Enhancing aluminum casting solidification via cooling fins.

This experiment relays the advantages of cooling fins and how they can improve solidification and mechanical properties, while reducing shrinkage and microporosity.

Cooling fins have been incorporated onto aluminum castings to reduce solidification time, assist with the elimination of shrinkage cavities and microporosity, and improve the local mechanical properties. In addition, fins can extend the effective range of feeders and, in some cases, eliminate feeders altogether. The benefits cooling fins provide to castings often cause them to be considered with chills as viable options to improve casting design. Cooling fins, however, provide aluminum castings with solidification advantages not inherent to chills and must be, in most cases, the primary design option.

In design, a cooling fin is a small, thin appendage that is added onto the surface of a casting to enhance local cooling of the metal. This article relays the inherent advantages of fins over chills and also discusses the results of an experiment in which aluminum was cast into horizontal plates in chemically-bonded sand molds to determine the optimum cooling fin geometry. During the experiment, analysis was performed in actual-time with thermal analysis via thermocouples and also by simulation via solidification software.

Fins vs. Chills The basic advantages of cooling fins when compared to chills include:

* Cooling fins are an integral part of the pattern and therefore are accurately located and automatically placed on the casting. Chills, which are separate from the casting, are often inaccurately placed and can be forgotten in the mold building process.

* Cooling fins are formed as part of the casting, so there isn't an interface gap through which heat transfer must take place. As a result, heat transfer to the fin is efficient and reproducible. The efficient heat transfer is lost with the use of chills as an air gap exists between the chill and the solidifying casting that reduces the chilling action soon after it starts.

* In comparison to the use of chills, cooling fins: are not ordered and manufactured; are not retrieved from the shakeout and sand return system where they can cause damage to castings and to the shakeout line and suffer wear; and do not have to be stored for later use. Cooling fins are an element of design and are not subject to worker error on the plant floor.

In addition, cooling fins play an important role when flash is produced. Flash can be considered an accidental cooling fin attached to a casting. But because flash is one of the first things removed during finishing, its presence and effect on the casting is overlooked by the customer or the manufacturer. Therefore, failure to consider the heat transfer properties of flash can lead to defects in the casting, such as cutoff of the feed paths in the section adjacent to the flash. In the same way that cooling fins are designed into the process, flash must also be considered in design.

Cooling fins have only one major disadvantage: they must be removed from the casting in the finishing room, which increases finishing cost. Also, the incorporation of extra casting material leads to a minor reduction in casting yield. But this disadvantage is usually more than offset by a reduction in the feeding requirements of the casting.

Due to the inherent advantages of cooling fins, an experiment was conducted to determine optimum geometry. Through it, foundries may be able to make use of the beneficial properties of cooling fins and, at the same time, be aware of their potential detrimental effects.

Experiment

Figure 1 illustrates the pattern dimensions of the casting being poured. The dimensions of the cast plate were approximately 4 x 4 x 0.8 in., onto which a variety of fin configurations were attached. The various lengths and the widths of the cooling fins that were tested were derived from half the thickness of the plate being cast or 0.4 in. This relationship is crucial to the optimum cooling fin geometry.

The dimensions and configuration of the filling system were chosen to promote quiescent filling of the mold cavity and the plate is oriented horizontally so that convection currents are minimized. (In previous work on cooling fins, the primary plate casting was placed vertically. This alignment, however, can lead to convection currents in the molten metal that redistribute the cooling liquid throughout the casting and increase overall casting solidification rate.)

Two vertical fins are designed symmetrically on either side of the test plate to form a cruciform. This symmetrical arrangement was chosen to avoid placing a single fin where the final freezing point (the thermal center) would not be on the centerline of the casting. If this freezing point were not on the centerline, a thermocouple could not be accurately placed prior to the experiment. With the symmetrical two-fin geometry, the central thermocouple unambiguously recorded the freezing time of the casting.

Castings were poured with 99.9% aluminum. This material was chosen because of its high thermal conductivity, high latent heat of fusion and single solidification temperature.

Phenolic-urethane molds were prepared using silica sand of AFS GFN 60. A 0.08-in. diameter steel rod was introduced externally from the molding box and located precisely at the pattern face. Once the binder hardened, the rods were retracted before stripping the pattern from the mold. This left a hole through which the thermocouples could be located.

Thermal analysis was carried out using thermocouples at the geometrical center of the casting. The solidification times cited in the following sections are the normalized total and local solidification times, respectively. The normalized solidification time is the solidification time of the casting divided by the solidification time of an equivalent casting with no fins.

Thermal analysis was done via the data received from the thermocouples. The sampling rate for the castings was 250 samples/channel/sec. The resultant data were then averaged to 10 readings a second and a further running average across 51 points was carried out.

The thermocouples were subject to a small amount of random movement during casting. As a result, their actual positions were determined by taking slices of the casting for X-ray radiography in plan and elevation views. The thermocouple junction then could be located with precision using image analysis software. When thermocouples have become significantly displaced from their intended positions, the specimen has been cast again.

Effect of Fins

The local solidification times for each casting were determined from the temperature-time graphs. As shown in Fig. 2, the local solidification time for all given lengths of fin drops to a minimum when the thickness of the fin is at the minimum of 0.04 in. The local solidification time then increases with an increase in fin thickness. Also, there is a drop in solidification time with an increase in the length of fin from 0.8 to 1.6 in. for a constant thickness. But, increasing the fin length further to 3.2 in. provides little, if any, reduction in solidification time.

Computer Simulation

The same casting dimensions and mold setup were used on computer solidification software for comparative analysis. The solidification simulations were carried out to mimic the actual tests and then were translated into meaningful data. The solidification software allowed the operator to pull the results and plot graphically onto the casting outline representations of the solidification wave front at the given times. The simulation achieves this by simply joining all the nodes together that have solidified at the given instant in time. Due to this, it is possible to map out the progress of the solidification front with time and specify the time at which local solidification events have taken place, including the area between the two opposite fins that has solidified.

The solidification times of the aluminum casting obtained from solidification simulation are plotted for varying fin thickness and lengths in Fig. 3. These graphs show that the local solidification time is shortened with an increase in fin length until a fin length of 1.6 in. is reached. Increasing the fin length above 1.6 in. does not reduce the solidification time further. Increasing the fin thickness, however, does increase the solidification times. Also, a fin thickness of 0.08 in. to 0.4 in. has a cooling effect on the casting, whereas a fin thickness of 0.6 in. introduces a hot spot into the casting. Therefore, a neutral fin geometry must have a fin thickness that lies somewhere between 0.1 in. and 0.6 in.

The modeling of fin thickness less than 0.08 in. was not possible using this computer simulation. This was due to the lack of definition in the enmeshment stage of the modeling. Experimental work has been carried out to a fin thickness of 0.04 in. Since casting sections over 0.8 in. long but less than 0.04 in. thick in a non-pressurized system approach the limits of fluidity of the metal, a larger pattern would need to be constructed to investigate fins of thickness less than 0.04 in.

Conclusions

By using thermal analysis and solidification software verification, the experiment was able to establish optimum cooling fin geometry. The observations from the experiment include:

* a decrease in fin thickness reduces the solidification time to a computed limit of approximately 0.08 in., but an experimental. limit of 0.04 in. Even thinner fins may be more effective, although the scope for further improvement appears strictly limited;

* the solidification rate increases with an increase in fin length to a limit of 1.6 in., beyond which negligible further effect exists;

* thermal conductivity of the casting material has a significant effect on the chilling power of a fin. For instance, fins cast in copper- and aluminum-based alloys are considerably more effective than those cast in gray iron and steel;

* a prediction of solidification time using Chvorinov's rule will not account for the fin effect;

* fins have a number of advantages over chills: fins are not forgotten; fins are always located accurately; and fins have a reproducible effect (being relatively unaffected by air gap formation problems). In addition, cooling fins do not need to be manufactured, placed, retrieved, cleaned, coated and stored. They also are free from the problems of condensation and blowing defects and the problem of wear for shaped chills.

For a free copy of this article circle No. 342 on the Reader Action Card.

This article was adapted from a paper (9 7-10) published in the 1997 AFS Transactions and is available from the AFS Library at 800/537-4237.