel/rail contact, fatigue, shape memory alloy.

Choong-hwan Lim received his Ph.D from the University of Science and Technology, Korea, 2010 and has been a researcher at the R&D Institute, STX Metal Co., Ltd since 2010. His main research area includes fatigue/fracture, alloys.

Thermal fatigue of cast iron brake disk materials -...

Documents

Transcript of Thermal fatigue of cast iron brake disk materials -...