Study of solidification and volume change in lamellar cast iron with respect to defect

formation mechanisms

Péter Svidró

Licentiate Thesis

Stockholm 2013

Division of Applied Process Metallurgy Department of Materials Science and Engineering School of Industrial Engineering and Management

Royal Institute of Technology SE-10044 Stockholm

Sweden

Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm,

framlägges för offentlig granskning för avläggande av Teknologie Licentiatexamen,

torsdag den 5 december, kl. 13:00 i Sal M131 Sefström,

Brinellvägen 23, Kungliga Tekniska Högskolan, Stockholm

ISBN 978-91-7501-938-3

ROYAL INSTITUTEOF TECHNOLOGY

ii

Péter Svidró Study of solidification and volume change in lamellar cast iron with respect to defect formation mechanisms

Division of Applied Process Metallurgy Department of Materials Science and Engineering School of Industrial Engineering and Management Royal Institute of Technology SE-100 44 Stockholm Sweden ISBN 978-91-7501-938-3 © Péter Svidró

iii

This thesis has been conducted in cooperation with

Research area of Materials and Manufacturing - Foundry Technology

School of Engineering, Jönköping University

SE-55111 Jönköping

Sweden

iv

v

To Edina & Gabriella

vi

vii

ABSTRACT Lamellar cast iron is a very important technical alloy and the most used material in the casting production, and especially in the automotive industry which is the major consumer. Beside the many great properties, it is inclined to form casting defects of which some can be prevented, and some may be repaired subsequently. Shrinkage porosity is a randomly returning problem, which is difficult to understand and to avoid. This defect is a volumetric deficiency which appear as cavities inside the casting in connection to the casting surface. Another frequent defect is the metal expansion penetration. This defect is a material surplus squeezed to the casting surface containing sand inclusion from the mold material. Shrinkage porosity is usually mentioned together with metal expansion penetration as the formation mechanism of both defects have common roots. It is also generally agreed, that these type of defects are related to the volumetric changes occurring during solidification. Additionally, the formation of these defects are in connection with the coherency of the primary austenite dendrites. The purpose of this work was to develop knowledge on factors affecting a volume-change related casting defect formation in order to minimize the presence of these defects in engine component production. This was done by extending the existing solidification investigation methods with novel solutions. Introduction of expansion force measurement in the determination of dendrite coherency combined with multi axial volume change measurement refine the interpretation of the solidification. Comparison of registered axial and radial linear deformation in cylindrical samples indicated an anisotropic volume change. Different methods for dendrite coherency determination have been compared. It was shown that the coherency develops over an interval. Dependent on the added inoculant the coherency is reached at different levels of fractions of a solidified primary phase. It is also shown, that inoculation has an effect on the nucleation and growth of the primary phase. Quantitative image analysis has been performed on the primary phase in special designed samples designed to provoke shrinkage porosity and metal expansion penetration. It was found, that the inter-dendritic space varies within a casting. This was explained by the coarsening of the primary dendrites which originates from differences in the local time of solidification. Keywords : lamellar cast iron, shrinkage porosity, solidification, volumetric change, dendrite coherency, inter-dendritic space

viii

ACKNOWLEDGMENTS I would like to give special thanks to my supervisor Professor Attila Diószegi for his guidance both in science and also other areas of my life. I would like to express my deepest appreciation to my supervisor Professor Pär Jönsson for his support and encouragement. I am deeply grateful to Lennart Elmquist my assistant supervisor and coauthor for the good advices and interesting discussions. I am grateful to Volvo Powertrain AB, Sweden and the Foundry Department at University of Miskolc, Hungary for their collaboration in the experimental part of the present work. I am also grateful to the Swedish Knowledge and Competence Foundation financing the project JÄRNKOLL 2C, and the cooperating parties (School of Engineering at Jönköping University, Volvo Powertrain AB and Scania CV AB) within the analyzing part of the present work. Special thanks are given to Lars Johansson and Tony Bogdanoff for technical assistance and advices. Thanks to all my colleagues at the department of Mechanical Engineering, Materials and Manufacturing - Casting at the Jönköping University. Thanks to Gabriella Fekete for the german and Katalin Gémesi for the french translations. Thanks to Richárd Szepesi. “Ne nyomd, mert leég !” I would like to thank my parents Adrien and József for the material and mental support, and my brother for his helpfulness. Last, I would like to thank my beloved wife Eszter for all her support and help during these years, and for giving us Bence.

ix

SUPPLEMENTS The present thesis is based on the following papers:

Supplement 1.: “On the problems of measuring volume change in Lamellar cast iron,” Péter Svidró and Attila Diószegi, Accepted for publication in the International Journal of Cast Metals Research under manuscript nr. IJC1295, 06. August. 2013.

Supplement 2.: “Determination of the columnar to equiaxed transition in hypoeutectic lamellar cast iron” Péter Svidró and Attila Diószegi, Presented at The 3rd International Symposium on Cutting Edge of Computer Simulation of Solidification, Casting and Refining (CSSCR2013), KTH and Aalto University, May 20 - 23, 2013. Accepted for publication in the ISIJ International under manuscript nr. ISIJINT-2013-371, 12. June. 2013.

Supplement 3.: “Investigation of cooling rate dependent dendrite morphology in hypoeutectic lamellar cast iron” Péter Svidró, Lennart Elmquist, Izudin Dugic, Attila Diószegi Research report, School of Engineering, Jönköping University Apr. 2013.

The contribution from the author to the different supplements of the thesis :

Supplement 1.: Literature survey, evaluation of data, calculations, major part of the writing

Supplement 2.: Literature survey, evaluation of data, calculations, major part of the writing

Supplement 3.: Literature survey, measurements, evaluation of data, calculations, major part of the writing

x

CONTENTS

ABSTRACT .................................................................................................................... vii

ACKNOWLEDGMENTS ......................................................................................... viii

SUPPLEMENTS ............................................................................................................ ix

CONTENTS .................................................................................................................... x

Introduction ...................................................................................................................... 1

1.1. Primary solidification of lamellar cast iron ................................................ 1

1.2. Defect formation mechanisms .................................................................... 3

Objective and overview of the work............................................................................. 6

Experimental methods .................................................................................................... 7

3.1. Volume change measurements .................................................................... 7

3.2. Calculation of fraction solid by Thermal Analysis ................................... 9

3.3. Determination of columnar to equiaxed transition .................................. 9

3.4. Stereological investigation of dendrite morphology .............................. 11

3.5. Image analysis and stereological definitions ............................................ 12

Results and discussion ................................................................................................... 14

4.1. On the problems of volume change measurements .............................. 14

4.2. Determination of the columnar to equiaxed transition ......................... 18

4.3. Investigation of cooling rate dependent dendrite morphology ............ 19

Concluding discussion ................................................................................................... 21

Conclusions ..................................................................................................................... 24

Future Work ................................................................................................................... 25

References ....................................................................................................................... 26

Supplement I ................................................................................................................... 30

Supplement II ................................................................................................................. 45

Supplement III ............................................................................................................... 55

1

Introduction Lamellar gray cast iron (LGI) is one of the most important cast materials. Mechanical, thermal, vibration damping properties and low production costs keep its position in the technological forefront. Along many of the advantages of this alloy, the tendency to form casting defect is one of the drawbacks. Foundries producing complex castings of LGI used by the marine or the automotive industry such as cylinder heads or blocks are more exposed to this failure. Some of the casting defects can be prevented and some may be repaired subsequently. Rejected castings increase the scrap ratio and have a negative impact on the production and environment.

1.1. Primary solidification of lamellar cast iron Solidification is an essential part of the casting process since it determines the microstructure and hence the physical and mechanical properties [1], [2]. A liquid to solid phase transformation occurs as atoms in a short-range order of the liquid phase are rearranged into regular positions in the crystallographic lattice of the solid phase as heat removed from the melt [3]. Heat is transported through the mold into the environment. The heat transfer property of the mold material being the dominant factor. Since the solidification generally is driven by the heat transport, the inspection of the temperature changes during cooling provides a lot of information [4], [5]. Registering temperature against time results in a cooling curve, which is a common way to observe solidification [6], [7]. During the solidification of hypoeutectic gray cast iron two main events can be noticed, a primary solidification and an eutectic solidification [8–10]. These can be easily identified on a cooling curve like in Figure 1.

Figure 1. Primary and eutectic solidification on a cooling curve.

primary solidification

eutectic solidification

2

Solidification of hypoeutectic gray cast iron is accompanied by a volume change, both a shrinkage and an expansion. Expansion is a volumetric surplus explained by the graphite precipitation, that takes place during the eutectic solidification [11]. Shrinkage is a volumetric loss due to the contraction of the alloy both in a liquid, a liquid-to-solid and in a solid phase [2]. Consequently the overall volume change is a complex progress of shrinkage and expansion and represents the character of solidification. The first event during the solidification of a hypoeutectic cast iron is the nucleation of the primary austenite [12–14]. A schematic representation of the primary grain development from nucleation to grain impingement is shown in Figure 2. Primary austenite forms two types of grains named after the preferable growth direction of the main dendrite arms. One type is the columnar grain, nucleated at the mold/melt interface. The heat flows from the melt toward to the mold. Hence, the main arm of the dendrite grows in opposite direction in a columnar manner [8], [12], [15]. The second type of grain is called equiaxed, and it is nucleated in the overcooled liquid front ahead of the columnar grains (Figure 2a). The growth conditions regarding to the even heat flow from the growing crystals to the surrounding liquid assist the main dendrite arms to develop equally in every direction [8], [15] (Figure 2b). The columnar crystals grow in competition with each other and perpendicularly to the mold surface. The network of the columnar crystals forms a pattern comparable with a container. However, the coherent grains are permeable with respect to the inter-dendritic liquid phase. (Figure 2c) In foundry practice, this columnar crystal zone is called a casting skin. The stiffness of this container like aggregate is expected to be very low, easily adapting the shape of the mold. The dominant morphological volumetric increase of the austenite crystal acquires due to the advancement of the dendrites lead by the hemispheric dendrite tip until the grains collide with the neighboring crystals [1] (Figure 2d). The moment when these crystals touch each other is called a dendrite coherency [16-27], which can be determined by different methods [28], [29]. On macro scale, the coherent aggregate behaves rather like a solid than a liquid [29]. As the crystals impinge on each other, the volumetric growth is blocked. Thus, an increase of the austenite phase is possible only by deposition of solid on the existing dendrite surfaces. The development of the primary dendrites after coherency can be characterized by dynamic coarsening. [29]. It is a combination of a phase increase due to the heat extraction process and a surface energy driven morphological change by the intention to minimize the free energy of the solid and liquid phase, originated from the thermodynamic imbalance of the phases [30]. In other terms, this is a reduction of the interfacial area of the primary dendrites over time [31]. Translated to the dendrite morphology it has been shown that by increasing the time in the solid-liquid state the inter-dendritic space will increase.

3

Parallel to the primary dendrite development, the chemical composition of the remaining liquid becomes eutectic due to carbon segregation. This leads to an eutectic nucleation and growth of eutectic colonies [9]. Until the dendrite coherency, liquid feeding was responsible for supplying the solidifying regions with liquid. If the liquid mobilized the floating equiaxed grains the literature named the phenomena “mass feeding” [32]. Thereafter, a movement of the mass feeding slurry is hindered as dendrites impinge on each other. Before the creation of the three-dimensional structure of a coherent dendrite network, a volume change of the casting is determined due to the movement of a casting skin. This movement is influenced by the mechanical properties of the columnar zone and the dimensional changes of the thermally loaded mold. After the columnar to equiaxed transition, the movement of the casting surface represents the volume changes of the whole casting (Figure 2d).

a. b. c.

d.

Figure 2. Schematic representation of the primary grain development from nucleation to grain impingement. The primary phase in Figure 2a. and 2b. is represented as dendrite crystals (white) surrounded by the liquid phase (gray). When the grain density becomes more compact only the theoretical boundary of the austenite grains are represented in Figure 2c. and Figure 2d. Within the grains the austenite crystals are still in mixture with a considerable fraction inter-dendritic liquid (white). The gray phase between the grains is considered the intergranular liquid phase. The thick line represents the theoretic boundary between the columnar zone (casting skin) and the equiaxed zone.

From this point until the end of the solidification, an inter-dendritic feeding in combination with an intergranular feeding is responsible for the liquid supplement [32]. This is also a complex procedure. The characteristic of the primary dendrite network is an important influencing factor on this type of feeding, since the liquid should travel in the remaining inter-dendritic and intergranular space.

1.2. Defect formation mechanisms Most challenging types of casting defects are related to the volumetric changes accompanying the solidification. One is shrinkage porosity, a randomly recurring problem in the everyday casting practice. In the case of LGI, the porosity is likely to form inside the casting. Often near to the surface and sometimes remaining hidden until the part is machined.

4

The shrinkage porosity is proposed to be formed on the primary austenite grain boundaries (Figure 3). An existing liquid depression in the intergranular area could be the drowning force when the external gaseous environment is sucked into the grain boundaries. The formed porosity can penetrate through the casting wall (Figure 4) which results in a leakage [9]. Foundries producing pressure tight parts spend a lot of efforts to prevent this type of casting defects. The shrinkage porosity is usually mentioned together with a metal expansion penetration, as they are believed to have a common cause in the creation mechanism. A metal expansion penetration defect is likely to form on the surface of the casting and appears as a metal adhesion with a sand inclusion, as seen in Figure 5 and Figure 6. The formation mechanism of metal expansion penetration is supposed to be comparable to the one of a shrinkage porosity formation. However, this time the liquid is pressurized in the intergranular space, which squeezes the excess of liquid metal to the casting/mold surface [33]. A repair of the affected area is an additional operation cost. The presence of penetration is worsened when it is situated in any of the fluid channels of a cylinder head. According to the literature, these types of defects are formed between the dendrite coherency and at the end of the solidification. The volume change during solidification and the conditions of the dendrite coherency are believed to be the main affecting factors on the formation of these defects. During solidification of lamellar cast iron the main phases are the austenite and graphite. The precipitated austenite has a higher density while the graphite has a lower density compared to liquid iron. The combined effect of precipitation will be a volume decrease due to the austenite precipitation and a volume increase due to the graphite precipitation. The overall volume change will be dependent on the primary austenite structure and the mixture of the austenite and graphite phases.

Figure 3. Porosity on grain boundary.[9]

Figure 4. Porosity interconnected through a thin wall casting. [9]

5

Figure 5. Penetration on casting surface.

Figure 6. Penetration and sand inclusions on casting surface.

5 mm 200mm

6

Objective and overview of the work The present thesis focuses on the factors influencing the volume change related defect formation mechanisms during solidification. The first specific objective for the thesis was to study the characteristics of the volume change during solidification. The second objective was to investigate the development of the macrostructure. More specifically, the transition between the columnar and equiaxed austenite grains. The third objective was to find a relation between the primary austenite morphology and the local solidification times. An overview of the thesis is presented in Figure 7. In supplement 1, an extensive literature survey has been performed to review the evolution of volume change measurement methods. The results of linear displacement based indirect measurements of volume change have been extended and compared to expansion force measurements. The characteristics of a volume change have been pointed out. In supplement 2, expansion force measurement has been introduced as a CET determination method and compared to methods known from the literature. The usefulness of expansion force measurement was described and the effect of inoculation on the primary austenite was pointed out. In supplement 3, the hydraulic diameter of the inter-dendritic space has been investigated as a function of the local solidification time on samples where shrinkage porosity and metal expansion penetration were present.

Supplement 1 Volume change measurements Evaluation of volume change and expansion force measurements Study of the volume change character.

Supplement 2 Columnar to equiaxed transition Determination of CET Study of the CET and the effect of inoculation on the solid fraction

Supplement 3 Investigation of dendrite morphology Quantitative image analysis of microstructure Study of the coarsening and the development of the inter-dendritic space as a function of local solidification time

Increased knowledge on factors affecting the formation

of shrinkage porosity and metal expansion penetration in lamellar gray cast iron

Figure 7. Schematic overview of the thesis work.

7

Experimental methods The experiments in Supplement 1 and Supplement 2 have been performed in cooperation with Volvo Powertrain AB and the Foundry Department at University of Miskolc, Hungary. The experimental concept and the setup were also developed at University of Miskolc, Hungary [34]. The experiments in Supplement 3 have been performed in collaboration between Volvo Powertrain AB and the School of Engineering at Jönköping University (JTH). The aim was to produce samples where a metal expansion penetration was present. The making of samples with the presence of shrinkage porosity was performed in collaboration between JTH, Volvo Powertrain and Scania CV AB. The investigation part of the experimental material used in all three supplements was performed within a collaboration project involving JTH, Volvo Powertrain AB and Scania CV AB.

3.1. Volume change measurements Volume change measurements were initially aimed to track a dimensional deviation, but later they became a method to study solidification [35]. Numerous experimental methods have been developed for linear volumetric change measurements [36-52], including the one of interest for the current work [53]. Together with the weaknesses, indirect measurement of volume changes based on a linear displacement have been found to be a good way to follow solidification in a macro scale [54]. If the measuring method is extended to include expansion force measurements, it will also be possible to study the solidification kinetics. The scope of the measurement was to find a relation between the volume changes that occur during solidification, based on a linear displacement and the expansion force. The innovative part of these measurements was the introduction of the multi-axis and directional linear displacement measurement. Moreover, the introduction of a two-thermocouple thermal analysis method and a measurement of the expansion force. Two cylindrical samples of Ф50 and length 350 [mm] were cast in parallel from the same alloy. The samples were molded in a sand shell mold. The dimensions of the drag and cope were 45 x 100 x 407 [mm]. One of the molds was used for the displacement and the temperature measurements. Also, two thermocouples of type “S” (Pt-PtRh10) were placed perpendicular to the longitudinal axis of the cylindrical casting. The measuring spots of the thermocouple were placed in the thermal centrum, which is equivalent with the cylinders rotational axis ( ). Furthermore, at a distance which is located 10 mm from the axis to the cylindrical surface ( ). The temperature sensors were protected by a quartz tube (Φ5 external and Φ3 internal diameter [mm]), which was welded to a hemisphere and a 1 mm thick protecting wall at the end of the tube. Two quartz rods (Φ10 mm) were built in the mold’s parting line, which coincides with the axis of the cylinder. The purpose was to transmit the axial linear displacement of the test bar.

8

Two quartz rods were built in the mold’s parting line perpendicular to the cylinders rotational axis, to transmit the radial linear displacement of the test bar. The mold with the quartz rods and the displacement sensors were assembled by a special frame on a table. The second mold, aimed for expansion force measurements was equipped with two (Φ10 mm) quartz rods built in the mold’s parting line, which coincides with the axis of the cylinder. In this case, the quartz rods connected to a force measuring unit were also assembled by a special frame on a table. The experimental assembly is shown in Figure 8.

Figure 8. Experimental assembly for volume change measurements and CET determination.

Gray cast iron was prepared in a 60-kg medium frequency induction furnace using scrap from commercial casting components. A typical chemical composition is shown in Table 1. C Si Mn P S Cr Mo Cu Sn Min. 3.30 1.65 0.54 0.05 0.09 0.15 0.21 0.77 0.05 Max. 3.78 1.82 0.64 0.05 0.11 0.17 0.25 0.92 0.08

Table 1. Chemical composition in samples used for volume change measurements and CET determinations During an experiment, the melt was poured into a ladle, while a Sr based commercial inoculant was added into the tapping stream. The amounts of inoculant, the pouring temperature and the time interval between inoculation and pouring were changed for the different heats. Overall, 11 heats were cast. The metallurgical parameters used in these experiments are shown in Table 2. After melting, the iron was heated to 1460 °C. Thereafter, it was held at the maximum temperature for 20 minutes, before being transferred to a preheated hand ladle of a 40 kg size to be casted. In the case of the highest pouring temperature T3=1440 °C and the shortest inoculation time t1=0 minute, the melt was cast immediately.

1. inlet2. frame

3. quartz rod4. shell mold

5. thermocouples6. force measuring unit

7. displacement transductor

1

2 3

3

3

4

56/7

7

9

Casting temperature [°C]

Inoculation time [min]

Inoculation content [w%]

T1=1320 t1= 0 c1=0.05 T2=1380 t2= 5 c2=0.10 T3=1440 t3=10 c3=0.15

Table 2. Metallurgical parameters used in the experiments In the other cases, the melt was poured back into the furnace to maintain the designated pouring temperatures and inoculation times. The signals from temperature and expansion force measurement were registered by a HBM-Beam U10 software system.

3.2. Calculation of the fraction of solid by Thermal Analysis Cast alloys often have more than one solid phase, which precipitate from the liquid phase. Here, the transformation releases the latent heat of solidification. This heat is assumed to be proportional to the fraction of solidified metal. Based on temperature measurement during solidification, different types of numeric algorithms are available to predict the release of latent heat during solidification. The thermal analysis is performed based on the registered temperature from two thermocouples and a numerical algorithm. The latter is known from the literature as the Fourier Thermal Analysis method (FTA) [55], [56]. It uses two measuring points in a nearly 1-D thermal field and needs a tabulation of the volumetric heat capacity of the phases taking part in the solidification. Several work found in the literature have been devoted to prove the accuracy of the Fourier Thermal Analysis method [57], [58].

3.3. Determination of columnar to equiaxed transition The solidification kinetics is dependent on the thermal behavior of the solidifying material. Therefore, the temperature data contains precious information. In this experiment, the two-thermocouple method [28], [29] was used to measure the temperature. The determination of the columnar to equiaxed transition based on temperature differences is based on the fact, that the thermal conductivity between the central part and the outer shell of the casting increases at the dendrite coherency. Conversely this method uses two thermocouples, one placed at the thermal center and one located slightly displaced from the thermal centrum toward the mold . Solidification of the cylindrical sample in the mold starts as heat transferred from the melt into the mold. Thereafter, columnar crystals growing from the mold wall create a container like columnar zone compassing the remaining melt. The temperature of the outer thermocouple is lower than at the center

. In the columnar zone, the temperature decreases as the heat flows in the direction of the mold. At the same time, the released latent heat increases the temperature of the melt. Thereafter, the solidification continues with the nucleation and growth of equiaxed crystals.

10

The primary crystals that are formed fills up the remaining room, while more latent heat is released. This, in turn, leads to an increased temperature. At the last moment before the crystals impinge, the temperature difference ∆ reaches its maximum value. Also, at a certain point advancing crystals touch each other, which is called a dendrite coherency. As soon as the primary crystals create a coherent dendrite network in the volume, the heat conductivity of the system suddenly increases. Therefore, heat can be taken away through this network which causes ∆ to decrease. Thus, the dendrite coherency can be determined at the maximum ∆ value. It should also be noted that the present description to determine the coherency point based on temperature differences, is supported by results presented in the open literature [28], [29]. The concept of the determination of the dendrite coherency based on registered expansion forces has been introduced in Supplement 1. The principle of the concept is based on the parallel measurements of two identical cylindrical samples, which derive from the same heat. One of the samples was connected to a linear displacement transducer in both the axial and the radial direction. Moreover, the other sample was connected to a force measurement unit by using quartz rods. After pouring the samples, a container-like columnar zone is expected to develop and to incorporate the quartz rods linking the columnar zone with the measuring device. The displacement transducer indicates an expansion of the columnar zone in both the axial and the radial direction from the end of the mold filling. This was interpreted to be a result of the interaction between the container-like columnar zone with the thermally loaded molding material and the absence of any resistance from the linear displacement transducer against the quartz rod displacement. The quartz rod in the second sample is supposed to be affected in similar way by the solidifying columnar zone, but the force measuring unit did not indicate any signal. This was only possible only if the solid columnar zone was gliding over the incorporated quartz rod, despite that an expanding displacement was expected. If the different force measuring unit devices are compared, the current one is the only one that develops a resistance versus the displacement of the rods. Consequently, the start of registering force signals should coincide with the existence of a coherent network inside the container-like columnar zone. By comparing the start of the force accumulation moment with the calculated sold fraction, it was possible to identify CET as function of the registered expansion forces. As the amount of released latent heat is proportional to the fraction of the solid phase, the released latent heat increases with the nucleation and growth of the solid fraction. From a morphological point of view, the equiaxed austenite grains start to develop as a small sphere until the dendrite arms develop the primary and the lower order dendrite arms.

11

The dominant morphological volumetric increase of the austenite crystal is acquired due to the advancement of the dendrites. This, in turn, is led by the hemispheric dendrite tip until the grains collide with the neighboring grains. The free surface of the austenite increases linearly with an increased volume fraction. From the CET measurements a volumetric increase of the austenite grain is possible to determine, if only a deposition of a solid phase has been formed on the existing surfaces. The phenomenon indicates an increase of the internal density of the grain, while the available free surface to continue the solidification decreases time-dependent released latent heat. From the present description the maximum of released latent heat indicates the dendrite coherency.

3.4. Stereological investigation of dendrite morphology The scope of the measurement was the quantitative analysis of the primary austenite morphology in regions affected by a shrinkage porosity and a metal expansion penetration. Samples taken from castings previously developed to represent the thermal conditions like in a complex shaped casting, hence promote the formation of the designated casting defects. The material of the casting alloys was a hypoeutectic gray cast iron, of a similar composition as those alloys used in ordinary production of automotive cast components. A casting called the cylinder head simulator (CHS) has been developed to simulate the thermal conditions like in a cylinder head casting (Figure 9) and to provoke a shrinkage porosity [59]. The specimens were cast into an Epoxi-SO2 hardened sand mold. Thereafter, the casting were cut into half at the middle cylinder. Thereafter, investigations were been made on the areas where defects were found (Figure 10).

Figure 9. Cylinder head simulator

casting.

Figure 10. Cross section and

investigation area of cylinder head simulator casting.

A casting called a metal expansion penetration simulator (MEP) casting has been developed to provoke penetration defects. Specimens were cast from the top into an organic binder-SO2 hardened sand mold (Figure 11).

12

After filling, the sample was covered by a sand lock to create an even heat extraction in all directions out from the sample. Thereafter, the casting was cut into half and an investigation on the presence of defects on the surface was made (Figure 12).

Figure 11. MEP mold with cover on top.

Figure 12. Cross section and

investigation area of MEP simulator casting.

3.5. Image analysis and stereological definitions After sample preparation, a color-etching was performed on the surfaces. The color-etching is a very useful technique to reveal the microstructure [60] (Figure 13). On the etched surface both primary austenite and eutectic colonies appear as blue tones, the color distribution is dependent on the Si segregation in both the primary and secondary austenite phase. Pictures for investigation were taken from the defected areas as well as from the surrounding areas. At a 2,5x magnification, one frame represents an approximately 5 [mm2] area. Automated commercial image analysis software products cannot distinguish the phases that appear within a same color range, therefore the phases had to be separated before a quantitative analysis could be carried out. For this purpose, an interactive pen display was used and the phases were marked by hand (Figure 14). Due to the Si segregation in both the primary austenite and the secondary austenite forming the eutectic cells in collaboration with graphite, it is difficult to distinguish the exact morphology of the primary austenite. Thus, for those areas where the contour of the primary phase could not be distinguished, the data were excluded from the investigation. These data are represented by a black color in Figure 14. The well distinguishable primary austenite surfaces are colored white and the remaining gray area represents the inter-dendritic space (Figure 14).

13

Figure 13. Color-etched micrograph from the CHS sample.

Figure 14. Picture preprocessed for image analysis.

White – primary austenite Gray – inter-dendritic space

Black – excluded from measurement

Two morphological parameters with respect to the primary dendrite were investigated. The secondary dendrite arm space (SDAS) given in mm were measured as well the surface area of the inter-dendritic phase along with the perimeter between the observed primary austenite phase and the inter-dendritic space. A recent research work [61] discussed the fraction between the surface area of inter-dendritic space and its perimeter. The results showed that it corresponds to the Hydraulic diameter of the inter-dendritic space with a unit of [mm]. In addition, the Hydraulic

diameter / inter-dendritic space has been shown to wary with respect to the local solidification time as SDAS do. Therefore, it can be interpreted as a mechanism showing a dynamic ripening. The local solidification time was calculated to compare the results to measured data. More specifically, the morphological parameters were simulated by a commercial simulation software MAGMASoft®, which considers the casting temperature and the chemical composition.

200mm

14

Results and discussion 4.1. On the problems with volume change measurements The combined effect of contraction and expansion of the precipitating phases at solidification of lamellar cast iron makes it difficult to understand the individual contribution of the phases. The resulting volume change becomes even more difficult to interpret, since the solidifying aggregate interacts with its thermally loaded mould environment. Throughout the years, many different measurement methods have been developed to investigate the volume change. The obtained results from the literature were interpreted mostly without taking into consideration the sample – mould interaction and the individual contribution of the precipitated phases. The reason was that such results resulted in many controversial interpretations of how the volume change took place. The chosen measuring method including both a multi-axial displacement measurement combined with an expansion force measurement and a cooling rate measurement in two positions, reveals some of the general problems of volume change measurements. Similar problems were also found in the literature [54]. The problem of an early linear expansion was found for the measured displacements in both the axial and radial directions, as seen in Figure 15 and Figure 16. The presence of the early linear expansion could be explained only by the metallo-static pressure that appears in the liquid iron. This is created by the level differences between the quartz rod and the top of the downsprue after mold filling. The duration of the metallo-static pressure responsible for the expansion displacement depends on how long the top of the downsprue and the quartz rods are interconnected with the liquid metal. This moment corresponds to a local coherence of the primary austenite network, which interrupts the liquid communication.

Figure 15. Cooling and displacement curves registered for Heat I.

15

Figure 16. Cooling and displacement curves registered for Heat II. The main benefit of the selected experimental technique was the possibility to determine the columnar to equiaxed transition based on the registered expansion forces. The theories explaining how the CET is determined was described earlier in Chapter 3.3. A comparison of the registered axial and radial displacements within the same casting reveal the anisotropy of the volume change, as seen in Figure 17 and Figure 18.

Figure 17. Accumulated force and displacement curves registered for Heat I.

16

Figure 18. Accumulated force and displacement curves registered for Heat II. The radial displacement is always larger than the axial displacement. The radial displacement is a result of the interaction between the internal equiaxed grains, the cylindrical mantle of the columnar zone, and the mold material. The axial displacement is more complicated, because in addition to the mold material and the equiaxed grains, the axial deformation of the columnar mantle is influenced by the columnar zone. This results in a formation of a disk-shaped zone at both ends of the cylinder. The equiaxed grains are expected to have a random orientation, which should have an equally distributed influence on the total radial and axial displacements. The existence of the disk-shaped columnar zone at the ends of the cylinder may hinder the displacement during both expansion and shrinkage. The volume change and consequent density variation calculation of cast iron is performed according to the literature [44] based only on axial linear displacement measurements. Most of the standard dilatation measurement techniques are also based on only axial linear displacement measurements. Calculations of the volume change were also done based on using both radial and axial linear displacements to compare with predictions performed by only using the axial linear displacements. This comparison showed that clear differences could be found, as seen in Figure 19 and Figure 20. The outcome of this observation indicates the drawback of basing density variation measurements only on a linear displacement measurement.

17

Figure 19. Volume change based on the axial displacement as a function of the volume change based on axial and radial displacements in Heat I.

Figure 20. Volume change based on the axial displacement as a function of the volume change based on the axial and radial displacement in Heat II.

4.2. Determination of the columnar to equiaxed transition The columnar to equiaxed transition was investigated using different identification methods while the process parameters were varied with respect to the added amount of inoculant, the time of inoculant addition before casting, and the pouring temperature.

18

The different identification methods were based on the inverse thermal analyses and the calculation of the fraction solid according to the Fourier Thermal Analyse method. By comparing the CET determination methods for all cases, the earliest CET as a function of the precipitated solid fraction could defined. This happened when the rate of the released solidification heat reached a maximum value, as seen in Table 3.

fs determined by Sample 14 Sample 19 12.97 7.89

∆ 20.29 13.82 20.29 13.82

Table 3. Fraction of solid at CET determined by different methods. This moment is interpreted as the moment when the first equiaxed grains impinge on each other. However, the whole interspace enclosed by the columnar network (also associated with a container) is not completely filled by the equiaxed grains. The CET value determined based on the expansion measurement and the maximum temperature gradient ∆ occur at the roughly same fraction of the solidified phase. This indication of the CET value is interpreted as the moment when the whole primary austenite network impinges. Furthermore, when the stiffness of the grains is capable to withstand a compressive tensile load. The differences in the fraction solidified from the different identification methods indicate that the CET value represents an interval rather than a fixed point. The differences due to the metallurgical parameters are summarized in Figure 21. The results indicate the role of an inoculant addition on the CET formation. Overall, a dominant tendency is that an addition of an inoculant leads to a decreased fraction of materials that is solidified at coherency. Indirectly, the present observation tells us that the addition of an inoculant with the purpose to influence the graphite nucleation also influences the nucleation and growth of the primary phases.

Figure 21. Fraction of solid at CET determined by released heat of crystallization.

19

4.3. Investigation of cooling rate dependent dendrite morphology Beside traditional morphological parameters like the Secondary Dendrite Arm Space (SDAS), new morphological parameters to characterize the primary austenite dendrite network have also been introduced recently. One of these parameters refers to the inter-dendritic space and it is also called the hydraulic diameter ( ). Furthermore, it has been demonstrated that the primary austenite dendrite network is also suffering a dynamic coarsening. In the present research, the hydraulic diameter was also considered. It was found that by increasing the local solidification time, the hydraulic diameter could also be increased. With other words, the space for liquid mass flow could be enlarged by increasing the solidification time. Colour etching techniques were used to reveal the morphological parameters and numerical simulation were carried out to determine the local solidification time for samples known to provoke shrinkage porosity and metal expansion penetration. The result showed that the same dynamic coarsening phenomena was found as reported in the literature [61]. More specifically, the measured inter-dendritic space ( ) is related to the simulated local solidification time. With other words, the largest inter-dendritic space is found were the local solidification time was the longest. The studied samples were specially designed to have the longest solidification time positioned close to the casting mould interface. Consequently, there is a decreasing from the casting surface to the rest of the casting. Worth to note is that the investigated samples contained shrinkage porosities respectively metal expansion penetrations in connection to the wide inter-dendritic space caused by the long local solidification time. The present observations indicate that the liquid mass flow could be facilitated by the wide inter-dendritic space, which contributes to the defect formation. However, at this moment it is still not understood why the inter-dendritic liquid state can alter between pressure and depression. The obtained relations indicate that a dynamic coarsening takes place, as presented in Figure 22 and Figure 23.

20

Figure 22. Comparison of literature data and [61] and CHS sample results and .

Figure 23. Comparison of literature data and [61] and MEP sample results and .

21

Concluding discussion In Supplement I., an extensive literature study of direct and indirect volume change measurement techniques is presented. Literature on advances in the interpretation of the solidification with respect to the primary austenite grains has also been included. This in order to support the discussion of phenomena that appear concurrently to the volume change, taking place during the solidification of gray cast iron. Both the current measured results and data found in the literature indicated that an early expansion can be observed in all measured direction (Figure 15, Figure 16). These expansions are related to the interaction between the molding material, the columnar zone forming a container such as a solid shell, and the mixture of liquid and equiaxed crystals. Consequently, a displacement indicated by transmitters represents a misleading movement until the development of the coherent austenite dendrite network has been reached. The introduction of a force measuring unit highlights the weakness of the displacement measurements. Moreover, it refines the interpretation of the solidification by indicating when the columnar to equiaxed transition takes place. The observation of a continuous force increment on the force measuring unit is related to the existence of a coherent dendrite network in the equiaxed zone, supporting the load on the instrument (Figure 17, Figure 18). A comparison of the registered axial and radial linear deformation indicates that an anisotropic volume change takes place in the cylindrical samples (Figure 19Figure 20). This volume change anisotropy should be influenced by the size of the columnar zone. In addition, it should reveal the problems with the classical measurement of a linear expansion or shrinkage measured in only one direction. Also, small differences in the metallurgical behavior due to the additions of different amounts of inoculants resulted in differences in the measured and calculated volume changes. In Supplement II, an expansion force measurement has been implemented into the dendrite coherency determination. This measuring approach has been compared to other measurement methods. According to the literature, the thermal analysis method is a common way to determine the dendrite coherency. Temperature difference indicates the occurrence of the first touch of the primary crystals. It is useful to assign the initiation of the interconnected dendrite network. Compared to thermal analysis method, the expansion force measurement detects the point when columnar and equiaxed crystals are welded together to form a coherent system. In addition, the determination of the CET by the released heat highlights the solidification differences related to the solid fraction (Table 3). Also, a microstructure modification was performed by using the same commercial inoculation material as commonly used in the foundry practice. The effect of these materials on the nucleation and growth of the eutectic colonies are well investigated in the literature.

22

The experimental results show that an inoculation has an effect also on the nucleation and growth of primary austenite (Figure 21). In Supplement III, samples with a tendency to form shrinkage or expansion defects were investigated with respect to the primary dendrite morphology. It was found that the inter-dendritic space increases as a function of the local solidification time. A gradient of a decreased inter-dendritic space was observed as the local solidification time became shorter. In addition, solidification times were simulated and the results were compared with the modulus of the inter-dendritic space at the same location. The present results were also compared with SDAS measurements and literature data [61], as shown in Figure 22 and Figure 23. It was pointed out, that when the casting surface coincides with an area where the solidification time is long, the dendrite grains which have a large inter-dendritic space possibly promotes the liquid flow between the dendrites. Hereby, they create the condition to form a shrinkage porosity or an expansion penetration. The overall findings in this works related to solidification and volume change with respect to defect formation are summarized in Figure 24. The columns from left to right represent the development of primary austenite and the circle represents the thermal center. The first row represents the ideal thermal behavior. In this case, the thermal center which consequently has the longest solidification time is situated in the middle of the casting at approximately an equal distance from the columnar zone. The consequence of this arrangement is a uniformly developed columnar zone with approximately an even permeability. The second and the third rows represent “complex” casting situations. The second row represents the case when an internal core forms a concave casting surface. Due to the thermal saturation of the internal core the hot spot and the longest local solidification time is expected to be situated at the metallic interphase in the centrum of the casting mold aggregate. The thermal case shown in the third row indicates a thin core or mold material section between the casting with a cylindrical and quadratic cross section. The thermal saturation of the mold / core will position the hot spot and the longest solidifying metal over the casting-mold-casting section included in the circle. The presence of the hot spot and the longest local solidification time in connection to the casting – mold surfaces will influence the columnar crystals belonging to this zone. The different thermal arrangements will result in the known dynamic coarsening phenomena. This, in turn, will lead to the consequent enlargement of the inter-dendritic space inside the marked circles.

23

Different primary-dendrite coarseness or different inter-dendritic spaces at high magnifications are included in the figure. The thermal condition of the second row is known to provoke metal expansion penetration from castings. Furthermore, the thermal condition in the third row is known to provoke shrinkage porosity during casting. One conclusion from the presented work is that special thermal conditions facilitate the creation of a flow path for the liquid mass flow due to the dynamic coarsening of the primary austenite network. If the initial liquid state is a depression, the enlarged network will facilitate the intake of a gaseous phase from the surrounding atmosphere which is present between the primary dendrite network. On the contrary, if the liquid is pressurized the enlarged inter-dendritic space will facilitate the squeezing of an inter-dendritic liquid to the casting surface.

Figure 24. Schematic of primary austenite development.

24

Conclusions This thesis focuses on the solidification of lamellar cast iron with respect to the defect formation mechanisms. More specifically, the volume change during solidification is studied. Moreover, the columnar to equiaxed transition and the as-cast microstructure are studied. The most specific conclusions from the study can be summarized as follows: Conclusions of Supplement I:

1. An early linear expansion is a weakness in the displacement measurements and it is caused by an aggregation of several movements.

2. A linear displacement represents the real volume change, which occurs after the columnar to equiaxed transition has taken place.

3. A columnar to equiaxed transition during solidification can be determined by the expansion force measurement method.

4. A volume change shows an anisotropy within a casting. 5. An investigation of a volume change during solidification based

only on displacement measurements in one direction may lead to wrong results.

Conclusions of Supplement II:

1. Different experimental methods can determine the columnar to equiaxed transition during different moments and in the same sample.

2. A columnar to equiaxed transition occurs rather during an interval than at a fixed moment.

3. An addition of an inoculant decreases the fraction of solid material in the casting when columnar to equiaxed transition is achieved.

Conclusions of Supplement III:

1. The hydraulic diameter of the inter-dendritic space is proportional to the local solidification time and goes through a morphological coarsening process during solidification.

2. The largest inter-dendritic spaces in the casting with tendency to form shrinkage porosity or metal expansion penetration are found in connection to the metal-mold interfaces where the casting defects are also located.

3. The large inter-dendritic spaces on the metal-mold interface are suggested to be the assistant condition for the liquid mass flow necessary to initiate shrinkage porosity or metal expansion penetration.

25

Future Work Solidification of lamellar cast iron and formation of volume-related casting defects represents very complex phenomena. Summarizing the results of the present work, it can be concluded that detailed additional studies are necessary in this field in order to understand the formation mechanism of casting defects like the shrinkage porosity and the metal expansion penetration. In addition, volume change measurements and solidification kinetics should be investigated in deeper details. A development of a new volume change measurement method, which eliminate or at least minimize the drawbacks of the experimental layouts know today is an interesting future challenge. Also, carefully planned volume change measurements in an improved layout coupled with up-to-date thermal analysis and expansion force measurements may provide further information on the solidification of lamellar cast iron and indirectly on the defect formation mechanism. Interrupted solidification experiments like direct austempering after solidification are expected to bring new information on the macrostructure development of the primary austenite. Also interesting results are expected from the comparison of stereological investigation on defect free castings with defected ones.

26

References [1] D. M. Stefanescu, Science and Engineering of Casting Solidification. Kluwer

Academic / Plenum Publishers, 2002. [2] J. Campbell, Castings, Second Edition. Elsevier, 2003. [3] T. S. Piwonka, “Casting,” in Metals Handbook, 2nd ed., Asm International,

1998, pp. 727–730. [4] G. Tamman, “Über die Ermitelung der Zusammensetzung chemischer

Verbindungen ohne Hilfe der Analyse,” Zeitschrift anorganische und allgemeine Chemie, vol. 37, no. 2, pp. 303–313, 1903.

[5] L. Bäckerud, 350 606.1970. [6] K.-H. Caspers, “Der Unterkühlungsquotient bei der termischen Analyse als

Größe zur Beurteilung von Gußeisenimpfungen,” Gießerei, vol. 61, pp. 611–615, 1974.

[7] E. F. Ryntz, J. F. Jonowak, A. W. Hochstein, and C. A. Wargel, “Prediction of nodular iron microstructure using thermal analysis,” AFS Transactions, vol. 82, pp. 141–144, 1974.

[8] P. Beeley, Foundry Technology, 2nd ed. Oxford: Butterworth-Heinemann, 2001.

[9] L. Elmquist, “Defect formation during solidification in grey iron castings,” Doctoral Thesis, Chalmers University of Technology, Göteborg, 2009.

[10] D. M. Stefanescu, “Cast Iron,” in ASM Handbook, vol. 15: Casting, 1998, pp. 168–176.

[11] I. L. Svensson, M. Wessén, F. Bonollo, and A. Diószegi, “Component Casting with Simulation,” Jönköping University, 2002.

[12] R. W. Heine and C. R. Loper, “Dendrites and Eutectic Cells in Gray Iron,” Modern Casting, vol. 56, no. 4, pp. 185–191, 1969.

[13] G. L. Rivera, R. E. Boeri, and J. A. Sikora, “Solidification of gray cast iron,” Scripta Materialia, vol. 50, no. 3, pp. 331–335, Feb. 2004.

[14] A. Diószegi, K. Z. Liu, and I. L. Svensson, “Inoculation of primary austenite in grey cast iron,” International Journal of Cast Metals Research, vol. 20, no. 2, pp. 68–72, 2007.

[15] W. Kurz and D. J. Fisher, Fundamentals of Solidification, 4th ed. Switzerland: Trans Tech Publications, 1998.

[16] B. Chalmers, Principles of Solidification. John Wiley & Sons, New York, 1964. [17] R. Morando, H. Biloni, G. S. Cole, and G. F. Bulling, “The development of

macrostructure in ingots of increasing size,” Metall and Materi Trans, vol. 1, no. 5, pp. 1407–1412, May 1970.

[18] M. H. Burden and J. D. Hunt, “Mechanism for the Columnar to Equiaxed Transition in Castings or Ingots,” Metall. Trans. A, vol. 6A, no. 1, pp. 240–241, 1975.

[19] J. D. Hunt, “Steady-State Columnar and Equiaxed Growth of Dendrites and Eutectic,” Mater. Sci. Eng., vol. 65, no. 1, pp. 75–83, 1984.

[20] H. Fredriksson and A. Olsson, “Mechanism of transition from columnar to equiaxed zone in ingots,” Materials Science and Technology, vol. 2, no. 5, pp. 508–516, 1986.

[21] S. C. Flood and J. D. Hunt, “Columnar to equiaxed transition,” in ASM Handbook. Vol. 15. Casting, ASM International, 1988, pp. 130–136.

[22] M. Rappaz, C.-A. Gandin, J.-L. Desbiolles, and P. H. Thevoz, “Prediction of grain structures in various solidification processes,” Metallurgical and Materials Transactions A, vol. 27A, pp. 695–705, 1996.

27

[23] C. Y. Wang and C. Beckermann, “Prediction of Columnar to Equiaxed Transition during Diffusion-Controlled Dendritic Alloy Solidification,” MMTA, vol. 25, no. 5, pp. 1081–1093, May 1994.

[24] A. Hellawell, S. Liu, and S. Z. Lu, “Dendrite fragmentation and the effects of fluid flow in castings,” JOM, vol. 49, no. 3, pp. 18–20, Mar. 1997.

[25] H. Jiang, W. Kierkus, and J. H. Sokolowski, “Dendrite coherency point determination using thermal analysis and rheological measurements,” presented at the 1st International Conference on Thermophysical Properties of Materials, Singapore, 1999, pp. 120–125.

[26] W. Kurz and C. Bezençon and M. Gäumann, “Columnar to equiaxed transition in solidification processing,” Science and Technology of Advanced Materials, vol. 2, no. 1, p. 185, 2001.

[27] J. A. Spittle, “Columnar to equiaxed grain transition in as solidified alloys,” International Materials Reviews, vol. 51, no. 4, pp. 247–269, 2006.

[28] J. Tamminen, “Thermal analysis for investigation of solidification mechanism in metals and alloys,” Chemical Communications, University of Stockholm, 1988.

[29] G. Chai, “Dendrite coherency during equiaxed solidification in aluminium alloys,” Chemical Communications, University of Stockholm, 1994.

[30] P. W. Voorhees, “The theory of Ostwald ripening,” Journal of Statistical Physics, vol. 38, no. 1–2, pp. 231–252, Jan. 1985.

[31] D. Kammer and P. W. Voorhees, “The morphological evolution of dendritic microstructures during coarsening,” Acta Materialia, vol. 54, no. 6, pp. 1549–1558, Apr. 2006.

[32] J. Campbell, “Feeding Mechanisms in Castings,” AFS Cast Metals Research Journal, pp. 1–8, Mar. 1969.

[33] A. Diószegi, L. Elmquist, J. Orlenius, and I. Dugic, “Defect Formation at Casting of Gray Iron Components,” International Journal of Metalcasting, vol. 3, no. 4, 2009.

[34] G. Nándori and P. Jónás, “A tágulási erő mint az öntöttvasak és acélok minőségi jellemzője a kristályosodás folyamán,” Bányászati és Kohászati Lapok, Öntöde, vol. 12, pp. 169–173, Dec. 1988.

[35] P. Svidró and A. Diószegi, “On the problems of measuring volume change in Lamellar cast iron,” Submitted to IJCMR, Apr. 2013.

[36] W. J. Keep, “Cooling curves and test for cast iron,” Journal of the Iron and Steel Institute, vol. 2, pp. 227–248, 1895.

[37] T. Turner, “Volume and Temperature Changes during the Cooling of Cast Iron,” Journal of the Iron and Steel Institute, vol. 1, pp. 48–74, 1906.

[38] F. Wüst, “Die Schwindung von Metallen und Metalllegierungen,” Metallurgie, vol. 6, pp. 769–792, 1909.

[39] A. Diefenthäler, “Die Ursachen der Lunkerung und ihr Zusammenhang mit Schwindung und Gattierung,” Stahl und Eisen, vol. 32. II., pp. 1813–1819, Oct. 1912.

[40] F. Wu ̈st and G. Schitzkowski, “Einfluß einiger Fremdkörper auf die Schwindung des Eisens,” Mitteilungen aus dem Kaiser-Wilhelm-Institut für Eisenforschung zu Düsseldorf, vol. 4, pp. 105–124, 1922.

[41] K. Bornemann and F. Sauerwald, “Dichtemessungen von Metallen und Legierungen,” Zeitschrift für Metallkunde, vol. 14, pp. 145–159, Apr. 1922.

[42] F. Sauerwald, “Über die Gesetzmässigkeiten der Volumengestaltung und Schwindung von Metallen und Legierungen,” Giesserei-Zeitung, vol. 20, pp. 391–393, 1923.

28

[43] P. Bardenheuer and C. Ebbefeld, “Beitrag zur Analyse des Schwindungsvorganges von weißem und grauem Gußeisen,” Mitteilungen aus dem Kaiser-Wilhelm-Institut für Eisenforschung zu Düsseldorf, vol. 6, pp. 45–60, 1926.

[44] G. Nándori, “Measurement of the volumetrical and linear shrinkage of gray-iron and testing some shrinkage affecting factors,” Kohászati Lapok, Öntöde, vol. 11, pp. 241–247, Nov. 1960.

[45] D. M. Stefanescu, L. Dinescu, S. Crăciun, and M. Popescu, “Production of vermicular graphite cast-irons by operative control and correction of graphite shape.,” in Proceedings of the 46th International Foundry Congress, CIATF, Madrid, Spain, 1979, pp. 37–37/15.

[46] A. Gazda and E. Zemlak, “Apparatus for Measurement of Dimensional Changes of Solidifying Casting Alloys,” Pr. Inst. Odlew., vol. 33, no. 2–3, pp. 159–173, 1983.

[47] Y. Yang and J. Alhainen, “Functional Improvement of Dilatation Analysis on Thermal Analysis for S.G. Iron Quality Estimation,” in Proceedings of the 5th International Symposium on the Physical Metallurgy of Cast Iron, 1994.

[48] Z. Gedeonova, S. Bodi, J. Dúl, G. Nándori, and L. Vigh, “Displacement on the surface mould and metal during the solidification of nodular graphite iron castings,” Materials Science Forum, vol. Solidification and Gravity II (215–216), pp. 391–398, 1995.

[49] J. Mutwil, “Device for the investigation of casting stresses,” Solidification of Metals and Alloys, vol. 2, no. 44, pp. 503–508, 2000.

[50] P. Mrvar, M. Trbižan, and J. Medved, “Examination of Dimensional Changes of Casting and Mould During the Solidification of Modified and Unmodified Aluminium Cast Alloys,” presented at the Metal 2002, Hradec nad Moravicí, 2002, pp. 1–8.

[51] M. Chisamera, I. Riposan, S. Stan, and M. Barstow, “Simultaneous cooling and contraction / expansion curve analysis during ductile iron solidification,” in Transactions of the American Foundrymen’s Society, vol. 120, 2012, pp. 375–388.

[52] D. M. Stefanescu, M. Moran, and S. Boonmee, “Use of Combined Liquid Displacement and Cooling Curve Analysis in Understanding the Solidification of Cast Iron,” in Transactions of the American Foundrymen’s Society, vol. 120, 2012, pp. 365–374.

[53] P. Bardenheuer and W. Bottenberg, “Einfluß der Schmelzbehandlung auf den Gasgehalt und die Schwindung von weißem und grauem Gußeisen,” Mitteilungen aus dem Kaiser-Wilhelm-Institut für Eisenforschung zu Düsseldorf, vol. 13, pp. 149–159, 1931.

[54] S. Engler, D. Boenisch, and B. Kohler, “Metal and Mold Wall Movements During Solidification of Cast Iron,” Cast Metals Research Journal AFS, vol. 9, pp. 20–30, Mar. 1973.

[55] L. Bäckerud, G. Chai, and J. Tamminen, Solidification characteristics of aluminium alloys, vol. 2: Foundry Alloys. 1990.

[56] E. Fras, W. Kapturkiewicz, and H. F. Lopez, “The solidification kinetics of cast iron using an improved thermal analysis technique,” International Journal of Cast Metals Research, vol. 6, no. 3, pp. 137–142, 1993.

[57] A. Diószegi and I. L. Svensson, “Interpretation of Solidification by Thermal Analysis of Cooling Rate,” Transaction of the Indian Institute of Metals, vol. 58, no. 4, pp. 611 – 616, Aug. 2005.

29

[58] A. Diószegi and V. Fourlakidis, “Validation of different numerical algorithms to interpret registered cooling curves for prediction of volume fraction primary austenite at solidification of lamellar cast iron,” presented at the The 3rd International symposium on Cutting Edge of Computer Simulation of Solidification, Casting and Refining, Finland, Sweden, 2013, p. 84.

[59] L. Elmquist and A. Diószegi, “Shrinkage porosity and its relation to solidification structure of grey cast iron parts,” International Journal of Cast Metals Research, vol. 23, no. 1, pp. 44–50, 2010.

[60] G. L. Rivera, R. E. Boeri, and J. A. Sikora, “Revealing solidification structure of nodular iron,” Cast Metals, vol. 8, no. 1, pp. 1–5, 1995.

[61] R. Lora, A. Diószegi, V. Fourlakidis, G. P. Vilaseca, and A. D. Osona, “Characterization of the Primary Austenite Dendrite and Interdendritic Space in Lamellar Cast Iron,” Jönköping University, Research Report, 2011.