As casting geometries become increasingly complex, resin selection will be more important in producing a solid casting. An aluminum foundry was experiencing issues with a dendritic-type shrinkage defect that was exposed during machining. Wanting to lower the scrap rate and not wanting to add another component to the molding package, such as a specialty sand, chill, or refractory coating, testing was carried out to determine how SigmaCure resin percentages on the silica sand, and the type of SigmaCure Part1/Part2 used in the casting process, affected the solidification of an aluminum A316 alloy.
The evaluation occurred on two phenolic urethane cold box systems. Both systems were HA’s SigmaCure phenolic urethane cold box (PUCB) resins. The experimental set-up included analytical testing and actual casting tests to generate cooling curves. The results, in conjunction with the temperature-dependent properties, were used to create custom sand datasets for solidification simulations to replicate the thermal properties of the mold and cores. Simulations were used to see the effect of resin changes before moving forward with sampling at the foundry.
Impact of resin percentage. The first round of testing studied the influence of reducing the resin percentage. Previous research has documented that polymers have approximately two times higher heat-capacity values than ceramics and metals. For the first experiment using the original PUCB package, a reduction of resin from 1.00% to 0.75% was studied.
The resin-reduced sand mixes were first tested on a Differential Scanning Calorimeter (DSC). This analytical test measured the amount of energy required to increase the temperature of the sample. The DSC showed that the 1.00% sand mix had a lower heat flow than 0.75%, taking less energy to heat up the sand mold, Figure 1.
Reducing the resin percentage to 0.75% resulted in increasing the heat flow peak from 3.3 w/g to 3.6 w/g. Pouring sample molds equipped with thermocouples in the metal and mold showed a similar trend during solidification. Figure 2 and Figure 3 display the temperature curves collected from the molding media.
Thermocouples were placed in the molding media to analyze the heat transfer from the liquid metal into the mold. These curves in Figures 2 and 3 were created by placing a thermocouple one-eighth-inch distance from the mold/metal interface and one-quarter inch from the mold/metal interface in the mold. The curves from the test castings correlated with the DSC results.
There was a change in the rate at which the heat transferred from the liquid aluminum into the sand mold, showing a decrease in the rate at which the heat left the liquid metal and transferred into the mold at 0.75%, resulting in the aluminum remaining in the fluid state longer.
Impact of resin selection. The second round of testing considered modifying one part of the resin package. An alternative Part 2 was implemented to understand how a change in the chemistry of the resin system would affect heat transfer into the mold. Using a Part 2 with a different solvent blend showed a change in heat-flow characteristics of the mold as well. The DSC testing showed an increase in the peak heat flow from 3.3 w/g to 4.0 w/g, Figure 4. The mold temperature curves also showed a change. The influence was not detected readily at the one-eight-inch location, Figure 5. However, as this heat energy is transferred deeper into the mold, Figure 6, the temperature of the sand mold is increasing at a slower rate with the new Part 2. Keeping resin levels the same and changing the formula makeup of the PUCB package influenced heat transfer from the metal to the mold.
Confirmation of results. Throughout the testing process, pouring temperatures were collected along with cooling curves of the A316 alloy by placing a thermocouple directly into midpoint of the casting. The start of solidification and end of solidification were calculated from the cooling curves. The solidification differences can be found in Table 1. When reducing the resin from 1.00% to 0.75% using the original resin package, solidification time was extended by 5.3%. In the casting experiments changing the Part 2, solidification time was extended by 0.76%.
Implementing results in production. The experiment revealed the amount of resin on the sand and resin makeup can change the amount of heat energy being transferred from the liquid metal into the sand mold.
The data from the experiment was implemented into solidification software via a custom sand-mix dataset. These newly developed, sand-mix datasets more accurately simulated the thermal properties of the mold and core. The casting producing scrap was simulated using the custom sand datasets. The simulations revealed a sounder casting could be made by moving away from the original sand mix of 1.00% resin and reducing to 0.75%. The simulation results showed that the critical feed paths to fill the thin-walled casting remained open longer with the adjustment in resin content.
A production trial was carried out adjusting the resin content to a lower percentage. A reduction in shrink related defects was found immediately.