Combine training and SPC for quality improvement.

Combine Training and SPC for Quality Improvement

Process control, SPC, casting simulations and plant quality improvement programs dominated the Engineering Division this year. Presentations demonstrated the importance of combining SPC and other process control programs with employee training and involvement procedures when plant-wide productivity and quality improvements are sought.

Quality Programs

An extensive, award-winning quality program was described by K. Leach, Globe Metallurgical, Inc. Every employee has received SPC training, covering control chart maintenance and interpretation. Documented elements of the quality program include task procedures; job work instructions; critical process variables; product parameters; and failure mode effects analyses.

Employee quality circles and daily control chart audits are other aspects of the program. Globe Metallurgical also has trained its suppliers in SPC and has instituted a certification program for them. According to Leach, "The program has succeeded because top management is committed; the program is multi-disciplinary, involving salaried and hourly personnel in every department and emphasizes prevention instead of detection."

Employee involvement is a big part of the quality program at Hitchcock Industries, Inc. Three types of employee teams are used. They are: part number improvement teams, product group improvement teams and specific process improvement teams. In addition to problem-solving, examples of team achievements cited by Butler include capacity improvement and labor-savings by eliminating non-productive hours; overhead reductions and other cost savings.

SPC

D. McCloskey, Future Systems International, advised low-volume foundries to make use of several SPC techniques suitable to low-volume and short run production environments. He recommends first using the short method for variable gage studies to determine the degree of measurement error.

Once measurements are proven capable, the average in observed measurements for one product and one process can quickly measure the degree of process error. "Traditional first-article inspection is both a poor measure of location (centering) and dispersion (variation) of a process," he said.

"Setting-Up Process" is an SPC method which utilizes control limits and only five samples. Ranges and averages for the samples are calculated and compared to the control limits. "Pre-Control Limits' are used for more aggressive continuous improvement," according to McCloskey. The process is measured for conformance to [+ or -] 2 [Sigma] of the nominal dimension as a pre-control measure.

A "Mini-Capability Study," which utilizes the range of a ten-piece sample, provides the [Sigma] capability of the control limits. The "Simple Test for Normality" is a quick and easy check of a histogram to learn if a process is centered and evenly distributed.

Modeling & Simulations

"Current and Potential Use of Process Modeling for Foundry Process Control" was discussed by T. Piwonka, University of Alabama. He commented on the potential of computational, analog and empirical process models for use in foundry process control. He pointed to the success of empirical models such as expert systems in particular, suggesting that the information generated in an SPC system can be used as an empirical model of the process.

"As the SPC system is refined and more use is made of on-line sensors to monitor the process, the combination of SPC and expert systems has the potential to provide rapid and very accurate real-time process control," Piwonka concluded.

T. A. Phelps, University of Missouri-Columbia, presented a knowledge-based expert system which analyzes internal unsoundness casting defects, choosing between defects caused by geometric shrinkage, gas and mold wall movement. The program has the potential for use detecting other types of defects, including a variety of surface defects.

Two new 3-D mold filling and solidification simulation programs were presented. Y. Nagasaka, Komatsu Ltd., Osaka, Japan demonstrated a system which considers heat transfer phenomena in three dimensions using linear-interpolation, explicit and finite-difference methods. Further, the program incorporates a new method for quantitatively predicting both macroshrinkage and microshrinkage.

A paper by H. J. Lin and W. S. Hwang, National Cheng Kung Univ, Taiwan, showed the development of a 3-D fluid flow analysis system which simulates the filling phenomena of castings. The analysis is based on computational fluid dynamics (CFD) techniques known as Marker-and-Cell (MAC) and Solution Algorithm-Volume of Fluid (SOLA-VOF). Results of comparisons to test castings proved favorable and the technique has been adapted to an inexpensive 32-bit computer workstation.

Sand System Design

The AFS Plant Engineering Committee (1-D) presented "Guidelines for Improved Green Sand Systems." A. Filipic, Casting Equipment, Inc, opened the discussion with the comment, "Green sand systems need updated equipment with attention paid to adherence to recommended sand specifications. This will help improve green sand quality, but it is not enough. Improved methods are also needed."

Filipic stressed the importance of sand cooling, which many foundries omit from their sand system. "Don't underestimate the value of a relatively low, constant sand temperature," he said.

J. Schoen, Kohler Co. discussed the increased importance of sand testing in today's sand systems. "Modern equipment innovations affect system sand properties. New SPC and PLC controls require constant testing," he noted. R. Mrdjenovich, Casting Div/Ford Motor Co. described vibratory deck, rotary drum and vibratory drum shakeouts.

Covering trends in sand system layout, B. McMellon, Vulcan Engineering Co, described new sand launder equipment. The launder provides off-line screening of the system sand, a small batch at a time, to gradually remove the smallest tramp particles from the system.

R. McMillan, Lynchburg Foundry Div, reviewed the evaporative cooling technique, which is dependent on sand/moisture/air contact to remove heat from the sand. An effective sand cooling system should incorporate a lump breaker, surge hopper and automatic metering of moisture addition. An over-the-belt aerator can be used to mix sand and water. A fluidized bed, which blows air through the wet sand, will remove the heat. Some cooling units combine moisture and air additions to provide both aeration and cooling.

R. Smith, Sand Mold Systems, Inc, discussed muller selection factors. For vertical wheel, continuous horizontal wheel batch and high shear batch type mullers, calculations have been made to determine mixing times and horsepower requirements for desired green compression strengths. Grain fineness and sand-to-metal rations are other considerations.

PLCs (programmable logic controllers) work very well in sand system controls, commented W. Schulze, Caterpillar, Inc. A PLC should be able to handle sensor inputs and command outputs to machine controls. But a critical part of such systems is control of properties, he stressed.

Moisture control is one of the most convenient means of automatic control of a sand system. Temperature sensors are used for on-line moisture analysis and automatic addition both at the cooling stage and at the muller. An compactibility controller with on-line PLCs, installed "downstream" from the muller, provides automatic final water adjustment. "Adding state-of-the-art equipment to a sand system can pay for itself in less than one year in lower binder use, reduced scrap and reduced downtime," Schulze concluded.

PHOTO : K. Leach, Globe Metallurgical.