PSI - Issue 54

Xingling Luo et al. / Procedia Structural Integrity 54 (2024) 75–82 Xingling Luo et al. / Structural Integrity Procedia 00 (2023) 000–000

76

2

Relative displacement between two faces of a crack 0 Critical fracture displacement 1. Introduction

Compacted graphite iron (CGI) is known for its excellent mechanical and fatigue strengths as well as competitive price. It is used extensively in petrol and diesel cylinder heads and blocks (Cao et al., 2023; Dawson, 1999). The microstructure of CGI comprises irregularly shaped graphite particles randomly embedded in a metallic matrix. Three main shapes of graphite particles were identified in CGI: nodular, vermicular and flake. At the microscale, fibre matrix debonding and matrix microcracking are the most likely damage modes. There are several studies on the probable damage modes at the microscale, which help to comprehend these phenomena and identify strategies to decrease their detrimental effects (Jivkov, 2018; Naghdinasab et al., 2018). Generally, two approaches are used to model crack initiation and growth in CGI or other two-phase composite materials: direct numerical simulation and micromechanical modelling. In the latter studies, unit-cell models usually consist of a single particle surrounded by the matrix, with the volume fraction of inclusion equivalent to the overall volume fraction in the composite (Papakaliatakis and Karalekas, 2010). Nevertheless, some damage models cannot be used in representative volume element (RVE) models. For example, due to the size problem in the simulation model, the GTN damage model might not be adequate for ductile-failure modelling for ferrite in CGI or two-phase steel (Vajragupta et al., 2012). Based on the statistical concept, a combination of the cohesive finite-element method and the Beremin local criterion, which uses the Weibull stress to describe the probability of fracture, seems to be a good method for investigating the brittle fracture in martensite (Johnson and Cook, 1985). However, this method requires a post-processing algorithm, which makes it hard to investigate the interaction between failure modes. Chen et al. (2022) combined the Johnson-Cook (JC) damage model and cohesive zone elements (CZE) to create a mesoscale numerical model to explore the material removal mechanism. Gad et al. (2021) also combined the JC and CZE models to study the effect of the shape of reinforcement particles. However, JC is more accurate in large-deformation problems (Johnson and Cook, 1985). So, inserting cohesive zone elements globally would get more precise results in a single-inclusion unit-cell problem. Hence, a novel simulation strategy, combining CZE and periodic boundary conditions (PBCs), to investigate the effect of graphite morphology on mechanical behaviour and crack growth is presented in this paper. The cohesive zone elements were assigned with three different sets of properties. This study provides a new strategy to solve problems with varying constituent geometries and simulate the initiation and propagation of cracks without any predefined cracks. The effect of boundary conditions, graphite morphology and orientation are discussed below. 2. Methodology 2.1. Microstructure-based modelling In general, a unit cell of CGI encloses a large number of graphite inclusions embedded in the metallic matrix. The effect of the interaction of graphite inclusions has a significant impact on its fracture behaviour. To study this fracture behaviour systematically, it is crucial to develop a model that separates the effects of individual graphite inclusions, thus eliminating the influence of their interaction. The effect of graphite particles in CGI with a single graphite inclusion was investigated in this work. Perfectly spheroidal-shaped graphite was used in this study since a large portion of graphite included in the material displays a spheroidal morphology with a high circularity (Palkanoglou et al., 2020). In the authors’ previous study (Palkanoglou et al., 2020), scanning electron microscopy was used to characterise the microstructure of CGI specimens and the resulting scans were analysed using image-processing software. The statistical values of geometrical parameters of graphite particles in CGI are listed in Table 1. The dimensions of both graphite particles and metallic matrix are based on Table 1. This finite-element model was developed in Abaqus/Explicit using two assumptions: (i) the metallic matrix was assigned the effective CGI properties, and (ii) the volume fraction of graphite was 8.3%, chosen as the average volume fraction of graphite in Table 1.