2016 Volume 88 Issue 8 Pages 485-491
This report describes the application of quality engineering for reducing inconsistencies of system sand to attempt the use of data-driven approach (decision-making using data). The influx of core sand was assumed to be one of the factors causing inconsistencies of system sand, and this was set as the noise factor. In addition, the bentonite amount, bentonite type (blended rate), kneading time, coal powder, starch, and wetting agent were selected as control factors to conduct direct product experiments. Experimental results confirmed that bentonite amount, bentonite type, coal powder, and wetting agent are effective for reducing the inconsistencies. Most of these factors are current levels, and the only realistic and modifiable factor was change in bentonite type.
These results indicate that change of bentonite type has little effects for reducing inconsistencies, but conversely, it means that factors for reducing costs without affecting inconsistencies can be determined with this approach.
This theme was conducted in 2009 and reported at the Kansai Branch meeting of Japan Foundry Engineering Society in 2010. The approach has been improved through subsequent efforts and experience, and these shall be reported in the following as well.