Production of Iron Castings Utilizing a New Generation of Feeding Systems

Greg Loving, Philip Dahlstrom,
Trevor Tackaberry, Tony Midea

ABSTRACT
A new generation of feeding systems has been developed to increase the pattern yield and reduce casting scrap by providing a consistent feed profile combined with ease of application and removal.

The KALMINEX K sleeve/core feeding system design was developed for use in medium-pressure, automated, horizontal, greensand molding lines. The design objectives were for the system to have the feeding performance of an exothermic/insulating insert sleeve, the strength of a ram-up sleeve, and the knock-off characteristics of a spot feeder. The result is a novel feeding system development.

This paper describes the development, optimization and testing of the sleeve and compressor core designs and integration. Foundry trials were conducted at Rochester Metal Products to further optimize the application of this novel feeding system. The casting layout and trial results are also discussed in detail.

INTRODUCTION
There is a distinct set of challenges the foundry encounters when producing iron castings on automated horizontal greensand molding machines. Most castings require the use of feeders to deliver liquid metal to the casting as it solidifies. In some cases, feeding systems such as riser sleeves can deliver the metal more economically than a traditional greensand riser. The application of the riser sleeve plays a major role in this economic equation.
Typically, a “riser bob” makes an impression on the cope face of the mold. The riser sleeve is then inserted into the formed cavity, and the mold is closed. This application requires adequate cope access and dimensionally accurate sleeves and riser bobs.

A different, more robust approach is to ram-up the feeding system during the molding. Sleeves used in this manner must be strong enough to withstand high molding pressures without damage. Cost justification for these types of applications can be difficult, due to the requirement for specialized feeding systems.

Standard insert sleeves can fail in medium pressure ram-up applications. Stronger sleeves are generally required for these applications. Increased sleeve strength is typically achieved by increasing the density of the sleeve material, potentially resulting in reduced insulating capability and increased cost.

The compromises inherent in using standard insert sleeves in ram up applications highlighted an obvious market need for an innovative feeding system design. A carefully engineered project was initiated in response to this market need. The three principal goals were to develop a feeding system: (a) to feed as well as a standard exothermic/insulating insert sleeve, (b) to withstand medium-pressure molding machines during ram-up, and (c) to allow for easy riser removal.

DEVELOPMENT
The design of the sleeve had to be optimized to obtain the necessary strength and feed performance for a standard exothermic/insulating insert sleeve recipe. It was hoped that the strength of the sleeve could be increased not only through the use of a compressor core, but also through a novel design of the sleeve itself. Engineering fundamentals such as strength of materials, mechanical design, and heat transfer were rigorously applied to develop a feeding system capable of serving the identified market.

The development efforts are divided into two sections. The sleeve and core designs were optimized individually. Additional engineering was required to integrate the individually optimized sleeve and core.

Sleeve Design Optimization
The design optimization of the sleeve geometry was carried out using MAGMASOFT casting simulation software. Several sleeve designs were evaluated at multiple wall thicknesses, shapes and height to diameter ratios, and compared against the baseline performance of a standard-shaped 2.5”x3.75” exothermic/insulating insert sleeve. The simulated test casting was a 4” steel cube. Some example results are shown in Figure 1.

Figure 1: Example feed safety margin results for several sleeve designs

The optimized sleeve design is the domed sleeve displayed in Figure 1. This sleeve design is predicted to have an initially flat feed and a broad feed pipe. The thicker walls at the base of the sleeve provide greater insulation in this area. This design slightly outperforms the standard insert sleeve.

Core Design Optimization
Breaker cores used in ram-up applications have several requirements. First, they must not fail during molding, and they should prevent the sleeve from failing as well. Secondly, they must provide a notch or stress concentration near the contact of the riser to allow the riser to be easily removed.

Traditional breaker core technology can meet some of these requirements effectively, but may suffer from breakage and are limited in some applications due to their relatively large contact requirements. Ideally, the core footprint should be minimized to increase applicability. To meet all these requirements, recently developed metal core technology was used.

The metal core is designed to collapse as the mold is compacted. Figure 2 shows a picture of a compressed and uncompressed metal core.

Figure 2: Compressed and uncompressed metal breaker cores

Typically, a flat flange is used to attach the metal breaker core to the sleeve material. As was the case in the sleeve design, it was desired to optimize the design of the metal compressor core for this application. This was accomplished by conducting a stress-strain analysis using Abaqus, a finite element analysis program that models the stress-strain behavior of geometries.

As a baseline for the study, a first generation metal breaker core with a flat flange was modeled. The compression of this core and the stresses imparted to the sleeve material were simulated. Material properties of the core and sleeve were included in the model. Figure 3 illustrates these stress results as a result of core compression.

Figure 3: Stresses in the sleeve during collapse of the metal core

The blue area represents the sleeve material, and the thin gray line is the cross-section of the metal compressor core. The green and yellow colors represent areas of elevated stress in the sleeve material near the inside corner. As the core compresses, the stress is concentrated on the inside corner of the sleeve. This “point stress” could cause the sleeve material to fracture in this area.

For an ideal design, the force would be evenly distributed throughout the entire bottom surface of the sleeve rather than in a localized area. Evenly distributing the force across the bottom surface area reduces the maximum pressure imparted to the sleeve, thus effectively increasing the strength of the design. Because the imparted pressure is equal to the force/area, increasing the area in which the force acts upon is critical in achieving a lower pressure at the base of the sleeve.

A further refinement was to engineer the design such that the forces imparted to the sleeve material resulted in material compression rather than tension or shearing. This can be achieved by changing the flat interface between the core and sleeve to a beveled design. As shown in Figure 4, a 30-degree angle was incorporated in the core/sleeve design.

Figure 4: Stresses in the sleeve with angled design

The stress results in Figure 4 show more evenly distributed stresses on the outer side of the sleeve. It should be noted that the model of the core was drawn with an angle slightly greater than that of the sleeve, and did not touch the sleeve material at all points along the sleeve/core interface. Stresses would have been more evenly distributed had the angle of the core and sleeve been equal. The angled core imparts approximately one half of the stress to the sleeve base as compared to the first generation metal core design.

Integrated Design – KALMINEX K

The simulation analyses run on the sleeve and the compressor core provided a theoretical solution. Solidification modeling showed that the sleeve design having a rounded top and thick walls at the base provided the optimal balance between performance and practicality. Stress simulations of the metal core showed that angling the flange of the core greatly reduce the maximum stress on the base of the sleeve.

A 30-degree bevel was added to the base of the thick-walled sleeve and metal core. This design was thought to provide a good combination between thermal performance and strength. Figure 5 below shows a prototype of this design.

Figure 5: Sleeve/Core Prototype

TESTING
A series of tests were developed and conducted to confirm the compression and feed performance of the concepts developed and optimized using the various simulation tools.

Strength testing at the University of Northern Iowa’s Metal Casting Center (UNI) was done to determine if the new sleeve/core prototype could withstand molding pressures as well as existing ram-up, slurry sleeve technology. Feed performance testing was also conducted to determine if the sleeve could perform as well or better than an existing exothermic/insulating insertable riser sleeve technology.

Strength Testing
Molding tests were conducted to determine the relative strength of the sleeve/core prototype. Molds were made using an adjustable pressure Herman molding machine. Figure 6 shows a picture of the machine.

Figure 6: Herman molding machine used for trial at UNI

A wooden pattern plate was bolted onto a steel carriage assembly. Four separate sleeves/cores were set on the pattern plate, as shown in Figure 7.

Figure 7: Pattern layout: 4 different sleeves prior to molding

After the sleeves were set on the pattern, approximately 2” of sand was riddled into the flask and evenly distributed over the pattern plate. Then the remainder of the flask was filled with sand. At this point, the mold was compressed to a prescribed pressure. The mold was then excavated in order to observe the condition of the sleeves.

Initially, a low pressure was used for the experiment. The pressure was then systematically increased over a defined range to simulate medium pressure horizontal molding applications, as shown in Figure 8.

Sleeve/Core Prototype Existing Ram Up Sleeve

Sleeve/Core Prototype	Existing Ram Up Sleeve

Figure 8: Sleeves after medium molding pressure

The metal core compressed as designed, and both sleeves maintained structural integrity, thus satisfying the strength and molding project requirements.

The sleeves were also tested at molding pressures that exceeded the defined target market and objectives for this project. These tests were run for completeness, and damage occurred in both sleeves at similar high pressures. Figure 9 shows these results.

Sleeve/Core Prototype Existing Ram Up Sleeve

Sleeve/Core Prototype	Existing Ram Up Sleeve

Figure 9: Sleeves after high molding pressure

Feed Testing
Feeding tests were conducted to determine the relative thermal performance of the sleeve/core prototype. A standard 2.5”x3.75” exothermic/insulating insert sleeve with a traditional silica sand breaker core was used as a baseline comparison.

Feeding tests were conducted at UNI. Both insert and ram-up sleeves were tested on several cube sizes. The castings were poured with ductile iron. The cubes and risers were sectioned in half, and the feed safety margin was measured.

The figure below shows a representative sample of sectioned cubes and risers for both insert and ram-up sleeves. In the example shown, the 4.9” cube was the largest casting tested, and was selected to evaluate the maximum performance of the sleeve.

Standard Insert Ram-Up Prototype Ram-Up Prototype

-6.3 safety margin	11.5% safety margin	6.3% safety margin
Standard Insert	Ram-Up Prototype	Ram-Up Prototype

Figure 10: 4.9” cube feeding results

In all cases, the sleeve/core prototype ram-up feeding system performs slightly better than the standard insertable riser sleeve. In addition, the prototype design fed the cubes consistently with no negative safety margins.
Results from the strength and feed tests confirmed that the sleeve/core development objectives were met successfully. Production tooling for sleeve and core was designed and the resultant product will be manufactured with the name KALMINEX K.

PRODUCTION TESTING

After testing was completed in a laboratory-foundry setting, a production facility was sought for conducting further testing. Rochester Metal Products, a world-class iron foundry, was approached for beta testing.

Rochester Metal Products began manufacturing gray iron castings in 1937 as a captive supplier to two manufactures of hand-push lawn mowers, and is currently selling both ductile and gray iron castings to over 250 non-captive customers. At present the foundry employs approximately 365 people and the facility spans more than 200,000 square feet.

The foundry is comprised of two specialized manufacturing areas; the Hunter Molding area is dedicated to the production of Gray Iron castings and the Disamatic Molding area is dedicated to the production of ductile iron castings. Total melt production from both areas is in excess of 80,000 tons per year.

The Hunter Molding area is comprised of four molding lines, a HMP 10 (14” x 19” flask), a HMP 20H, a HMP 20E and a HMP 20D (all 20” x 24” flasks). The iron is melted in two Ajax 1,500 KW Jet Flow 3.0 ton/ hour Electric Channel Induction Furnaces and one 3,000 KW Ajax/Duca 6.0 ton/ hour Electric Channel Induction Furnace. Gray Iron alloys of Class 25, Class 30 and Class 35 are all produced with a per day casting capacity of 70 + Net Tons. The Hunter Molding area produces diverse casting types ranging from Water Pump and Compressor Housing to Pulleys and Bearing Housings. Casting sizes range from 1 lb to 50 lbs.

The Disamatic Molding Facility is comprised of two Disamatic MK5B Molding Machines. The iron is melted in two Brown Boveri 11 ton per hour Coreless Melting Furnaces and is poured using Duca Pressure Pour Furnaces with Selcom automatic pouring controls. Ductile Iron grades of 65/45/12 (as cast), 80/55/06 (as cast) and 60/40/18 (heat treated) are produced with a capacity of 166 + Net Tons of castings per day. Casting types range from Brackets, Clamps, Fittings, Pulley and Housings and have a size range of 1 lb to 40 lbs.

Casting
A bearing housing casting produced in the Hunter molding area was selected for beta testing the new sleeves. This bearing housing is produced in Class 30 gray iron and has a cast weight of 44.08 lbs. The pour weight of the system is 79.08 lbs., resulting in a yield of 56 %.

The molding machine used is a Hunter HMP 20D. (The production target pouring temperature is 2525 +/- 25 deg. F. and the target pouring time is 16 seconds). The production pattern utilizes a pressurized gating and filtration system. The filter is positioned horizontally in the drag mold and utilizes a cope to drag cross-over to allow the metal to flow through the filter. A drag and cope runner bar is utilized to connect the gating system to the casting cavity and to the single greensand riser.

Figure 11 shows the current layout of the casting and rigging.

	Original Drag
	Original Cope

 Figure 11: Original pattern layout

This pattern was modified to accept a KALMINEX K123 feeding system for beta-testing. The drag pattern was modified to include a non-pressurized filtered gating system. A stepped drag runner bar was utilized to ensure equal laminar metal flow through each ingate and into the casting cavity. A locating pin for the sleeve was attached to the cope pattern.

Modified Cope Modified Drag

	Modified Cpe
	Modified Drag

Figure 12: Modified pattern layout including locating pin

During testing, the drag fill and squeeze process was carried out as per normal production molding methods. After the match plate rotated to present the cope, the machine operator placed the feeding system onto the locating pin. Figure 13 shows the sleeve on the cope pattern plate.

Figure 13: KALMINEX K123 sleeve placed on pattern plate

The pattern plate was then shuttled to the right and the cope fill and squeeze process was completed. During the mold squeeze, the design of the compressor core allowed the metal core and sand to compress. This formed a breaker core of densely packed sand and a metal notch. The initial height of the metal core was 1.007”, and the compressed breaker core thickness was measured as 0.69”, or 31% compression. Figure 14 shows the fracture plane from the compressed metal core.

Figure 14: Contact area of metal core

During shakeout the reduced breaker core diameter combined with the unique under cut design of the compressor core notch allowed for easy riser removal. In many instances the riser falls off during shakeout. Figure 15 shows the feeding system after shakeout with core fully compressed. Care was taken to prevent the riser from falling off during shakeout. A sectioned feeder is also presented and shows good quality, flat feed performance.

Figure 15: Riser at casting surface

During the trials several castings were retained for sectioning, and the castings were defect-free with no evidence of primary or secondary shrinkage.

The pour weight for the modified system was 67.0 lbs (reduced from 79.08 lbs) producing a pattern yield 65.8% (increased from 56%). Figure 16 shows a picture of the casting, gating, and riser.

Figure 16: Modified casting layout

The beta-testing trial at Rochester Metal Products confirmed that KALMINEX K could be successfully applied in a production setting. The feeding system met the objectives set forth at the onset of the project. It withstood the ram-up molding pressures, successfully fed the casting, and was easily removed from the casting during shakeout.

CONCLUSIONS
Automated, horizontal, greensand molding machines present a special challenge for feeding systems. In the past, sleeve recipes were altered to make conventional sleeve shapes denser, stronger and more robust. These compromises often resulted in sleeves that did not perform as well and/or were more expensive than their insert sleeve counterparts.

This market need was recognized and a project was designed to develop a sleeve that would have the feed performance of an insulating/exothermic insert sleeve, the strength of a ram-up sleeve and the knock-off characteristics of a spot feeder. The project resulted in a fully engineered sleeve called KALMINEX K that is a truly a novel development, with patent pending.

This new generation of feeding systems has been extensively tested for strength, feed and knock-off performance and has met and/or exceeded all of the design criteria both in laboratory tests and in foundry trials.

ACKNOWLEDGEMENTS
The authors would like to thank Colin Powell from Foseco International, as one of the original inventors of the KALMINEX K concept, for all his assistance and support on this project, including providing metal core prototypes and participation in trial activities. Thank you to Dave Drda (Foseco NorAm) for all his assistance preparing hundreds of sleeve samples for trial activities, and to Todd Steele (Foseco NorAm) for supporting prototyping and production activities with tooling drawings of the sleeves and cores. Keith Sweetman (Foseco Account Executive) was instrumental in liaising with foundry customers and provided critical assistance with trial activities for which we are very grateful. Thank you Keith.

Special thanks to Don Jervis, pattern shop supervisor at RMP, for facilitating pattern changes and assisting with trial activities. Finally, the team recognizes the Hunter production department for all their efforts on this project. Thank you all.

Back to top