2021 年 61 巻 5 号 p. 1556-1566
Important large components, such as rotors for power generation steam turbines, pressure vessels and reaction vessels, are manufactured through ingot casting. However, it is quite tasking to manufacture the materials with sufficient properties because of the macro-segregation of ingots.
To develop effective and versatile macro-segregation countermeasures in the casting of large steel ingots for manufacturing large parts for power plants, the insert casting in vacuum atmosphere, in which a core material similar in composition as the base steel, is placed at the center of the mold, was studied. The effectiveness of the proposed insert casting as a macro-segregation countermeasure was evaluated in the casting experiments with 0.5 mass% carbon steel using cast iron mold with a 150 mm square inner cross section, insulated to reduce solidification rate. In addition, it is now confirmed that good bonding between the core material and base material can be achieved even under conditions where bonding by normal insert casting in air atmosphere is hard to achieve.
The behavior of the core material melting and the solidification of the molten steel in the experiments of macro-segregation reproduction casting and insert casting were investigated using the direct finite difference method. The mechanism by which this method suppresses the macro-segregation formation and solidification conditions for the suppression, the reasons, and conditions for good bonding in this insert casting are clarified by the analyses. Furthermore, the cause of internal crack formation in the insert casting was investigated and guidelines for preventing the crack formation were presented.
Although important large-scale components such as rotors for steam turbines of power generation, pressure vessels, and reaction vessels are manufactured using the ingot casting, macro-segregation of steel ingots harms the soundness and uniformity of the internal quality, deteriorating the mechanical properties. Since it causes temper brittleness, embrittlement due to neutron irradiation, embrittlement of welded parts, etc.,1,2) and also leads to a decrease in yield and an increase in manufacturing cost, drastic measures are required.
In power plants, the increase in power generation capacity and steam temperature are pursued from the improvement of power generation efficiency with regards to economic improvement and environmental problems such as CO2 reduction.1,3) For the purpose, it is necessary to increase the size of parts and improve the operating temperature. As a result, the steel ingots utilized in manufacturing components of power generation are increasing in size,1,2) and macro-segregation level is worsening, and the allowable segregation level is getting higher.1,2,4) Since the solidification rate decreases with the increase in ingot sizes, the formation of macro-segregation is promoted, and the more prominent measures for macro-segregation are required. Currently, the ingot casting has not developed an effective and versatile countermeasure for macro-segregation, and thus it is eager to create a drastic segregation countermeasure.
Macro-segregation of steel ingot includes V segregation,1,2,5,7,8,9,12,13) A segregation (including freckle),1,2,5,6,7,10,11,12,13,14,15,16) negative segregation at the ingot bottom1,2,5,6,10,15,16,17) and concentrated segregation at the ingot top (hereinafter referred to as upper concentrated segregation)1,2,7,8,10,15,16,17) and positive segregation at the center of cross section (hereinafter referred to as center segregation).6,7,10,18,19,20)
Negative segregation at the bottom of steel ingot is caused by sedimentation and pile up of equiaxed crystals.2,5,10,16,17) The formation of a part of V segregation and A segregation is as a result of the floating or sedimentation of molten steel between dendrites due to the density change caused by solute concentration.1,2,5,6,7,10,11,12,13,14,15,16) The formation of upper concentrated segregation results from thermal solute convection.1,2,5,6,7,10,12,13,14,15,16,17,19,20) V segregation and center segregation in the axial center of ingot are caused by solidification shrinkage2,5,6,7,9,16,18,19) or bridging.18,19) Furthermore, the flow of residual molten steel promotes solute partitioning between solid and liquid, which is one major cause of this macro-segregation, and the effect of flow is promoted as the solidification rate decreases.6,9)
Countermeasures were investigated to prevent macro-segregation caused by these mechanisms, and as a result, a method of effectively suppressing the formation of macro-segregation was developed. This method involves insert casting in a vacuum atmosphere, in which a core material with the same composition as the base material was placed in the center of the mold and cast-wrapped.
In this method, the generation of center segregation and V segregation is reduced by suppressing the flow and negative pressure due to bridging and solidification shrinkage in the steel ingot axial part.7,8,9) It is also possible to suppress the generation of negative segregation at the bottom of the ingot through curbing the sedimentation and pile up of equiaxed crystals. Further, in this method, the flow of the residual molten steel can be suppressed by reducing the solid-liquid coexisting region and dispersing the solidification shrinkage amount, and the solidification rate is enhanced by the heat absorption to the core material. These effects can prevent the formation of V and A segregation as a result of molten steel floating or sedimentation. Further, it is possible to suppress the reduction of the effective distribution coefficient due to thermal solute convection and the formation of upper concentrated segregation resulting from the accumulation of residual molten steel.
In this study, a casting experimental method using small ingots that enables the reproduction of macro-segregation such as A segregation and upper concentrated segregation that tends to form in large ingots, were developed. The above-mentioned effect of suppressing the macro-segregation by insert casting was confirmed through this method. In this study, the influence of the size of the core material on the improvement effect of macro-segregation and the influence of the bonding state between the core material and the base material were also investigated.
In addition, in order to confirm the macro-segregation improvement mechanism of this method, the behavior of the core material melting and the solidification of the poured molten steel was studied by melting/solidification analyses. Furthermore, through these analyses, the thermal hysteresis of the interface was estimated, and the thermal conditions under which the core material and the base material were bonded were examined.
Insert casting is utilized in compounding so as to enhance the functionality and performance of metal materials, and the conditions to achieve good bonding has been studied, insert casting for the improving macro-segregation however, has not been studied. In addition, there has been no study on insert casting in a vacuum atmosphere and bonding conditions for this casting, including insert casting to bond the same steel type and to bond different metals and different steel types.
Figure 1 shows a schematic diagram of the mold used in the macro-segregation reproduction experiment (hereinafter referred to as SRPE) using a 150 kg vacuum melting furnace and the macro-segregation prevention experiment through insert casting. In these experiments, the composition and temperature of molten steel were adjusted using a 150 kg vacuum melting furnace, and cast in a cast iron mold. In casting using this mold, so as to effectively reduce the solidification rate and reproduce the macro-segregation, a cast iron mold for a 150 kg steel ingot is processed, and a heat insulating material (Denka Alsenboard) is formed on the mold inner surface with a thickness of 30 mm. Using this mold, S50C (0.5 mass% C, JIS) molten steel where macro-segregation is likely to occur is top-poured and cast in a vacuum atmosphere. The state of macro-segregation formation was observed on the longitudinal cross section of the resulting ingot.
Schematic view of experimental methods of ingot casting for macro-segregation recreation and vacuum insert casting as countermeasure for macro-segregation by using 150 kg vacuum melting furnace. (Online version in color.)
The cross-sectional size of the ingot is 150 mm × 150 mm and the height is 470 mm, and the height under the riser is secured for more reliable reproduction of the macro-segregation while preventing the extreme formation of shrinkage holes. The size of the riser was limited to a cross-sectional size of 260 mm × 260 mm and a height of 150 mm.
In a SRPE (No. 1 experiment), on using this mold to cast S50C molten steel, it was confirmed that remarkable macro-segregation such as clear A segregation and upper concentrated segregation can be reproduced. Later, using the same mold, some experiments were carried out in which molten S50C steel was top-poured and cast in a vacuum atmosphere around the S50C square bar placed in the center of the mold. The effect of preventing macro-segregation formation by the vacuum insert casting was examined. In these experiments, a square bar having a cross-sectional size of 60 mm square (No. 2 and No. 4 experiments) or 35 mm square (No. 3 experiment) cut out from a commercially available S50C square steel was utilized as a core. The core length was limited to 400 mm in such a way that the molten steel pouring throat would not become narrow. Round steel is also an option for the core material, but this time, square steel was selected with priority given to the reduction of the solidification area in the corners of the mold. Table 1 shows the compositions of molten steel and core material. Hereinafter, the ingot No. will be referred to as the experiment No. No. 1 experiment was a segregation reproduction experiment, and the cross-sectional size of the core material used in each insert casting experiment was 60 mm square in No. 2 and No. 4 experiments, and 35 mm square in No. 3 experiment.
| Exper. No. | C | Si | Mn | P | S | TLL (K) | TSL (K) | Tap Temp. (K) |
|---|---|---|---|---|---|---|---|---|
| 1 | 0.495 | 0.245 | 0.744 | 0.018 | 0.021 | 1756 | 1681 | 1857 |
| 2 | 0.494 | 0.249 | 0.745 | 0.020 | 0.017 | 1756 | 1682 | 1893 |
| 3 | 0.492 | 0.245 | 0.745 | 0.020 | 0.02 | 1757 | 1681 | 1903 |
| 4 | 0.495 | 0.248 | 0.746 | 0.021 | 0.022 | 1756 | 1679 | 1903 |
| S50C Core | 0.501 | 0.248 | 0.745 | 0.022 | 0.025 | 1756 | 1686 | − |
In experiments other than No. 4, the R thermocouple was fixed at the position inside the mold shown in Fig. 1 to indicate temperature change in the steel ingot during solidification. The apparent heat transfer coefficients between the molten steel or ingot side surface and the atmosphere where the temperature transition of the solidification analysis coincides with the observed result were estimated and used for melting/solidification analyses.
In experiments No. 1 to No. 3, the etch prints (EP)26) of the longitudinal cross section of the cast ingot were sampled, and the state of macro-segregation and the solidification shell development were investigated by the prints. Some samples of 35 mml × 35 mmw × 10 mmt were cut out from the position of the upper concentrated segregation (E-1) and A segregation (E-2) after longitudinal section EP of No. 1 experiment. In the longitudinal section EP of No. 2 ingot, samples (35 mml × 35 mmw × 10 mmt) cut out from the intermediate position between the core material and the mold inner surface (E-3), since A segregation and upper concentrated segregation were not observed in the EP. Using these samples, the concentration distribution of C, Mn and P was investigated by EPMA elements mapping in the range of 20 mml × 20 mmw at the sample center. In the investigation, the characteristic X-ray intensity was converted into each solute concentration by the calibration curve method.
For ingots of insert casting, some micro samples were taken in order to include the interface between the core material and the solidified layer (S-1, S-2), and the surface etched by nital was observed using a scanning electron microscope (SEM) at a magnification of 20 to 2000 times to examine if there were voids at interface and to investigate the state of interface bonding.
Furthermore, the No. 4 ingot was hot forged into a round bar with a diameter of 20 mm, 30 mm, and 60 mm, and the cross section and longitudinal section of the round bar were investigated by EP, EPMA, or SEM as the investigation about ingot. It was verified that there was no macro-segregation and good bonding was achieved, but details are omitted.
To investigate the macro-segregation formation behavior in the SRPE and the experiments of insert casting, and the preventing conditions the formation and the bonding conditions between the core material and the base metal in the insert casting, it is important to clarify the melting behavior of the core material and the solidification behavior of the molten steel and the thermal hysteresis at the bonding interface.
In order to analyze these melting and solidification behaviors and the thermal hysteresis of the bonding interface, a two-dimensional melting/solidification analyses were performed employing the outer node method of the direct finite difference method (DFDM).27) Latent heat of solidification and melting were taken into consideration through the equivalent specific heat method.28)
In some kinds of insert casting between various metals, melting of the core material proceeds by simultaneous heat and mass transfer, like in the case of scrap melting in carbon-saturated molten iron.29) Therefore, it is necessary to have a combined analysis of heat and mass transfer for analysis of its melting behavior. Since the composition of the core material and the base material is similar in this insert casting, it is unnecessary to consider mass transfer in the melting/solidification analysis and thus the melting and solidification process is taken to be rate-controlled only by heat transfer.
The basic equations used for melting/solidification analysis by the DFDM are shown below. The sum of the amount of heat that flows in by heat conduction from each adjacent region, and the amount of heat that flows in and out of the boundary of heat transfer and radiation equals the change of the heat accumulation amount of node region i. Then, this formula is derived.28)
| (1) |
| (2) |
| (3) |
| (4) |
T: Temerature, Δt: Time step, ρ: Density, Cp: Specific heat at constant pressure, V: Volume, S: Area, h: Heat transfer coefficient, l: Distance between nodes, λ: Thermal conductivity, εe: Emissivity, Γ: Stefan−Boltzmann’s constant
Suffix e: emission, i: Node i, j: Node j, h: Boundary of heat transfer
Furthermore, the latent heat absorption during melting or solidification was treated by the equivalent specific heat method,28) and Cp in Eq. (2) was replaced with Cpe defined by the following equation.
| (5) |
Regarding the heat transfer behavior on the outer surface of molten steel or ingot, the apparent heat transfer coefficient ha at which the temperature change measured by the R thermocouple and the change estimated by the analyses coincide is searched by trial and error method, and it was decided. Since ha decreased exponentially as time elapsed, its behavior was approximated by Eq. (6), and parameters A and B were determined by the above method.
| (6) |
In these analyses for the 1/4 cross section of the ingot, the half of the ingot width and thickness in the ingot width and thickness direction is equally divided into 30, and numerical calculations were done using the explicit method employing the formula (1), (2), (3), (4), (5), (6). When rectangular node elements are used in the DFDM, the equations for the analyses are almost similar with the difference approximation equation of the finite difference method (FDM). However, it is difficult to analyze complex shapes with the FDM, so an analysis model was developed with regards to the DFDM for future use.
Table 2 shows the analysis conditions and the properties used for the analyses.
| Calculation region: 75 mm × 75 mm (1/4 Cross section) |
| Core size: 60 mm × 60 mm, 35 mm × 35 mm |
| Liquidus, Solidus Temperature: 1756, 1681 K |
| Tap Temperature: 1857, 1893, 1903 K |
| Atmosphere Temp: 288 K |
| Density: 7400 kg/m3, Specific heat: 711.3 J/(kg·K) |
| Thermal conductivity: 28.96 J/(m·s·K) |
| Latent heat of Solidification: 282 kJ/kg |
| Discrete time: 0.25 s |
| Division number Width: 30, Thickness: 30 |
The liquidus temperature (TLL), solidus temperature (TSL) which were estimated based on each composition using the previous study’s formula30) and tapping temperature (Tap) of each experiment are shown in Table 1.
The tapping temperature was set to 1857 K in the No. 1 experiment. But in the No. 2 to No. 4 experiments, the heat absorption by the core material improves solidification, making it hard to supply molten steel between the core and the inner wall of the mold. Owing to danger, the tapping temperature was raised to 1893–1903 K.
Figure 2 shows the longitudinal section EP at the ingot width center of the ingots produced by the SRPE (No. 1) and the segregation prevention experiments (No. 2, No. 3). In the EP of the No. 1 experiment shown in Fig. 2(a), streaks of A segregation are evident in the peripheral part of the ingot. In the center of the upper part of the ingot, there is a shrinkage hole and a remarkable upper concentrated segregation can be seen just below the hole. For ordinary small cast iron molds, no A segregation and remarkable upper concentrated segregation as shown in Fig. 2(a) is observed. Therefore, it was verified that various macro-segregation can be formed as a result of reducing the solidification rate with applying the heat insulation to the mold.
Etch prints of longitudinal cross section at width center of ingots of macro-segregation recreation experiment and vacuum insert casting experiments. (Online version in color.)
Figure 2(b) shows the EP of the ingot of No. 2 experiment using a 60 mm square S50C core material. The A segregation and the remarkable upper concentrated segregation noted in Fig. 2(a) disappeared and the shrinkage hole observed in Fig. 2(a) was significantly reduced in this EP. From this result, it was also realized that the insert casting proposed in this study can suppress not only the formation of macro-segregation such as A segregation and upper concentrated segregation but also the generation of the shrinkage hole. So, the soundness and homogeneity of the internal quality of steel ingots were significantly enhanced, and a large improvement in yield could be expected by this casting.
As a result of observing the columnar crystal structure development in the EP of Fig. 2(b), in this case, the columnar crystal structure grew from the core surface and from the inner surface of the heat insulating material, almost in the heat flow direction. The following resulted from such crystal growth. Solidification progressed from both core surfaces and heat insulating material in this insert casting. The growth rate of the solidified shell from the core side was higher, and the solidified shell grew from the core surface to the range of about 30–40 mm horizontally, and also developed from the insulating surface in the remaining range about 10–15 mm. Such crystal growth also indicates that in the insert casting of No. 2 experiment, the effect of heat absorption by the core material was large, and the development of the solidified shell resulting from that absorption was dominant than the development by heat removal thorough heat insulating material. In the EP of No. 1 experiment (Fig. 2(a)), the development of columnar crystals as like the insert casting was quite indefinite on observation.
EP of longitudinal section of No. 3 ingot is shown in Fig. 2(c). In this experiment, the core size was reduced to 35 mm square, and the heat capacity of the core decreased. As a result, the core was partially melted and the core bent by the decreasing rigidity because of heating to near the TSL. It was also observed that one streak of A segregation was generated due to the disappearance and displacement of core from the center of ingot and the expansion of solidification region in the upper part of ingot. The number of streaks of A segregation observed in the EP was reduced from 7 in Fig. 2(a) to 1 in Fig. 2(c). Also, in this case, despite the state of core, the upper concentrated segregation disappeared and the shrinkage holes in the ingot upper part were greatly reduced.
Furthermore, a columnar crystal structure grown in the range of about 30–35 mm in the horizontal direction from the core surface in the core material undissolved part within about 270 mm from the ingot bottom. From the above, this experiment concludes that the solidification acceleration by the insert casting suppressed the formation of A segregation, upper concentrated segregation and shrinkage holes.
In addition, in the EP of Fig. 2(c), the structure of the central part of the ingot in the range of about 270–330 mm from the ingot bottom became unclear and differed from the structure of solidified base material and the core material. This seems like suggestion that the core material in this region was once in a solid-liquid coexisting state but did not completely melt, and did not flow and, after that, solidified. It is assumed that that region became an obstacle to the flow of the residual molten steel in the liquid rich region by the high viscosity and the low fluidity of the coexisting phase and suppressed the formation of A segregation and upper concentrated segregation.
Figure 3 shows the composition image of secondary electron beam (CP image) (a), concentration distributions of C (b), Mn (c) and P (d) in E-1 sample. The color bar in the figure indicates concentration; the higher the color, the higher it is. In each mapping image, the profile of the average concentration of the region sandwiched by the two parallel lines is shown by the solid line. Figure 3(b) shows that the C concentration in the upper concentrated segregation greatly exceeds 1.0%, and that C is significantly concentrated. A large number of high-concentration regions where the Mn was 1.0% or more and the P was more than 0.13% were also observed. It was revealed that the solute enrichment was significantly advanced in the upper concentrated segregation of No. 1 ingot. The solute concentration and the degree of segregation in the upper concentrated segregation are extremely higher than the solute concentration and the degree of segregation in the A segregation described later. It is assumed that such high-concentration segregation emanates from the significant decrease in the effective distribution coefficient of each element by great reducing of solidification rate and by the increased effect of thermal solute convection,9) and the accumulation of concentrated molten steel because of solidification shrinkage.
Result of EPMA elements mapping (No. 1 Ingot, E-1, Upper concentrated segregation part). (Online version in color.)
Figure 4 shows the solute concentration distribution in the 20 mm × 20 mm range including the streak of A segregation observed (E-2) in No. 1 ingot, as a result of EPMA elements mapping. Segregation streaks were also evident in this elements mapping, and C, Mn, P plateau-like high-concentration areas were formed in the streaks. Furthermore, a sample (E-3) was taken from the position of 25 mm from the core surface of the No. 2 ingot. The concentration distribution of C, Mn, and P in the sample was investigated by EPMA elements mapping. Like the observation by EP of Fig. 2(b), the high-concentration segregated portion such as Figs. 3 and 4, was not noticeable in the elements mapping. From this quantitative evaluation of the solute concentration distributions in the No. 2 and No. 3 ingots as described above, it was verified that insert casting suppressed the formation of the macro-segregation such as the upper concentrated segregation and A segregation.
Result of EPMA elements mapping (No. 1 Ingot, E-2, A segregation part). (Online version in color.)
In conventional insert casting for combination of different metals to improve the performance of materials and adding functions, the bonding conditions between unlike metals and between different steel types such as cast iron and steel have been studied.22,23,24,25) However, there are few studies on insert casting of the same metal or the same steel type as in this study, and insert casting in a vacuum atmosphere and bonding conditions in these insert castings have not been studied.
In this study, the insert cast materials of No. 2 and No. 3 experiments were visually observed by EP for the bonding condition. In the observation, there was no void at the bonding interface, which may indicate defective bonding. Furthermore, from the vertical cross section of the No.2 ingot, at the positions S-1, S-2 in Fig. 2(b), a 25 mml × 20 mmw × 15 mmt sample is cut out so that the polished surface is etched by Nital, and the bonding interface is observed at a magnification of 20 to 2000 times by SEM to investigate the bonding state.
Figure 5 shows the secondary electron beam images (SEI) by SEM, obtained on observing the bonded interface at ×20 and ×100 magnification. Similar to Fig. 5, no voids, oxide inclusions or oxide scales were observed at the bonding interface in the SEM observation range at other locations or other samples. Also, some prior austenite grains grown with the bonding interface taken into the grains were observed as in the upper part of Fig. 5(a). From these observation results by EP and SEM, a conclusion could be drawn that good bonding state between the core material and the base material were realized at the casting stage by the insert casting in the vacuum atmosphere.
Secondary electron images near bonding interface etched by nital, observed by SEM (No. 2 Ingot, S-2 sample, Inset casting).
Similar results as the ingot investigation results were obtained in the investigation on the formation of macro-segregation and the bonding state of the interface with the round bar forged from the No. 4 ingot.
To prevent macro-segregation by applying insert casting to the ingot making, it is important to accelerate the solidification of molten steel by absorbing heat to the core material while preventing excessive melting of the core material installed in the center of the steel ingot. For that purpose, it is important to understand the melting behavior of the core material and the solidification behavior of the molten steel. In particular, regarding solidification behavior, understanding the characteristics of the solidification behavior in insert casting, in comparison with the case of normal casting, is considered necessary for elucidating the proper insert casting conditions to prevent macro-segregation. In this study, for the above purpose, the solidification behavior of the molten steel pored or produced from core melting and the melting behavior of core material in the No. 1, No. 2 and No. 3 experiment, were analyzed using the melting/solidification analysis model. In this model, the apparent heat transfer coefficient ha on the surface of the ingot is grasped and arranged in the form of the Eq. (6) employing the aforementioned technique.
Figure 6 is a contour diagram of the temperature distribution on the cross section of the ingot after 200 s, 400 s, 600 s, and 800 s from the solidification start in the No. 1 experiment. The contour diagrams that follow were created using graphing and analysis soft ORIGIN 2019 of Light Stone® based on the numerical data of the temperature or solid fraction of each node got from numerical calculation. In this figure, the broken lines represent the TLL and TSL of the cast material. The higher color of the color bar in each contour diagram indicates that the temperature or the solid fraction is higher.
Calculated distribution of temperature on 1/4 cross section of ingot at each time from solidification start (No. 1 experiment, Normal casting). (Online version in color.)
The following was estimated from Fig. 6. In this experiment, solidification progressed from the mold corner toward the center of the cross section, and the region of the solid-liquid coexisting phase decreased slowly as time elapsed, and finally solidification was completed in the center of the cross section. The solidification is not completed in 800 s after the solidification start, it is completed at 875 s. Further, in this case, the temperature gradient decreasing from the central portion of the cross section toward the peripheral portion was maintained. The temperature gradient became gentle in the solid-liquid coexisting region and it became larger in the peripheral region compared to the coexisting region, and the cooling rate increased at each position where the solidification was completed.
Figure 7 shows contour diagrams of the estimated distributions of temperature and solid fraction after 200 s and 300 s after solidification start in the insert casting of No. 2 experiment. As depicted in the figures, in the insert casting, the solidification and the temperature decrease were more advanced in the center of the cross section rather than in the corners as a result of the heat absorption of the core material. The behavior of solidification and the temperature distribution significantly differed from the SRPE. Moreover, the followings are estimated. Solidification was accelerated in the entire cross section as compared with the SRPE, and was completed in most of the cross section 300 seconds after it started. The ingot completely solidified after 368 s, and the solidification time was shortened to less than 1/2 compared with the SRPE. In this case, the core material did not melt and solidification proceeded continuously from the core surface.
Calculated distribution of temperature and solid fraction on 1/4 cross section of ingot at each time from solidification start (No. 2 experiment, Insert casting). (Online version in color.)
In addition, from Fig. 7, it was estimated that solidification was completed at about 35 mm from the core surface on the center axis of thickness or width of the cross section of the ingot. It almost agreed with the shell growth observed by EP. This agreement indicates that the analysis results by numerical calculation by the model almost correctly estimated the melting and solidification behavior in the experiments.
From this analysis result, the final solidification position on the cross section was estimated to be the corner side on the corner diagonal. So as to verify that macro-segregation did not occur near this position, the segregation was investigated by the EP of the longitudinal section sampled along the diagonal line including the position. At this position, neither the concentrated segregation zone nor the A segregation was confirmed.
Figures 8 and 9 show the temperature distribution and the solid fraction distribution in the cross section every 200 s from the solidification start to 1000 s later of No. 3 experiment. The findings, based on the figures, are as follows. At first time, a solidified phase was generated from the core surface, but after 200 s, the central part of the core material was heated to a temperature above TSL and entered solid-liquid coexisting state. In addition, the region with a solid fraction of 0.3 or more was expanded to a range of about 25 mm from the core surface because of the heat absorption by core material and partial melting of the core material.
Calculated distribution of temperature on 1/4 cross section of ingot at each time from solidification start (No. 3 experiment, Insert casting). (Online version in color.)
Calculated distribution of solid fraction on 1/4 cross section of ingot at each time from solidification start (No. 3 experiment, Insert casting). (Online version in color.)
The solid-liquid coexisting phase spread over a wide area at the center of the cross section after 400 s and most of the region had a solid fraction of 0.3 or more. After 600 s, the temperature gradient that decreased toward the center of the cross section disappeared and the temperature distribution flattened. The solid fraction in the coexistence region reached about 0.45 or more, and after that, solidification proceeded from the periphery to the center of the cross section, and solidification ended at the center. It was presumed that once the original position of the core material became the solid-liquid coexisting state and the solid fraction was maintained at 0.47 or more, the solidification finished. In the No. 3 experiment, the core material was bent due to a decrease in rigidity resulting from heating to a high temperature and the impact by the pouring flow. Therefore, although the core position is different from the analysis condition with the core fixed, it was confirmed from the EP in Fig. 2(c) that the core partially melted and a region where the solid-liquid coexistence state was maintained until solidification completely finished. This is the point is considered to be the outcome supporting the calculation results of the melting/solidification analysis model.
The final solidification portion in the ingot of this experiment was the center of the cross section, and the solidification completion time was estimated to be 1060 s, which was estimated to go beyond the time of 875 s in the No. 1 experiment. The extension of the time could have resulted from the decrease in the amount of heat removed from the ingot surface due to heat transfer coefficient reduction compared to the No. 1 experiment and the decreasing of the amount of heat conduction by reduction of the temperature gradient in a wide range in the center of the cross section. As is clear from Figs. 8 and 9, in this experiment, the solidified shell was once generated around the core material, and the solidified shell and the core melted, so that the solid-liquid coexisting phase expanded in the center of the cross section. As a result, the temperature gradient reduces and flattens in a wide range in the center of the cross section, so that the gradient caused the decrease in the amount of heat conduction.
Figure 10 shows the estimation results of the temperature transition at each position in the thickness direction on the center axis of ingot width in the SRPE and the two insert casting experiments. In the figure, the TLL (1756 K) and the TSL (1686 K) of S50C are represented by two dotted lines. From Fig. 10(a), it can be seen that in the SRPE, solidification progresses sequentially from the surface and solidification is completed at 875 s. Moreover, the following was estimated in the insert casting experiment of Fig. 10(b) in which the core material was 60 mm square. The core portion does not reach the TSL during solidification, that is, it does not melt at all. The molten steel at a position 45 mm from the center of the cross section solidifies due to heat removal by the core material, and is further cooled to 1623 K. After that, a temperature gradient is formed in which the temperature decreases toward the center of the cross section, and the position is reheated to near the TSL by heat transfer from the higher temperature peripheral side. In addition, there was an estimate of solidification being complete at 235 s on the width center axis of the ingot. It was found that the solidification time at the part was significantly shortened compared to the SRPE. In the insert casting experiment of Fig. 10(c) in which the core size is 35 mm square, the following was estimated. The core material is heated rapidly after pouring, reaching a maximum of about 1723 K, and the solid fraction decreases to about 0.47, and the coexistence state is maintained for 1000 s or more from the solidification start, and solidification is complete after 1060 s.
Changes of temperature at each position of width center in cross section of ingot. (Online version in color.)
From the results of the melting/solidification analyses and the comparison of the generation behavior of of each experiment, the following estimates were made. In No. 2 experiment, the solidification finish time is significantly reduced from 875 s in No. 1 experiment to 368 s, and the solid-liquid coexistence region also disappears early. It is thought that the above-mentioned accelerated solidification and early disappearance of the liquid phase as a result of insert casting suppressed the flow of residual molten steel and the reduction of the effective distribution coefficient due to flow, and this in turn prevented the formation of macro-segregation such as A segregation and upper concentrated segregation. Further, in this experiment, as earlier discussed, the final solidification position moves to four positions on the diagonal side. Therefore, it is considered that the solid-liquid coexisting region and the amount of solidification shrinkage were dispersed and the thermal solute convection and solidification shrinkage flow were suppressed, which also contributed to the improvement of macro-segregation.
On the other hand, in the No. 3 experiment using 35 mm square core, the time until complete solidification increased from 875 s in No. 1 experiment to 1060 s. Though there was an increase in the time, the formation of A segregation and upper concentrated segregation was suppressed, and there seems to have been no improvement on employing the same mechanism as in No. 2 experiment. The macro-segregation improvement mechanism in this case is considered as follows. It is presumed that the flow of residual molten steel and the formation of macro-segregation were suppressed by the early transition to a solid-liquid coexisting state with high viscosity and low fluidity in certain region inside the ingot although the liquid phase did not completely disappear. From this view point, comparing Figs. 10(a), 10(b), and 10(c), it can be seen that the time taken for a wide area in the ingot to reach the solid fraction 0.4 (temperature 1728 K) is shown in No. 1 experiment (Fig. 10(a)), it is 780 s, whereas in No. 2 and No. 3 experiment (Figs. 10(b) and 10(c)), it is shortened to 125 s and 400 s, respectively. An assumption is made that the early decrease in fluidity due to the early transition to a solid-liquid coexistence state with a solid fraction of about 0.4 or more causes the suppression of macro-segregation formation by insert casting. In addition, when the above-mentioned region having high flow resistance is solidified covering a wide range, that region becomes an obstacle to flow of residual molten steel due to change of liquid density caused by solute concentration, thermal solute convection and solidification shrinkage. It is estimated that these flows were suppressed in a wide range within the ingot, and the formation of A segregation and upper concentrated segregation was suppressed.
5.2. Bonding Conditions for the Same Steel Type by the Insert Casting in Vacuum AtmosphereConventionally, in the production of composite materials by the insert casting, the bonding conditions between the core material and the base material have been examined, the bonding conditions between different steel types and between different metals have been studied. The effect on the conditions of the volume ratio or weight ratio, the core material and the base material, pouring temperature, coating of low melting point metal on the interface, effects of atmospheric oxidation and presence of oxides on the interface, have been examined. The relationship between the bonding conditions and the thermal conditions including dissolution and solidification behavior in this casting have been studied.22,23,24,25) However, very little or no study has been made on the bonding conditions for insert casting between steel types having the same composition or for insert casting in a vacuum atmosphere, as this study does. Therefore, with a basis on the investigation results of the bonding conditions and the melting/solidification analysis results for the insert casting experiments, the bonding conditions between the same steel type by insert casting in the vacuum atmosphere were examined.
The No. 2 and No. 3 insert casting experiments, in which the bonding conditions were investigated, almost had good bonding conditions. The volume ratio (weight ratio) of the base material to the core material was 5.3 in the No. 2 experiment and 17.4 in the No. 3 experiment. It has been reported that good bonding was realized in some experiments in which mild steel was cast wrapped in cast iron in the atmosphere.22) In the case, the volume ratio for good bonding was about 10 or more without thermal spraying the self-melting alloy and about 5 or more under the condition with the spraying. In this study experiments, unlike the insert casting between mild steel and cast iron, lowering the melting point of the core material surface due to mass transfer of C and formation of liquid at the interface29) cannot be expected. Under these conditions, and even under the condition that there is no thermal spraying, good bonding could be realized in the experiments even with a volume ratio of about 5.3. It has been found that even under the disadvantageous conditions in which the melting point of the core surface does not decrease and the liquid does not intervene, greatly relaxing the bonding conditions in the insert casting in which mild steel was cast wrapped in cast iron in the atmosphere without thermal spraying, by applying of the insert casting in vacuum atmosphere is possible. It has already been found that, in the case of insert casting, the oxides generation due to the atmospheric oxidation at interface and the oxides entrainment results in bonding defects, and good bonding can be achieved by preventing them.24,25) It seemed that good bonding could be achieved even with a volume ratio of 5.3 because the formation of oxides and the oxides entrainment at the interface never occurred during insert casting in a vacuum atmosphere.
In addition, the following was estimated from the outcomes of melting/solidification analyses. In No. 3 experiment, the solidified shell was generated and remelted near the interface, and the vicinity of the interface was for a long time kept in a solid-liquid coexisting state. Due to this, it was presumed that good bonding was realized because the conventional thermal conditions for good bonding22) were achieved. On the other hand, in the No. 2 experiment, good bonding was realized despite the fact that it was estimated that no liquid was generated at the interface where bonding was difficult.22) Insert casting in a vacuum atmosphere prevents the generation of oxides and oxides entrainment at the interface, and also suppresses gas and gap generation at the interface. Therefore, the contact between the core material and the solidified shell is not hindered. In addition to these conditions, in the No. 2 experiment, the maintaining the temperature at 1623 K or higher for 200 s or more at the interface as shown in Fig. 10(b), promoted diffusion bonding,31,32) resulting in good bonding.
5.3. Causes and Countermeasures of the Internal Cracking in the Insert CastingInternal cracking occurred in the No. 2 experiment (Fig. 2(b)). It was estimated from the outcome of melting/solidification analysis as shown in Fig. 10(b) that due to the restrictions peculiar to this experimental method, internal cracks were generated because the solidified shell on the core surface was pulled by the thermal contraction of the shell itself while being restrained by the core, and the thermal expansion of the core. To prevent this, it was presumed that relaxing the restraint of the solidified shell and optimizing the temperature transition of the solidified shell and core was important by, for example, optimizing the core size and devising the pouring method to decrease the tensile stress on the shell.
Effective and versatile countermeasures were investigated to prevent macro-segregation in large size ingot casting for the production of large-scale parts for power plants. As a result, a method for the effective suppression of the formation of macro-segregation was developed. It is the method using insert casting in a vacuum atmosphere, in which a core material with the same composition as the base material is placed in the center of the mold and cast-wrapped.
The effectiveness of this method as a macro-segregation countermeasure was confirmed through laboratory experiments. In addition, it has been clarified that good bonding between the core material and base material can be realized even under conditions where bonding by normal insert casting in air atmosphere is difficult. The analysis of the melting behavior of the core material and the solidification behavior of the molten steel of the base material in the experiments of macro-segregation reproduction casting and the insert casting were completed using the DFDM.
The mechanism by which this method suppresses the formation of macro-segregation and the solidification conditions for the suppression, the reasons, and conditions for good bonding in this insert casting are clarified by the analyses.
Furthermore, the cause of internal cracking of the insert casting experiment was investigated, and guidelines to prevent the formation of internal cracks were presented.
Part of this research was supported by JSPS KAKENHI Grant Number JP16K06818 and the 26th ISIJ Research Promotion Grant. We sincerely thank each research grant.
