Proposition of new yield criterion for green sand mold and its experimental validation by FEM stress analysis of triaxial compression test
Document Type
Article
Publication Date
9-1-2023
Abstract
Prediction of the deformation during shape casting by FEM (finite element method) thermal stress analysis has been desired. For such prediction, validated constitutive models are essential for both the casting alloy and the mold. In this paper, two fatal defects in the modified Cam-Clay (MCC) model, which has conventionally been employed for a green sand mold, are clarified. The aim of this paper is to concurrently resolve these defects by a new yield criterion. The defects were those that cause an unreal distinct yielding and dilatancy peak on the stress strain curve, which was clarified by comparing the true stress-strain curves obtained in the experiment and the FEM stress analysis of a triaxial compression test. A device capable of measuring the true stress during vertical compression was newly developed for the triaxial experiment. These defects inevitably led the MCC model to overestimate the mold strength by approximately 200% of the experimental value. This paper provides a new yield criterion, which employs an additional yield surface with the Drucker-Prager inherent surface, to rectify these defects of the MCC model. As a result of the stress analysis of the triaxial test with the new criterion, the smooth stress-strain curve with neither a peak nor a distinct yielding was reproduced. The calculated compressive stress traced the experimental value with less than a 15% error. Therefore, the new criterion should remarkably improve the accuracy of calculating the mold strength in the FEM thermal stress analyses during casting when compared to the MCC model.
Keywords
Constitutive modeling, Yield criterion, Finite element method, Green sand mold, Casting
Divisions
mechanical
Funders
Komatsu Ltd.,JST SPRING, Grant Number JPMJSP2128
Publication Title
Journal of Materials Processing Technology
Volume
318
Publisher
Elsevier
Publisher Location
PO BOX 564, 1001 LAUSANNE, SWITZERLAND