Development of New Feeding-Distance Rules Using Casting Simulation: Part I. Methodology

casting yield in a general sense must begin with these rules. THE consequences of a low casting yield are well known Furthermore, casting simulation does not provide the initial in the foundry industry: lower profits due to increased proriser design for a casting, nor does it automatically optimize duction costs and decreased capacity. Additional melted the risering. Thus, even if simulation is used, feeding rules metal and heats, as well as the increased labor and materials are useful to develop a reasonable starting point for simulacosts required for production, are the primary reasons for tion and to shorten the iterative optimization cycle by the increased costs. Furthermore, it is recognized that a providing insight into, for example, the maximum distance higher casting yield has the side benefit of lowering casting between risers. cleaning costs. Essentially, when a foundry achieves the The first rules governing the riser feeding distance were highest possible yield, it can operate at maximum capacity, developed empirically, based on extensive casting trials that maximizing its revenues. were performed in the early 1950s. Most notably, there is Computer simulation of the casting process is becoming the substantial body of experimental work from the Naval an indispensable tool in the effort to achieve increased castResearch Laboratory (NRL) by Pellini, Bishop, Myskowski, ing yield. Simulation enables more experimentation with the et al. for determining feeding distances[3–7] and adequate casting process, since a design can be modified, simulated, riser dimensions.[8] In these studies, castings were considered and analyzed to determine the effects of the modifications, sound if there was no evidence of shrinkage detected by without having to actually produce the casting. In addition, radiography at 1.5 pct sensitivity. Rectangular sections of a simulation enables better use of feeding aids such as sleeves thickness of T 5 1.27, 5.08, and 10.2 cm (0.5, 2, and 4 in.), and chills, which can sometimes be difficult to apply effecwith width-to-thickness ratios of W/T 5 2, 3, 4, and 5, were tively without numerous trial runs in the foundry. However, cast for these studies, as well as bar-shaped cases (“bar”: computer simulation is applied on a case-by-case basis. Also, W/T 5 1) with 15.2 and 20.3 cm (6 and 8 in.) sections. The to be used effectively, simulation requires expertise and accucastings were made of plain-carbon steel of 0.25 to 0.30 wt rate data for numerous process variables. Therefore, in spite pct carbon content, with a pouring temperature of approxiof the power of simulation, a recent survey indicates that mately 1621 8C. In addition, the SFSA conducted compresimulation is used for less than 10 pct of the tonnage of hensive casting trials involving about 50 foundries of steel castings produced.[1,2] The same survey indicated that member companies.[9] In the SFSA studies, the feeding disfeeding rules (or rule-based software) were used to rig about tance required to cast a section to commercial soundness 80 pct of the tonnage produced. This 80 pct was comprised (better than or equal to ASTM class 2 radiographic testing of 50 pct “rules developed in-house;” 20 pct feeding rules (RT) soundness at 2 pct sensitivity)[10,11] was determined for given in the Steel Founders’ Society of America (SFSA) T 5 1.27, 2.54, 5.08, and 10.2 cm (0.5, 1, 2, and 4 in.) plate handbook, Risering Steel Castings; and 10 pct “non-SFSA (“plate”: W/T $ 3) and bar (W/T 5 1) sections. The steel published rules.” Due to the prevalence of rules-based cast in the SFSA studies was reported as a carbon steel