You are on page 1of 33

On the analysis of cast structure and its

changes during hot working of forging ingots
Jan Sarnet
Casting of Metals
Royal Institute of Technology (KTH)
SE-100 44 STOCKHOLM, Sweden

Forging ingots are hot worked to consolidate structural imperfections and refine
the microstructure. Possible imperfections are segregations, porosities, cracks,
and inclusions. In this thesis, techniques and methods for the characterisation of
behaviour and properties of ingots and forgings. The root causes for quality
issues in open-die forgings are shown to be numerous. Ingot structure in cast
tool steel was analyzed, and the main imperfections were centre
macrosegregation, and mid radius A-segregations.
For overheated steel forgings and low reductions after reheating, a high
ultrasonic attenuation and low Charpy-V toughness was found. It could be
related to the coarse grain structure found. Only a high forging reduction after
reheating will break down the coarse structure. Shorter reheating times and
lower forging temperatures gave higher toughness and lower ultrasonic
attenuation. Some reduction in toughness was also found from the inclusion
field from the bottom of the ingot. Accurate attenuation measurements require a
lathe turned surface, complicating in-between-measurements in heat treatment
sequences. But on carefully surface prepared forgings, attenuation measurement
can be used to determine the success of grain refinement in the heat treatment. A
new method for ultrasonic macrography of cast ingot samples is presented. In
addition, a new method for hot compression testing of cylindrical metal samples
is presented.

This page is left intentionally blank.

This Doctor Thesis is based on the following supplements:
Supplement 1
Macrosegreation in Ingot Cast Tool Steel
A. Lagerstedt, J. Sarnet, S. Adolfi, and H. Fredriksson
TRITA-MG 2004:09
Supplement 2
Causes of High Ultrasonic Attenuation and Reduced Mechanical Toughness of a
Forged Rotor Shaft
J. Sarnet and B. Widell
TRITA-MG 2005:04
Supplement 3
Influence of Various Heating and Forging Routes on the Microstructure of
J. Sarnet and B. Widell
15th International Forgemasters Meeting, Kobe, Japan, 2003
Supplement 4
The possibility of evaluation of defects and grain structure by ultrasonic
techniques in large forgings and test pieces of ingots
J. Sarnet
16th International Forgemasters Meeting, Sheffield, United Kingdom, 2006
Supplement 5
Ultrasonic Evaluation of Imperfections and Defects in Cast Steel Ingot Samples
Using Immersion Tank Testing
Jan Sarnet, Ketil Törresvoll, Hasse Fredriksson
TRITA-MG 2009:01
Supplement 6
Hot Deformation Testing of Cast Metal Samples
Jan Sarnet, Hani Nassar, and Hasse Fredriksson
TRITA-MG 2009:02

.This page is left intentionally blank.

............................................................................................................................................................... 12 Grain size and toughness properties ................. 14 Ultrasonic attenuation ............................................................................................................................................................................................................ 10 Ingot structure by sulphur printing .................... 5 Sample preparation ... 16 Flow stress curves ........ 15 Ingot structure ............................ 27 ......................................................................................................... 5 Forging process ............................................. 6 Sulphur printing .. 5 Ingot casting and analysis ........................................................... 1 Methods..................................................... 23 References ..................................................................................................................... 20 Discussion ............. 13 Grain size and ultrasonic attenuation...................................................... 6 Ultrasonic attenuation measurement.................................................................................................................................................................................................. 8 Results .......................... 10 Temperature measurements ......................................... 7 Hot compression testing.............................................................................................................................................................................................................. 21 Conclusions .................................................................................................................................................Introduction ..... 18 Hot worked grain size ................................. 25 Acknowledgements .........................

.This page is left intentionally blank.

and also vary with respect to the zones within the component.Introduction / 1 Introduction The reliability of open die forged products is of the greatest importance. Ai+1.95). This is tested by mechanical and non destructive testing. This criterion is highly dependent on fracture toughness and on the use of the component. The surface zones have the highest demands placed on them. Forging reduction = Ai Ai +1 (1) Forging two or several products from one forging ingot is an old method of increasing the forging reduction for a given product. Many of the defects can be traced upstream through the process. and tool steels. and usually calculated as forging reduction. moreover variation be low. The permissible sizes and numbers of imperfections for a component are given as a criterion in the specifications or in agreements between producers and customers. ball bearings. and be found originating from the cast or forged structure. The dendritic structure of the cast ingot deforms into a banded one as the hot working ratio increases [3]. of the product as given by Equation 1. though most times at the end of the manufacturing. If the product is of insufficient quality. Though there are few papers of cooperation between the mentioned disciplines [2]. a simple ratio between the original transversal area. The idea of solving quality issues at the source in manufacturing processes is well-established since already Pearson contributions [1]. of the ingot or intermediate product. and therefore cast structure is still of importance for the final product and the porosity consolidation is of importance. Banded structures have anisotropic properties and imply high working ratios. It was early acknowledged that transversal variation in cast ingot macrostructure and variable local deformations during hot working influence the properties of products. In large forging products plastic deformation is low.0 mm for heavy structural steel components such as rotors. The size criteria for single imperfections are typically 0. Samples from . meaning that only small imperfections can be tolerated. In addition. the root causes can be numerous. both economical and logistical goals will be endangered. and the final transversal area. Defects are material imperfections that endanger the integrity of a component as crack initiators. an example can be seen in Figure 1 where the forging reduction was close to three (2.3 mm up to 3. there could most certainly be improvements done taking several disciplines into account for each quality issue. Ai. This ratio only of the order of three. The forgings must not only be as strong and also ductile.

when calculated from laboratory data. visual observation of segregation pattern [4]. The strain. ε= A0 A (2) High strain and low forging temperature produces small grain sizes as can be seen in Figure 3 [7]. Sawing in length direction. Figure 1. There are several types and sizes of defects and . which is the locadeformation is sometimes approximated as given by Equation 2 [8]. ε. [6] and german researchers investigated the influence of forging temperature and strain on the grain size already in the 1920s [7]. Sample position relative to the ingot and forging schedule for a trial forging [4]. Figure 2.2 / Introduction such forging practice were visual inspected and revealed strong A-segregations after processing as can be seen in Figure 2 [4]. The mechanical and ultrasonic properties are strongly dependent on the grain size [5].

However. these techniques failed. Influence of temperature and strain on the grain size [7]. [10]. which renders quantitative analysis difficult. such as segregations. the attenuation is also strongly dependent on the frequency of the probe [6]. The size and statistical occurrence of such defects will also depend on the local solidification conditions within the ingot [9]. Historically. and inclusions. Figure 3. as shown in Equation 3 given by Krautkrämer [13]. the surfaces of an ingot are arched and non-parallel.Introduction / 3 imperfections that may exist within cast ingots. both methods have lost ground. and the coarse microstructure it have to penetrate [6]. α. both sulphur printing and deep etching are qualitative methods. and negative-exponentially proportional to the attenuation. . however. The poor ratio resulted from the large distance the signal would have to travel [13]. [11]. today. p. In addition. Attenuation values increase to 10. from the probe. the ingot macrostructure was visualised either by sulphur printing on photographic paper or by deep etching. quantifying the size and the number of defects and their distribution in respect to the ingot macrostructure are important for public safety. Ultrasonic characterisation of ingot structures would enable quantification of defects. d.100 dB/m as this ratio approaches unity or above [13]. which further reduces the amount of the signal returning to the probe [6]. p ∝ p0 ⋅ 1 −α ⋅d ⋅e d (3) The influence of attenuation on signal amplitude increases for longer sound paths. In addition. The amplitude. as can be seen in Figure 1. due to poor signal-to-noise ratios. Ultrasonic techniques using normal probes for defect characterisation in ingots were evaluated in the 1950s [12]. of the ultrasonic signal for a diverging beam is inversely proportional to the distance. Therefore. However. porosities.

For example. some focus was given to the number of austenitic transformations [17]. Detailed investigations of and comparisons between flat bottom hole. mechanical and ultrasonic properties related to industrial forging and heat treatment processes [25]. Chromium and Molybdenum content. structure analysis [22]. Work on the forging process has very much focused on the tool/workpiece geometry [21]. A method for ultrasonic macrography was outlined in a previous thesis [4]. though this was probably not realisable then [15].4 / Introduction Cutting sections of ingots to reduce the sound path would lead to a more favourable signal-to-noise ratio. The relative attenuation of the ingot ranged from 30. the development of the large turbogenerators for the nuclear industry in 1970s focused on the Nickel. and the cooling rate during quenching. [23]. [26].150 dB/m for the 110 mm thick samples [14]. . porosities. In addition. Various remelting processes have been developed and refined for the ingots [20]. and it showed a reduction in centre porosities for the big end up-ingot. [24]. This was tested by Japanese researchers in the 1970s who measured the absolute attenuation as an indication of ingot porosities. and inclusions [16] are scarce. Much research of open die forgings has been done regarding alloying content and heat treatment. but has been greatly improved and extended in this thesis. Interest has also been turned to reducing the A-segregations by ingot design and alloying content [19]. Furthermore there has been much focus of the cleanliness of the steel. The objective of their investigation was to compare big end downwith big end up-ingots. Canella argued that the raw data should be immediately stored and statistically analysed. This thesis is composed of six papers and treats a broad spectrum of issues related to structure formation [22]. and hot deformation of cast structure [27]. meaning both lower inclusion content and a lower content of harmful residual elements [18]. The developments were incorporated in a computerised analysis program.

4-0.12 [30] [21]. [24].2-0.93 0.20 n. 0. [25] Rotors 26NiCrMoV145 Forged 0. Cast 0. tool Diam.45 [30] [26] Tools 40CrMnNiMo8-6-4 Cast 0.18 0. Table 2.6:1 9.29 0.21 Forging reduction 7.0:15.36 0. Table 1. Reheating Reheat. [29].0:1 Forging Forging reduction Rel.a.4 840 2.56 [29] [23].Methods / 5 Methods Ingot casting and analysis The ingots and forgings for this thesis [22].6:1 n. [26] were cast by uphill casting at three Swedish steel plants. The alloy content of the forging and samples is shown in Table 1.2:1 9. [30].008 0. [23].370. [25].006 0.43 [30] [24]. temp time (ºC) (h) Other [18] 1250 18 Old practice 1250 2-5 Rejected shafts 1260 6 New practice 1 1225 <2 New practice 2 1225 <2 Plant trials 12001250 4 .2-0.5 715 2. 1000 4. and compared with a literature reference.n. a. [24] Rotors 26NiCrMoV145 Forged 0.51 0. [27] Forging process The measurement presented in this thesis have been performed on various industrial forgings. Reheat.46 0.27 0.055 0.19 [28] [23] Bearings 100Cr6 (modified) Cast 0.8:1 0. and the steels were refined by ASEA-SKF ladle processes.03. Use Steel Grade Structure %C %Si %S %Mo Steelmaking overview Thesis supplement Tools n.2:1 3. The ingots solidified in cast iron moulds that had a relative weight 1:1 compared with the ingot.6:1 0.36 0.a.0:1 0. Thermomechanical details of different forging sequences.0:1 0.4 840 1.4-0. 0.0 159300 .a. In Table 2 some process parameters are displayed. last heat width (mm) 4.5-1. These ladle processes are described elsewhere [28]. Alloy content of the test samples from the two ingots.0:1 0.5:1 3.5 1005 2.26 0.

The vertical cross sections were printed and for a octagonal ingot a horizontal section was also printed. and a bearing steel ingot ingots. The thickness of the test samples was chosen to be 60 mm in accordance with standard industrial practice [31]. The locations of the test positions for the test specimen for Charpy-V testing from the rejected rotor shaft [25]. and the diverging angle between the upper and lower surface of each sample was between 0. In order to obtain as detailed macrostructure as possible some surfaces was milled since the readily sawed surface proved too rough to give good sharpness. Sulphur printing In supplement 1.05° to 0.6 / Methods Sample preparation The test specimens for Charpy-V mechanical toughness testing in supplement 2 [25] were taken as shown in Figure 4. . All test samples were machined to plane parallel surfaces.5°. The samples with sides that face toward the ingot surface are easily identified. longitudinal and transversal test samples were cut from ingots. Photo paper was soaked in 5% sulphuric acid and thereafter rolled on to the metal surface with a roller to ensure good contact. the sawed surfaces of two tool steel ingot were sulphur printed in order to obtain macrographs of the ingot structure. These samples represent half the width and the full height of a tool steel. and oriented in axial direction of the shaft. In addition. Figure 4.

on the other hand. VS . Figure 5. is the amplitude loss of through the divergence. 4 und 5 MHz sampling frequency.Methods / 7 Ultrasonic attenuation measurement The relative ultrasonic attenuation was measured in supplements 2 . where V1-V2 corresponds to the difference of the logarithmic amplitude of the echo’s and s is the covered signal path. Amplitude measurement of the first and second bottom echo in the far field for the attenuation calculation. α= 1 (V1 − V2 − ∆V s ) 2s dB/m (4) . in which the imperfection will have the largest extension. In order to calculate the accuracy of the measurements. Since imperfections present are drawn out in the axial direction during forging processing. The measuring equipment was a Krautkrämer USLT2000 and normal straight probes as in Figure 5 B1SE to B5SE with 1. most testing is performed in the perpendicular direction. The relative attenuation was calculated from Equation 4.4. The variation was less than ±1 dB. This loss is 6 dB in the far field. but in random order at the same position. the attenuation was measured several times. 2.

The press is a hydraulic 810 MTS® testing system. Figure 6. The temperature was also measured at 4 locations in the lower and upper tools using K-type thermocouples. . The tools are made of Nimonic 75 (80Ni/20Cr). and S-type thermocouples in high temperature tests (>1000°C). The samples’ temperatures were measured using K-type thermocouples in low temperature tests (<1000°C). The samples were heated by concentrated light from three 700W lamps placed inside highly-polished parabolic reflectors as can be seen in Figure 6.8 / Methods Hot compression testing Compression experiments were conducted on pure copper and bearing steel. The electrical power supplied to the lamps is controlled by a Eurotherm® control system. Teststar® software was used to control the hydraulic motion of the press. Amplitude measurement of the first and second bottom echo in the far field for the attenuation calculation. The portions of the thermocouples placed inside the samples were shielded inside alumina tubes.

and temperature is Jonas and Sellars’ modification of Garofalo’s equation stated in Equation 6 [32]. and the change in height. strain rate. α and n are material constants. h and h0 are the instantaneous and original heights of the sample and ∆h=h-h0 is the change in height. A and A0 are the instantaneous and original areas of the cylindrical sample. A common model that has been used to express the relation between peak flow stress.Methods / 9 The flow stress for a uni-axial state was calculated using Equation 5 as a function of the original area. (6) . ε& = A sinh(ασ ) n exp(−Q / RT ) where A. original height. F is the force applied by the tools. σ= F F ⋅h F  ∆h   = = 1 − A A0 ⋅ h0 A0  h0  (5) where σ is the stress.

and the larger ingot in the present research might exert larger forces on the thermocouples. calculated 600 Mould outside surface.7 tonnes. calculated 50 100 150 200 250 300 Time [minutes] 350 400 450 500 Figure 7. Thermocouple measurement of the cooling and solidifying of the 12 t octagonal ingot .10 / Results Results Temperature measurements The temperature measurement in the hot top of the ingot was not successful. 1600 1400 Melt Temperature [°C] 1200 1000 800 Mould outside surface. Those ingots had a smaller weight of 1. even though similar equipment was used as in an earlier investigation [33]. measured 200 0 0 Hot top outside surface. measured 400 Hot top outside surface.

where the first forging heat takes 15 minutes. forging trials of 300 mm steel bars were performed. Curve 1 and 2 are normalizing treatments [34]. Thermocouple measurement of the cooling and solidifying of the 12 t octagonal ingot In supplement 3 [26] miniature trial forgings were according to the cooling conditions in large rotor forgings. to investigate the influence of forging and heat treatment on the properties of rotor forgings. Curve 3 is a quenching [17]. Temperature (°C) 1000 Experiments Theoretical 800 600 400 1 2 200 3 0 0. In the second and last heat.1 1 10 Time (h) Figure 9. and is followed by a short reheating and tool changes. [17] and these can be seen in Figure 9 as dashed lines. Heat treatment diagram. surface temperature was measured in an industrial forging process and is presented in Figure 8. For comparisons. 1200 1225°C Ø715mm Ø 715mm Ø 480mm 900 800 700 600 Too l ch ang e 1000 Forging of grip end Square 1050mm Temperature (°C) 1100 Side Centre Corner 500 00:00 00:30 01:00 01:30 02:00 Time (h) Figure 8. Solid curves are measured temperatures of the tests and dashed curves are theoretical ones. . The cooling conditions were taken from the literature [34]. surface temperature falls off within 10 minutes.Results / 11 In supplement 3.

Sulphur print of the 12 t tool steel ingot. The cut sample is taken from the upper part of the ingot. . surface to centre. only the hot top was discarded. Supplement 1 describes states that the chemical composition is even in that zone.12 / Results Ingot structure by sulphur printing In Figure 10 the columnar crystal zone of the tools steel ingot is seen. extending inwards 150 mm from the surface. Centre porosity A-segregations Ingot surface Figure 10.

. a strong correlation between the grain size and the toughness is seen. The spread of grain size represents maximum and minimum of the measured values. the toughness rises [5]. 10 grains were measured at each position. Grain size distribution of a 756 mm rotor forging in 3. When comparing that curve to the toughness. When the grain size decreases.Results / 13 Grain size and toughness properties The results of the grain size measurements of the discarded rotor in supplement 2 [25] are displayed in Figure 11 as a function of the transverse position from centre and out to the surface. Figure 11.5%Ni-steel in quenched and tempered condition.

. The attenuation increases for the identical testing position for increasing frequency. Frequency dependence of the attenuation of a fine-grained rotor with known grain size from a centre bore.14 / Results Grain size and ultrasonic attenuation Figure 12 shows the attenuation for the ultrasonic signal in a fine grain rotor at the largest diameter. Figure 12.

. The frequency dependence of the attenuation. a reduced reheating temperature can decrease the ultrasonic attenuation. As can be seen in the figure. Figure 14. In Figure 14 it is seen that the normalizing heat treatment lowers the attenuation for the rotor. Ultrasonic attenuation of three test piece step forged to different forging reductions. especially at high frequencies. Figure 13. no heat treatment were performed at this stage. higher forging reduction continuously reduces the attenuation.Results / 15 Ultrasonic attenuation Figure 13 shows the influence of reheating temperature and forging reduction on the attenuation at 2 MHz in as-forged state. In addition.

the structure changes abruptly to a coarse equiaxed dendritic structure as was described in an earlier paper[10]. a step process was used to analyse the resolution for 1:1 printing size.25 . The expected coarsening of the columnar dendrite crystals is clearly visible in this tool steel sample. in the zone identified as A. At a distance of approximately 130 mm from the ingot surface. In area D. The increased visual information can be seen in Figure 15 for a tool steel ingot sample. Ultrasonic C-Scan of a tool steel sample. The indications just to the left of this area were not anticipated.16 / Results Ingot structure The technique of converting amplitude values to a continuous greyscale should lead to a smoother and more detailed image of the indications within sample. typical A-segregations are clearly indicated. A fine equiaxed macrostructure is seen close to the cooling surface. Further away from the cooling surface of the ingot in the zone identified as B. When generating the image. It was confirmed that the indications actually originate from the surface roughness of the original ingot. the columnar dendrite crystals are revealed. and the sample was carefully measured and compared to a real size printing of the C-scan image. The stepping interval used in this investigation was 0. The qualitative impression of Figure 15 is close to the well-known ingot macrostructure [9]-[11] reported in the literature. Comparison of image and steel sample is easy at 1:1 as can be seen in Figure 16. Figure 15. characteristically pointing upward due to the downward convection of the liquid passing the growing dendrite tip.

Figure 16. yielding an image resolution of 4 pixels/mm or 100 pixels/inch. The absolute printing resolution will be given by the width of the beam’s focal point. at a scale of 1:15. Testing with 5 MHz would produce approximately 20 dpi and would print with excellent sharpness.Results / 17 mm/step. the metric gives some idea of the approximate printing resolution. Although pixels/inch is not directly related to dpi (dots/inch). Comparison of steel sample and 1:1 printing . .

the strain rate exponent could decrease. since the temperature sometimes decreased during compression. The strain rate exponent. In his original data. . particularly in low strain rate experiments.18 / Results Flow stress curves The actual instantaneous centre sample temperature was used as criteria for selecting comparable results. but still predicts a continuous increase in m above these temperatures. Boulger used no temperatures higher than 80% of the melting point. which can be seen as a dashed line in Figure 17. based on unpublished result from Boulger. Figure 17. based on results from other materials. at about 85% of the melting temperature. Strain rate exponent as a function of the relative temperature. The present results actually show a stronger increase of m with relative temperature than in that reference. The results are compared with a part wise linear fitting proposed by Hosford[35]. the results in the present work indicate that at higher temperatures. m is expected to increase as the relative temperature T/Tm increases. In addition. T/Tm.

Peak flow stress for different hot working parameters for the bearing steel.006.6s-1) as can be seen in the values for the bearing steel in Figure 18.Results / 19 Activation energy.06 and 0. Figure 18. . is calculated from the slope of the linear fitting of the temperature reciprocal plotted against the transformed peak flow stress at three different strain rates (0. 0. Q.

Correlation between observed and predicted grain size. The predicted dependence of grain size on strain and temperature is shown in the contour diagram in Figure 20. Figure 19.20 / Results Hot worked grain size The correlation between the model and the measured grain size can be seen in Figure 19. the model fits the central values the best. Contour diagram of grain size model as a function of strain and temperature. . As awaited. Figure 20.

The attenuation depends on absorption and scattering processes. but the amount seem dependent on the forging and reheating sequence. largely independent of the ingot position. where relationships between geometrical forging parameters and local strain was given. This can lower the demands of forging reduction. Relation between the inverted square root of the grain size of various positions in the rotor shaft and the Charpy-V toughness. For the larger grain sizes. the toughness does not seem to be linear dependent with the inverted square root usually stated [5]. . strain rate and temperature in a large number of recent papers. Previous research has also shown the forging tool width to the workpiece height. and therefore the peak flow stress is determined by the momentary process parameters. but these can be closed by appropriate forging reduction. The grain size in overheated steel forgings has a strong gradient and coarse grains abruptly appear close to the surface. The ultrasonic attenuation measurement decreases during each heat treatment step. It is concluded that the toughness is determined by the grain size. The peak flow stresses were measured and the constitutive constants were evaluated for the relation between the stress. However. In the present work it was shown that the flow stress. greatly influences the properties [36]. most of the authors avoided an evaluation of the associated peak strain. Figure 21.Discussion / 21 Discussion A-segregations in ingots give defects detectable by ultrasonic testing. Figure 21 shows the toughness versus grain size for a single rotor shaft from supplement 2 [25]. A-segregations can be reduced by choosing a reduced Si-content and thereby reducing defects in ingots.

1 to ε=0. .22 / Discussion To evaluate the effect of cooling of the sample during pressing on flow stress.5 compared with a scatter of values obtained from different experiments at strain ε=0. a flow stress curve from supplement 6 was compared with literature data [37]. Flow stress as a function of temperature for steel sample EA3 deformed from ε=0.1 as seen in Figure 22. Figure 22. where experiments were conducted at different temperatures and the flow stress value was obtained at strain value ε=0.1 [33].

The deformation process can be interrupted rapidly and the sample are quenched into water. and the levels are low in the outer and bottom parts of the ingots. These can be visualised by greyscale images from C-scan data files. A new experimental method has been developed to study plastic deformation of metals at high temperatures. It has been shown in supplements 2-4 that for industrial processes forging reduction and forging temperature have a decisive influence on the mechanical properties and the ultrasonic attenuation[25]-[27]. This influence is measurable after forging [26].Conclusions / 23 Conclusions The present thesis has shown that immersion tank testing is suitable for characterising cast steel macrostructures such as A-segregations and macroporosity. . The samples are evaluated by well-known theoretical models. In addition. provided a low frequency probe is used and a sufficient volume is tested. [25]. [26]. The mechanical properties of a scrapped rotor forging deteriorated in axial direction towards the centre. [24]. but with slightly lower resolution due to the finite dimension of the focused ultrasonic beam. The method is capable to distinguish between the cooling of the sample by the tools and temperature changes due to deformation and microstructural changes in the sample. and the ultrasonic attenuation only decrease to reasonable levels after a high degree of reduction. but also indirect influencing the success of the heat treatment. These images can give visual impressions of the cast structure that are comparable to sulphur prints or deep-etching. quantitative analysis of the distribution of 200 to 800 micrometer imperfections in the ingot is possible. A technique allowing for the elastic deformation before the plastic deformation has been used and it effects the measured peak strains. The imperfections are unevenly distributed in radial and axial direction of the ingot. The temperature inside the samples is measured during the deformation process.

24 / Conclusions This page is left intentionally blank. .

28. in Physical Metallurgy Ed. 13 (1986) No. p. Mitchell et al. [21] E. pp. R. [10] H. pp. Calmon. Heil and D. Sakata. 305. M. Phragmen: Transactions of the American Institute of Mining and Metallurgical Engineers. Elsevier. 32 (1968). Amsterdam.H. Sheffield. 1117. Nagaoka and Y.C. December. No. Piehl: Stahl und Eisen. Hanemann and F. [12] A. Hultgren and G. Sarnet: Licentiate Thesis. Cahn and P Haasen. Iwata: Transactions Japan Institute of Metals. Review of Quantitative Nondestructive Evaluation. P. 1. Kopp: Stahl und Eisen. [14] T. and J. 3. 22. UK.I. American Institute of Physics. No. and R. Suzuki. 105. 1985. [5] W. 24 (1984). Ito. [6] D. 2 (1935). Darmon. Meyer. No.S. 322. [19] H. No. 62 (1976). 13 (1980). Springer-Verlag.H. Siemer. 101. [11] A. No. Ultrasonic Testing of Materials. [13] J. 1990. Stahl und Eisen. No. 75 (1989). Fredriksson: ISRN KTH-MG-INR-09:01SE TRITA-MG 2009:01 (Supplement No. 1606. 8. (1919). No. 2. Pedicelli: NDT International. Stockholm. J. 285. 106 (1986). [3] G.C Tippett. [17] K. 45 (1925). 383.References / 25 References [1] E. [18] R. 2. Kraukrämer. 1668. [16] M. Tetsu-toHagane.W. 15-16. No. [2] H. 133. K. Nieschwitz. 2005. W. Larén: Scandinavian Journal of Metallurgy. 837. 59. Berlin. T. Charpy: Iron Age. Y. Aurich and E. [9] A. 103 (1983). Trans. submitted for publication in Steel Research International [24] J. 135 (1939). Adolfi and H. Jaffe. [15] G. 26. Krautkrämer and H. and H. [8] G. April. London. 4) . 13. 2006 (Supplement No. P. 2003. Fredriksson and I. Kawawa. [4] J. No. 1079. [7] H. Bèle: Modelling of the ultrasonic response of inclusions in steels. 4th edition. Yamada et al. Martin: Archiv für das Eisenhüttenwesen.P. Canella and L. Sarnet. 5). Lohr: Stahl und Eisen. CP657. Jpn. 1) [23] J. [22] A. 547. 1996. Discussion on a paper by L. Y. and B. Sarnet. 74 (1954). Törresvoll. 1301. Suzuki. Fredriksson: ISRN KTH-MGINR-04:09SE TRITA-MG 2004:09 (Supplement No. 95 (1975). 27. No. 108-116. S.E. Tetsu-to-Hagane. P. Michalski. 733. 6. Pearson: Supplement to the Journal of the Royal Statistical Society. 1. 41 (1970). Lucke: Stahl und Eisen. [20] A. Leslie: Physical Metallurgy of Steels. Lagerstedt. Miyoshi. Iron Steel Inst. 1 (1972). Machner. 1. N. Steiner: Ironmaking and Steelmaking. The Institute of Metals.J Richardson: Worked Examples in Metalworking. Sarnet: Proceedings of the 16th International Forgemasters Meeting. Hosoda. and S.

Adapted from F. 30 (2003). Fredriksson: ISRN KTH-MG-INR-09:02SE TRITA-MG 2009:02 (Supplement No. S. . New York.W. Battelle Me. [37] R. p. DMIC Report 226. H. Japan. Widell: Proceedings of the 15th International Forgemasters Meeting. [29] T. Kobe. R. 201. 15 (1986). No. Verlag der Augustinus Buchhandlung. Cheng. H. 1977. R. Atkins: “Atlas of continuous cooling transformation diagrams for engineering steels”. 54. [33] A. pp. H. Sarnet and B. Jung. 2003 (Supplement No. 1998. Sheffield. Olsson. [28] J. 6). Lund and K. Fredriksson: Scandinavian Journal of Metallurgy. Inst (1966). Tegart: Metall. Caddell: Metal forming mechanics and metallurgy.J. [31] T. Nassar. Jönsson: Ironmaking and Steelmaking. C. control and properties of bearing steels intended for demanding applications” ASTM Special Technical Publication. 66.M. [34] M. 104. Sjöqvist. Bearing Steels: Into the 21th Century. Wiegels: Einführung in die Umformtechnik. 32.M. Boulger. 1. ASTM STP 1327. No. 1998. 2007. 3. No. Törresvoll: Quantification of Large Inclusions in Bearing Steels. Rev. 2) [26] J. 373. STP 1361 (1999). 27 (2000). Jonsson. W. [30] T. No.26 / References [25] J. P. [36] H. Andreasson: Ironmaking and Steelmaking (UK). p.P. Ölund: “Improving production. 46 (1975). R. Sarnet. Sellars. Sarnet and B. 14 (1969) 1. 3) [27] J. Jonas. [32] J. Widell: ISRN KTH-MG-INR-05:04SE TRITA-MG 2005:04 (Supplement No.F. West. 13-37. A. Hosford.McG. Kopp.J. H. 5. Schütz: Archiv für das Eisenhüttenwesen. submitted for publication in Materials Science and Technology. M. Eriksson and P. 52-55. Heil. pp. Lund and P. 2. [35] W. Cambridge University Press.

Early on. Göran Wahlstedt. Joachim Furuberg. Dr. Dr. Claudio Pecorari. Mattias Skrinning. Paul Nieschwitz. I truly appreciate the support and practical help from my parents Tönu and Eva. Alf Bohlin. Roger Jävergård. Erik Wedin. Bruksägare Anders Henrik Göranssons Fond. but important critical appreciation on my work in connection with meetings and conferences from Dr. The finishing of this work was supported by the Swedish Steel Producers’ Association (Jernkontoret) and Lars-Henrik Österholm therefore deserves special mentioning. Ketil Törresvoll. Hans Thörnsten. but still having a broader view of what to achieve. and my father Tönu. Egon Erlenkamp. Udo Lehmann. Peter Lidegran. Christoph Heischeid. Yomei Yoshioka. Hani Nassar. my grandfather Hadar Björklund. Astrid Granberg. Additional grants was received from Sven and Astrid Toresson’s Foundation (Stiftelsen för Sven och Astrid Toressons Fond). my grandmothers Linda Sarnet and Kerstin Björklund. and Professor Pavel Huml. Magnus Brännbacka. and Stiftelsen Jernkontorsfonden för Bergsvetenskaplig Forskning. Joachim Furuberg. Robert Gustafsson. Johannes Federer. Rickard Gustafsson. inspiration and help from many people. Anders Lagerstedt. Magnus Brännbacka and Kristin GranathBrännbacka. Dan Thorsén. Dr. I received supervision from Björn Widell. Dr. Harmut Vogel. and Thore Lund at SKF. In addition. Arno Frank. my brothers Johan and Joakim. especially Christer and Katarina Furuberg. but also from many others at various stages of the work. Verena Schneider. and Dr. Hiroyuki Nagasako. and Axel A:son Johnson’s Foundation (Stiftelsen Axel A:son Johnsons Forskningsfond). I had the chance of interchanging information and reflections about issues in this field with Kent Öberg. Patrik Olsson. Heinz-Jürgen Peters. Dr. and has had to use his full creativity to finance this work. My supervisor Professor Hasse Fredriksson supported me already when I was an undergraduate. Srecko Milenkovski. Peter Björklund. and I was encouraged and truly inspired by my uncle Hans Björklund. Robert Gustafsson. I was interested in science and engineering. . Håkan Lundbäck. Rickard Gustafsson. Mattias and Anna Skrinning. I also acknowledge brief. Gunilla Olsen. the managers and personnel at Scana Steel Björneborg were very supportive of this research. Lena Magnusson.Acknowledgements / 27 Acknowledgements During my research education. Fritz Oesterhelt. Lars Holmström. Jan Terhaar. Ove Carlsson. especially Mikael Andreasson. Anders Eliasson. Additional support was arranged by Ewa Persson at Uddeholm Tooling. and Professor Reiner Kopp. Ralf Luce. Gustaf Janssons Jernkontorsfond. Anita Mattson. Geir Grasmo. I have received guidance. Jonas Åberg. Professor Toshishiko Emi. In addition to the supervision from professor Hasse Fredriksson.