Since the mid-1960s, the hottest parts of the engine, the blades (rotating) and vanes (non-rotating), have been manufactured by investment casting. As the alloy content of nickel based superalloys was continually increased to obtain better creep and stress rupture capability, the alloys became increasingly difficult to forge. To allow even higher contents of alloying elements, it became necessary to change the fabrication process to casting. Investment casting became the process of choice because it is amendable to the fabrication of hollow blades with intricate cooling passages, which allows higher operating temperatures. Since the mid-1980s, turbine inlet temperatures have increased by 500° F. About half of this increase is due to more efficient designs, while the other half is due to improved superalloys and casting processes. The introduction of directional solidification allowed about a 50° F increase in operating temperature, while the single crystal process produced another 50° F increase.11 The reader may want to refer back to Chapter 4 on Titanium for the detailed steps involved in investment casting.
The original cast blades and vanes were fine grained polycrystalline structures made using conventional investment casting procedures. These blades were then heat treated to coarsen the grain structures for enhanced creep resistance. Eventually it became possible to produce directionally solidified (DS) structures with columnar grains oriented along the longitudinal axis of the blade. The columnar grain structure enhances the elevated temperature ductility by eliminating the grain boundaries as failure initiation sites. The DS process also creates a preferred low modulus texture, or orientation, parallel to the solidification direction that helps in preventing thermal fatigue failures. An extension of the DS process was the development of the single crystal (SX) process in which a single crystal grows to form the entire blade. Since there are no grain boundaries in the SX process, this allowed the removal of grain boundary control alloying elements that were detrimental to high temperature creep performance, allowing even higher operating temperatures. However, the SX process is significantly more expensive than the DS process because yields are lower; therefore, both processes are used extensively for fabricating blades and vanes. The macrostructures of polycrystalline, DS and SX blades are shown in Fig. 6.16, and the dramatic improvements in creep resistance are shown in Fig. 6.17.
Although DS and SX castings have replaced polycrystalline castings for applications like blades and vanes, polycrystalline castings are still used for structures such as compressor housings, diffuser cases, exhaust cases and engine frames. Some of these parts are very large (e.g., 60 in. in diameter) and the castings can weigh up to 1500 lb. The casting of large structural parts allows reduced manufacturing costs with more unitized structure by eliminating the casting of many smaller parts and then welding them together. Since many of these parts are strength and fatigue critical, innovations have been developed to produce castings with finer and more uniform grain sizes. Examples include mold agitation
Fig. 6.16. Cast Turbine Blades25
Fig. 6.16. Cast Turbine Blades25
Mar-M200 30 ksi 1800° F
Mar-M200 30 ksi 1800° F
during solidification (Grainex process) and the use of low superheat during pouring (Microcast-X process).
6.7.2 Directional Solidification (DS) Casting29
To develop a directionally solidified structure, it is necessary for the dendrites (grains) to grow from one end of the casting to the other. This is accomplished by creating a sharp temperature gradient, by removing the majority of the heat from one end of the casting. As shown in Fig. 6.18, a water chilled mold is slowly withdrawn from the furnace, setting up a strong temperature gradient in the freezing metal. A thin wall investment casting mold, that is open at the bottom, is placed on a water-cooled copper chill plate. The mold is then heated to the casting temperature. The alloy is heated under vacuum in an upper chamber of the furnace and then poured into the heated mold. After a couple of minutes to allow the grains to nucleate on the chill plate, the mold is slowly withdrawn from the hot zone and moved to the cold zone. The chill plate insures that there is a good nucleation of grains to start the process. Although the grains nucleate with random orientations, those with the preferred growth direction normal to the chill surface grow and crowd out the other grains.
The thermal gradient is established in the zone between the liquidus and solidus temperatures, and is passed from one end of the casting to the other at a slow rate that maintains the steady growth of the grains. Temperature control and extraction rate are critical. If the progression of the thermal gradient is too fast, grains will nucleate ahead of the solid/liquid interface, while if the
movement is too slow, excessive macrosegregation will occur along with the formation of freckles, which are defects of equiaxed grains of interdendritic composition. Typical temperature gradients are 170-330° F/in. with withdrawal rates of around 12in./h. In general, higher temperature gradients produce better quality castings.
Since a very high degree of process control is necessary to achieve proper grain growth, the entire process is automated. One reason why automation is necessary is that the withdrawal rate is not necessarily constant. For example, large differences in section size change the solidification rate, and the withdrawal rate must be changed to compensate. In addition, the effect of cooling rates due to the ceramic filled cores must be included since the cores lengthen the time to preheat the mold and slow the withdrawal rate.
Typical defects in DS castings include equiaxed grains, misoriented grains, grain boundary cracking, excessive shrinkage, microporosity, and mold or core distortion. Increasing the thermal gradient during casting is used to control the formation of equiaxed grains, misoriented grains and microporosity. Hafnium, in the amount of 0.8-2.0%, can be added to the alloy composition to avoid grain boundary cracking. Shrinkage on the upper surfaces (the last surfaces to solidify) is sometimes encountered, because risers are not used since they interfere with the radiation heat transfer. Inverting the casting can sometimes help to eliminate this problem. Due to the thin mold cross-sections used and the long exposure times at high temperatures, ceramic mold or core distortion can also be a problem. Careful control of their compositions, uniformity and firing conditions are required to prevent mold/core distortion problems.
6.7.3 Single Crystal (SC) Casting29
After the development of the DS casting process, it was recognized that if all but one of the growing columnar grains could be suppressed, it would be possible to cast parts with only a single grain, thereby eliminating all of the grain boundaries. In addition, alloying elements that were necessary to prevent grain boundary cracking, but were detrimental to creep strength, namely boron, hafnium, zirconium and carbon, could be eliminated.
Single crystal castings are produced in a manner similar to DS castings with one important difference; a method of selecting a single properly oriented grain is used. In the most prevalent method, a helical section of mold (Fig. 6.19) is placed between the chill plate and casting mold. This helix, or spiral grain selector, acts as a filter and allows only a single grain to pass through it. Because superalloys solidify by dendritic growth that can occur in only three mutually orthogonal directions, the continually changing direction of the helix gradually
constricts all but one grain, resulting in a single crystal emerging from the helix into the mold. Seeding is another method of creating a single grain. Instead of the helical section to control grain growth, a single crystal of the alloy is placed on the chill plate. The crystal is oriented so that its orientation will be repeated in the alloy that fills the mold. The temperature of the seed is controlled so that the seed does not completely melt, allowing the molten metal to solidify with the same orientation as the seed crystal. One advantage of the seed crystal method is that the orientation of the metal that grows from the crystal can be controlled, to a degree, by the orientation of the crystal that is placed in the mold.
With the exception of grain boundary cracking, SX castings can experience the same defects as those found in DS castings. In addition, SX castings can form low angle grain boundaries. Although the misorientation is usually less than 15°, they can act as crack initiation sites. The misorientation allowed depends on the specific application, but boundaries above 10° are usually not permitted. Single crystal castings are usually inspected by X-ray diffraction to check for crystallographic orientation.
Was this article helpful?