Three-dimensional thermal analysis of multi-layer metallic deposition by micro-plasma transferred arc process using finite element simulation

Accepted Manuscript Title: Three-dimensional thermal analysis of multi-layer metallic deposition by micro-plasma transferred arc process using finite element simulation Authors: Sagar H. Nikam, Neelesh K. Jain PII: DOI: Reference:

S0924-0136(17)30218-2 http://dx.doi.org/doi:10.1016/j.jmatprotec.2017.05.043 PROTEC 15251

To appear in:

Journal of Materials Processing Technology

Received date: Revised date: Accepted date:

10-10-2016 26-5-2017 30-5-2017

Please cite this article as: Nikam, Sagar H., Jain, Neelesh K., Three-dimensional thermal analysis of multi-layer metallic deposition by micro-plasma transferred arc process using finite element simulation.Journal of Materials Processing Technology http://dx.doi.org/10.1016/j.jmatprotec.2017.05.043 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Three-dimensional thermal analysis of multi-layer metallic deposition by microplasma transferred arc process using finite element simulation Sagar H. Nikam1, Neelesh K. Jain2 @ 1

Research Scholar; 2Professor, Discipline of Mechanical Engineering

Indian Institute of Technology Indore, Simrol 453 552 (MP) India @

Corresponding author: [email protected]; Phone: +91 731 2438 731; Fax: +91 731 2438 721

Abstract Manufacturing of complex 3D-parts by micro-plasma transferred arc (µ-PTA) powder deposition process involves repeated heating and cooling at the same location causing thermal distortion and residual stresses in the substrate and deposited material. Simulation of temperature distribution, thermal cycles and temperature gradient can be helpful to predict thermal distortion and residual stresses. This paper describes 3D-analysis of temperature distribution and thermal cycles in multi-layer metallic deposition by µ-PTA process by finite element simulation using temperature dependent properties of the deposition material. Analysis was also done to study influence of deposition direction on temperature distribution and temperature gradient in multi-layer metallic deposition. The simulated results were experimentally verified on the µ-PTA process experimental apparatus developed for temperature measurement depositing powder of titanium alloy (Ti-6Al-4V) on substrate of the same material. Good agreement is observed between the simulated and experimental results. The results showed that temperature increases with increase in the deposition height and the temperature gradient in parallel deposition is higher than that in back and forth deposition. This implies that parallel deposition exhibits better heat diffusion than back and forth deposition. This work will be helpful in selection of optimum heat input, process parameters and deposition direction in multi-layer metallic deposition by µ-PTA process. Keywords: Mutli-layer metallic deposition; Micro-plasma; 3D-thermal analysis; Finite element simulation; Ti-alloy; Additive layer manufacturing.

Research highlights: 

3D FES of temperature and thermal cycles in multi-layer deposition by µ-PTA process



Study of effect of deposition direction on temperature and thermal gradient



Experimental verification using Ti-6Al-4V powder deposition on substrate of same material



Parallel deposition showed better heat diffusion than back and forth deposition



Useful in optimizing heat input, process parameters and deposition direction



Will help in minimizing thermal stresses and thermal distortion

1

Nomenclature C Correction factor Cps Specific heat of the substrate material (J/Kg K) E Electric field for micro-plasma arc (V/m) Ees Electric field for substrate material (V/m) hconv Convective heat transfer coefficient (W/m2 K) J Current density of micro-plasma (A/m2 ) Ks Thermal conductivity of the substrate material (W/m K) Ks* Modified thermal conductivity of the substrate material after taking into account the Marangoni effect (W/m K) L Standoff distance between substrate material and torch (mm) n Unit vector normal to the surface of the molten pool in terms of i, j, k P Power of heat source (W) Qlosses Heat loss due to convection and radiation (J) Qp Heat generated by micro-plasma arc (J) Qs Heat required to melt the substrate material (J) q Actual volumetric heat flux density = η q(x,y) (W/m3) q(x,y) Theoretical volumetric heat flux density at a point having coordinates x and y (W/m3) ro Radius of micro-plasma arc (m) T Instantaneous temperature in the molten pool (K) Ti Ambient temperature (K) Tms Melting temperature of the substrate material (K) Time at time at which micro-plasma arc strikes the substrate material (s) t V Micro-plasma arc voltage (volts) εs Emissivity of the substrate material Ƞ Thermal efficiency of micro-plasma transferred arc (%) ρs Density of the substrate material (Kg/m3) σes Electrical conductivity of the substrate material (ohm -1 m-1) σsbc Stefan-Boltzmann constant (5.67 x 10-8 W/m2 K4) 1. Introduction Additive layer manufacturing (ALM) has become an important industrial process having capability of near net-shape manufacturing, adding delicate features to the existing products and repair/remanufacture/refurbish the components used for industrial applications. Generally, high energy beam processes such as laser beam and electron beam or arc-based processes are used for the ALM applications of metallic materials. High energy beam based processes can be used only for low volume deposition with higher accuracy but, these processes are very costly.Though, arc-based deposition processes are cheaper and can be used for bulk deposition of metallic material but these processes may lead to higher dilution, heat affected zone and thermal distortion. Therefore, a wide gap exists between capabilities of high energy beam based and arc-based processes forcing compromise between deposition quality and deposition volume particularly for ALM applications of metallic materials. Efforts have been made by

2

researchers either to combine two different heat sources (Baufeld et al., 2011) simultaneously or to develop a new heat source to fulfill these gaps. Continuing in this direction, efforts have been made to develop a novel technique, referred as micro-plasma transferred arc (µ-PTA) process for ALM applications of metallic materials at IIT Indore, India (Jhavar et al., 2014). This process has potential to produce good quality deposition of metallic materials and at the same being cost-effective, material efficient, energy efficient and environment friendly. Jerby et al. (2015) used localized microwave heating (LMH) for melting bronze-based powder and its stepwise additive deposition to manufacture its rod using ceramic support a tungsten electrode. They also used iron-based powder for magnetic fixation to provide a contact-less means to hold the deposition powder instead of ceramic support. Study of distribution of temperature and thermal stresses, temperature gradient and thermal cycles during metallic deposition is essential in optimizing and selecting the required amount of heat input which helps in prediction of thermal distortion and residual stresses developed in the substrate and deposited material. All these aspects depend on type of substrate and deposition material. Optimizing the required heat input that will result in minimum thermal distortion and thermal stresses and selection of optimum process parameters for each combination of substrate and deposition material will require extensive experimentation which is time consuming, costlier and most of the time infeasible. Simulation of deposition in any ALM process can solve these issues to a large extent. Studies have been conducted in past on multi-layer and multi-material deposition using finite element simulation (FES) to study and predict distribution of temperature and stress, effect of deposition process parameters on them and modification mechanism of properties of deposition material. Zhao et al. (2011) carried out a 3D transient heat transfer simulation and investigated the thermal characters and effects of deposition directions on the thermal process for single pass multi-layer deposition weldbased rapid prototyping. They optimized the deposition parameters and deposition direction through simulation. Dai and Shaw (2001) built a 3D finite element model to study the temperature field in multi-material deposition by laser and effects of laser processing conditions such as deposition direction and scanning rates. They concluded that the temperature distribution and thermal distortion in multi-material deposition are dependent on laser processing conditions and material properties. Roberts et al. (2009) did 3D finite element analysis to study transient temperature field during multi-layer deposition by laser melting of metallic powders. Their study revealed that the thermal interactions between successive layers in multi-layer deposition affect the temperature gradient and that laser scanning of the successive and subsequent layers undergoes rapid thermal cycles producing temperature spikes

3

in previous layers. Kolossov et al. (2004) have done 3D-finite element simulation to predict the temperature distribution in selective laser sintering of titanium powder and studied numerically and experimentally the effect of sintering on temperature evolution. Gan et al. (2004) did finite element based thermo-mechanical analysis to predict residual stresses generated in plasma sprayed coating and observed good agreement between the simulated and experimental results. Alimardani et al. (2007a) developed a transient numerical model to study the temperature distribution and thermal stresses generated during laser solid free form fabrication. Their model investigated the effect of pre-heating and clamping of the substrate to the positioning table. Their results revealed that pre-heating of the substrate material leads to its smooth heating and cooling which causes reduction in thermal stress. Neela and De (2009) analyzed three-dimensional heat transfer to study influence of process parameters on peak temperature field, thermal cycles and melt pool size of the deposited layers using laserengineered net shaping (LENS) process. Their study explained that selection of appropriate values of deposition process parameters result in steady dimensions of the melt pool with good inter-layer bonding. Costa et al. (2005) simulated laser based powder deposition process using finite element analysis to study effects of thermal cycles on various metallurgical phenomena involved. They also studied effects of the substrate dimensions and idle time between two successive deposition layers on microstructure and hardness of the deposition. Their result showed that the thermal properties, mechanical properties and microstructure of the final partsare significantly affected by the process parameters. Tian et al. (2010) modeled the temperature field and stress distribution in the deposited layers by laser sintering. Their stress relieving mechanism explained influence of the process parameters and metallurgical properties on thermal distortion of the deposited layers. Ding et al. (2011) performed transient and steady state thermo-mechanical analysis of multi-layered deposition by wire and arc additive layer manufacturing process. Their study predicted temperature and stress distribution in the substrate and deposition material across the deposited wall and found them to be uniform with less influence of adding of a new successive layer. Matsumoto et al. (2002) did finite element analysis of single metallic layer deposited by selective laser melting to calculate the temperature and stress distribution. Their model considered the shrinkage due to solidification and its effects on cracking during layer forming. Limited work has been done on thermal analysis of metallic deposition by µ-PTA process. Nikam et al. (2016) developed a one dimensional thermal model to predict the deposition width and deposition height in the term of µ-PTA process parameters for single track deposition of metallic materials. The model was used for predicting deposition geometry for any combination

4

of substrate and deposition materials with good agreement between predicted and experimental values. But, no work has been done on three dimensional analysis of temperature distribution and thermal cycles in multi-layer metallic deposition by µ-PTA process. Therefore, objective of present work is to do 3D thermal analysis of temperature distribution and thermal cycles in multi-layer deposition of metallic materials by µ-PTA deposition process using finite element simulation in terms of temperature dependent properties of the deposition material. Since, selection of optimum deposition direction is important to improve the heat diffusion therefore 3D-analysis was also extended for studying the influence of deposition direction on temperature distribution and temperature gradient in multi-layer deposition by µ-PTA process. The simulated results were experimentally verified on the µ-PTA process experimental apparatus developed for temperature measurement depositing powder of titanium alloy (Ti6Al-4V) on substrate of the same material. 2. Finite element Simulation of Multi-layer deposition Three dimensional thermal analysis of multi-layer metallic deposition by µ-PTA process was done using finite element simulation by ANSYS (version 13.0) software ( ©ANSYS Inc., 2010). The following paragraphs describe the simulation approach mentioning the assumptions, energy balance equation, governing equation and the boundary conditions used. 2.1 Assumptions 

Micro-plasma arc travels perpendicular to the substrate material and constant value of stand-off distance is maintained between the torch and substrate material.



At the start of deposition process, the substrate material is considered at the room temperature (i.e. at 298 K) and the boundary conditions are applied to the geometry of the substrate material.



Deposition is done at middle of rectangular shaped substrate material.



Geometry of the deposited layer and size of the molten pool remains constant during deposition.

2.2 Energy balance equation The energy balance equation according to 1stlaw of thermodynamics is (Roberts et al., 2009) 𝑄𝑝 = 𝑄𝑠 + 𝑄𝑙𝑜𝑠𝑠𝑒𝑠

(1)

Where, Qp is the input heat from the micro-plasma source (J); Qs is the heat required to melt the substrate material (J); and Qlosses is the heat loss due to convection and radiation (J). 2.3 Governing equation

5

Since, µ-PTA is a thermal type ALM process consequently heat conduction is dominant mode of heat transfer in this process. The heat conduction equation described by Bejan and Karus (2003) for a moving heat source has been used as the governing equation in the present work to describe conductive heat flow from the substrate material. Consequently, the governing equation can be written as 𝜌𝑠 𝐶𝑝𝑠 (

𝜕𝑇 ) = 𝑞 + 𝛻 ∙ (𝐾𝑠∗ 𝛻𝑇) 𝜕𝑡

(2)

In which, ρs is the density of the substrate material (Kg/m3); Cps is the specific heat of the substrate material (J/Kg K); t is the time at which micro-plasma arc strikes the substrate material (s); q is the actual volumetric heat flux density (W/m3) [i.e = η q(x,y) in which q(x,y) is given by Eq. 5]; Ks* is the modified thermal conductivity of the substrate material taking into account the Marangoni effect (W/m K); and T is instantaneous temperature in the molten pool (K). Though, Eq. 2 can be used to describe most of the thermal deposition processes but some additional details about initial condition, boundary condition, type and model of the heat source, amount of heat input to the substrate material, properties of the substrate material, etc. and assumptions related to them are to be included in the governing equation for a specific deposition process. Following additional details and assumptions have been included for 3Dthermal analysis of multi-layer metallic deposition by µ-PTA process: 

Initial condition: It describes initial temperature of the substrate material i.e. at the start of the deposition process. It has been assumed that the substrate material is at ambient temperature i.e. 𝑇(𝑥, 𝑦, 𝑧, 𝑡) = 𝑇𝑖

(3)

Where, Ti is the ambient temperature of the substrate material. 

Heat input to substrate material: Ellipsoidal heat source model and Gaussian heat source model have been used in most of the metallic deposition processes. Gaussian heat source gives symmetric distribution of heat flux density about its center which has its maximum value as shown in Fig. 1. According to it, theoretical volumetric heat flux density q(x,y) (W/m3) at a point having coordinates x and y with respect to the center of a heat source, which has radius ‘ro’ (m) and power ‘P’ (W), is given by 𝑃 𝑞(𝑥, 𝑦) = 2 𝑒𝑥𝑝 𝜋𝑟𝑜 𝐿

−(𝑥 2+𝑦 2 ) 𝑟𝑜2

(4)

Where, L is the stand-off distance (m) between the substrate material and micro-plasma torch.

6

Fig.1. Distribution of heat flux density for Gaussian heat source. In the present work, Gaussian heat source model has been modified to take into account current density ‘J’ (A/m2) in µ-plasma arc and electric field ‘E’ (V/m). Consequently, theoretical value of Gaussian volumetric heat flux density q(x,y) at a point having coordinates x and y with respect to center of the micro-plasma arc having radius ‘ro’ is given by 𝑞 (𝑥, 𝑦) = 𝐽 𝐸

−(𝑥 2 +𝑦 2 ) 2 𝑒𝑥𝑝 𝑟𝑜

(5)

Here, the micro-plasma current density ‘J’ and electric field for the substrate material ‘Ees’ are given by following equations 𝐽~𝜎𝑒𝑠 𝐸𝑒𝑠

(6)

Due to small value of stand-off distance between substrate material and the microplasma torch used in the µ-PTA process, the electric field for substrate material ‘Ees’ can be considered approximately equal to electric field for µ-plasma arc ‘E’ 𝐸 ≈

𝑉 𝐿

(7)

Where, σes is the electrical conductivity of the substrate material = 5.8 x 10 5 ohm-1 m-1 (NDT resource center, 2014); V is the micro-plasma arc voltage (volts); and L is the standoff distance between substrate material and the micro-plasma torch (m). Since, the microplasma arc passes through the atmosphere before striking to the substrate material therefore, some of its heat flux density is lost to the atmosphere which is referred as radiation losses and remaining amount is used to melt the substrate and deposition materials. Therefore, actual value of volumetric heat flux density ‘q’ (W/m3) can be obtained by multiplying theoretical volumetric heat flux density ‘q(x,y)’ by thermal efficiency of micro-plasma arc ‘ƞ’

7

𝑞 = ƞ 𝑞 (𝑥, 𝑦)

(8)

Dupont and Marder (1995) have mentioned that thermal efficiency for arc based deposition processes lies in the range from 40 to 90%. Therefore, in the present work, 11 simulations was done using values of thermal efficiency at 40%; 45%; 50%; 55%; 60%; 65%; 70%; 75%; 80%; 85% and 90%. Based on these, it was observed that simulation using 40% thermal efficiency over predicted the thermal cycles while simulation done using 90% thermal efficiency under predicted the thermal cycles. Minimum error was observed between the simulation predicted and experimental values of thermal cycles for 60% thermal efficiency. Therefore, this value has been used in the present work. 

Marangoni flow: This is a phenomenon of motion of molten pool fluid due to thermocapillary action in which molten pool fluid having high surface tension forces to move away the molten pool fluid of lower surface tension due to temperature gradient. In the present work, this effect has been considered by modifying thermal conductivity of the substrate material as suggested by Alimardani et al. (2007b). 𝐾𝑠∗ = 𝐶𝐾𝑠 𝑖𝑓 𝑇 > 𝑇𝑚𝑠

(9)

Where, Ks* is the modified thermal conductivity of the substrate material after taking into account Marangoni flow (W/m K); Ks is the thermal conductivity of the substrate material (W/m K); C is the correction factor; T is instantaneous temperature in the molten pool (K); and Tms is the melting temperature of the substrate material. Lampa et al. (1997) have mentioned that the effective thermal conductivity in the presence of thermo-capillary action is at least twice the thermal conductivity of the molten pool. Therefore, value of 2.5 of correction factor has been used in the present work. 

Heat loss from the substrate material: The heat is transferred through conduction to the deposition material and substrate material is lost to the environment after the deposition by convection and radiation from the heated areas of the substrate material. This has been expressed as boundary condition in the present work by equating heat transferred to deposition and substrate materials through conduction (i.e. right hand side of Eq. 8) is equated to the combined loss by convection and radiation also mentioned by Abid and Siddique (2005). It can be described as follows (𝐾𝑠∗ 𝛻𝑇) · 𝑛 =

4 𝜀𝑠 𝜎𝑠𝑏𝑐 (𝑇𝑚𝑠 − 𝑇𝑖4 ) + ℎ𝑐𝑜𝑛𝑣 (𝑇𝑚𝑠 − 𝑇𝑖 )

(10)

Here, εs is the emissivity of the substrate material; σsbc is the Stefan-Boltzmann constant (5.67 x 10-8 W/m2 K4); Tms is melting temperature of the substrate material (K); Ti is the ambient temperature (298 K); hconv is the convective heat transfer coefficient (W/m2 K); n

8

is the unit vector normal to the surface of the molten pool in terms of unit vectors i, j, k along x, y and z axes respectively. 

Temperature-dependent properties of material: In the multi-layer deposition by the µPTA process, the deposition material in the powdered form is added to the molten pool on the substrate material. During the deposition of very first layer temperature-dependent properties of the substrate material are relevant and during deposition of the successive layers, the properties of previously deposited and solidified layers are relevant. Since, substrate and deposition material are same therefore, temperature-dependent properties in bulk form have been considered. The accuracy of simulated results depends primarily on thermo-physical properties of the substrate material. Figure 2 presents the variation of thermal conductivity (Fig. 2a); density (Fig. 2b) and specific heat (Fig. 2c) of the substrate material (i.e. titanium alloy Ti-6Al-4V) with temperature used in the present work. It can be seen from these graphs that these properties exhibit almost linear variation up to melting point of the substrate material.

(a)

(b)

(c)

9

Fig.2.Variation of properties of the substrate material (i.e. titanium alloy Ti-6Al-4V) with temperature: (a) thermal conductivity; (b) density; and (c) specific heat (Mills, 2002). 2.4 Finite element simulation of multi-layer deposition Substrate material of titanium alloy Ti-6Al-4V having rectangular geometry of size 80 mm x 80 mm x 20 mm was used in finite element simulation. It was discretized into 40,640 cubic (having 8-nodes) elements amounting to total 50,767 nodes. This element has only nodal temperature as the active degree of freedom. Edge length of cubic element was kept as 0.62 mm near the deposition area and it gradually increased away from the deposited layer towards the outer surface of substrate as shown in Fig. 3. The motion of the heat source (i.e. microplasma arc) was simulated by applying a volumetric heat flux density on the substrate material for the duration of micro-plasma arc. Based on change in temperature, the thermo-physical properties of substrate material were updated for the appropriate elements and this process was repeated for entire duration of micro-plasma arc. The multi-layer deposition process on the substrate material was simulated using the element “birth and death” technique. Gan et al. (2004) used this simulation technique in predicting the residual stresses of the deposited metallic material using plasma spraying. In this technique, the deposited layer elements have been considered to be present on the substrate material in the death state i.e. the elements do not get added to the overall stiffness value of the matrix. New elements get birth during travel of µ-PTA arc over the substrate material for the time duration of the micro-plasma arc. The nodal temperature distribution for each time duration of deposition is recorded until the element reaches the melting point. On completion of a deposition time duration, micro-plasma arc moves to the next spot and the continuous simulation take place for entire geometry of a deposition layer. After simulating first layer of deposition, the micro-plasma arc moves to new layer and the entire simulation process is repeated until the complete layer is deposited. Figure 4 presents the algorithm used for analyzing the temperature distribution obtained by FES.

10

Fig. 3. Geometry of 3D model used for finite element simulation of multi-layer metallic deposition by µ-PTA process with upper inset showing its enlarged view and lower inset depicting photograph of the actual multi-layer deposition obtained experimentally.

Start Define element type, material properties and other parameters

Building model and meshing Apply thermal boundary condition and heat source for first time step

Activate layer elements for first time step Compute and store nodal temperature Increment time step

End of all time-steps

Yes

End

Activate new elements, move heat No source and update boundary conditions

Fig. 4. Algorithm used in analysis of temperature distribution in multi-layer metallic deposition by µ-PTA process. 3. Experimental validation Figure 5 depicts the schematic of deposition of 8 mm high and 50 mm long straight wall using powdered titanium alloy Ti-6Al-4V on substrate of the same material and having dimensions 80 mm x 80 mm x 20 mm in four layers of equal thickness 2 mm by µ-PTA process. The process parameters used in the experimental validation were same as used in the finite

11

element simulation that were selected considering the quality of the deposition. Their values are: micro-plasma arc power: 400 W; travel speed of the worktable: 190 mm/min.; and powder mass flow rate: 3.5 g/min. The thermal cycles generated during µ-PTA arc heating were measured with three K-type thermocouples placed in the location as shown in Fig. 5. Schematic representation of experimental apparatus used for multi-layer deposition of metallic material by µ-PTA process is depicted in Fig. 6.

Fig. 5. Schematic of four-layer deposition of powdered titanium alloy Ti-6Al-4V on substrate of the same material by µ-PTA process showing deposition direction and location of three Ktype thermocouples.

Fig. 6. Schematic of experimental apparatus used for multi-layer deposition of metallic materials by µ-PTA powder deposition process.

12

4. Results and discussion 4.1 Thermal cycles in multi-layer metallic deposition Figures 7a to Fig. 7c compares the finite element simulated and experimentally obtained thermal cycles in four-layer deposition of powdered titanium alloy Ti-6Al-4V in the back and forth direction on substrate of the same material. The experimental data were acquired by three K-type thermocouples located in the substrate material as shown in Fig. 5. Figure 8 depicts the maximum temperature of each deposition layer recorded by the three thermocouples. It can be seen from Figs. 7 and 8 that 

FE simulated thermal cycles agrees with slight under prediction of thermal cycles as compared to experimental results for all the layers of deposition. Similar trends of thermal cycles were observed by Ding et al. (2011) from temperature field simulation of wire and arc additive layer manufacturing process.



While depositing the 1st layer, the temperature recorded by the thermocouple located near to the starting point of deposition (i.e. TC1 shown in Fig. 7a) reaches nearly a temperature of 456 K whereas, temperature recorded by other two thermocouples (i.e. TC2 and TC3) remained almost at ambient temperature.



After deposition of 1st layer, the temperatures recorded by the three thermocouples are in order of 566 K (by TC3 shown in Fig. 7c); 532 K (by TC2 shown in Fig. 7b) and 456 K (by TC1 shown in Fig. 7a). This increase in peak temperature is due to pre-deposition heat transfer effect on the substrate material.



Temperature drops suddenly after deposition of 1st layer as shown by all three thermocouples as shown in Figs. 7a to 7c. This is due to the fact that while depositing 1 st layer the substrate material is at ambient temperature therefore effect of cooling is higher as compared to effect of heating the substrate material. Deposition of successive layers smoothens the thermal cycles.



Maximum temperature of thermal cycles increases with increase in deposition height after deposition of each successive layer as shown in Fig. 8 i.e. peak temperature of 4 th layer > 3rd layer > 2nd layer > 1st layer of deposition. This can be explained by the facts that the (i) 1st layer is deposited near to the substrate material with the higher heat diffusion hence dissipation of heat from 1st deposition layer to the substrate material is higher. Successive deposition layers gets away from the substrate material hence heat diffusion is reduced and large amount of heat generated from the micro-plasma arc gets accumulated in the successive layers of deposition causing their pre-deposition heating; and that (ii) every

13

deposited layer is subjected to atmospheric cooling which is enough to solidify the deposited layer but not enough to bring deposition to ambient temperature hence leading to the post-deposition heating of the substrate material causing increase in temperature. The combined effect of pre and post deposition heating could result in the distortion and residual stresses in the substrate material on completion of deposition process.

(a)

(b)

(c) Fig.7. Comparison of simulated and experimental thermal cycles recorded by the three thermocouples located at (a) TC1; (b) TC2; and (c) TC3; in multi-layer metallic deposition.

14

Fig. 8. Peak temperature of each deposition layer recorded by different thermocouples in multilayer metallic deposition. 4.2 Temperature distribution in the molten pool of multi-layer metallic deposition Figures 9a to 9d present the finite element simulated temperature distribution within the molten pool for back and forth deposition of 1st, 2nd, 3rd and 4th layer of powdered titanium alloy Ti-6Al-4V on substrate of the same material by µ-PTA process. It can be observed from these figures while depositing, each layer experiences a simultaneous heating and cooling cycles which continue until last layer has been deposited and that during heating and cooling cycles, high temperature area of the molten pool gradually expands as height of the deposition increases. This is due to decrease in heat losses by convection and radiation and increase in heat accumulation in the molten pool zone between previously deposited layers.

(a)

(b)

15

(c)

(d)

Fig. 9. Temperature distribution within molten pool after deposition of (a) 1 st layer; (b) 2nd layer; (c) 3rd layer; and (d) 4th layer, in multi-layer metallic deposition by µ-PTA process. 4.3 Effect of deposition direction on temperature and its gradient in multi-layer metallic deposition Three dimensional finite element simulation was also done to study influence of deposition direction on temperature and temperature gradient in depositing ten layers of powdered titanium alloy Ti-6Al-4V in the parallel and back and forth depositions on substrate of the same material by the µ-PTA process. Figures 10a and 10b show simulated temperatures at the start and end point of deposition after completion of each deposition layer in parallel (Fig. 10b) and back and forth (Fig. 10a) depositions. Figure 11 depicts comparison of temperature gradient, computed at location C along the deposition height (i.e. Z-direction), for each deposition layer in parallel and back and forth depositions. It can be seen in Figs. 10a and 10b that temperatures at start points (shown in red color) and end points (shown in black color) of each deposition layer for parallel deposition are lower than the corresponding temperatures in the back and forth deposition i.e. at the end of deposition of 2nd layer, the temperatures at start and end points in parallel deposition are 878 K and 1021 K respectively, while their values in back forth deposition are 906 K and 1064 K. This implies that the heat diffusion conditions for the parallel deposition direction are better than the back and forth deposition direction. It can be observed from Fig. 11 that with increase in the deposition height the temperature gradient decreases due to reduced heat diffusion conditions. The 1 st deposition layer has very high temperature gradient as compared to other deposition layers. Temperature gradient becomes almost constant after 5th deposition layer. Temperature gradient in parallel deposition direction is found to have higher value (i.e. 3.10 x 105 K per meter) as compared to its value for the back and forth deposition (i.e. 3.05 x 105 K per meter). Therefore, it can be concluded that

16

temperature gradient at different deposition height and the heat diffusion condition in parallel deposition direction are better than that in the back and forth deposition direction.

(a)

(b)

Fig. 10. Simulated temperatures computed at the end of each deposition layer in multi-layer metallic deposition using (a) parallel; and (b) back and forth, deposition direction.

Fig. 11. Comparison of temperature gradient along deposition height (i.e. Z-direction) for each deposited layer in multi-layer metallic deposition using parallel and back and forth depositions. 5. Conclusions This paper described three-dimensional finite element simulation of temperature and thermal cycles in four-layer metallic deposition and study of influence of parallel and back and forth deposition directions on temperature and temperature gradient of each deposition layer in

17

ten-layer metallic deposition by the µ-PTA process. The simulation has taken into account the temperature dependence of thermo-physical properties of the substrate material, combined heat loses due to convection and radiation and Marangoni flow. The simulated results were validated experimentally depositing multi-layers of powdered titanium alloy Ti-6Al-4V on substrate of the same material using the experimental apparatus of µ-PTA powder deposition process with three K-type thermocouples placed in the substrate material. The simulation results slightly under-predict the thermal cycles as compared to experimentally recorded thermal cycles. Following conclusions can be drawn from the present work: 

Thermal model developed in present study can be used for any combination of metallic deposition and substrate materials and for any form of deposition material (i.e. wire or powdered) in optimizing heat input, process parameters and deposition direction which minimize thermal stresses and thermal distortion.



While depositing the 1st layer, the temperature near the starting point of deposition increases to a very high temperature whereas, towards the end of deposition it remains almost at ambient temperature. This results in sudden change in the temperature. During deposition of successive layers, the temperature towards the end point increases due to predeposition heat transfer effect on the substrate material leading the thermal cycles to become smoother.



Maximum temperature of thermal cycles increases with increase in deposition height after deposition of each successive layer i.e. peak temperature of 4 th layer > 3rd layer > 2nd layer > 1st layer of deposition. This is due to pre-deposition and post-deposition heat transfer effect on the deposited layer. Combined effect of this leads to thermal distortion and thermal stresses in the substrate material.



Each deposition layer experiences a simultaneous heating and cooling cycles which continues until last layer has been deposited. Increase in the deposition height leads to increase in temperature with maximum temperature at the topmost layer causing expansion in temperature area of molten pool.



Temperatures at start and end points of each deposition layer for parallel deposition are lower than the corresponding temperatures in the back and forth deposition implying that the heat diffusion conditions for the parallel deposition are better than the back and forth deposition.

18



Parallel deposition has higher temperature gradient than that of back and forth deposition implying that the heat diffusion conditions in parallel deposition direction are better than that in the back and forth deposition direction.



Temperature gradient of deposited layer decreases with increase in deposition height due to reduced heat diffusion conditions. The 1st deposition layer has very high temperature gradient as compared to other deposition layers. Temperature gradient becomes almost constant after 5th deposition layer.

References Abid, M., Siddique, M., 2005. Numerical simulation to study the effect of tack welds and root gap on welding deformations and residual stresses of a pipe-flange joint. International Journal of Pressure Vessels and Piping 82, 860-871. Alimardani, M., Toyserkani, E., Huissoon, J.P., 2007a. A 3D dynamic numerical approach for temperature and thermal stress distributions in multilayer laser solid freeform fabrication process. Optics and Lasers in Engineering 45, 1115-1130. Alimardani, M., Toyserkani, E., Huissoon, J.P., 2007b. Three-dimensional numerical approach for geometrical prediction of multilayer laser solid freeform fabrication process. Journal of Laser Applications 19 (1), 14-25. ANSYS13.0, 2010© ANSYS, Inc. Canonsburg, Pennsylvania (USA). Baufeld, B., Brandl, E., Biest, O., 2011. Wire based additive layer manufacturing: Comparison of microstructure and mechanical properties of Ti–6Al–4V components fabricated by laser-beam deposition and shaped metal deposition. Journal of Materials Processing Technology 211, 1146-1158. Bejan, A., Kraus, A.D., 2003. Heat Transfer Handbook, John Wiley & Sons, Inc., New Jersey, pp. 1480. Costa, L., Vilar, R., Reti, T., Deus, A.M., 2005. Rapid tooling by laser powder deposition: Process simulation using finite element analysis. Acta Materialia 53, 3987-3999. Dai, K., Shaw, L., 2001. Thermal and stress modeling of multi-material laser processing. Acta Materialia 49, 4171-4181. Ding, J., Colegrove, P., Mehnen, J., Ganguly, S., Almeida, P.M.S., Wang, F., Williams, S., 2011.Thermo-mechanical analysis of wire and arc additive layer manufacturing process on large multi-layer parts. Computational Materials Science 50, 3315-3322. Dupont, J.N., Marder, A.R., 1995. Thermal efficiency of arc welding processes. Welding Research Supplement. 406-416.

19

Gan, Z., Ng, H.W., Devasenapathi, A., 2004. Deposition-induced residual stresses in plasmasprayed coatings. Surface & Coatings Technology 187, 307-319. Jerby, E., Meir, Y., Salzberg, A., Aharoni, E., Levy, A., Torralba, J.P., Cavallini B., 2015. Incremental metal-powder solidification by localized microwave-heating and its potential for additive manufacturing. Additive Manufacturing 6, 53–66. Jhavar, S., Jain, N.K., Paul, C.P., 2014. Development of micro-plasma transferred arc (µ-PTA) wire deposition process for additive layer manufacturing applications. Journal of Materials Processing Technology 214, 1102-1110. Kolossov, S., Boillat, E., Glardon, R., Fischer, P., Locher, M., 2004. 3D FE simulation for temperature evolution in the selective laser sintering process. International Journal of Machine Tools and Manufacture 44, 117-123. Lampa, C., Kaplan, A.F.H., Powell, J., Magnusson, C., 1997. An analytical thermodynamic model of laser welding. Journal of Physics D: Applied Physics 30, 1293-1299. Matsumoto, M., Shiomi, M., Osakada, K., Abe, F., 2002. Finite element analysis of single layer forming on metallic powder bed in rapid prototyping by selective laser processing. International Journal of Machine Tools and Manufacture 42, 61-67. Mills, K.C., 2002. Recommended values of thermo-physical properties for selected commercial alloys, Woodhead Publishing, Cambridge, pp. 211. NDT resource center, 2014: https://www.nde-ed.org/GeneralResources/MaterialProperties/ ET/Conductivity_Ti.pdf Neela, V., De, A., 2009. Three-dimensional heat transfer analysis of LENS process using finite element method. International Journal of Advanced Manufacturing Technology 45, 935943. Nikam, S.H., Jain, N.K., Jhavar, S., 2016. Thermal modeling of geometry of single-track deposition in micro-plasma transferred arc deposition process. Journal of Materials Processing Technology 230, 121-130. Roberts, I.A., Wang, C.J., Esterlein, R., Stanford, M., Mynors, D.J., 2009. A three-dimensional finite element analysis of the temperature field during laser melting of metal powders in additive layer manufacturing. International Journal of Machine Tools and Manufacture 49, 916-923. Tian, X., Sun, B., Heinrich, J.G., Li, D., 2010. Stress relief mechanism in layer-wise laser directly sintered porcelain ceramics. Materials Science and Engineering A 527, 16951703.

20

Zhao, H., Zhang, G., Yin, Z., Wu, L., 2011. A 3D dynamic analysis of thermal behavior during single-pass multi-layer weld-based rapid prototyping. Journal of Materials Processing Technology 211, 488-495.

21

List of Captions of Figures Fig.1. Distribution of heat flux density for Gaussian heat source. Fig. 2. Variation of properties of the substrate material (i.e. titanium alloy Ti-6Al-4V) with temperature: (a) thermal conductivity; (b) density; and (c) specific heat (Mills, 2002). Fig. 3. Geometry of 3D model used for finite element simulation of multi-layer metallic deposition by µ-PTA process with upper inset showing its enlarged view and lower inset depicting photograph of the actual multi-layer deposition obtained experimentally. Fig. 4. Algorithm used in analysis of temperature distribution in multi-layer metallic deposition by µ-PTA process. Fig. 5. Schematic of four-layer deposition of powdered titanium alloy Ti-6Al-4V on substrate of the same material by µ-PTA process showing deposition direction and location of three Ktype thermocouples. Fig. 6. Schematic of experimental apparatus used for multi-layer deposition of metallic materials by µ-PTA powder deposition process. Fig. 7. Comparison of simulated and experimental thermal cycles recorded by the three thermocouples located at (a) TC1; (b) TC2; and (c) TC3; in multi-layer metallic deposition. Fig. 8. Peak temperature of each deposition layer recorded by different thermocouples in multilayer metallic deposition. Fig. 9. Temperature distribution within molten pool after deposition of (a) 1st layer; (b) 2nd layer; (c) 3rd layer; and (d) 4th layer, in multi-layer metallic deposition by µ-PTA process. Fig. 10. Simulated temperatures computed at the end of each deposition layer in multi-layer metallic deposition using (a) parallel; and (b) back and forth, deposition direction. Fig. 11. Comparison of temperature gradient along deposition height (i.e. Z-direction) for each deposited layer in multi-layer metallic deposition using parallel and back and forth depositions.

22

Figr-1List

of Figures with captions

Fig.1. Distribution of heat flux density for Gaussian heat source.

23

(a)

(b)

(c) Fig.2.Variation of properties of the substrate material (i.e. titanium alloy Ti-6Al-4V) with temperature: (a) thermal conductivity; (b) density; and (c) specific heat (Mills, 2002).

24

Fig. 3. Geometry of 3D model used for finite element simulation of multi-layer metallic deposition by µ-PTA process with upper inset showing its enlarged view and lower inset depicting photograph of the actual multi-layer deposition obtained experimentally.

25

Start Define element type, material properties and other parameters

Building model and meshing Apply thermal boundary condition and heat source for first time step

Activate layer elements for first time step Compute and store nodal temperature Increment time step

End of all time-steps

Yes

End

Activate new elements, move heat No source and update boundary conditions

26

Fig. 4. Algorithm used in analysis of temperature distribution in multi-layer metallic deposition by µ-PTA process.

Fig. 5. Schematic of four-layer deposition of powdered titanium alloy Ti-6Al-4V on substrate of the same material by µ-PTA process showing deposition direction and location of three Ktype thermocouples.

27

Fig. 6. Schematic of experimental apparatus used for multi-layer deposition of metallic materials by µ-PTA powder deposition process.

28

(a)

(b)

(c) Fig.7. Comparison of simulated and experimental thermal cycles recorded by the three thermocouples located at (a) TC1; (b) TC2; and (c) TC3; in multi-layer metallic deposition.

29

Fig. 8. Peak temperature of each deposition layer recorded by different thermocouples in multilayer metallic deposition.

30

(a)

(b)

(c)

(d)

Fig. 9. Temperature distribution within molten pool after deposition of (a) 1st layer; (b) 2nd layer; (c) 3rd layer; and (d) 4th layer, in multi-layer metallic deposition by µ-PTA process.

31

(a)

(b)

Fig. 10. Simulated temperatures computed at the end of each deposition layer in multi-layer metallic deposition using (a) parallel; and (b) back and forth, deposition direction.

32

Fig. 11. Comparison of temperature gradient along deposition height (i.e. Z-direction) for each deposited layer in multi-layer metallic deposition using parallel and back and forth depositions.

33