Feasibility study for rework minimization.
Tiang, Yu An.
Date of Issue2009
School of Mechanical and Aerospace Engineering
Singapore Institute of Manufacturing Technology (SIMTech), Hamworthy Pte. Ltd.
Sand casting is one of the oldest and most prevalent manufacturing methods for producing large and complex metal products today. However, this production method is notorious for its high defect and rework rates. Casting defects can cause serious detriments to the product’s mechanical properties, cosmetic appearance, manufacturing cost and can potentially cause complete service failure. In this study, the defects of C2G100HB marine pump were investigated. Because of the nature of this product, defects cause the pump to leak and rendered the pump useless unless repaired. The rework rate for an eight piece production run of the pump stands at 43.71 hours (or 5.40hrs/piece), an exceptionally high figure by any standards. The primary aim of this study is to reduce this rework figure. This study utilized finite difference modelling (FDM) software - MagmaSoft®, to predict and prevent potential defective sites. The use of this simulation tool enabled a robust, systematic and cost effective way to improve casting quality. The top half of the casting (with respect to casting orientation) was found to exhibit the most severe porosities problem. Numerous porosity sites – mainly in the pump leg to body intersections, were also found. These simulation results were tally with foundry rework statistics and were established to be largely accurate. Metallographic analyses were also carried out and ascertained “oxide films” and “design related shrinkages” as the root cause to porosities. Defect rate for three (out of five) sectors shown improvements. Defects also occurred at regions that were easier to access and repair. On the whole, defect rate went down by 7%. The rework hours after design modifications increased marginally by 0.37%. With the presence of process variability, it may take a few more casting batches before rework hours can be accurately determined. These variations include but not limited to; melt quality, worker’s fatigue, machining sequence and inconsistent reporting and classification of defects.
Final Year Project (FYP)
Nanyang Technological University