Abstract
Large-scale steel ingots are critical raw materials for heavy equipment manufacturing, yet their quality is significantly compromised by shrinkage defects such as shrinkage pipe, shrinkage cavity and shrinkage porosity. This study investigates the spatial distribution and formation mechanisms of these defects in an 18-ton steel ingot with a large height-to-diameter ratio (height/diameter = 4.23) prepared under vacuum conditions through a combined approach using experimental sectioning and finite element method (FEM) simulations. The ingot was longitudinally sectioned, and its axial plane was analyzed to map shrinkage defect distribution. This revealed a deep shrinkage pipe at the top of the ingot, extending to approximately 25% of the ingot's height. Additionally, severe shrinkage cavity and porosity were observed at the axis of the ingot body. Based on industrial temperature measurements and sectioning analysis results, a validated FEM model was employed to simulate solidification behaviour. Two criteria were established: (1) the critical solid fraction combined with critical feeding distance for predicting shrinkage cavity, and (2) G/R0.5 < 17K1/2s1/2cm−1 (G and R are the temperature gradient and cooling rate, respectively) for predicting shrinkage porosity. Subsequently, the effect of different casting parameters on distribution of shrinkage pipe, shrinkage cavity and shrinkage porosity was explored. The use of an insulating brick significantly reduces shrinkage pipe depth by enhancing liquid feeding efficiency, and the critical height of the insulation brick is recommended to be 600 mm. Pouring temperature and rate slightly affect shrinkage pipe depth, but their impact on centreline shrinkage cavity and porosity distribution is negligible. The cooling intensity of outer surface of mould has minimal influence on the distribution of shrinkage defects. This work provides practical insights into mitigating shrinkage defects in heavy ingots, with implications for improving industrial forging quality.
Keywords
Get full access to this article
View all access options for this article.
