ISIJ International, Vol. 50 (2010), No. 12, pp. 1981–1984

1. Introduction

Cast iron is a prime example of materials where theproperties achieved depend on the characteristics of the mi-crostructure. In ductile iron, the number and size distribu-tion of graphite nodules, as well as the shape of the nod-ules, affects the mechanical, thermal properties.1) The nod-ules to be considered should be round or nearly so. When agraphite particle’s length is two or more times its diameter,it is usually no longer considered a nodule.2) As the numberof the nodules increases the structure and properties become more uniform, segregation is reduced and carbidesgenerally will be minimized. Higher counts will also gener-ally produce more uniform nodule size.

It is well know that in ductile iron, the graphite noduledensity (nodule count) can be significantly influenced bythe charge materials, alloy additions and metal processingincluding treatment and of course, inoculation and castingsection modulus. Among those factors, modulus, carbonequivalent and inoculation quantity play a more importantrole on the shape and count of graphite particles in ductileiron castings. Many researchers studied the relationship ofsection size, carbon equivalent and inoculation quantity tothe shape and count of the graphite particles separately. Inthose works, nodule count was characterized by solidifica-tion time, cooling rates at the eutectic temperature, sectionsize or maximum undercooling.3–6) Few models include thecarbon equivalent though it plays a very important role onnodule count. In order to discovery how these three factorsaffect on the nodules synthetically, a series of orthogonalexperiments was designed below.

2. Experimental Procedure

A series of experiments which were designed based onthe Orthogonal principle with three factors and three levelsas shown in Table 1, were carried out in order to study theeffects of carbon equivalent, solidification time and maxi-mum undercooling comprehensively. The levels of castingmodulus and inoculation quantities were designed for thedifferent levels of solidification time and maximum under-cooling respectively.

The experimental melts were made in a medium fre-quency induction furnace of 500 kg capacity. Steel scrap,low impurity pig iron and foundry returns were used ascharge materials. The melts were spheroidized with 2 wt%of Mg(7.5wt%)–Re alloys and were inoculated with0.8 wt% of Fe75%Si alloys, using the sandwich method andlate inoculation. The melts was poured at approximately1 400°C. Hypoeutectic, eutectic and hypereutectic alloyswere produced. The final chemical composition of the experimental ductile cast iron is listed in Table 2.

The molten cast iron was poured into 20, 40 and 60 mmthick plate molds. Both the length and height are 300 mm.The shape and dimensions are shown in Fig. 1. All of the

An Evaluation Model for the Nodule Count of Graphite Particlesin Ductile Iron Castings

HongLiang ZHENG,1,2) YuCheng SUN,1) Ning ZHANG1) and XueLei TIAN1)

1) Key Laboratory for Liquid–Solid Structural Evolution and Processing of Materials (Ministry of Education), ShandongUniversity, Jinan 250061 China. E-mail: honglz@sdu.edu.cn 2) Mechanical Engineering Post-doctoral Research Station,Shandong University, Jingshi Road 73, Jinan 250061 P. R. China.

(Received on April 20, 2010; accepted on June 25, 2010 )

An evaluation model on the nodule count of graphite particles in ductile iron castings was developedbased on a series of orthogonal experiments. The model is proposed for evaluating the count of graphiteparticles in ductile iron castings with different carbon equivalent (CE), local solidification time (ts), and maxi-mum undercooling. According to the nodule count spacial distribution theory, the graphite nodule density Nvequals nuclei density after solidification. The measurement of NA values were carried out on experimentalductile iron castings designed according to the orthogonal principles. Nv was calculated from area densitiesNA using Owadano rules. The results indicate that the count of graphite particles has a close relationship toCE, ts and instead of one of these three parameters.

KEY WORDS: nodule count; ductile iron; graphite particle; evaluation model.

1981 © 2010 ISIJ

Table 1. Three experimental factors and factor levels.

plates have a common gating system. And the molds wereprepared using green sand with enough hardness to avoidthe mold wall movement during the solidification.

The cooling curves corresponding to the geometricalcentre of each section of the specimens were recorded usingK type thermocouples covered with Al2O3 and Na2SiO3 anda 14 channel measurement device named Thermo-scan withrunning recording software on the computer. All the datawere transmitted to the computer via a high speed CAN-Bus to improve the measurement accuracy. Approximatelyevery 100 ms (millisecond) the data were updated. The tem-peratures were recorded until solid-state transformationtemperature was overcome.

Metallographic specimens were taken from center ofeach experimental casting. They were prepared using con-ventional polishing techniques. The area nodule count NAwas measured by a QiuTie quantitative analyzer, only thediameter of the nodule particles beyond 2 mm are countedat a magnification of 100�.

3. Experimental Results

Several cooling curves are shown in Fig. 2. From thecooling curves, the minimal temperature Tm at the onset ofeutectic solidification was determined. The maximum un-dercooling in the individual casting was calculated formformula (1), (2)4)

DTm�TE�Tm ................................(1)

TE�1 153.9C�5.25Si�14.88P..................(2)

where TE is the stable equilibrium temperature of thegraphite eutectic and C, Si, P are the weight percent of Car-bon, silicon and phosphorus in cast iron.

The local solidification time at a given sample location(ts) is calculated from the cooling curves using the princi-ples of computer-aided differential thermal analysis (CA-CCA) of spheroidal and compacted graphite cast irons,7,8)

which is shown in Fig. 3.The metallograph of the speciments in plates of 60 mm

thickness with different carbon equivalent are shown in Fig.4.

The nodule count in volume unit NV can be calculatedfrom the nodule count in area unit NA using the equationproposed by Owadano9)

NV�(p /6VG)1/2(aNA)3/2 ........................(3)

Where, VG is the graphite volume fraction and a is a coeffi-cient of materials. Using values a�1.2 and VG�0.15, NV

can be written as NV�2.46NA3/2.10) The calculated of each

melt is shown in Table 3.

4. Discussion

It is very obvious that there is a very close correlationbetween carbon equivalent and nodule count as shown inFig. 5. The nodule count increases with an increase in CEin each section size. The nodule count is markedly sensitive

ISIJ International, Vol. 50 (2010), No. 12

1982© 2010 ISIJ

Fig. 1. Shape and dimensions of the experimental castings.

Fig. 2. Cooling curves of the experimental castings.

Fig. 3. Specify the Local solidification time through the CA-CCA analysis.

Table 2. Chemical composition of ductile iron.

to CE for thinner samples.Local solidification time is used as variable parameter for

analysis of results, instead of thickness. Because the givenmodel based on it can be also applied to other systemswhich may include different mold conditions. Figure 6shows the effect of local solidification time on nodule countof S.G. Iron castings, it can be concluded that the shorter

solidification time of the ductile iron castings shows thelarger nodule count.

Figure 7 shows the experimental results for the measure-ments of the maximum degree of undercooling DTm andnodule counts Nv. The results indicate that for a given melt,as the thickness of the castings increases the maximum de-gree of undercooling decreases and as a result the nodulecount decreases the maximum degree of undercooling de-creases.

5. Regression Model

When considering the carbon equivalent, local solidifica-tion time and undercooling, the following regression modelwas obtained to evaluate the nodule count of graphite parti-cles in ductile iron castings:

NA�168.8121CE�15.36173DTm�74.43494DT2m

�0.1682272ts�9.910833ts2�1 464.135

R2�0.924413 ...........................................(4)

ISIJ International, Vol. 50 (2010), No. 12

1983 © 2010 ISIJ

Table 3. Graphite nodule counts of each melt.

Fig. 4. Graphite nodule structures in plates of thickness 60 mm with different carbon equivalent. a) CE�4.1, b)CE�4.33, c) CE�4.61, d) CE�4.81.

Fig. 5. Effect of carbon equivalent on nodule count of S.G. Ironcastings.

Fig. 6. Effect of local solidification time on nodule count of S.G.Iron castings.

NV�11 273.19CE�2 332.12DTm�12 454.7DT 2m

�3.127672ts�268.2986ts2�41 855.77

R2�0.908179 ...........................................(5)

Where: R is the determination coefficientIn Eq. (2), NA is the nodule count in area unit, NV is the

nodule count in volume unit CE is the carbon equivalent,DTm is the maximum undercooling, ts is the local solidifica-tion time.

Using the regression model, the nodule count can be cal-culated from the carbon equivalent, maximum undercoolingand local solidification time. Comparison of the experimen-tal data and calculated data is shown in Fig. 8 and Fig. 9.The calculated data is very close to the experimental dataand the regression model is acceptable.

6. Conclusions

(1) The effects of metallurgical and processing parame-ters on the nodule count of graphite particles in ductile ironcastings have been studied, considering the parameters ofcarbon equivalent, maximum undercooling and local solidi-fication time within specific ranges of these variables.

(2) The nodule count increases with an increase in CEin each section size. The nodule count is markedly sensitiveto CE thinner samples. The shorter solidification time ofthe ductile iron castings shows the larger nodule count. Andas the thickness of the castings increases the maximum degree of undercooling decreases and as a result the nodulecount decreases the maximum degree of undercooling decreases.

(3) An evaluation model was obtained with consideringthe valuables of carbon equivalent, maximum undercoolingand local solidification time. The calculated data is veryclose to the experimental ones. The model provides a newway to calculate the nodule count of the graphite particlesapproximately.

Acknowledgement

This work was supported by Komatsu Ltd. and Promo-tive research fund for young and middle-aged scientisits ofShandong Province (BS2009ZZ005).

REFERENCES

1) T. Skaland and O. Grong: AFS Trans., 91 (1983), 153.2) J. D. Mullins: Ductile Iron News, Issue 2, Ohio, (2003),3) A. Javaid, J. Thomson and K. G. Davis: AFS Trans., 108 (2002), 889.4) E. Fras, K. Wiencek, M. Górny and H. F. López: Int. J. Cast Met.

Res., 18 (2005), 156.5) J. M. Borrajo and R. A. Martínez: ISIJ Int., 42 (2002), 257.6) H. D. Zhao and B. C. Liu: Int. J. Cast Met. Res., 16 (2003), 281.7) U. Ekpoom and R. W. Heine: AFS Trans., 81 (1973), 27.8) I.-G. Chen, D. M. Stefanescu and K. Lieu: AFS Trans., 89 (1981),

947.9) T. Owadano, Dr. Eng: Imono (J. Jpn. Foundarymen’s Soc.), 45

(1973), 193.10) A. Almansour, K. Matsugi, T. Hatayama and O. Yanagisawa: Mater.

Trans. JIM, 36 (1995), 1487.

ISIJ International, Vol. 50 (2010), No. 12

1984© 2010 ISIJ

Fig. 9. Comparison of the experimental and calculated NV.

Fig. 7. Effect of undercooling degree on nodule count. Fig. 8. Comparison of the experimental and calculated NA.