Discovering hot distortion properties in PUCB cores.
* PUCB cores made from silica sand have the tendency to produce casting defects due to hot distortion errors.
* Previous investigations have not evaluated how the cores perform under molten aluminum fill conditions.
* By changing the sand or weight of cores, metalcasters may reduce distortion defects.
Since it was introduced to the metalcasting industry in 1968, the phenolic urethane coldbox (PUCB) process has been beneficial by providing a low-cost process that has the ability to be used with all types of sand. However, the process has low hot strengths, and, consequently, hot distortion may occur. The low hot strength can result in an increased frequency of casting defects, such as veining and core breakage, and hot distortion from thermal expansion and resin breakdown under high temperatures. These distortion characteristics become more important when castings require thin cores.
Unfortunately, evaluations of such core distortion are scarce, and the hot distortion curves from previous tests represent resin characteristics, but not necessarily under casting conditions. Almost all efforts in prior investigations have been directed toward optimization of resin system characteristics by manipulating the coremaking process, amount of binders, part 1/part 2 ratios, using additives, applying coating, etc. Thus, previous test castings were based on a "best guess" approach; however, the dynamics of molten metal flow around the cores were ignored.
Without a fundamental understanding of thermo-kinetics and fluid dynamics of molten metal-sand core interactions, it is difficult to make decisions in preventing defects. This article details the preparation and analyses of hot distortion in PUCB systems and how this dilemma may be relieved from metalcasting facilities.
Preparation and Properties
The investigations conducted involved the evaluation of an amine-cured, two-part phenolic urethane resin system. The sand cores were made predominately out of silica sand, which contained a significant amount of aluminum oxide as well as ferric oxide and calcium oxide. Cylindrical core specimens measuring 25 mm long and 10 mm in diameter were produced to determine the thermal expansion/contraction of sand cores, and the sands were bonded with 1.4% resin at a 50/50 part 1/part 2 ratio.
The thermal expansion/contraction of sand core specimens was measured within a temperature range from 25-800C (77-1,472F) with a heating rate of 20 degrees/min. Due to the combined effect of expansion and contraction of sand and resin binders at temperatures up to 300C (572F), there were no significant changes in the specimens' volumes. At temperatures above 300C, the resin binders started melting and subsequently began to burn off. When the resin binders completely burned off at 360C (680F), the contraction process stopped, and the sand grains, free of resin binders, started to expand. This continued up to 560C (1,040F) when the expansion process halted. This is common for silica sand cores prepared by the PUCB process (Fig.1). After the tests, it was found that the specimens lost a mass of about 1.4%, which was the mass of resin binders in the sand cores.
[FIGURE 1 OMITTED]
Next, the effects of the part 1/part 2 ratio, bench life (BL) and temperature were determined on the tensile strength of the test sand cores with 1.4% resin binders. Sand core test specimens used to evaluate test material were prepared of a size and shape defined by standard test methods--in this case, the samples were 13 mm thick. Tension tests were conducted using a computer-controlled testing machine equipped with a heating furnace, which allowed conduction of tension tests at up to 2,000C (3,632F).
The test results showed that the tensile strengths of sand cores made of silica sand were significantly affected by bench life (Figs. 2 and 3).
[FIGURES 2-3 OMITTED]
After 1 hr. of the bench life, the tensile strengths decreased by 15%. Test temperature more dramatically affected the strengths. At 204C (399F), the tensile strength decreased by 20%, and at 247C (477F), it decreased 50%. The part 1/ part 2 ratio also affected tensile strength. Decreasing the percentage of part 1 from 55% to 50% resulted in a 10% decrease of tensile strength.
Once the test cores were established and fully tested, the high temperature hot distortion investigations were conducted using 300 mm long and 19 mm diameter half-cylindrical cores with a mass of 66.5 g and a density of 1.537 g/cc. Both non-coated and coated silica sand cores were tested in these studies. Aluminum foil, aluminum spray paint, boron-nitride and aluminized plastic film were used as coatings, and tests were conducted for silica sand cores with a 1.4% 50/50 binder level. Two furnaces were used for high-temperature hot distortion test, which allowed conducting the tests at temperatures up to 500C (932F).
The tests showed the relatively higher resistance of the coated (both aluminum foil and aluminum spray) sand cores to hot distortion. At 300C, non coated cores were broken, while coated sand cores were not distorted. But at 500C, some distortion of coated sand cores was observed. Based on heat transfer simulations and experimental observations, it was assumed that aluminum coating radiated the significant amount of heat from the core surface, delayed the burning of the resin binders due to the less contact with air and created a "shell" around the sand core, which prevented the fall of decomposed sand at high temperatures.
Under Molten Conditions
Although the tests on the sample cores helped determine the high-temperature properties, investigations were further carried forward in creating sample castings to generate distortion during pouring and solidification of the molten aluminum alloy. The test specimens for hot distortion test castings were produced with parameters identical to those of the oven-heating tests. Two sets of the test castings were poured. In the first set, the three different coating materials (refractory spray paint, aluminum foil and aluminum spray paint) were applied to estimate the hot distortion and the degradation of sand core specimens.
With the test cores placed on the drag side of the mold, the aluminum A356.2 samples were cast via the V-process technique using a side gating system, A space of 50 mm was allowed on the top and bottom of the core to be surrounded by cast metal. With a pouring temperature of 770C (1,418F), an open riser was used to prevent any shrinkage and air entrapment inside the casting. Temperature variations were measured inside the spree and riser. When the cores were examined dimensionally for distortion, it was found that regardless of the application of coating and the coating type, all three cores were distorted (Fig. 4).
[FIGURE 4 OMITTED]
Recently, synthetic iron oxides (SIO) have been used extensively for reducing nitrogen and veining defects. Therefore, in the second set of the trial castings, SIOs were used as an additive to the PUCB system. In this set, the three sand cores were tested as either non-coated, with a 1% SIO additive or aluminum foil coated plus a 1% SIO additive. Using the same casting procedure and parameters identical to those of the first set of the casting trials, examinations showed that all three cores were significantly distorted. In fact, the addition of the SIOs enhanced the hot distortion.
Calculating Defect Prevention
Along with the actual casting trials, it needed to be determined why such properties occurred with the silica cores. Therefore, numerical simulations of the sand core cylinder in the molten metal and the flow dynamics around the cylinder were conducted. These simulations revealed that for the materials considered in the real castings, the buoyancy force is the major contributor to the core distortion. Thus, investigating physical characteristics of molten metal flow around a core was necessary.
During pouring, before the liquid metal reached the sand cores, only gravitational forces acted on each cylindrical core. When the liquid front reached the bottom of the cores, the buoyancy forces moved opposite to the forces of gravity. The buoyancy forces also increased the more the cylinders were submerged into the liquid and reached their maximum when the molten surface reached the top surface of the cylinder (Fig. 5). The further upward propagation of the liquid level generated the molten head pressure on the cylinders (Fig. 6).
[FIGURES 5-6 OMITTED]
These forces resulted from the distribution of forces over the cross-section of the core cylinder at a built-in or fixed support. This type of support was capable of holding an axial force, a transverse force (shear force) and a couple (bending moment) to prevent rotation. In this process, the fluid flow effect must be taken into consideration.
As seen in Fig. 7, until the molten metal front touched the bottom of the core, only gravity force acted on the sand core cylinder (-0.68 N). As soon as the liquid metal submerged the cylindrical core, the buoyancy force on the cylinder began moving in the opposite direction to the gravity force. When the core cylinder was completely submerged into the molten metal, the buoyancy force achieved its maximum value (1.15 N), hence, the resulting force (0.467 N) acted upward.
[FIGURE 7 OMITTED]
Increasing the Weight
The high-temperature tensile strength and hot distortion furnace rests demonstrated that at temperatures far below the melting point of the aluminum alloys, the PUCB sand cores were losing their rigidity and exhibited same plasticity. Therefore. at molten metal temperatures under the applied resulting force, the fixed-supported cylindrical core could distort upward. From the previous simulations and the experimental results, it was concluded that one of the solutions in preventing the hot distortion of the sand cores is increasing their weight, which would balance the buoyancy force and bring the resultant force to the minimum (or to zero). Thus, hot distortion test castings were conducted using heavier sand cores.
Here, zircon sand was used instead of silica due to zircon's popularity in the industry, and zircon cores have similar densities as molten aluminum A356. Again, PUCB systems were used to prepare the 300 mm long and 19 mm diameter semicircle cylindrical cores, and the binder used was a 1.5% resin mixture (by weight) at a part 1/ part 2 ratio of 55/45. Each core had a 126.6g mass and the density of core (2.925 g/cm3) was selected according to the simulations described above.
The casting procedure and parameters were identical to those of the first two sets of the casting trials. Two different coating materials--93% pure zinc spray and aluminum foil--were applied to estimate hot distortion and degradation of the sand cores. An examination of the transversally sliced casting showed that all three zircon cores (Fig. 8) didn't bend as previous silica sand cores, thus, no hot distortion was observed in all three cores regardless the application of the coatings.
Keep the Density Up
As the performances of the zircon cores demonstrated, higher-density cores especially with densities near or equal to the molten metal--will greatly decrease the chance of hot distortion defects. Although silica sand also could be used in preventing hot distortion detects, it is incapable of doing so due to its low density. Thus, additives would need to be incorporated within the PUCB process to increase the silica density to reduce the chance of defects. Metalcasters also may seek to prevent hot distortion defects by making cores with high hot strengths and also low core decomposed gases. Investigations are being refined to discover the hot distortion properties of cores in magnesium castings, and these will be made public within the next year. MC
This article was adapted from a paper (04-027) presented ell the 2004 Metalcasting Congress.
For More Information
"Effects of Aluminum Fill Temperatures on the Distortion of Chemically Bonded Band Systems," S. Ramrattan, S. Cheah and S. Zandarski, 2004 AFS Transactions, Paper No. 03-154.
"Thermal Distortion of Green Sand and Chemically Bonded Sand at Cast Iron Fill Temperatures," M. B. Krysiak, T. Keener, S. Ramrattan and S. Cheah. 2002 AFS Transactions Paper No. 02-037.
About the Authors
Sayavur Bakhtiyarov and Ruel A. (Tony) Overfelt are professors in the Mechanical Engineering Dept. at Auburn Univ., Auburn, Ala. Chuck H. Sherwin is a senior processing manager at the Citation Corp. facility in Bay Minette, Ala.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Phenolic Urethane Coldbox|
|Date:||Oct 1, 2004|
|Previous Article:||M/G vs. Z: comparing refractory coatings on shell sand systems.|
|Next Article:||History holds keys for future environmental, health & safety issues.|