Temperature and strain rate effects in cold spray investigated by smoothed particle hydrodynamics V. Lemiale a, ⁎, P.C. King b , M. Rudman a,c , M. Prakash a , P.W. Cleary a , M.Z. Jahedi b , S. Gulizia b a CSIRO Computational Informatics, Clayton, Victoria, Australia b CSIRO Materials Science and Engineering, Clayton, Victoria, Australia c Department of Mechanical and Aerospace Engineering, New Horizons Centre, Monash University, Clayton, Victoria, Australia abstract article info Article history: Received 11 December 2013 Accepted in revised form 22 May 2014 Available online 10 June 2014 Keywords: Cold spray Smoothed particle hydrodynamics Strain rate sensitivity Copper Impact deformation Cold spray is a manufacturing process that has proved to be a valuable technique for producing high strength metallic coatings. It has been extensively studied in recent years, both experimentally and computationally. Among the various modelling investigations in the literature, a large number have considered the problem of a single particle impacting a substrate by means of a continuum technique, most commonly the finite element method. These models that have been used previously are generally based on two inadequate assumptions, namely 1) both particle and substrate are assumed to initially be at room temperature, and 2) the experimentally observed increase in strain rate sensitivity of the flow stress at high strain rates is ignored. To investigate the combined effect of temperature and strain rates, a three-dimensional model of a single particle impact with a metallic substrate has been developed using smoothed particle hydrodynamics. This meshless method is ideally suited to the simulation of cold spray as very large material deformations can be readily accommodated, whereas this is often a significant difficulty in mesh based techniques. A Cu-on-Cu impact was considered and quantitative comparisons with experimental cross-sections were conducted. It was found that predictions within 5% of the experimental data could be achieved only when both the softening effect due to the initial thermal field and the hardening effect due to the very high strain rates were included in the model. In contrast ignoring these effects led to results that had as much as 50% variation compared to experiments. The model described in this paper is a robust tool that will enable quantitative predictions of cold spray impacts and help further optimise the process. Crown Copyright © 2014 Published by Elsevier B.V. All rights reserved. 1 . Introduction Cold spray (CS) is a promising deposition and manufacturing technique that has gained much interest in recent years [1]. One of the key advantages of CS compared to other deposition techniques such as thermal spraying is its relatively low operating temperature which substantially reduces the problems associated with thermal stresses, phase changes (e.g. melting) and particle reaction with the treatment environment (e.g. oxidation). CS has often been investigated using computational methods [2–4] as a complementary tool to experiments to understand the fundamental principles of the process and assist with its optimization. In the present paper only the models focussing on single or multiple mechanical impacts of particles onto a substrate will be discussed. This class of models has been used to examine many important aspects in CS such as the interfacial bonding mechanism between particles and substrate [3,5,6], the suitability of various materials for CS [7,8], the formation of nanocrystalline structures upon impact [9], the prediction of the critical velocity for bonding [10] or the effect of process parameters such as the preheating of materials [11]. Table 1 provides a comparison between models presented in the literature and the present work. The majority of published modelling work in CS has been carried out using the commercial finite element (FE) package Abaqus/Explicit both in its Lagrangian (see e.g. [2,3,13,14]) and Arbitrary Lagrangian Eulerian (ALE) forms [6,11,13,14]. The main drawback of the Lagrangian FE is its inability to handle large deformations which result in severe mesh distortion. In CS this problem produces artificially large plastic strains at the particle/substrate interface [23]. This can be mitigated to some extent by ALE adaptive remeshing schemes similar to that available in Abaqus/Explicit however such approach may introduce numerical diffu- sion leading to unphysical results in regions of high strain gradients as observed in CS [14]. Some authors have proposed to use a failure model in conjunction with the above approaches in which the failed elements no longer contribute to the overall solution [13,14]. While potentially effective such strategy adds another level of complexity to the model by introducing new material parameters to calibrate. Other grid based software used in CS rely on the application of Eulerian techniques in which the material flows through a fixed mesh [5,12,23]. As with adaptive meshing schemes, the problem of mesh Surface & Coatings Technology 254 (2014) 121–130 ⁎ Corresponding author. Tel.: +61 3 9545 2980. E-mail address: Vincent.Lemiale@csiro.au (V. Lemiale). http://dx.doi.org/10.1016/j.surfcoat.2014.05.071 0257-8972/Crown Copyright © 2014 Published by Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat