Foundrymen Discuss Improving Production, Preventing Penetration.
The Steel Founders' Society of America (SFSA) offered the 235 attendees of 53rd Annual Technical & Operating Conference, held November 4-6 in Chicago, a wealth of technical and practical presentations. In addition to a well-attended hands-on workshop on improving foundry operations, the conference boasted more than 30 speakers on subjects ranging from ferrite measurement and improving steel casting properties to scheduling software, yield improvement and the largest steel casting produced in the U.K.
Presentations focusing on equipment installations, shop-floor efficiency and defect reduction were especially well received. This conference wrap-up covers three such papers that offer insight into the installation of a flow coating operation at West Michigan Steel, production improvements at Durametal and remedying penetration problems in steel castings.
Scott Counselor of West Michigan Steel Foundry Div. (TIC United Corp.), Muskegon, Michigan, discussed how the steel jobbing foundry made the transition to flow coating of molds and cores.
In 1994, the operation decided to eliminate green sand molding and silicate and oil sand coremaking and upgrade to a state-of-the-art phenolic urethane nobake molding and coremaking system. Up to this point, the facility's only nobake molds (produced using a continuous mixer and manual mold manipulator for stripping and closing) were sprayed with two coats of alcohol-based refractory coating, which was torched for drying. Because this process was slow, it was used only for low-production runs. The addition of a molding system capable of producing 42 x 42 x 17-in. molds at 30/hr required a more efficient refractory application method--each mold half had to be handled and coated in less than 1 min. Because of inherent problems with spraying (clogging guns, differences in operators' techniques and mold coverage), flow coating was determined to be the best choice considering productivity and surface quality.
Nobake sand cores and molds contain at least 40% void space, even when compacted to maximum density. The flow coating process produces 0.04-0.06-in. subsurface refractory deposits, which reduce the void space between sand grains to approximately 28% and the size of the pores by a factor of 20.
When flow coating, a mold is presented at an angle to the operator. A coating then is flowed (under modest pressure) over the mold. To be effective, the coating must penetrate the surface of the mold, filling the voids between sand grains and providing a barrier to the metal, and also cover the surface of the mold, enhancing the casting finish while further preventing penetration.
The process requires a method of mixing and dispensing refractory (some type of air-driven mixer coupled with a diaphragm-type pump); a method of manipulating and presenting the mold to the operator (automated equipment); and a means of catching excess refractory as it flows off the mold back into the mixing/dispensing unit. In addition, testing equipment, such as a mud balance and an American Petroleum Institute (API) funnel, must be available.
The mold should be preheated prior to application. In addition, the application wand must be able to deliver a steady stream of coating in a fan-shaped pattern. The coating should be applied from the top of the mold down, allowing the coating farthest from the drainage point to have the maximum time to exit the mold cavity. This will prevent puddles and tear drops from forming (Fig. 1).
A refractory applied through flow coating must act as a gel, remaining in suspension for long periods when in a static mode and acting as a liquid when agitated or pressure is applied with the wand.
Thixotropic additives, surfactants and wetting agents are added to achieve the desired characteristics. Generally the refractory chosen for steel applications is zircon, although bauxite has been used successfully. The refractory also must use a carrier, generally isopropyl alcohol or water, to disperse it over the mold cavity. Alcohol washes are either burned off or oven-dried, while water-based coatings are oven-dried or allowed to evaporate.
The decision was made to use water as the refractory carrier, and initial flow coating trials were performed using an existing molding manipulator, a portable pump system equipped with a flow coating wand, and a plastic children's swimming pool. After several trials, a coating that produced an acceptable casting surface finish was selected.
Installation of the automated molding line, including the flow coating equipment, began in 1995. Flow coating equipment included two infrared core ovens with a slat conveyor, a hydraulic flow coating mold manipulator and a flow coat station to catch, mix and dispense the coating. In production, the mold was removed from the flask by a Rollover and transferred into the first oven, which warmed the surface and helped drive out remaining solvents. The mold then moved to the flow coating station where it was clamped and presented to the operator for coating. The coated mold then was transferred to the second infrared oven for drying.
Two issues that arose were the inability of the oven to dry the mold in the time allotted and contamination of the coating by loose sand. The drying issue was solved with the installation of a gas-fired oven in place of the first coresetting belt. At full production speed, the mold spent 1 mm in the infrared oven and 4 mm in the gas-fired oven. However, the ovens still were unable to dry the molds quickly enough. To combat this problem, a new coating was selected that would allow the mold to be coated and drained within 1 mm.
Because the flow coat tank was an open tank with a large exposed surface area, evaporation also was a problem. The foundry purchased a second enclosed mixing and dispensing tank. The open tank was used to catch the coating as it ran out of the mold, at which point it was pumped through a Y-strainer into the enclosed tank. An automatic compressed air blowing station was incorporated to remove any loose sand that remained as the molds were stripped.
Because the coating is returned to the mixing tank and continuously is being contaminated, strict process controls must be maintained. Contamination by resin, catalyst and parting agents can cause changes in the viscosity characteristics of a coating, such as flocculation (refractory particles separate from the carrier and form clumps). These clumps may fill the voids in the sand surface, preventing subsurface penetration.
In addition, small amounts of water may be absorbed by the sand, and the surfactant level may reduce. When the surfactant level has become so low that it will no longer wet the sand surface, subsurface penetration will no longer occur.
Process control must include using specific gravity to measure water content; using the API funnel to determine viscosity; and checking surfactant level by testing the depth of coating penetration into the sand. At West Michigan Steel, Baume is recorded at least once per hour, specific gravity is recorded at least once every 2 hr, viscosity is recorded and measured daily and a determination of subsurface penetration effectiveness is done periodically throughout the week.
Brett Johnson, Durametal Corp., Muncy, Pennsylvania, discussed projects that have improved the manufacturing infrastructure of his facility by increasing capacity, shortening lead times, improving productivity and reducing material handling effort. "Efforts in facilities that are using or implementing lean production systems usually start with shop floor improvement activities," Johnson said.
Johnson said that to improve shop-floor productivity, Durametal has focused on four activities: foundry process improvement, labor reallocation, plant layout/facilities planning, and strategic capital acquisition/planning.
An example of process improvement at Durametal is its success with a heat treatment rack redesign. As demand for heat-treated product lines increased, the heat treatment department became a bottleneck, leading to long lead times, interrupted production flow and excessive overtime, Johnson said. The solution came from shop floor employees and manufacturing engineers.
Heat treatment furnaces had room for additional refiner plates, however the racks were able to hold limited load sizes of plates and the base of the racks was "overdesigned." The old fixture was retrofitted with two tiers of holding racks, and the racks were made to one-half their previous thickness without jeopardizing strength. This resulted in a rack that weighed the same (1100 lb) but was able to hold twice the weight (Fig. 2).
An example of a labor allocation improvement at Durametal is its finish grinding area, which handles several different refiner plate production lines. Labor requirements and demand are different for these product flows, and management offered some solutions to better operate the system.
Because of the department's haphazard layout, it was difficult to visually determine proper labor allocation, Johnson said. In addition, there was a lack of understanding of capacity differences between operations and each production flow. Little data was taken on throughput of individual operations, and almost no data was gathered on a continual basis, he said. Now, on a daily basis operators fill out a small form with their name, shift operation, total hours at operation and number of pieces completed. This data then is entered at the end of a shift into a spreadsheet database that can be used to determine relative capacity of operations within a production line and the relative capacity between two production lines.
In addition, there was a lack of response in terms of labor allocation to changing demands across the two product flows, according to Johnson. "Very little had been done to monitor changing production demands over time by flow," he said. "Historically, each line was staffed in a certain fashion and left that way. The waste in this approach was realized by comparing the relative volatility of monthly production demands over the past year." This indicated that a modification of staffing levels in each production line could sometimes be desirable. Continual monitoring of production demand by product flow has allowed supervisors to make informed staffing decisions such as when to switch operators from one line to another based on demand, when to hire temporary workers and when to plan overtime work.
Another improvement project was the redesign of Durametal's finishing department. The department had been designed with similar machines being placed together, and over the years, capital decisionmaking had relied upon "where the machine will fit" rather than "where the machine should go" based on product flow, Johnson said. Durametal used an equipment addition as an opportunity to make some significant layout changes that would reduce part travel distances, improve throughput of one production line, create an additional overflow production line with existing equipment, and create a layout that would allow modularity for future expansion.
The redesign was successful, with average lead times for one product line falling 70% and potential shop capacity increasing through the use of redundant equipment in the second overflow line.
What began as a simple project to replace two worn-out shotblast machines became an opportunity to significantly improve production flow and productivity through the rough grinding area. The project evolved to achieve the following objectives:
* conversion from a batch-type process to a continuous cleaning process;
* reduction of man-hours by purchasing equipment capable of cleaning refiner plates with their gating;
* elimination of a forktruck between blasting and rough grinding;
* improved throughput.
With a new blast cleaner and conveyor system, Durametal cut man-hours in half and doubled throughput.
Von Richards, Tri-state Univ., and Ray Monroe, SFSA, covered control of metal penetration in steel casting production. Because penetration is not easily predictable or controllable, the project set out to examine regions where steel penetrates into a section of sand and forms a localized composite mixture of metal and sand.
Penetration is explained as an imbalance of forces in which the net pressure of liquid steel (difference between gas pressure and static and dynamic head pressure) overcomes the surface tension resistance at the mold or core. Momentum is imparted from the elevation in the steel being poured, and head pressure at a particular location is due to the force of gravity on the steel above it (density of the steel times the, height of the steel above it). Penetration of liquid metal into, the mold's porous refractory aggregate is resisted by surface tension of the metal and gas pressure in the pores.
The three mechanisms thought to be responsible for penetration are liquid state/mechanical penetration (head pressure vs. surface tension), chemical reaction penetration (pore size increased and tension reduced due to a reaction at mold/metal interface) and vapor phase penetration (reduction of surface tension through vapor deposition of metallic elements across the interface).
Case studies conducted in foundries, including the application of zircon washes under typical conditions for heavy-section steel castings (Table 1), revealed why some failures occur where they do. Four mechanisms of coating failure were illustrated in the trials: exceeding the coating toughness limitations during drying, core material thermal expansion or thermal shock, oxide eutectic liquid phase hole-drilling, and material handling effects.
Thermal strain from the core, shrinkage during drying and elastic strains from contact stresses due to core handling devices can exceed the strain-energy limit of the coating toughness, causing cracking. Silica imparts the greatest thermal strain. Coatings also shrink as they dry, and at a greater than critical thickness, they tend to form cracks because the elastic strain energy exceeds the critical amount to propagate cracks. In addition, rapid drying or thermal drying can be a problem.
Chemical attack may breach the coating--on cope surfaces, core overhangs or vertical surfaces, reoxidation products rich in manganese or iron oxide can attack the coating by dissolving or "fluxing." The precursor to this coating failure looks like an array of coarse pores with a glassy surface, which suggests oxide liquid formation including iron, manganese and aluminum.
Liquid phase sintering caused by decomposition of zircon grains requires the coating to shrink, but the coating is stuck to a core that is growing through thermal expansion. So instead, the finer pores in the coating shrink and the coarser pores grow, eventually coalescing into a hole large enough for a metal stream to penetrate the core. Thus, a relatively small volume of reoxidation inclusions can bore holes in a refractory coating.
The results of the study indicate that skin formation and rapid solidification prevent penetration in most steel castings. Hot spots that keep steel liquid and heat the mold or core above the steel solidus cause serious penetration.
In areas prone to penetration, coatings are required. Erratic occurrence of penetration is the result of a coating breach. Coatings can be breached by handling damage, excessive coating thickness, cracking while drying, poor mold or core strength, thermal strain from silica expansion or chemical attack from reoxidation inclusions.
Chromite and zircon sand reduce the thermal strain on a coating-reducing the tendency of the coating to fail mechanically. They also speed solidification, limiting the degree of penetration.
Improving coating toughness can provide better performance and reduce the frequency of penetration.
The project also found that:
* large castings are most likely to exhibit core penetration;
* head height tends to aggravate penetration;
* pouring temperature variation has only a small effect on penetration;
* venting can decrease penetration.
Summary of Case Study Conditions at Four Foundries Foundary A B C Sand Silica + Additives; Silica; 3 Screen Olivine; AFS 50-55 AFS 58 AFS 55 LOI N/A 1.0-1.2% 0.5% Binder Proprietary 1.25% Resin; Sodium 28% Co-reactant Silicate Casting Weight 75,000 lb 664 lb 2144-2348 lb Melt Practice Basic Arc Basic Arc Basic Arc Pouring Temp. 2810-2850F 2840-2860F 2620-2680F Coatings 75% Solids; Zircon 80 Be; Magnesite; Zircon-based Chromite 6810 62-67 Be 50 Be Alcohol-based Alloy 0.43% C, 0.85% Mn, 1.1% C, 13% Mn, 0.03% max P, 0.025% 0.75% Si, 0.035% P, max S, 0.6% max Si, 0.4% Cr, 0.025% Al, 0.85% Cr, 0.7% Ni, balance FE 0.47% Mo Foundary D Sand 3 Screen Silica; Average GFN 59 LOI 0.88% Binder Phenolic Urethane Nobake Casting Weight 2573 lb Melt Practice Acid Arc Pouring Temp. 2851-2969F Coatings Spray 75-79 Be % Solids 71.5 [plus or minus] 1.5%; Brush 69 Be Alloy 8630
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Steel Founders' Society of America|
|Comment:||Foundrymen Discuss Improving Production, Preventing Penetration.(Steel Founders' Society of America)|
|Date:||Jan 1, 2000|
|Previous Article:||Outsourcing Casting Cleaning & Finishing: A Cost-Effective Solution.|
|Next Article:||1999 The Year in Review.|