Role of heat-flux density and mechanical loading on the microscopic heat-checking of high temperature tool steels under thermal fatigue experiments
Highlights
► Thermal fatigue tests are performed on a X38CrMoV5 steel under various conditions. ► The crack density is independent of the maximum temperature of the thermal cycle. ► A higher heating rate leads to a higher microscopic heat-checking density. ► The saturated crack density is linearly dependent on the maximum heat-flux density. ► The micro-crack initiation life depends on the plastic strain amplitude in the steel.
Introduction
Thermal fatigue (TF) is a major life-limiting factor of several key industrial components, like hot-forming tool (rolling mill rolls [1], [2], moulds for die casting or forging [3], [4], [5], [6], [7], [8]), engine cylinder heads and blocks in automotive industry [9], gas-turbine blades in aircraft engines [10] and pressurized water reactors in nuclear industry [11], [12]. These components, as with the laboratory samples, experience the variation of temperature with time, during start-up (heating) and shutdown (cooling) operations [2], [9] or thermal fluctuations in service [3], [4], [5], [6], [7], [8], [10]. The repetition of these thermal fluctuations induces transient thermal gradients, which result in thermal strains and subsequent stresses by self-constraining of component or sample. It should be remembered that the surface of the tools is also subjected to additional efforts required to form parts. The cyclic thermo-mechanical solicitations are further coupled with reactive environments (such as oxidation in hot forging and/or corrosion in die-casting), which can lead to a dissolution (“wash-out”) of the dies [3], [13].
In high temperature forming processes (e.g. High-pressure Die Casting, hot-stamping or forging), the heat exchange between the molten alloy (or hot part) and the die increases the temperature of the surface. However, the heat-flux density and thus the maximum temperature (Tmax) may vary from one region to another, depending on tool design and dimensions (e.g. small or large), and/or processing conditions [6], [8]. Therefore, the resulting thermo-mechanical stress–strain hysteresis loops differ locally. One can assume that each sector of a tool is a TF specimen experiencing specific test conditions (applied temperatures Tmin and Tmax, transient thermal gradient, heating and cooling rates, etc.). According to the amplitude of the thermo-mechanical solicitation, thermo-plastic or thermo-visco-plastic yielding can occur and lead cracks to initiate and propagate under non-isothermal fatigue conditions. Near geometrical singularities constituting stress raisers (e.g. corners, holes, etc.), a uniaxial thermo-mechanical fatigue state prevails and the amplitude of the solicitation is in general high enough to initiate and propagate very early the first cracks [4], [6]. On planar surfaces, a network of interconnected cracks, commonly named “heat-checking”, is initiated and propagated under multi-axial (or at least bi-axial) thermo-mechanical loading [3], [4], [5], [6], [14]. As the tool is facing cyclic oxidation and/or corrosion reactions, a multi-layered material (oxide or others scales) can be formed. The diffusion of chemical elements (like oxygen, chromium, and iron) changes the composition of the surface and may alter its mechanical properties and strength [7], [13]. The cracking of this “multi-layered” material results from mechanical loading induced by the mismatch of thermal expansion between an oxide layer (with elastic behaviour) and the substrate steel (with an elasto–visco-plastic behaviour) [3], [5], [6], [8]. During hot forming, the “free” thermal expansion of the die surface is (self-)constrained by the bulk, which is at a lower temperature. Therefore, the surface undergoes compressive strains and subsequent stresses [4], [5], [6], [14]. During ejection of parts, the working surface is rapidly cooled, in particular when a cooling system is employed. The “free” contraction of the surface is again constrained by the bulk, that stands now at higher temperatures. Thus, the surface is stretched under tensile thermal stresses. However, as the modulus of elasticity increases when the temperature decreases, if any plastic shielding occurs, its amplitude would be lower than at higher temperature (e.g. Tmax) and can be neglected in many conditions.
Testing of real components under service conditions is expensive. Laboratory in-house TF experiments are generally developed to assess the behaviour of materials [4], [7], [15]. Different unconstrained specimens (with various sizes and geometries) and heating and cooling systems are used. Also, tests are conducted on the basis of a “standard in-house” thermal cycle, i.e. a variation of the temperature vs. time compatible with the design requirements and test facility available [3], [4], [5], [6], [7], [12]. The direct measurement of thermal and mechanical strains in operation conditions is difficult [16]. The thermo-mechanical history seen by the specimen must be calculated by appropriate constitutive laws and an elasto–visco-plastic thermo-mechanical modelling [4], [5], [6], [9]. However, a non-linear thermo-mechanical analysis of industrial tools is often time-consuming, and one can use linear thermo-elastic calculations. We have previously proposed a normalized approach for assessing the heat-transfer and thermo-elastic stresses under transient conditions. It was shown that our laboratory TF test and the data produced can be reliably up-scaled to analyze and/or design tools in HPDC [17].
It is very common and straightforward to use the maximal temperature Tmax as a damage parameter. In fact, the temperature per see is not a cracking criterion. The temperature (and more specifically Tmax) can undoubtedly contribute to form oxidation and corrosion products, and also, by coupling with mechanical loading, initiate and propagate cracks. By considering a uniaxial thermo-elastic case for the sake of simplification, the mechanical strain amplitude (Δεm) and the stress amplitude (Δσm) can be calculated as:where α is the mean thermal expansion, ΔT is the temperature amplitude (ΔT = Tmax − Tmin), and E is the mean modulus of elasticity. Therefore, these are the “mechanical” strains and stresses that must be considered as criteria for crack initiation and propagation [15].
This paper deals with the influence of the TF test conditions on the microscopic heat-checking pattern formed on the surface of a X38CrMoV5 steel specimen. The effects of the initial hardness of the steel, the maximum temperature and the heating rate of the thermal cycle on the crack density and the network morphology, are reported. We have previously shown that the evolution of the microscopic heat-checking density vs. number of cycles is sigmoidal, and the micro-crack density achieves rapidly an asymptotic saturated value in the “stabilized regime” [18]. The saturated heat-checking density is here related to the maximum heat-flux density and to the TF strains and stresses generated in the base steel. Crack propagation aspects and detailed post-mortem cross-sections observations are reported in [6], [19].
Section snippets
Thermal fatigue experiments
Thermal fatigue (TF) tests were performed on a X38CrMoV5 (AISI H11) tool steel, quenched and double tempered to achieve a martensitic structure with either 42 or 47 HRC hardness. The chemical composition is given in Table 1. The hollow cylindrical shaped specimens, with an inner diameter of 10 mm and an outer diameter of 30 mm, have a wall thickness of 10 mm in the useful area of 40 mm length (Fig. 1). The external surface was mechanically polished to a mean arithmetic roughness (Ra) close to 0.02
Effect of the maximum temperature (Tmax) of the thermal cycle
Fig. 4a shows the effect of the maximum temperature (Tmax, investigated in the range between 550 and 685 °C) on the microscopic heat-checking aspect in the “gauge area” of the specimen. It should be specified that the heating period (ht) of these tests increases from 0.9 to 1.6 s by increasing Tmax from 550 to 685 °C. However, the estimated maximum heat-flux density (Φmax) imposed in the mid section of the specimen during the heating period is nearly constant at 4.85 ± 0.1 MW m−2 (Table 2). The
Dependence of the size of the heat-checking cells to heat-flux density
It is usually claimed that there is a critical temperature range ΔTc for the crack initiation of “brittle materials” (elastic rupture behaviour), such as oxide scale [20] or ceramics [21]. Korneta et al. investigated the cracking of thermal shocked ceramics from various maximum temperatures ranging from 300 to 700 °C, and observed that the crack density changes with the temperature range ΔT, when ΔT > ΔTc [21]. However, Fig. 11a shows that the saturated heat-checking density ρsat measured on our
Conclusions
Heat-checking of X38CrMoV5 hot work tool steel was investigated under thermal fatigue experiments. The specimen surface was subjected to cyclic induction heating and natural cooling, under various test conditions. Thermal cycles, with a minimal temperature of 100 °C and a maximal temperature (Tmax) ranging from 550 °C to 685 °C, were performed. The heating time (ht) was also changed for Tmax = 650 °C, leading to various maximum heat-flux densities applied on the specimen surface. The tests were
Acknowledgments
The authors would like to thank the CTIF (French Technical Centre of Foundry), and especially Dr. Hairy, for supporting a part of this activity. The French National Program on Forging (ACRII and Simulforge) and CETIM are also acknowledged for financial support.
References (25)
- et al.
The thermal fatigue behavior and cracking characteristics of hot-rolling material
Mater Sci Eng A
(2007) - et al.
Thermal fatigue cracking of surface engineered hot work tool steels
Surf Coat Technol
(2005) - et al.
Computer modeling and prediction of thermal fatigue cracking in die-casting tooling
Wear
(2004) - et al.
Thermal fatigue properties of differently constructed functionally graded materials aimed for refurbishing of pressure-die-casting dies
Eng Fail Anal
(2012) - et al.
Thermal stresses in aluminum alloy die casting dies
Comput Mater Sci
(2008) - et al.
Failure analysis of a gas turbine blade made of Inconel 738LC alloy
Eng Fail Anal
(2005) - et al.
A study on the evolution of crack networks under thermal fatigue loading
Nucl Eng Des
(2008) - et al.
Crack initiation under thermal fatigue: an overview of CEA experience. Part I: Thermal fatigue appears to be more damaging than uniaxial isothermal fatigue
Int J Fatigue
(2009) - et al.
Image analysis of microscopic crack patterns applied to thermal fatigue heat-checking of high temperature tool steels
Micron
(2013) - et al.
Conditions for the initiation of oxide-scale cracking and spallation
Corros Sci
(1984)
Evaluation of the crack-initiation strain of a Cu–Ni multilayer on a flexible substrate
Scripta Mater
Critical thickness for cracking of Pb(Zr0.53Ti0.47)O3 thin films deposited on Pt/Ti/Si(100) substrates
Acta Mater
Cited by (0)
- 1
Present address: Airbus France, Site de Saint Martin du Touch, 316 Route de Bayonne, 31060 Toulouse Cedex 9, France.
- 2
Present address: NOPSEMA, Level 08 Alluvion Bld., 58 Mounts Bay Road, Perth, 6000 WA, Australia.