Computer-Designed Feeding Systems for Iron Castings

Many iron foundries do not take into account the solidification characteristics of iron when designing feeding systems; feeders for iron castings are designed essentially as feeders for steel castings, resulting in defects in production castings.  Often the suggested remedies for these defects worsen the situation, due to the same lack of understanding.

The design methods are generally quite easy to implement and require only a minimal investment of time. Most of these problems could have been prevented had the foundry engineers applied correct design methodology from the beginning. Spending a short amount of time up front to prevent ongoing problems in foundry production over months or years results in an extremely high return on investment.

Design principles for cast iron

The biggest difference between iron and other alloys is graphite expansion during solidification.  This is significant because in most situations the casting becomes “self-feeding” after the onset of expansion, and no further feeding is required.  The objects of feeding systems for iron castings are 1) to provide feed metal only for contraction of the liquid alloy and 2) the contraction of the solidifying iron prior to the start of expansion. Once expansion begins, a feeding system should control the expansion pressure to ensure that the casting is self-feeding from that point forward.  With shrinking alloys, feed metal must be supplied during the entire solidification time.

Another difference has to do with the “piping” mechanism in the feeder.  Cast irons (particularly ductile iron) do not readily form a solid skin during solidification; the freezing mechanism is often described as “mushy” or “pasty”. This renders atmospheric cores ineffective with these alloys.

For blind feeders to pipe effectively, atmospheric pressure must be able to collapse the weak plastic skin after the internal pressure drops below atmospheric. Once one riser punctures, the pressure is equalized so there is no longer a higher external pressure to cause other feeders to pipe.  This means that only one feeder should be used on each “feeding zone” in an iron casting; if multiple feeders are used, one feeder will begin piping while the others will not.  Often, porosity will be seen at the contact point of non-piping feeders.

The requirement for a single feeder within a single zone of the casting is the rule that is violated most often in iron foundries.  Designs where two or more feeders are feeding the same zone a casting results in porosity, often at the contact point of one of the feeders.  The tendency of many foundry engineers to add more feeders to try to resolve the porosity issue is, in fact, exactly the wrong approach and will worsen the situation.

To correctly design a feeder system for iron castings, it is necessary to analyze the cast shape and determine the location and size of feed zones in the casting. To make this determination, we use the Transfer Modulus.

Feed zones are defined by knowing where it is possible for liquid metal to flow from one point to another in response to expansion pressures.  If metal cannot flow from one area of the casting to another, then each of these areas forms a separate feed zone and will require its own feeder (but no more than one.)

Begin with M_c

Analysis of a casting begins with the Casting Modulus (M_c).  This is defined as the volume-to-surface area ratio for various areas of the casting, and has been used for many years to estimate the order of solidification of different parts of the casting.  The Casting Modulus allows us to estimate which part of the casting will solidify first, and which will solidify last.  In steel castings, the modulus of the feeder should be greater than the modulus of the casting.  In iron castings, the Casting Modulus is used to estimate when expansion will begin.

Prior to development of computers and software, calculating M_c was time-consuming and not very accurate; it required the foundry engineer to estimate volumes and surface areas by approximating various parts of the casting with relatively simple shapes. With casting simulation software, solidification of a casting can be predicted in a matter of minutes. The data from this simulation can be converted to Modulus values in the casting.  Modulus data is available at every point in a the casting and the data is more accurate, as time-related effects such as local superheating of the mold material are effectively taken into account by the simulation, which is not possible with manual methods.

With the Modulus data for the casting, as well as the chemistry and temperature data, the point at which expansion begins can be calculated.  The point at which expansion begins is expressed as a percent of full solidification and is referred to as the Shrinkage Time (ST).

Knowing the ST point for a casting, you can calculate the Modulus value at which contraction stops and expansion begins.  This is known as the Transfer Modulus (M_TR), because it defines the areas of the casting where liquid metal transfer is possible.  M_TR is calculated as:
M_TR  =  SQR ( ST /100) * M_C

By plotting the value of M_TR in our simulation, we can determine whether the entire casting is a single feed zone (M_TR is continuous throughout the casting) or whether there are multiple zones (M_TR  is discontinuous).  This allows us to determine the number of required feeders, using the rule of one feeder per feed zone.

M_TR can be understood as the point at which the iron becomes self-feeding due to expansion.  M_TR is critical in designing the feeding system for the casting.  The basic premise for feeding iron castings is that the expansion pressure must be controlled.  Assuming the mold is rigid enough, all contacts with the casting (gates and riser contacts) should be solid enough to ensure that the expansion pressure is contained in the casting.  This leads to another rule:  The Modulus of the feeder contact neck should be equal to M_TR.  This ensures that feeding of the liquid contraction will be able to occur, and that expansion pressure will be contained in the casting due to freezing of the feeder contact at just the correct point in solidification. In-gates should be thin so that they freeze off shortly after filling.

Case study: Ductile iron control arm

Let’s consider the example of ductile iron control arm shown above. The foundry originally approached the feeding design for this iron casting by placing two symmetrical feeders. This was, perhaps, understandable as the two sections to which these feeders were attached are the heaviest sections of the casting.

During initial production of this casting, porosity occurred consistently at one feeder contact. The porosity does not always show at the same contact, but on all castings one contact showed porosity and the other did not.  No acceptable castings were produced with this pattern design.

Next, this casting was analyzed using the approach described previously to determine feeding requirements.  First, an unrigged simulation was performed.  Then, the simulation data was converted to Modulus data so that feeding calculations could be performed.  Figure 4 shows a plot of the areas of highest Modulus in the casting.  From this plot, the foundry engineer might be tempted to conclude that the original feeder design was correct, as there are two areas of high Modulus value in the casting and these are adjacent to the feeder contacts in the original design. 

However, it is necessary to analyze this casting further to determine the Shrinkage Time, and from this the Transfer Modulus (M_TR), in order to understand the location and size of the feeding zones in the casting. 

Analysis of the iron characteristics for this casting indicated that the value of the Transfer Modulus is 0.645 cm.  Plotting this value in the casting indicates the single feed zone.

The areas of higher modulus are connected by a section of the casting in which the Modulus is above the value of M_TR, thus allowing liquid transport for feeding throughout the casting.  This means that only a single feeder should be used on this casting.  With the two-feeder design, both feeders were connected to the same zone of the casting; when this is done, one feeder will pipe and the other feeder will not, resulting in porosity at the contact of the non-piping feeder.

It should be noted that the computer simulation in this case took 16 minutes to perform, and less than 5 minutes for the calculation of ST, M_TR, and plot of simulated casting.  This means that it took just about 20 minutes of analysis to determine the correct feeder design.  Had this been done before the original pattern equipment was created, several months of time involved with production of defective castings would have been avoided. 

After this information was presented to the foundry, the pattern was revised to reflect a single feeder.

Note that the feeder in this case is not connected to the casting at one of the areas of high Modulus.  This illustrates that with iron castings, the location of the feeder is not as critical as in steel castings, due to the expansion pressure that acts throughout the casting once precipitation of graphite begins.

In the final design, the contact area with a single feeder showed no porosity, and no porosity elsewhere in the casting.  Thus, a simple and quick analysis of the casting has produced the correct feeder design for making a sound casting.

Understanding the solidification mechanisms of graphitic iron alloys in terms of expansion/contraction behavior, feeding mechanisms and control of expansion pressure is critical to correct design of feeding systems.  Quick and simple analysis is available that will help the foundry engineer to design the production process correctly at the beginning of production, thereby avoiding the potential for major costs involved in production of defective castings.

Larry E. Smiley is the president of Finite Solutions Inc., the developer of the SolidCast series of PC-based casting modeling software programs.  David C. Schmidt is the vice president of Finite Solutions. Visit www.finitesolutions.com