Thermo-Mechanical Fatigue of Compacted Graphite Iron in Diesel Engine Components

Abstract

Cast iron components in combustion engines, such as cylinder blocks and heads of trucks, are exposed for long periods of time to elevated temperatures. Moreover, the engines are started and stopped frequently during their operational life, constituting a large number of heating and cooling cycles. In geometrical complex components the sudden heating (starting the engine) and cooling (stopping the engine) lead to thermal gradients and thermal mismatch within the material, resulting in the local development of high stress levels. The many start-stop operations and their associated alternating stress levels can lead to a localized cracking phenomenon known as Thermo-Mechanical Fatigue (TMF). Compacted Graphite Iron (CGI) is a common material of choice for diesel engine cylinder heads of heavy trucks and is a type of graphitic cast iron, consisting of vermicular graphite particles embedded in a metal matrix of pearlite. This material provides a suitable combination of thermal and mechanical properties, satisfying the functional requirements of these engine components. The main aim of this research is to identify and understand the damage micro-mechanisms that control thermo-mechanical fatigue phenomena in cast iron (CGI). The acquired knowledge is of relevance for predicting the lifetime, improving the properties and increasing the reliability of diesel truck cylinder heads. The work of this study can roughly be categorized into three main subjects: (i) Microstructural evolutions of CGI at elevated temperatures, (ii) TMF crack growth characterization and (iii) precise microstructural analysis of the TMF-crack path. Microstructural Evolutions of CGI at Elevated Temperatures In a first series of experiments, time and temperature induced microstructural changes in CGI were characterized, in view of their possible role in the TMF behavior of CGI. During open air annealing of CGI at 420 °C microstructural changes take place in the material, which gave rise to volume expansion and weight increase. The weight increase can be explained by considering the formation of an oxide scale whereas the volume expansion can be attributed to the decomposition of pearlite into ferrite and graphite. It was observed that the atmosphere is of crucial importance in this process. Annealing in an open-air atmosphere produced ten times less volume expansion as compared to annealing in vacuum conditions. Internal oxidation was observed during annealing under atmospheric conditions and the presence of an internal oxidation layer largely inhibited the progress of pearlite decomposition. The observed oxide layers at the internal metal/vapour interface of cavities (left behind by denuded graphite) cause the obstruction of carbon diffusion and thus the suppression of the pearlite decomposition process. In addition it was found that the depth of the oxidized zone near the surface (the oxide penetration depth) was of the same order of magnitude as the eutectic cell size, i.e. the volume in which graphite particles are interconnected. This suggests that the interconnectivity of the graphite has a dominant influence on the kinetics of the oxidation process. The microstructural dependence of tensile and fatigue properties of CGI at room temperature were evaluated by an extended annealing treatment of 720 h at 420 °C. This extended annealing treatment leads to better tensile (increase in yield strength, ductility) and dynamic properties (fatigue lifetime) at room temperature. The variations of mechanical properties were observed both after annealing under atmospheric and vacuum conditions, but were more pronounced after vacuum annealing. This can be explained by the decomposition of the pearlite phase during annealing and the formation of new ferrite at the graphite/metal interface. It is assumed that the ferrite/graphite interface exhibits a stronger bond than the pearlite/graphite interface. As this stronger bond will be better resistant to delamination, it will strengthen the material both in static and dynamic loading. Such effects were far less pronounced in the open-air annealed material, which could be associated with the fact that it was shown that internal oxidation strongly reduced the kinetics of decomposition. TMF Crack Growth Characterization An important part of this study was to measure and analyze the TMF lifetime of CGI. For smooth and notched specimens, the TMF lifetime was measured in TMF tests under total-constraint conditions, with temperatures cycling between 50 °C and 420 °C. By considering the notch depth as the initial crack length, TMF lifetimes were reproduced numerically using the Paris equation for fatigue crack growth (da/dN = C (?K)m). The calculated lifetimes were found to be in good agreement with all experimental results, covering a wide range of TMF lifetimes from 30 to 1400 cycles. Also for smooth specimens the Paris model worked well by considering the typical graphite particle size as notch depth. It is one of the main conclusions of this work that graphite particles act as internal notches from which a TMF crack almost immediately starts to grow during the first TMF cycles. Hence, it was established that TMF lifetime in CGI is governed by crack growth and not by crack initiation. The relevance of the Paris growth law was further confirmed by meticulously measuring the actual crack growth rates for three typical values of the stress intensity factor. The resulting crack growth rates proved to be in reasonable agreement with the predicted values according to the Paris model. It was further shown that the cyclic plasticity of the bulk material, accumulated during TMF cycles, does not have a noticeable effect on TMF lifetime (i.e. crack growth rates are not affected). The notched dog-bone specimen geometry is proposed in this work as a valid alternative for monitoring the TMF behavior of CGI. By applying standard TMF tests with notched specimens, it was possible to significantly reduce both testing time and experimental data scatter, whilst preserving a realistic estimation of the lifetime of the smooth sample. The effect of prolonged holding times (HT) on TMF lifetime was studied by using notched specimens and a clear effect was observed. Extended holding times were accompanied by an increased relaxation of compressive stresses, causing higher tensile stresses to develop in the subsequent low temperature stages of the TMF cycles. So, extended HTs had an adverse effect on the sample lifetime with a saturating effect for HTs above 1800 s. The Paris fatigue-crack-growth model was used also to estimate the impact of extended HTs. According to the Paris growth law, using a fixed value of tensile stress at low temperature, it was estimated that an increase of holding time from 30 s to 18000 s (5 h) produced a drop of 45% in lifetime. In reality a 60% drop in lifetime was measured, though, which implies that a combined effect of (tensile) stress and microstructural evolution during TMF is responsible for the reduction of lifetime. Precise Microstructural Analysis of the TMF Crack Path To the purpose of precise characterization of the complex TMF-crack-path morphology in CGI in relation to local microstructural features and to find out how and by which mechanisms the cracks predominantly develop, 2D and 3D orientation contrast imaging was carried out on wide field sample volumes, covering several mm3 of imaged material. The data analysis revealed that the crystal planes that are parallel to the (local) crack plane are essentially of a random orientation. Conversely, it was found that graphite particles do not only play a crucial role in the crack initiation, but also are of primary significance for crack propagation. Quantitative analysis of the EBSD data in 2D and 3D showed that the distribution of graphite particles is very important for the crack propagation, as it was revealed that graphite particles enhance crack growth. It was statistically proven that the density of graphite particles in the crack plane is more than double of the density in an arbitrary plane. Our materials knowledge, based on the interpretation of test results in terms of quantifiable microstructural data functions, is of crucial importance to develop a microstructurally based TMF model.