On microstructure formation and mechanical properties in grey cast iron

O

N MICROSTRUCTURE FORMATION

AND MECHANICAL PROPERTIES

IN GREY CAST IRON

Attila Diószegi

Department of Mechanical Division of Component

Jönköping Engineering Technology - Castings

2004 Linköping University Jönköping University 581 83 Linköping, Sweden 551 11 Jönköping, Sweden

SCHOOL OF ENGINEERING

Jönköping University

Linköping Studies in Science and Technology Dissertation No. 871

Linköping Studies in Science and Technology

Dissertation No. 871

ON MICROSTRUCTURE FORMATION AND

MECHANICAL PROPERTIES

IN GREY CAST IRON

Attila Diószegi

Division of Engineering Materials Division of Component

Department of Mechanical

Technology-Castings

Engineering

School of Engineering

Linköping University Jönköping University

S-581 83 Linköping, Sweden S-551 11 Jönköping, Sweden

Jönköping May 2004

ON MICROSTRUCTURE FORMATION AND

MECHANICAL PROPERTIES IN GREY CAST IRON

© 2004 Attila Diószegi

Division of Component Technology – Castings School of Engineering, Jönköping University

Linköping Studies in Science and Technology, Dissertations, No.871 ISBN 91-7373-939-1

ISSN 0345-7524

On microstructure formation and mechanical

properties in grey cast iron

Attila Diószegi

Jönköping University, School of Engineering / Component Technology, Box 1026, SE-551 11 Jönköping, Sweden

Abstract

A major user of cast components is the automotive industry, where the functionality of the components is related to environmental demands. Internal combustion engines are constantly being improved to emit less pollution. A vital part in this development is to increase the material properties of engine components during their life cycle. In particular, cylinder heads, cylinder blocks and piston rings for diesel engine are produced in grey cast iron. Cast iron is expected to be in use far into the foreseeable future, due to favorable properties and low production costs. This work has been devoted to study microstructure formation, the tensile properties of cast iron and to some extent defect formation.

The microstructure develops during solidification and solid state transformations. An inverse thermal analysis method was developed to study the kinetics of the microstructure formation. The inverse thermal analysis used, the Fourier method, analyses the cooling curves of two thermocouples to study the solidification or transformation. To decrease experimental errors, simulations have been done and the cooling curves were analyzed. The best results were obtained when the thermocouples were placed close to each other.

With the help of the thermal analysis a time dependent and fading nucleation law of the eutectic cells was found to fit the experimental results best. The experiments were made by multiple thermal analyses, and six different types of inoculants were investigated. The eutectic growth behavior during solidification was evaluated with inverse thermal analysis, and it was found that commercial inoculants not only affect the eutectic nucleation but they also control the eutectic growth rate.

Models of densities and volume changes are an integral part of a microstructure simulation of cast irons. These models are important for the inverse thermal analysis and an understanding of the porosity and expansion penetration in cast iron.

The tensile strength of grey cast iron has been discussed by examining the fracture mechanism of the material at failure. The ultimate tensile strength is a result of the intimate collaboration between the graphite flake and the primary phases. Several parameters, including the graphite morphology, carbon content, inoculation and cooling conditions influence the ultimate tensile strength by offsetting the equilibrium between the major constituents, the graphite flakes embedded in the primary metallic matrix. A model to predict the ultimate tensile strength is developed based on the interpretation of the stress intensity behavior in a eutectic cell.

The models developed for nucleation, eutectic growth and prediction of tensile strength were introduced into a casting simulation program. Mould filling, solidification, microstructure development and tensile strength of a complex shaped cylinder head were simulated.

Keywords: Grey iron, nucleation, primary austenite, eutectic cell, growth rate, inoculation, thermal analyze

My charge is to transmit a message:

» On the move signify happiness. Achieve the goal denote passing. « So long!

Let me go my way!

Hobo / Földes László , 1988

To my loved ones

ON MICROSTRUCTURE FORMATION AND MECHANICAL PROPERTIES IN GREY CAST IRON

The thesis includes the following supplements:

Supplement I

A. Diószegi and M. Wessén, Measurement and simulation of thermal condition and mechanical properties in a complicated shaped cylinder head cast in gray iron. (Modelling of Casting, Welding and Advanced Solidification Processes – IX, 20-25 August, 2000, Aachen, Germany, pp. 869-876)

Supplement II

A. Diószegi, A. Millberg and I.L. Svensson, Microstructure evaluation and simulation of mechanical properties of a cylinder head in cast iron (Conference on The Science of Casting and Solidification, 28-31 May, 2001, Brassó-Brasov, Romania, pp. 269-277.)

Supplement III

A. Diószegi, J. Hattel, An inverse thermal analysis method to study the solidification in cast iron. . (Journal of Cast Metals Research, 2004, volume 17, p311-318.)

Supplement IV

I.L. Svensson, A. Dioszegi; On modelling of volume related defect formation in cast irons. Proc. (Modelling of Casting, Welding and Advanced Solidification Processes IX, eds. PR Sahm, PN Hansen and JG Conley, Shaker Verlag GmbH, Germany, Aug. 2000, 102-109).

Supplement V

A. Diószegi, I.L. Svensson, Inverse kinetic analysis method to study eutectic growth . (Journal of Cast Metals Research, 2005, volume 18, p41-45.)

Supplement VI

A. Diószegi, I.L. Svensson, A comparison of Fourier vs. Newtonian thermal analysis and its influence on the inverse kinetic growth calculation (Submitted to International Cast Metal Journal 031202).

Supplement VII

A. Diószegi, Evaluation of eutectic growth in grey cast iron by means of inverse modelling. (Journal of Cast Metals Research, 2003, volume 16, p301-306.)

Supplement VIII

I.L. Svensson, A. Millberg, A. Diószegi; A study of eutectic inoculation in gray iron by addition of Fe-Si-Ca-Al-,Sr, Ba, Zr, Ti, RE, and C. (Journal of Cast Metals Research, 2003, volume 16, p29-34.)

Supplement IX

A. Diószegi, Vasilios Fourlakidis and I.L. Svensson: Microstructure and tensile properties of grey cast iron (Research report 04:1 ISSN 1404-0018, School of Engineering, Jönköping University, Sweden, 2004.)

Supplement X

A. Diószegi: Microstructure and tensile property simulation of grey cast iron components. (Research report 04:2 ISSN 1404-0018, School of Engineering, Jönköping University, Sweden, 2004.)

Division of work between paper authors

The papers included in the supplements were written in collaboration with other researchers according to the following account:

Supplement I

This work was performed by the respondent, the co-author contributed with the temperature measurements conducted at Volvo Powertrain, The Skövde Foundry. Supplement II

This work was a continuation of the work presented in supplement I. The work was performed by the respondent. Adam Millberg contributed with the investigation of microstructure and Ingvar L Svensson was the adviser regarding the simulation of microstructure.

Supplement III

The respondent wrote the program codes for both the direct simulation and inverse analysis. The co-author provided the feedback for the formulation of both direct and inverse simulation and contributed with solutions regarding the iteration procedure. Supplement IV

The original idea on modeling the density variation in cast iron was introduced by Ingvar L. Svensson. The respondent contributed by dilatometer measurements on density variation.

Supplement V

This work was performed by the respondent; the co-author was the adviser regarding the kinetic formulation of eutectic growth

Supplement VI

This work was performed by the respondent; the co-author was the adviser regarding the kinetic formulation of eutectic growth

Supplement VII

The work was carried out by the respondent based on research for development of material properties in cast iron, supported by the Volvo Powertrain Division Foundry and Daros Piston Rings AB.

Supplement VIII

The main author of this paper was Ingvar L Svensson. A.Millberg contributed with the microstructure investigation. The respondent carried out the thermal analysis involved in the calculations.

Supplement IX

The work was performed by the respondent. Vasilios Fourlakidis contributed by performing the microstructure and tensile test investigations. Ingvar L Svensson was the adviser regarding the connection between microstructure and tensile properties. Supplement X

This work was performed by the respondent based on research for development of material properties in cast iron, supported by the Volvo Powertrain Division Foundry and Daros Piston Rings AB. The simulation work was performed using a commercial simulation code. The calculation models for microstructure formation and tensile properties prediction were implemented in the program MAGAiron with help from Foundrysoft AB.

CONTENTS

1. INTRODUCTION

1.1. Cast Iron/Gray cast iron

1.2. Simulation of microstructure development in gray cast iron

1.3. Simulation of tensile properties in gray cast iron

1.4. Aims and contents of the thesis

1.5. Experimental techniques

2. SURVEY

2.1 Inverse thermal analysis

2.2 Nucleation of the eutectic phase

2.3 Growth of the eutectic phase

2.4 Fracture mechanics of gray cast iron

2.5 Modeling the tensile strength

2.6 Simulation of microstructure formation and tensile strength in grey cast iron

3. CONCLUDING REMARKS 4. FUTURE WORK 5. ACKNOWLEDGEMENTS 6. REFERENCES 7. NOMENCLATURE 8. SUPPLEMENTS

1 INTRODUCTION 1.1 Gray cast iron

Metallic materials constitute the framework of our civilization. The estimated weight of extracted and refined metallic materials in the world is three billion tons. From extraction of the ore to a final product, casting is the most widely used method to produce metallic parts. Casting is popular due to its simplicity. The molten metal is poured into a mould cavity, followed by solidification, which leads to a geometry very close to the final product shape. The simple principal of casting conceals a large number of interdisciplinary scientific phenomena. Scientists have for a long time been occupied in studying a range of phenomena related to casting. During the

manufacturing stages of a cast component, metals undergo various phase and property transitions before reaching a final state and are taken in use.

A large user of casting components is the automotive industry, where the functionality of components used is connected to environmental safety. Internal combustion

engines are continuously being improved to emit less pollution. A vital part of the development is to increase the material properties of engine components during their life cycle. In particular, cylinder heads, cylinder blocks and piston rings for diesel engine are produced in grey cast iron. An important trend of development is the increase of tensile properties of the components, allowing higher operating pressures. This is followed by more efficient combustion and less resulting pollutants. The engine components are exposed to cyclic loads at elevated temperatures. An accepted procedure is to derive relationships between the dynamic tensile properties the static tensile properties, however it is not really understood how tensile strength is linked to the microstructure.

Cast iron is an alloy where the major alloying element, i.e. carbon, precipitates as graphite during solidification. The metallic matrix is tough and gives the cast iron its strength, while the graphite particles are brittle and lower the strength properties. The shape of graphite has a decisive influence on material properties. Graphite can be lamellar, compact or nodular. The tensile properties are lowest in cast irons containing lamellar graphite, and highest in nodular graphite cast iron. Examples of the different graphite shapes observed are shown in Figure 1.

Lamellar Compact Nodular Figure 1. Graphite shapes in cast iron

The presence of graphite in cast iron also induces beneficial properties. The lamellar graphite contributes to indispensable good heat conduction and vibration damping

-properties. Transformation from a lamellar to a nodular shape decreases both heat conductivity and vibration-damping properties.

The fracture surface of cast iron containing lamellar graphite shows a grey appearance; hence it is called grey cast iron.

Gray cast iron used in foundry applications contains in addition to carbon also silicon, manganese, phosphorus and sulphur. In the binary Fe-C system there is a eutectic point at , and . At carbon contents lower than the eutectic composition solidification starts by precipitation of primary austenite. The austenite forms dendrites, which dissolve a limited amount of carbon, while the surplus segregates to the liquid phase. When the liquid reachs the eutectic composition, graphite precipitates and growths in collaboration with the austenite, creating eutectic cells. Figure 2 shows the microstructure of grey cast iron. The primary dendrite

direction is indicated by dashed lines and the eutectic cell contour by continuous lines.

% 3 , 4 w

C = T =11530C

Figure 2. Microstructure of grey cast iron1. (20x)

At lower temperatures (about 700 oC) the austenite transform to ferrite, pearlite or a mixture of those depending on composition and cooling conditions. Figure 3 shows the pearlite lamellas in the metallic matrix. Thanks to the color etching technique the pearlite can be distinguished according to its origin. The light areas were solidified as primary austenite while the dark areas solidified as part of the graphite-austenite eutectic.

Figure 3 Pearlite lamellas

1.2 Simulation of microstructure development in gray cast iron

The extreme complexity of the gray iron microstructure has challenged the efforts of scientists to predict the formation of microstructure. The mechanism of transformation from the liquid to solid state involves phenomena at the atomic level. The step from the atomic level to the macroscopic size level of cast iron components in industrial applications is large. One method used to overcome the size induced differences is the allocation of approximate properties to the material. In the field of gray cast iron the microstructure simulation is based mostly on deterministic models. The basic procedure uses macroscopic heat transport models, including the deterministic formulation of the transformation kinetics for the solidifying phases.

Hypo-eutectic cast iron starts solidification by nucleation and growth of the primary austenite. Recently Rivera at al.2 have shown that independent of the carbon content even alloys with eutectic and hyper-eutectic compositions develop primary austenite. Nucleation of primary austenite grains is considered instantaneous by Tian and Stefanescu3 who proposed a formulation dependent on the cooling rate

• T : (1) 2 087 . 0 33 . 5 12 . 48 • • + + = T T N_γ

Another expression for the number of austenite grains at continuous nucleation is given as a function of supercooling

γ N T ∆ : (2) γ γ γ n T k N = ∆ 3

-The nucleation constants suggested by Tian and Stefanescu are: kγ =2.45, n ,

for calculated in while Fras at al.

93 . 0 = γ γ N mm−2 4 suggest k_γ =500, n_γ =2, for calculated in γ N mm−3

Growth simulation models for primary austenite given in the literature are based on direct connection of the solidifying fraction primary austenite to a segregation and diffusion model. Depending on the assumptions made regarding the segregation and diffusion of elements, the Scheil5, the Brody and Flemings6 and the Clyne and Kurz7 models are most widely used. These models do not employ any kinetic assumption, however primary austenite is observed to develop both columnar and equiaxed grains. The growth of primary austenite dendrites decay when graphite nucleates. The

austenite continues to grow in collaboration with the graphite forming the eutectic cells. Nucleation of the eutectic phase used to be modeled as a continuous event. The classical Oldfield8 model is the most frequently used in gray cast iron simulation.

(3) e n e e k T N = ∆

The nucleation constants suggested by Oldfield are: , , while Fras at al.

3 10 12 . 7 × − = e k (_mm−3_K−ne) 2 = e n 4 suggest k_e =3.1 (_m−3_K−ne), =_6.4 e n .

The graphite austenite eutectic is considered spherical and the growth rate of the sphere is calculated by the formula:

n g eut T k dt dR ∆ = (4)

The growth constants and are experimentally determined values and vary between vide limits in the literature. Thorgrimsson

g

k n

9

suggested and for flake graphite;

and for undercooled graphite. At large supercoolings, instead of graphite- austenite eutectic the carbid- austenite eutectic develops. The model to simulate carbide eutectic growth is similar to that presented in equation 4. Reported growth coefficients by Hillert at al.

8 10 48× − = g k (ms−1Kn) n=0.66 8 10 2 . 2 × − = g k (ms−1Kn) n=2 10 are and . 3 10 30× − = c k (mms−1Kn) 2 = c n

The austenite obtained during solidification is stable over the eutectoid temperature. At temperatures below the eutectoid point the austenite decompose to pearlite, ferrite or a mixture of these. Models for pearlite nucleation and growth were proposed by Mehl and Dube11, and models for ferrite growth by Stefanescu and Kanetkar12. The supercooling ∆T plays a central roll in the microstructure simulation. Empirical relations are suggested to calculate the equilibrium temperature as a function of the chemical composition.

1.3 Simulation of mechanical properties in gray cast iron

The ultimate scope of microstructure simulation is to establish a relationship with the mechanical properties. Two major properties treated in the literature are the Brinell hardness HB and the ultimate tensile strengthσUTS. The reported relations between

hardness or tensile strength and the microstructure have been obtained after sophisticated experimental investigations and statistical calculations. Ruff and Wallace13 reported a relation for gray cast iron covering the carbon equivalent rangeCE =3.85÷4.45as follows: − × + × + × + × − × + = S pearlite UTS 11248 179.29 DAS 4949.08 D 2.74 Lγ 90.74 Aγ 77.51 f σ f gr E grD N dir L CE f + × + × − × − × × −30.18 0.79 287.74 _γ 9.79 5.45 (5)

The parameters included in the relation are the ultimate tensile strengthσ_UTS, the dendrite arm spacing , the section size of cylindrical samples , the average dendrite length , the dendrite interaction area , the pearlite content , the fraction D type graphite , the eutectic cell count , dendrite directionality , the average flake length and the carbon equivalent factor . The weakness of this type of relation from the simulation point of view is the presence of parameters characterizing the microstructure which are too complex to simulate.

DAS DS γ L Aγ fpearlite grD f N_E dir_γ gr L CEf

Bates14 reported relations similar in form to equation 5 where the main parameters were the alloying elements C, Si, Mn, S, Cr, Ni, Cu, Mo and the diameter of the cylindrical sample used in the experiment. These parameters were related to the ultimate tensile strength covering the carbon content rangeC =3.1÷3.5. Furthermore Bates expressed the ultimate tensile strength as a function of half of the maximum graphite flake length c observed on the sample surfaces examined.

c k_t

UTS =

σ (6)

The material coefficient when the parameter is measured in mils, and the ultimate tensile strength

2 . 4 = t k c UTS σ in ksi.

Goettisch and Dantzig15 implemented equation 6 in their simulations for prediction of the ultimate tensile strength. Their microstructure simulation model calculated the eutectic cell size wherefrom the maximum graphite flake length was approximated as 75 to 95 pct of the predicted maximum cell diameter. The simulated results were found to be consistent with published empirical data.

A more simulation friendly model was proposed by Junming16, where the Brinell hardness HB is calculated. 74 . 14 68 . 90 5 . 2 03 . 123 72 . 421 × 0.228 + × + × 0.5 + × − = − − aus P carb e f f D HB λ (7) 5

-The parameters included in this equation are the eutectic cell diameter , the fraction carbide , the pearlite lamellar spacing

e

D

carb

f λ_P and the fraction primary austenite .

The Brinell hardness was linked to the ultimate tensile strength by the following relation: aus f 7 . 322 273 . 2 × − = HB UTS σ (8)

The simulation model when compared to experimental data was reported to fit fairly well, but nevertheless the experimental work included only one chemical composition.

1.4 Aims and contents of the thesis

Casting simulation has become an invaluable tool for the advancement of casting technology in foundries. Appropriate application of mould filling simulation shortens the preparation time and thereby reduces expenses. After simulating mould filling, the next stage is to simulate the solidification process. Complex shaped casting

components such the cylinder head for diesel engines cast in gray iron were simulated using commercially available software.

The simulation results were compared to thermal measurements on a real casting. These results are presented in supplement I. It is shown that by careful selection of boundary conditions with respect to the geometrical complexity; the simulated thermal field was in good agreement with the measured thermal behaviour.

Furthermore, the experimentally cast cylinder head has been investigated with respect to the microstructure and Brinell hardness. These results are discussed in supplement II. It was found that the simulated microstructure was not in good agreement with the experimentally observed results. The number of eutectic cells and the measured time to the nucleation event were not consistent with the known nucleation laws. Cylinder heads alternate between extremely thin and extremely thick sections, which is

believed to be an important factor influencing the accuracy of the simulation. The available models for nucleation and growth were probably developed within small process parameters, between short cooling ranges, small variation of chemical contents and unilateral nucleation conditions.

At that time it was clearly necessary to improve insight into the nucleation and growth mechanism of the solidification process in gray iron. A fundamental characteristic of solidification is the latent heat release when the liquid transforms to a solid phase. All microstructure formation mechanisms are expected to be proportional to the latent heat release wherefore the examination of latent heat release was proposed. The challenge was to find a suitable method to measure the release of latent heat in conditions which reflect the industrial environment. An inverse thermal analysis method, called the Fourier Thermal Analysis (FTA), was chosen. The FTA method involves an inverse numerical solution of a 1-dimensonal heat transfer problem. At least two thermocouples are necessary to perform the calculations, which also needs a tabulation of the volumetric heat capacity of the phases taking part in the

solidification. The outcome of the analysis is the latent heat of solidification released. The FTA method, including the validation of its mathematical consistency is

The volumetric heat capacity of materials is the product of heat capacity and density. These quantities were found to be critical for the quality of the thermal analysis. The heat capacity was calculated using models developed for thermodynamic calculations, while for the density calculation a sublattice molar volume model is proposed. The density model is described in supplement IV.

Once the release of latent heat of solidification is known the growth mechanism of the microstructure may be modelled by an inverse kinetic analysis method. The eutectic phase has been chosen on which to perform the calculations. The eutectic phase is considered spherical and the outcome of the inverse kinetic analysis is a growth model where the growth rate of a eutectic cell is related to the supercooling. As a first step the validity of the method was established by processing simulated data. The inverse kinetic analysis method is reported in supplement V.

The validation procedure of the inverse kinetic analysis revealed the dependence of the results obtained on the position of thermocouples. The best result was obtained when the thermocouples were positioned close together. Previous experience made clear the practical difficulties to handle multiple thermocouples in an experimental environment. In metallurgical process control, thermal analysis systems using one thermocouple are widespread. An inverse thermal analysis method called the Newtonian Thermal Analysis based on a single thermocouple is known from the literature. A comparative study is reported in supplement VI on how well a thermal analysis with one ore two thermocouples reflects the phase transformation.

An evaluation of eutectic growth in gray cast iron by means of inverse modeling based on FTA is presented in supplement VII. The evaluation was done on data from casting gray iron inoculated with different commercial inoculants. A further outcome of the inoculation experiments was the investigation of eutectic nucleation. The nucleation model found to provide the best fit for all inoculants cases is reported in supplement VIII.

Establishing a valid connection between the microstructure and mechanical properties was the objective of a series of experiments performed with varied parameters.

Parameters including the graphite morphology, carbon content, inoculation and cooling condition were included in the experiments. The fracture mechanism of gray cast iron has been investigated to obtain a reasonable interpretation on how cast iron loaded in tension fails. The results of this investigation, including a model to interpret the stress intensity behaviour in a single eutectic cell, and a calculation model to predict ultimate tensile strength are reported in supplement IX.

The investigation and modeling work conducted until this point led to a new nucleation and growth model of the eutectic phase and a new model to connect the microstructure and tensile properties in order to simulate the ultimate tensile strength. The models obtained have been implemented into a commercial, finite difference method based, simulation software. A cylinder head cast in grey iron has been simulated. The results were compared to thermal measurements, microstructure investigation and tensile tests conducted on a cylinder head cast under conditions similar to those simulated. The comparison is presented in supplement X.

-1.5 Experimental techniques

The inverse thermal analysis, the subsequent investigation of eutectic nucleation and growth as well the basis for the tensile model were conducted using a multiple thermal analysis equipment. The experimental arrangement was composed of three cylindrical bodies (Figure 3), insulated at the end surfaces to achieve a 1-D cylindrical heat flow.

Figure 3. Cylindrical samples and gating system

The 1-D heat flow was needed to establish the mathematical models related to the solidification. The multiplicity of the equipment was provided by the different cooling condition of the cylinders. Different materials (sand, chill, insulation) surrounded each cylinder, to produce various cooling rates.

Two thermocouples were inserted in each cylinder (Figure 4), serving to record cooling rates during solidification. One thermocouple was placed in the central axis of each cylindrical sample while the second was displaced laterally in each cylinder close to the mould wall. The size of the cylinders was ø50x70 mm in the case when sand and chill constitute the cooling media and ø80x70 mm in the case of insulated media.

Cylinder heads (Figure 5) were cast in an industrial production environment and investigated with respect to microstructure and mechanical properties. The moulding material was green sand; the cores were prepared in quartz sand, bonded by an organic binder using sulphur dioxide-gas as catalyst.

Figure 5. A sectioned cylinder head

Thermocouples of S-type (Pt/Pt+10%Rh) were isolated from each other in a 2-hole alumina tube, and a thin alumina coating covered the soldering points. The

thermocouples were positioned at several locations of the metal, core and mould (Figure 6). The sectioned casting parts were x-ray investigated to check the final position of thermocouples, which had been displaced due to thermally induced deformation.

-Figure 6. Thermocouples assembled in the green sand mould

2 SURVEY

2.1 Inverse thermal and kinetic analysis (supplements III, IV, V, VI)

Solidification, from a thermal point of view, is a non-linear heat conduction problem expanded with a heat source corresponding to the latent heat of solidification released. The general case of heat conduction including a heat source term is written:

sol p k T q t T C =∇ ∇ + & ∂ ∂ ) ( ρ (9)

The heat source is an expression describing how the liquid transforms to solid phase. Considering the eutectic phase of grey cast iron to be spherical, a kinetic expression of the growing phase is given by the KJMA17,18,19 equation:

e

f_s = 1− 3 4πR_e3N_V −

(10)

Where Nv is the number of eutectic cells and Reis the radius of the eutectic cell. The inverse analyses of thermal and kinetic behavior of solidification are expected to identify the fraction of metal solidified and the growth characteristic of the eutectic phase from the known thermal behaviour and a known number of eutectic cells. The procedure to find a reliable inverse analysis method is started by a direct numerical simulation of the solidification. A control volume method is used in a 1-D polar formulation to simulate the solidification of a eutectic Fe-C alloy cast in a green sand

mould with known thermal and kinetic properties. The simulated geometry is shown in Figure 7. z ∆ ri−1 ri ri−1 1 − ∆r_i i r ∆ 1 + ∆r_i

Figure7. Cylindrical mesh used for direct simulation of solidification20. The cooling curves obtained for the case of an Ø50 mm cylinder are presented in Figure 8. 0 400 800 1200 Time [sec] 1000 1100 1200 1300 Te m p e rat u re [ oC] dcasting = 50 mm T1 ; r1 = 0; T10 ; r10 = 10 mm T15 ; r15 = 15 mm T20 ; r20 = 20 mm T25 ; r25 = 25 mm

Figure 8. Simulated cooling curves

The inverse thermal analysis of the known cooling condition according to the

simulated case is performed by using two different methods. The Newtonian method is based on an interpretation of a single thermal point with respect to solidification, while the Fourier method involves an inverse numerical solution of 1-dimensional heat transfer between two thermal points.

-0 0.2 0.4 0.6 0.8 1 Fraction solid 0x100 1x10-3 2x10-3 3x10-3 4x10-3 5x10-3 S o lid if icat io n r at e [ s -1] df_{s,inv,F(1-15)}/ dt dfs,inv,N(1)/ dt df_s,sim,7/ dt dfs,sim,20/ dt 0 0 20 40 60 80 Supercooling [o_C] 0x100 2x10-6 4x10-6 6x10-6 8x10-6 G row th r at e [ m *s e c -1] Direct simulation Inv(1-7) ; dcasting= 24 mm Inv(1-15) ; d_casting= 50 mm Inv(1-35) ; dcasting= 80 mm 0

Figure 9. Simulated vs. calculated solidification rate

Figure 10. Growth rate vs. supercooling

A comparison of the methods (Figure 9.) shows clear differences. When the Newtonian method is used, the solidification rate calculated based on the

thermocouple situated at the thermal centreline of a cylinder corresponds to the solidification rate simulated 20 mm displaced from the centreline . The Fourier method reproduces more accurately the solidification rate. The calculated solidification rate based on two thermal points displaced 15 mm corresponds to the simulated solidification rate at the half distance between the thermal points used. The accuracy of the Fourier method is dependent on the displacement of the thermal points. Closely displaced thermal points give the best results depending on the finite difference approximation used in the inverse solution.

dt df_s,inv,N(1)/ dt dfs,sim,20/ dt dfs,inv,F(1−15)/ dt df_s,sim,7 /

The inverse kinetic analysis uses the fraction solid calculated by the inverse thermal analysis method, the number of eutectic cells from the numerical simulation, and calculates the growth rate of the eutectic cell as a function of supercooling. How the growth kinetic is reflected by the Fourier method is given in Figure 10. The growth rate was subjected to inverse analysis on different sized cylinders. The growth rate obtained was in good agreement with the simulated rate, and demonstrates the mathematical consistency of the Fourier method in combination with the inverse kinetic analysis.

2.2 Nucleation of the eutectic phase (supplement VIII)

Nucleation of the eutectic phase in grey cast iron is assumed to take place

heterogeneously. A substrate of a solid crystalline particle with low crystallographic mismatch to graphite is believed to be required for eutectic nucleation. There exists a large variety of solid particles expected to behave as substrates for nucleation.

Nucleation is believed to be temperature dependent, where the various different nucleation sites (depending on their crystallographic properties) became active at

different temperatures. However the traditional Oldfield model given in equation 3 was not applicable for the cylinder heads investigated in the introductory work. In foundry practice there is widespread use of inoculants, granular crystalline powder, added to the molten alloy before casting to promote nucleation. Six commercial inoculants based on Fe-Si-Ca-Al with addition of Sr, Ba, Zr, Ti, RE and C were investigated. The eutectic structure obtained was analysed according to the Oldfield model and none of the investigated inoculants showed a typical supercooling behaviour for the whole solidification interval. Another nucleation model was

proposed by Lifshitz, Slyzov21 and Wagner22 (LSW). They used an idea from Ostwald, who was first to understood that the number of potential nucleants decreased with time. The mechanism, also called ripening, is based on the standard theory of nucleation which assumes that all the particles nucleate at the same time and have exactly the same size. As time proceeds without passing the critical nucleation size, neighboring substrates are in an unstable equilibrium and small fluctuations cause the smaller nucleants to dissolve and the larger ones to grow. This process leads to a situation where the number of nucleation sites decreases while the average size increase without passing the critical nucleation size. The time dependent relation is given as:

t K

Nv = / ₍₁₁₎

Investigating the results from the inoculation experiments a good relation according to the Ostwald/LSW ripening model was found. Figure 11 shows the typical relation between the number of eutectic cells nucleated and time. The phenomenon is also called fading. The small deviations at early nucleation times can be explained by the limited dissolution rate of the inoculants. This investigation showed that if a small cell size is desirable for a medium sized casting, the inoculants can be ranked in order of increasing potency, Sr, RE, Ba, Ti, Zr and Si-C-Ba.

1 10 100 1000 time ( s ) 1E+007 1E+008 1E+009 1E+010 1E+011 1E+012 1E+013 Nv ( N /m 3 ) 1 1 1 2 2 2 3 3 3

Figure 11 Number of eutectic cells for the Sr based inoculant. The curves show from bottom, 0.09 %, 0.39 and 0.90 % addition.

-2.3 Growth of the eutectic phase (supplement VII)

The eutectic cell of grey iron is considered as spherical particles growing from nucleation sites until they impinge. The growth rate

eut

dt dR

of the eutectic cell is related to the supercooling T∆ by a power equation,

n g eut T k dt dR ∆ = (12)

The inverse thermal and kinetic analysis based on the Fourier method has been used to investigate the eutectic growth of cast iron when different inoculants were added. The results are presented in Figure 12. The growth coefficients and exponents

obtained are consistent with those reported by Thorgrimsson, but a variation in growth kinetics can be observed when different inoculants are added. An investigation of eutectic cell shapes (Figure 13) makes evident that there is a deviation from the spherical shape assumed. The various growth coefficients, which depend on the inoculants actually hide a more complex shape of the eutectic cell. The relatively small addition of elements via inoculation contributes to a transition in the

transformation kinetics, resulting in different graphite shapes and thereby different shapes of the eutectic cells. The mechanism of transition between different lamellar graphite shapes can be described as a competition between faceted and non-faceted growth. 0 5 10 15 20 25 Super Cooling [o_C] 0 4 8 12 16 G rowth rate [µ m s -1] Fourier Fs = 0,3 - 07; n = 0,696 kg=107*10-8 Sr 0,06% kg=110*10-8 Sr 0,39% kg=100*10-8 Sr 0,90% kg=97*10-8 RE 0,39% kg=104*10-8 Ba 0,39% kg=95*10-8 Zr 0,39% kg=104*10-8 Ti 0,39% kg=78*10-8 C+Ba 0,06% kg=77*10-8 C+Ba 0,19% kg=70*10-8 C+Ba 0,45%

2.4 Fracture mechanics of gray cast iron (supplement IX)

The fracture mechanism of gray cast iron was investigated on tension loaded samples produced under different conditions. The parameters studied included the graphite morphology, the carbon content, the inoculation and the cooling condition. The observations made reveal the role of the microstructure on crack propagation. The cracks were found to always propagate parallel with the graphite flakes. Hardened primary arms of austenite dendrites adjacent to the crack surface have been observed. The hardening process is assumed to occur due to the development of a plastic zone prior to failure. The interaction between the primary austenite and graphite has been interpreted by a simplified model of the austenite reinforced eutectic cell presented in Figure 14.

The geometrical transcription gave a standard crack component configuration with known mathematical solution. The microstructure observed in the experiments has been analysed by means of a novel interpretation. The fictitious stress intensity at yield (Figure 15.) and the fictitious maximum stress intensity at failure (Figure 16.) are strongly related to the relative shape of the eutectic cell and primary austenite. A different slope is observed for the material cooled at high rate when the precipitation of primary carbide reduces the stress intensity. The observed relations indicate that the tensile strength of the grey cast iron is the result of the collaboration between the toughness of the primary austenite and the brittleness of the graphite phase. The shape and distribution of the primary austenite and graphite can be influenced by chemical composition, by inoculation or by the cooling condition, but they will maintain equilibrium with respect to the stress intensity.

-Graphite

Primary austenite

Figure 14. Transcription of austenite reinforced eutectic cell

1 1.2 1.4 1.6 1.8 2 D/d 0 5 10 15 20 25 30 35 KfI Fictitiou s st res s inten sity facto r [M P a *m 1/ 2] Cooling condition Chill 05 Chill 06 Chill 13 Sand 05 Sand 06 Sand 13 Insulation 05 Insulation 06 Insulation 13 1 1.2 1.4 1.6 1.8 2 D/d 0 5 10 15 20 25 30 35 KfC Fi c titious maxim u m st ress inten s ity facto r [M P a *m 1/ 2] Cooling condition Chill 05 Chill 06 Chill 13 Sand 05 Sand 06 Sand 13 Insulation 05 Insulation 06 Insulation 13

Figure 15 Stress intensity at yield Figure 16. Stress intensity at failure

d

D

P

P

2.5 Modeling the tensile strength (supplement IX)

Based on the model used to interpret the fracture mechanics of grey cast iron a new model to predict the ultimate tensile strength is proposed.

+ ∗ + ∗ + ∗ + ∗ + ∗ + = _{aus white} UTS b0 b1 C b2 Si b3 Cu b4 f b5 f σ Y b K b dt dT b D b fC C ∗ + ∗ + ∗ + ∗ + 8 9 750 7 6 0 (13) The authenticity of the model is supported by the existence of direct physical relations between the parameters involved. The main microstructure components, i.e. the primary austenite, the primary carbide and the eutectic cell are all parameters predictable in standard simulation software.

2.6 Simulation of microstructure formation and tensile strength in gray cast iron (supplement X)

The introductory investigation and casting simulation of a cylinder head revealed the necessity to improve the comprehension of eutectic nucleation and eutectic growth as well as to establish a link between the microstructure and tensile properties. The models developed in the field of eutectic nucleation, eutectic growth and tensile strength presented in the present thesis were implemented in MAGMAiron,

simulation software including kinetic models for calculation of solidification and solid state transformation. A series of simulations was performed using the modified model. Firstly the experimental arrangement used for inverse thermal analysis, including the cylindrical samples, was simulated. In spite of the wide range of cooling conditions the simulated microstructure and tensile properties are closely similar to the measured properties obtained from real casting experiments. (Table 1)

aus f D

[ ]

[

]

[

]

Meas. Sim. Meas. Sim. Meas. Sim. Meas. Sim Chill 0,396 0,31-0,32 157 110-400 0,9 0,75 368 378-385 Sand 0,313 0,26-0,28 775 750-780 0,26 0,28 254 257-262 Insulation 0,436 0,25-0,26 1765 1150-1200 0,11 0,15 211 216-218 Meas. = Measured on tensile bars; Sim. = Simulated

Table 1. Comparison of measured and simulated parameters in cylindrical samples

-The largest discrepancies are obtained between the predicted and measured fractions of primary austenite, which is consistent with the difficulty to measure the fraction of primary austenite, and the poorly known kinetic parameters included in the simulation. It is remarkable that the predicted ultimate tensile strength is in good agreement with the measured values.

A cylinder head has been also simulated using the improved simulation models. The simulation includes mould filling, solidification, microstructure and tensile property prediction. Comparing the simulation results to measurements on a cylinder head cast under identical conditions to those simulated, the predicted tensile strength is found to show good correlation with the measured tensile strength. Even in the case of the complex shaped cylinder head, the simulated versus measured fractions of primary austenite shows the largest differences.

aus f D

[ ]

[

]

[

]

Meas. Sim. Meas. Sim. Meas. Sim. Meas. Sim

TP2 0,38 0,28-0,32 1456 1200 - 0,095 281 260-270 Bu2 0,25 0,265-0,28 1598 1500 - 0,07 251 250 Bp2 0,38 0,265-0,295 1237 1500 - 0,075 254 250

Meas. = Measured on tensile bars; Sim. = Simulated

Table 2. Comparison of measured and simulated parameters in cylinder head

3 CONCLUDING REMARKS

Inverse analysis is a valuable tool to study the thermal and kinetic behaviour of solidification. The use of two thermocouples for thermal and kinetic analyses is the minimum demanded. Fewer thermocouples lead to unrealistic results, more then two thermocouples is superfluous, and the distance between the thermocouples becomes decisive.

A time dependent nucleation law best depicts the eutectic nucleation event in gray cast iron when different commercial inoculants are used. The inoculants do not exclusively control the number of nucleated eutectic cells, as the primary austenite network also has an influence. At constant inoculation the increase in fraction austenite increase the number of eutectic cells nucleated.

The eutectic growth is not perfectly spherical. Deviation from the spherical shape is attributed to the existing primary austenite network and the surface active elements added by inoculation.

The primary austenite is the ultimate strength carrier in grey cast iron. The tensile strength is dependent on the stress intensity induced by the graphite flakes. The interaction between the primary austenite and graphite can be optimized by

inoculation, lowering the carbon content, or increasing the cooling condition. However, the phases remain in stress intensity related equilibrium. Introduction of new tensile phases such the primary carbide acts to reduce the stress intensity. The models introduced to describe the interaction between the microstructure and tensile properties are comprehensive and are suitable to predict the tensile properties of extremely complex shaped gray iron components within the investigation limits.

4 FUTURE WORK

Primary austenite plays a prominent roll in determining the properties of gray cast iron. At the same time it is the least investigated phase probably due to the fact that it is stable only at elevated temperatures.

Comparison of simulated and measured primary austenite reveals the largest discrepancy when cast iron is simulated. The nucleation and growth of primary austenite require more insight. Observations reveal that primary austenite even determines the nucleation and growth kinetic of the eutectic phase. The Fourier method is a suitable tool to obtain kinetic parameters for simulation of primary

austenite. An improvement in modeling the primary austenite formation is expected to improve the whole simulation procedure of grey cast iron.

Since primary austenite has been found to be the most important phase for tensile strength, it is believed that strengthening the primary austenite will increase the tensile properties of gray cast iron. The model for prediction of tensile strength presented in the present thesis was based on experiments with unalloyed – very low alloyed grey cast iron. An investigation of alloying elements such Cr, Cu, Mo, Ni, etc will enable the extension of the model with respect to these strengthening alloying elements. This thesis treats gray iron properties under static loads and at ambient temperature. A natural extension of this work is to investigate the connection between static and dynamic properties, and adding a further parameter, that of elevated temperature.

5 ACKNOWLEDGEMENTS

The work performed and reported here is the outcome of a long process which traces back to my past. Lovely people paved my way and facilitated this performance. I would like to start by commemorating my Parents and Grandparents who encouraged me to become a scientist.

At the age of 10, I was given the opportunity to make real castings. Lath Gyula Bartha senior, gave me my first lessons in moulding and casting.

-Lath Lajos Moldován, teacher in mathematics gave the perspective of faithful work. Lath György Schmidt Dipl.ing. was the metallurgist reinforcing my enchantment of cast materials.

I would like to continue by expressing my gratitude to:

Mr. Gustav Lorenzoni who introduced me to the Swedish foundry industry. Mr. Olle Sjögren for being my mentor besides the research work.

Dr. Pál Jónás , the former and present members of the Foundry Department at University of Miskolc, Hungary, who provided me magnanimously, with valorous knowledge on cast iron.

Dr. Gyula Szalai , for his friendly encouragement during the years.

Mr. Sven-Erik Dahlberg, Mr. Leif Hultman, Mr. Sven-Eric Larsson and Mr. Rolf Shalberg, for promoting the establishment of the research project serving as base for my thesis.

Mr. Kent Eriksson, Dr Tony Liu, Mr. Hans-Gunnar Qvist,

Mr. Per Samuelsson and Mr. Torsten Sjögren for their invaluable contributions to the project team, working with cast iron development

Mr. Berndt Gyllensten and Mr. Per Oskarsson for their professional skill and endless patience in conducting the casting experiments.

Mr. Ola Agné, Mr. Leif Andersson, Mr. Lars Johansson for their inventiveness during the experimental preparations.

Mr. Allan Cleassen and Scandinavian Foundry School for their help in experimental preparation.

The staff at the Library of Jönköping University for their rapid delivery of technical literature.

All my colleges at the Volvo Foundry in Skövde for their support during the years. My colleagues and the new generation PhD students at the Department of

Mechanical Engineering and Component Technology at Jönköping University. for all their kindness and support.

Volvo Powertrain AB; Skövde Foundry, Daros Piston Rings AB and the Swedish Knowledge and Competence Foundation (KK-Stiftelsen) for financial support. Mr. Hans Standar and Elmia AB, for giving me The Elmia Scholarship 2003. Foundrysoft AB for the collaboration regarding the computer implementation of the improved simulation models.

Ms. Kristina Hellström , Mr. Adam Millberg and Mr. Vasilios Fouralikidis for their assiduous and altruistic contributions in the preparation, conduction and evaluation of experiments.

Assistant Professor Magnus Wessén for his professional attitude and his friendly support given to me during the years.

Associate Professor Jesper Hattel for his excellence in teaching numerical modeling of casting processes.

Professor Torsten Eriksson for giving me the opportunity to become a PhD-student at the Institute of Technology at Linköping University.

Professor Ingvar L Svensson for setting up The Svensson School of Microstructure Modelling and giving me the education in that spirit.

My brother Zoltán Diószegi for creating a precedent on how dreams became true. My mother-in-law Rozália Gergely for her invaluable help during the years.

My lovely daughter Tekla for her love and my wonderful wife Éva for her intellect.

-6 REFERENCES

1. P. Jónás, A. Diószegi, Microstructure investigation of a cylinder head cast in grey iron, Volvo Foundry report, 1999.

2. G.L.Rivera, R.E.Boeri, J.A.Sikora, Solidification of grey cast iron, Scripta Materialia 50 (2004) 331-335.

3. H. Tian, D.M Stefanescu, Modelling of Casting, Welding and Advanced Solidification Process VI, Piwonka, Voller, Katgerman (editors), TMS, Warrendale, p.639., 1993.

4. E. Fras, W. Kapturkiewicz, H.F. Lopez, Macro and Micro Modeling of the Solidification Kinetics of Casting, AFS Transactions 92-48, 583-591. 5. E. Scheil, Metallforschung, 2 (1947) 69.

6. H.D. Brody, M.C. Flemings, Transactions of the Metallurgical Society of AIME, 236 (1996) 615.

7. T.W. Clyne, W. Kurz, Metall. Trans. A. 12 (1981) 873 8. W. Oldfield, Trans ASM, vol. 59 (1966), pp 945-961.

9. J.H. Thorgrimsson: Licentiate thesis, KTH, Stockholm, (1986).

10. M.Hillert, V.V.Subba Rao:in The Solidification of Metals, The Iron and Steel Institute, London, 1967, pp.205-12

11. R.Mehl, A.Dube: Phase Transformation in Solids, Smoluchowski, Mayer, Weyl (editors), John Wiley and Sons,Inc., New York, NY, 1951.

12. D.M.Stefanescu, C.Kanetkar: Computer Simulation of Microstructure evaluation, D.J. Srolovitz, ed., Warrendale, PA, 1986, pp.171-88.

13. G.F.Ruff, J.F. Wallace: Effects of Solidification Structures on the Tensile Properties of Grey Iron, AFS Transactions, 77-56B, p.179-202.

14. C.E. Bates: Alloy Elements Effects on Grey Iron Properties: Part II, AFS Transactions, 86-154, p.889-912.

15. D.D.Goettsch, J.A.Dantzig: Modelling Microstructure Development in Grey Cast Iron., Metallurgical and Materials Transactions A, 25A, May 1994, 1063-1079.

16. L.Junming: Prediction of Microstructure and Mechanical Properties of Alloy V-0430 by Means of Computer Simulation, Diploma work, Chalmers

University, 1995

17. A.E.Kolmogorov: Akad. Nauk: SSSR. ISV. Ser:Mat (1937), vol.1, p.355 18. W.A.Johnson and R.F.Mehl: Trans.AIME, (1939), vol.135, p.416-458. 19. M.Avrami: J.Chem.Phys., (1940), vol.8, p.212-224.

20. J.Hattel (ed): Fundamentals of Numerical Modelling of Casting Processes (to be published)

21. I.M.Lifshitz, V.V. Slyzov: J.Phys. Chem. Sol., 1961, vol.19, pp35-50. 22. Wagner,: Z. Elektrochemie Vol. 65, 1961, p35.

7 NOMENCLATURE

Upper case symbols

γ

A Dendrite interaction area, %

CE Carbon equivalent, %

f

CE Carbon equivalent factor

P

C Heat capacity, Jkg−1K−1 D , D_e Eutectic cell diameter, m

S

D Section size of cylindrical samples, in=25,4*10−3m

DAS Dendrite arm spacing, mil =25,4*10−6m )

7 1 ( −

Inv Growth rate calculated by inverse kinetic analysis between node 1 and node 7, ms−1

) 15 1 ( −

Inv Growth rate calculated by inverse kinetic analysis between node 1 and node 15, ms−1

) 35 1 ( −

Inv Growth rate calculated by inverse kinetic analysis between node 1 and node 35, ms−1

HB Brinell hardness

fC

K Fictitious maximum stress intensity factor, MPa m

fI

K Fictitious stress intensity factor, MPa m K Fading coefficient

KJMA Abbreviation of the name of the scientists Kolmogorov, Johnson, Mehl and Avrami who developed and applied the equation at the very

beginning.

gr

L Average graphite flake length, mil =25,4*10−6m

γ

L Average dendrite length, mil =25,4*10−6m

γ

N Number of primary austenite grains, −3

m

E

N Number of eutectic cells, in 2m

1 3 2 1 ) 10 * 4 , 25 ( − − − = V

N ,Ne Number of eutectic cells,

3 −

m

-e

R Radius of the eutectic cell, m

T Temperature, 0C

•

T Cooling rate, 0Cs−1

i

T Spatial parameter for simulated temperature, 0C

Y Shape factor.

Lower case symbols

9 1 0,b ,...,b

b Coefficients for ultimate tensile strength calculation

c Half of the maximum graphite flake length, mil =25,4*10−6m

casting

d Diameter of cylindrical casting, m

γ

dir Dendrite directionality, Scale from 1 to 5.

dt df_{s,inv,F(1−15)}

Solidification rate calculated by Fourier thermal analysis, using simulated data from node 1 and node 15, s−1

dt dfs,inv,N(1)

Solidification rate calculated by Newtonian thermal analysis, using simulated data from node 1, s−1

dt df_s,sim,7

Solidification rate, simulated in node 7,s−1

dt df_s,sim,20

Solidification rate, simulated in node 20,s−1

C dt dT 0 750 Cooling rate at 7500C, 0C∗ s−1 eut dt dR

Growth rate of the eutectic cell, m∗ s−1

aus

f Fraction austenite

carb

f Sum of the fraction eutectic carbide and intercellular carbide

grD

f Fraction D-type graphite

pearlite

f Fraction pearlite

s

f Fraction solidified metal

white

k Thermal conductivity, Wm−1K−1

c

k Growth coefficient of carbide eutectic

e

k Nucleation coefficient of graphite-austenite eutectic

g

k Growth coefficient of graphite-austenite eutectic

γ

k Nucleation coefficient of primary austenite

t

k Material coefficient, Nm−1m−1/2

n Growth exponent of graphite-austenite eutectic

c

n Growth exponent of carbide eutectic

e

n Nucleation exponent of graphite-austenite eutectic

γ

n Nucleation exponent of primary austenite

sol

q

•

Heat released during solidification, in heat conduction equation, Wm−3

i

r Spatial parameters for cylindrical coordinate system, m

t Time, s

Greek symbols

i

r

∆ Space increment of ith control volume, m

z

∆ Space increment, equal to unity in 1-D polar formulation T

∆ Supercooling, 0C

P

λ Pearlite lamellar spacing, µm ρ Density, kgm−3

UTS

σ Ultimate tensile strength, MPa

-Supplement I

A. Diószegi and M. Wessén, Measurement and simulation of thermal

condition and mechanical properties in a complicated shaped cylinder

head cast in gray iron. (Modelling of Casting, Welding and Advanced

Solidification Processes – IX, 20-25 August, 2000, Aachen, Germany,

pp. 869-876)

MEASUREMENT AND SIMULATION OF

THERMAL CONDITION AND MECHANICAL

PROPERTIES IN A COMPLICATED SHAPED

CYLINDER HEAD CAST IN GREY IRON

Attila Diószegi and Magnus Wessén

Jönköping University, Div Component Technology

S- 551 11 Jönköping Sweden

E-mail: Attila.Dioszegi@ing.hj.se and Magnus.Wessen@ing.hj.se

The present paper was published in the conference proceedings: Modelling of Casting, Welding and Advanced Solidification Processes – IX, 20-25 August, 2000, Aachen, Germany, pp. 869-876

Abstract

Development and design of heavy truck engine parts require improved knowledge on solidification kinetics and development of material properties. A suitable tool to handle the complex shape and solidification kinetics is the computer simulation of the casting process. The quality of calculated results is dependent on the thermo-physical properties used, boundary condition and the quality of kinetic models implemented for calculation of resulting metallographical structure and material properties. A good quality means a good correlation between simulated and measured properties.

The paper will present a casting simulation of a cylinder head in complex shape for a diesel engine cast in grey iron together with results from measured cooling curves and investigated material properties. The values of heat transfer coefficient were calculated for a simple shaped sample, cast in a shell sand mould and used then for simulation of the cylinder head. The calculated and measured cooling curves correlate well, as well as the calculated and measured hardness value do.

Keywords

Cylinder head, cooling curve, casting simulation, heat transfer coefficient

Introduction

The cast material and the manufacturing technology of heavy truck engine parts become more and more sophisticated. A suitable tool to achieve this improvement is computer simulation of the casting process. Many efforts have been made to improve the quality of casting simulation results, where a good quality means a good correlation between simulated and measured properties. The quality of calculated results is dependent on the thermo-physical properties, boundary condition and the quality of kinetic models implemented for calculation of resulting metallographical structure and material properties. In an earlier work [1], a concept based on inverse modelling was suggested for evaluation of heat transfer coefficient and thermo physical properties, from experimental data. The aim of this paper is to present results from measurement and simulation of thermal condition and mechanical properties in a complex shaped cylinder head cast in grey iron. The thermo physical properties and boundary condition used were from reference [1]. The experiments and simulation were carried out at the Skövde Foundry of Volvo Truck Component Corporation in collaboration with the University of Jönköping, Div Component Technology.

Experimental procedure for measuring the thermal condition and mechanical properties

The investigated cylinder heads are produced on a standard moulding line. The moulding material is green sand; the cores are prepared in quarts sand, bounded by an organic bounder using SO2-gas as catalyst. The cores were prepared by drilling holes for the thermo

couples (TC) before assembling in the mould. The assembled mould was taken of from the moulding line. The TC’s were of S-type (Pt/Pt+10%Rh). The TC wires were isolated from each other in a 2-hole alumina tube, and a thin alumina coating covered the soldering point. The TC’s soldering points were positioned in several locations of the metal, core and mould; see Table 1. The mould was poured from a 1,5 ton ladle. The poured material was a grey iron with a CE = 4,0 %. (CE = C%+Si/3%+P/3%), inoculated in stream with a standard inoculant containing Strontium.

TC1 TC2 TC3 TC4 TC5 TC6 TC7 TC8

metal metal metal metal metal metal metal metal

TC9 TC10 TC11 TC12 TC13 TC14 TC15 TC16

metal metal metal core core core green sand

green sand

Table 1. Distribution of thermo couples (TC) in the mould assembly

The cooling curve data was collected by a 16-bit resolution data acquisition system, with a sampling rate of 100Hz. TC5 and TC13 were damaged during the mould filling. After cooling down to room temperature the casting was cut in sections. The cut parts were investigated by X-ray to check the final position of the top of the thermocouples. Due to thermally induced deformation, some of the thermocouples had moved from its original position. The final position of thermocouples was used in the simulation. The list of control points, shown in Table 1, can be compared to the pictures on the sectioned casting parts of the cylinder head in Appendix 1.

The sectioned parts were further cut to obtain samples for hardness measurements. The Brinell hardness value presented in Table 2 is mean values from three measurements in the neighbourhood of the top of the thermocouple.

TC1 TC2 TC3 TC4 TC5 TC6 TC7 TC8 TC9 TC10 TC11 Measured Brinell Hardness 211 202 202 211 202 202 207 202 207 198 202 Calculated Brinell Hardness 207 207,5 208 207,5 207,5 207,5 207,5 208 207,4 207,7 208,1

Table 2. Measured and simulated Brinell hardness

Simulation of thermal condition and mechanical properties Pre-processing

The simulation of the cylinder head was done using the commercial FDM simulation code MAGMAsoft®. The solid mould/core/casting geometry was created in a CAD program, ProEngeneer® and transferred to the simulation program as an STL-file. Every part of the mould/core/casting system was created as a separate part. Homogenous parts of mould and cores were also divided in different geometry files in order to allow setting of different heat transfer coefficients depending on the contact between the metal and mould/core. A number of 16 imaginary TC control points were created in the pre processor, which was compared after simulation to the experimentally registered ones. The control points were situated in different sized parts of the metal, core and mould. Pictures on the meshed casting, cores and gating system are shown in Figure 1 and Figure 2.

Figure 1. Meshed cylinder head+gating system Figure 2. Section through the cylinder head

Thermo – physical properties and boundary conditions

The materials described in the geometry files were assigned different thermo – physical properties. The thermo-physical properties of the cast metal were evaluated from experiments using an inverse calculation method based on the Fourier Thermal Analyse (FTA) described