Evaluation of through-thickness residual stresses and microstructure in SA516 Gr. 70 steel welds

Abstract

The critical working condition of nuclear power plant equipment necessitates meticulous determination of the welding process and parameters. In this work, some major influential factors of welding were investigated to observe their effects on the through-thickness residual stress distribution in multipass pressure vessel steel welds. In this regard, experiments were conducted to find the characteristics of residual stresses dispersed in SA516 Grade 70 steel welds of different groove geometries with distinct welding conditions. Three-dimensional finite element models of the weldments were developed considering a moving heat source with temperature-dependent material properties to simulate the welding thermal cycles and corresponding residual stress fields. Effects of weld groove geometry, number of weld passes, external constraints, and preheating on the through-thickness residual stress fields were studied. Additional attention was given to the evaluation of the heterogeneous microstructure and microhardness across the weld cross-section associated with their weld thermal history. Finally, the evolution of the through-thickness residual stresses attributed to subsequent weld passes was elaborated.

Keywords

Welding simulation residual stress thermo-mechanical analysis deep hole drilling thick multipass welding

Introduction

The safety aspect of pressure vessels in power generation industries requires a detailed quality inspection to avoid disastrous failure under high pressure and critical operating environment. The deterioration mechanism of powerplant machinery gets expedited with weld-induced residual stresses.¹ The tensile nature of residual stresses at the weld and heat-affected zone (HAZ) can accelerate the propagation of fatigue cracks under service conditions. Determination of residual stresses induced in welded joints of pressure vessels has been prioritized to enhance the durability of the structures. Difficulties arise in measuring the through-thickness residual stress magnitudes in significantly thicker components like reactors, steam pipes in boilers, etc., where the evaluation of residual stresses using non-destructive approaches are not always feasible. Application of destructive and semi-destructive techniques of residual stress measurement leads to repair welding or complete rejection of the component after evaluation. Finite element (FE) modeling of welded joints is trending now to overcome this problem.^2–5 The generalization of the factor of safety in pressure vessel design codes to account for the effect of residual stresses may not always be sufficient to maximize the load-bearing efficiency. Modifications in the weld groove geometries and process guidelines can revamp the existing data of residual stress distributions in the standardized pressure vessels and piping designs.

Residual stress fields strongly depend upon a few influential welding conditions, which can alter the distribution of stresses across the welded joints.^6–9 Leggatt¹⁰ mentioned influential factors affecting the residual stress profiles, which include material properties, residual stresses present prior to the welding operation due to the fabrication processes, weld geometries, restraint conditions, welding process being applied, and post welding employed thermal or mechanical operations. Mousavi and Miresmaeili¹¹ investigated the intensity of residual stresses present in different weld groove geometries. It was found that the welded joint of U-groove geometry exhibited a lower magnitude of both longitudinal and transverse residual stresses compared to the V-groove welds. Moreover, in a study of groove angles, the 50° angle of the V-groove exhibited the minimum amount of residual stresses. Sattari-Far and Farahani¹² developed a FE model of the pipe weld and validated it with the results obtained experimentally using the blind hole drilling method. The FE model was further used to investigate the effects of weld groove geometries and the number of weld beads on the residual stresses. It was observed that the axial tensile residual stress increased on the inner surface of thick pipe welds with an increasing number of weld passes. Chen et al.¹³ investigated the effect of the weld bead sequence on the welding residual stresses and distortion of stiffened plates. Six welding sequences were analyzed, and significant differences in welding residual stress distributions and distortions were observed. Javadi¹⁴ studied the effects of clamping on the welded joints of a nickel-based alloy and employed both numerical and experimental approaches to study the residual stress distribution. Longitudinal residual stress was observed to be higher with comparatively lower values of shrinkage in the clamped plate compared to the plate welded without clamping. Teng and Lin¹⁵ predicted the residual stress distribution in a single-pass butt-joint steel plate using ANSYS FE software. It was observed that the preheating temperature can effectively decrease the residual stress magnitudes in the weldment. Mitra et al.¹⁶ analyzed the effect of post-weld heat treatment (PWHT) on the residual stress distribution in a narrow gap weld of ASTM A516 Grade 70 steel plates. Egea et al.¹⁷ investigated the application of ball-burnishing process to induce in-depth compressive residual stresses in friction stir welded specimens of 2050 aluminium alloy. The kind of welding process is applied to a particular joint also influences the magnitude and distribution of residual stresses. Colegrove et al.¹⁸ performed six different welding procedures on 4-mm-thick ASTM A131 steel plates to find the effect of welding processes on the resultant residual stress and distortion. The pulsed gas metal arc welding (GMAW) process and hybrid laser process exhibited the best outcomes. Balasubramanian and Guha¹⁹ compared the shielded metal arc welding (SMAW) process and flux-cored arc welding (FCAW) process on the fatigue performance of welded joints. It was found that the fatigue crack growth rate is significantly less in welded joints produced by the SMAW process. Prajapati et al.^20,21 investigated the hybridization of weld filler wire in the course of gas metal arc welding of SA516 Grade 70 welds. The application of hybrid fillers exhibited enhanced mechanical properties and significantly reduced the angular distortion.

Most of the past works of the literature illustrating the influential factors affecting the welding residual stresses largely comprise only the surface residual stress evaluation. However, the magnitude and trend of weld-induced residual stresses may vary through the thickness of the specimen in thick multipass weld joints. In the present investigation, thick plates of SA516 Grade 70 steel were welded under different welding conditions to study the effects of weld groove geometry and the number of weld passes on the through-thickness residual stress distribution and angular distortion. Three-dimensional FE models were prepared and compared with the experimental results for better accuracy. One of the validated FE models was then used to study the effects of external constraints and preheating on the trend of residual stresses. Characteristics of inhomogeneous microstructure and microhardness dispersed across the weldments were studied with their corresponding thermal cycle. Lastly, the evolution of through-thickness residual stress fields was explored using the FE model.

Numerical modeling

In this present investigation, the thermal elastic-plastic FE analysis was performed using Abaqus code and Fortran User Subroutine linked together to predict the welding thermal cycles and corresponding residual stresses and distortion. In these numerical models, temperature-dependent material properties, along with material hardening model, moving welding heat source, and boundary conditions were defined. The temperature-dependent material properties are represented in Figure 1 and Table 1.²² The density of the material was kept constant as 7850 kg/m³.²³

Table 1.

Temperature-dependent yield strength.²¹

Temperature (°C)	20	100	300	500	700	800	1000	1200	1400
Yield stress (MPa)	398	379	305	192	41	36	28	20	12

Figure 1.

Temperature-dependent material properties.²¹

Coupled temperature displacement analysis

Three FE weld models were prepared, and the dimensions were kept same as of the experimental specimens. The sequence of each weld bead and its dimensions was in accordance with the macrograph of the weld cross-section, as shown in Figure 2. Fine meshing order was defined in the weld zone, and a biased meshing scheme was employed for the region away from the fusion boundary to reduce the total computational time and maintain a steep temperature gradient towards the weld centerline. Weld filler material deposition was simulated following the element birth and death technique to activate and deactivate the weld beads according to the respective deposition sequence. During the first weld bead deposition, all other subsequent weld passes were kept deactivated, and only the selective elements of the first pass were activated. After the activation of elements, the moving heat source was traversed from one end to another following the actual weld speed. The continuation of weld metal deposition followed by heating and cooling was divided into three steps. Cooling time after every successful application of heat flux was introduced to maintain the inter-pass temperature around 100°C–180°C.^24,25 C3D8T: 8-node thermally coupled brick, trilinear displacement and temperature elements were opted for the present FE modeling.²⁶ It satisfies the mechanical and thermal equilibrium equations simultaneously. The mechanical and thermal equilibrium conditions are given in equations (1) and (2), respectively.²⁷

div σ + f_{v} = ρ \overset{\cdot\cdot}{u}

(1)

k \nabla^{2} T - ρ c_{p} \overset{\cdot}{T} + \overset{\cdot}{q} = 0

(2)

where $σ$ is the Cauchy stress tensor, $f_{v}$ is the body forces, $\overset{\cdot\cdot}{u}$ is acceleration, $T$ is temperature, $ρ$ is the density of the material, $k$ is thermal conductivity, $c_{p}$ is the thermal capacity, and $\overset{\cdot}{q}$ stands for the heat flux. Both of these equilibrium equations are coupled with each other where stress tensor $σ$ actively depends upon the temperature variable $T$ ensuing the behavioral law.

Figure 2.

Weld bead sequence: (a) conventional groove (10 passes), (b) conventional groove (17 passes), and (c) narrow groove (6 passes).

The present FE model is composed of complex nonlinear interactions between the temperature-dependent material properties and welding process variables. So, the following assumptions were made during the modeling.

The workpiece was considered to be isotropic and homogenous in nature.

The effect of fluid flow due to melting was neglected.

The materials follow von Mises yield criterion and flow rule.

The welded specimens were considered to be stress-free at the initial stage.

The applied heat energy of the moving heat source was much higher compared to the heat dissipated due to plastic deformation. So, the latter one was neglected in the modeling steps.

Thermal analysis

The following equation gives the balanced relation of heat flow in a volume enclosed by an arbitrary surface:

\begin{matrix} - (\frac{\partial R_{x}}{\partial x} + \frac{\partial R_{y}}{\partial y} + \frac{\partial R_{z}}{\partial z}) + Q (x, y, z, t) \\ = ρ c \frac{\partial T (x, y, z, t)}{\partial t} \end{matrix}

(3)

where $R_{x}$ , $R_{y}$ , $R_{z}$ denotes the rate of heat flow per unit area, $T$ stands for current temperature and $Q (x, y, z, t)$ is the instantaneous rate of internal heat generation as a function of time $t$ in the local coordinates $x, y,$ and $z$ . Equation (3) can be further modified by introducing the Fourier heat flow as:

R_{x} = - k_{x} \frac{\partial T}{\partial x}

(4a)

R_{y} = - k_{y} \frac{\partial T}{\partial y}

(4b)

R_{z} = - k_{z} \frac{\partial T}{\partial z}

(4c)

where the thermal conductivities in the $x$ , $y$ , $z$ directions are $k_{x}$ , $k_{y},$ and $k_{z}$ respectively. Equation (3) can be rewritten by applying equation (4(a−c)) into equation (3) as:

ρ c \frac{\partial T}{\partial t} = \frac{\partial}{\partial x} [k_{x} \frac{\partial T}{\partial x}] + \frac{\partial}{\partial y} [k_{y} \frac{\partial T}{\partial y}] + \frac{\partial}{\partial z} [k_{z} \frac{\partial T}{\partial z}] + Q

(5)

Equation (5) is the governing differential heat transfer equation of the present analysis.

The volumetric heat source of uniform density was employed by applying heat flux ( $DFLUX$ ) on the elements of weld beads using the Fortran-based User Subroutine code. The applied volumetric heat flux can be determined as:

Q = UI η / V_{h}

(6)

where $U$ is voltage, $I$ is current, $η$ is arc efficiency and $V_{h}$ is volume of the heat source.²⁸ Arc efficiency was assumed to be 0.85 for the GMAW process.²⁹

The predefined temperature of the base materials was assumed to be the same as room temperature (21°C), and it got modified computationally with the increasing welding steps simultaneously. Convection and radiation were considered for the heat loss where radiation was employed for the elevated temperature fields limited to the weld zone, and convection was assumed for the base material away from the fusion boundary. Linear Newtonian convective cooling can be expressed as:

Q_{c} = - h_{c} (T_{s} - T_{i})

(7)

where $h_{c}$ is heat transfer coefficient assumed as 15 × 10^-6 W/mm²°C,^30,31 $T_{s}$ is the surface temperature and $T_{i}$ is the ambient room temperature (21°C). Radiation is expressed by Stefan-Boltzmann law as:

Q_{r} = - ϵ σ {{(T_{s} + 273.15)}^{4} - {(T_{i} + 273.15)}^{4}}

(8)

where $ϵ$ is emissivity (0.8), and $σ$ is Stefan-Boltzmann constant. Instantaneous amplitude was considered for both cases.

Mechanical analysis

In the mechanical analysis, the temperature history acquired from the multiple thermal cycles induces thermal stresses in the model. The rigid body motion of the model was restrained in terms of displacement. The total strain and strain rate can be broken down into three factors: elastic, plastic, and thermal strain. The sum of total strain rate can be expressed as below:

\overset{\cdot}{ε} = \dot{ε^{e}} + \dot{ε^{p}} + \dot{ε^{t}}

(9)

The influence of creep and solid-state phase transformation was ignored due to the short heating time. The elastic strains were evaluated following the isotropic Hook’s law with the temperature-dependent modulus of elasticity and Poisson’s ratio. The plastic strains were calculated considering a rate-independent plastic model that comprises Von Mises yield surface criterion, temperature-dependent mechanical properties, and isotropic hardening model. The thermal strains were calculated based on temperature-dependent thermal expansion coefficients.

Experimental procedures

The material used for the present investigation is ASME SA516 Grade 70 steel plates with a length of 300 mm, a width of 110 mm, and a thickness of 25 mm. Both conventional and narrow weld groove configurations are represented in Figure 3. Two welds of conventional groove configuration having different number of weld passes and a single weld of narrow groove geometry were prepared. The GMAW process with direct current electrode positive (DCEP) polarity was employed with a root gap of 1.5 mm using ER70S6 as filler material of 1.2 mm diameter. The welding process was executed as per the instructions provided in the American Society of Mechanical Engineers (ASME) Section IX. Initially, bead-on-plate trials were carried out following the parameters of previous literature.³² Suitable welding process parameters were obtained from the trial runs and then were applied to the prepared samples. The welding parameters are given in Table 2. The specimens have been named according to their geometry and the total number of weld passes for ease. CG-26 (10 passes), CG-26 (17 passes), and NG-18 denote the conventional groove joint of 10 passes, conventional groove joint of 17 passes, and narrow groove joint, respectively. A digital dial gauge of 0.01 mm least count was used to measure the angular distortion.

Table 2.

Welding parameters.

Specimen name	Groove geometry	No. of passes	Passes	Current (amp)	Voltage (volt)	Travel speed (cm/min)	Gas flow (l/min)
CG-26 (10 passes)	Conventional	10	Root	180	24	80	10
CG-26 (10 passes)	Conventional	10	Weld	196–208	28	17	14
CG-26 (17 passes)	Conventional	17	Root	180	24	80	10
CG-26 (17 passes)	Conventional	17	Weld	176–190	26	17	14
NG-18	Narrow	6	Root	185	24	80	10
NG-18	Narrow	6	Weld	196–204	28	17	14

Figure 3.

Groove geometries: (a) conventional groove and (b) narrow groove (all measurements are in mm).

Residual stress measurement using deep hole drilling technique

Evaluation of residual stress through mechanical strain relaxation involves the measurement of strains or displacement as a result of elastic unloading due to precisely controlled material removal.^33,34 Residual stress measurement techniques can be classified as non-destructive, semi-destructive, and entirely destructive techniques. All these techniques have their own advantages and limitations. The deep hole drilling (DHD) technique belongs to semi destructive category, and it is utilized to measure through-thickness residual stress fields. After the initial studies were done by Zhdanov and Gonchar³⁵ and Beaney,³⁶ efforts have been made by investigators to improve the accuracy of the conventional DHD technique; such as, increased the number of angles of measurement of the reference hole diameter and avoiding plastic yielding of the core hole.³⁷ DHD technique is notably appropriate for thicker components. Figure 4 shows the steps involved in the DHD technique. In step one, a small reference hole is made at the location of interest. In the present investigation, a reference hole of 5 mm diameter was made. In the second step, the internal diameter of the reference hole is measured through certain depth intervals at specific angles. An air plug gauge was used for the purpose of hole diameter measurement. The diametral measurements were performed at three different angles (0°, 45°, and 90°). In the third step, the core hole is trepanned coaxially up to a certain depth using an electro-discharge machining process. Later on, in the final step, the internal diameter of the reference hole is remeasured at the same previous locations. The change in diameter of the reference hole indicates the deformation caused under the influence of in-plane biaxial residual stresses.

Figure 4.

Schematic representation of steps involved in the DHD technique.

Plane stress condition is followed in the calculation of the DHD technique. The measurement of the radial deformation is carried out at N number of angular directions $(θ = θ_{1}, θ_{2}, θ_{3} \dots \dots \dots θ_{N})$ through specific depth intervals. The initial and final diameters of the reference hole after the trepanning operation are denoted by $d (θ)$ and $d' (θ)$ respectively. The diametral change $(Δ d (θ) = d' (θ) - d (θ))$ is then used to determine the normalized strain $ε (θ)$ for the residual stress fields in $σ_{x}$ , $σ_{y}$ _, $σ_{xy},$ and $σ_{z}$ directions. It can be expressed as below:

\begin{matrix} ε (θ) = \frac{Δ d (θ)}{d (θ)} \\ = - \frac{1}{E} [σ_{x} f (θ) + σ_{y} g (θ) + σ_{xy} h (θ) - μ σ_{z}] \end{matrix}

(10)

ε_{z} = - \frac{1}{E} [- σ_{z} - μ (σ_{y} + σ_{x})]

(11)

where $f (θ) = 1 + 2 \cos 2 θ$ , $g (θ) = 1 - 2 \cos 2 θ$ , $h (θ) = 4 \sin 2 θ$ , and $ε_{z}$ is the calculated axial strain obtained during the material removal process. According to plane stress condition $σ_{z}$ turns to zero. So, equation (10) can be rewritten as:

ε (θ) = \frac{Δ d (θ)}{d (θ)} = - \frac{1}{E} [σ_{x} f (θ) + σ_{y} g (θ) + σ_{xy} h (θ)]

(12)

Here, the $x$ direction matches with $θ = 0^{°}$ angular position. The above equation can be expressed in a matrix form with diametral distortion measurements in N different angular directions as below:

\bar{u} = - \frac{1}{E} M \times σ

(13)

where $E$ is Young’s modulus and

\begin{matrix} \bar{u} = [\begin{matrix} {\bar{u}}_{r} (θ_{1}) \\ ⋮ \\ {\bar{u}}_{r} (θ_{i}) \\ ⋮ \\ {\bar{u}}_{r} (θ_{N}) \end{matrix}], M = [\begin{matrix} f (θ_{1}) & g (θ_{1}) & h (θ_{1}) \\ ⋮ & ⋮ & ⋮ \\ f (θ_{i}) & g (θ_{i}) & h (θ_{i}) \\ ⋮ & ⋮ & ⋮ \\ f (θ_{N}) & g (θ_{N}) & h (θ_{N}) \end{matrix}], \\ and σ = [\begin{matrix} σ_{x} \\ σ_{y} \\ σ_{xy} \end{matrix}] \end{matrix}

(14)

Finally, the unknown magnitude of residual stresses can be found using the below equation:

\hat{σ} = - E M^{*} \times \bar{u}

(15)

where,

M^{*} = (M^{T} \times M)^{- 1} \times M^{T}

(16)

Here, $\hat{σ}$ is the most favorable stress vector that best fits the annular deformation and $M^{*}$ is the pseudo inverse of matrix $M$ .³⁸

Results and discussion

Microstructure

Figure 5 represents the macrographs and microstructure of the three weld samples. The weld samples were polished, followed by etching using nital solution (5% concentrated nitric acid in ethanol), and the microstructure of the different zones were acquired, as shown in Figures 5 to 7. The microstructure of the weld zones were obtained from the final weld passes, as shown in Figure 5. In the weld zone of CG-26 (10 passes) (see Table 2 for the specimen names), columnar elongated ferrite grains were seen along with acicular ferrite (AF) and Widmanstätten ferrite (WF) grains.³⁹ Fine proeutectic grain boundary ferrite (PGBF) and acicular ferrite grains were seen in CG-26 (17 passes) and NG-18. Precipitation of carbides were observed between the ferrite grains in the weld zone of all the weld samples.

Figure 5.

Macrographs and microstructure of the welded samples: (a) conventional groove joint (10 passes), (b) conventional groove joint (17 passes), and (c) narrow groove joint.

The comparison of different microstructures observed at the HAZs of the three weld samples is shown in Figure 6. The HAZ mainly consists of three different microstructural regions, viz., coarse-grained HAZ (CGHAZ), fine-grained HAZ (FGHAZ), and intercritical HAZ (ICHAZ). These subdivisions can be associated with specific temperature brackets. The CGHAZ nearly ranges from 1500°C−1100°C, whereas the FGHAZ and ICHAZ are associated with 1100°C to $A_{C 3}$ (790°C) and $A_{C 3}$ (790°C) to $A_{C 1}$ (670°C), respectively. The CGHAZ adjacent to the fusion zone endures a peak temperature appropriate for both austenitization and grain growth. The fine-grained HAZ (FGHAZ) next to the CGHAZ experiences temperature high enough for austenitization. However, in the FGHAZ, due to a faster cooling rate, substantial grain growth does not occur, which leads to a fine-grained microstructure. The microstructure of the intercritical HAZ (ICHAZ) show ferrite and austenitization due to the endured peak temperature, which lies between the lower ( $A_{C 1}$ ) and upper critical temperature ( $A_{C 3}$ ).⁴⁰ The probable thermal cycles corresponding to different microstructural regions of the HAZs have been presented in addition to the microstructures in Figure 6. These thermal cycles were obtained from certain nodal points of the FE models. Approximate peak temperatures 1300°C, 950°C, and 750°C were selected to signify the CGHAZ, FGHAZ, and ICHAZ, respectively. These thermal cycles indicate the cooling rate by means of change in temperature with respect to time, which significantly influences the evolution of microstructures.

Figure 6.

Microstructure of different sub-zones of the HAZs with their corresponding predicted nodal thermal cycles.

The final appearance of microstructures at the HAZ depends upon the chemical composition of the parent material, peak temperature, and cooling rate. The peak temperature determines austenitization, whereas the size of the grains is dependent upon the cooling rate. The microstructure of the CGHAZ of CG-26 (10 passes) was an accumulation of bainitic ferrite (BF), grain boundary ferrite (GBF), polygonal ferrite (PF), and Widmanstätten ferrite (WF) grains. The CGHAZ of CG-26 (17 passes) and NG-18 shared identical microstructures consisting of bainite, fewer martensite, and prior austenite grain (PAG) boundaries. The FGHAZ endured a peak temperature well above the upper critical temperature $A_{C 3}$ in all the three weldments. This temperature allows austenite grains to nucleate, but during cooling, austenite was converted to fine pearlite and ferrite grains. Here, the uniform dispersal of pearlite and ferrite grains was not seen owing to the insufficient diffusion time available for the carbon particles due to rapid heating and cooling during welding and inhomogeneous austenitization. In CG-26 (10 passes), the FGHAZ consisted of pearlites in a polygonal ferrite matrix. The FGHAZ of CG-26 (17 passes) and NG-18 had an appearance comprising granular bainite and some retained austenite grains. The ICHAZ or partial grain-refining region experienced thermal cycles within a temperature bracket which was above the lower critical temperature $A_{C 1}$ and below the upper critical temperature $A_{C 3}$ . During heating, the previously present pearlite grains transformed into austenite, and upon cooling, it got transformed into fine pearlite and ferrite. The ICHAZ of CG-26 (10 passes) had a microstructure similar to that of the parent material, but the carbon-rich region was found to be a combination of fine pearlite and ferrite grains. Irregular grain boundaries with spheroidized carbides were observed in the ICHAZs of CG-26 (17 passes) and NG-18.

No significant differences were observed in the microstructures of the parent material, as shown in Figure 7. Coarse ferrite and pearlite grains were seen in all the base region of the weldments.

Figure 7.

Microstructure of the parent material: (a) conventional groove joint (10 passes), (b) conventional groove joint (17 passes), and (c) narrow groove joint.

Microhardness

The maximum hardness was observed at the CGHAZs of all the weld specimens except CG-26 (10 passes). The variations in hardness values confirmed the inhomogeneous microstructural evaluation, as shown in Figure 8. Among the three weld specimens, the weld zone and CGHAZ of the narrow groove joint exhibited the highest hardness values. The higher hardness value at the weld zone of NG-18 can be attributed to the presence of a comparatively finer grain structure (Figure 5). The CGHAZs revealed the highest hardness region in CG-26 (17 passes) and NG-18 due to the presence of bainite and martensite. In the case of CG-26 (10 passes), the hardness value at the CGHAZ was relatively low; here, the CGHAZ endured re-austenitization, and the microstructure transformed into bainitic ferrite (BF) along with a few grain boundary ferrite (GBF), polygonal ferrite (PF) and Widmanstätten ferrite (WF) grains (Figure 6).

Figure 8.

Comparison of microhardness across the weld cross-sections at the middle of the thickness (FL is fusion line).

Temperature contours

The accuracy of any FE numerical model for the prediction of welding residual stresses and distortion is subjected to the accurate evaluation of the welding thermal history numerically. Technically, the main aim is to establish an appropriate fusion zone which comprises the weld zone and fusion boundary within the melting temperature. Figure 9 shows the weld cross-section of conventional and narrow groove weld pools at peak temperature. The peak temperatures obtained in these three models were observed to be reasonably similar to each other. It can be observed from Figure 9 that the weld pool reached above the melting temperature, and the fusion boundary also attained melting temperature.

Figure 9.

Temperature contours: (a) conventional groove weld (10 passes), (b) conventional groove weld (17 passes), and (c) narrow groove weld.

Residual stress contours

The longitudinal (S33) and transverse (S11) residual stress contour profiles of the three weld samples are represented in Figure 10. It can be observed from the figures that the zones of maximum tensile residual stresses were at the top layers of the weldments. The width of this zone is shorter in the narrow groove weld joint compared to the conventional groove weld (10 passes), as shown in Figure 10(a) and (c). It may happen because the opening section of the conventional V groove geometry at the top surface is larger compared to the narrow groove. So, the top portion of conventional groove weld is subjected to a higher magnitude of heat input to accommodate more number of weld passes and experiences more cycles of thermal stresses. On the other hand, among the conventional groove welds, that zone of tensile residual stresses is wider in the one with a higher number of weld passes, as observed in Figure 10(a) and (b).

Figure 10.

Stress contours: (a) conventional groove weld (10 passes), (b) conventional groove weld (17 passes), and (c) narrow groove weld.

As an example, for verification, the comparison of experimental and corresponding numerical results of through-thickness residual stress patterns observed in the conventional groove joint with 10 number of passes is shown in Figure 11. Notably, the numerical simulation results matched well with the experimental results except in a few regions where the experimental results seem overestimated. It may happen as a result of the presence of initial residual stresses in the base plates due to the application of other fabrication processes like machining, rolling, etc., before the welding operation.^41,42 These initial stresses influence the equilibrium state of residual stresses in the weld zone and can lead to intensified welding induced stresses. The regions away from the weld zone do not endure much higher temperature during welding, which could lead to stress relaxation. As the numerical models do not comprise the presence of initial stresses, so the results produced from the simulations exhibited only the welding induced residual stress fields. This might be the reason behind the mismatch of numerical and experimental results.

Figure 11.

Comparison of experimental and numerical results obtained in conventional weld (10 passes) specimen (S33 = longitudinal residual stress, S11 = transverse residual stress).

Effect of groove geometry

The through-thickness residual stress patterns obtained in CG-26 (10 passes) and NG-18 welds were further compared to understand better the effect of weld groove configuration on the residual stress distributions in terms of both magnitude and direction. It can be observed from Figure 12 that the conventional V groove weld exhibited a comparatively higher magnitude of residual stresses in both the longitudinal and transverse directions. The peak tensile residual stress obtained in conventional V groove weld was 308 MPa in the longitudinal direction, whereas narrow groove weld exhibited around 25% reduction in that peak tensile residual stress. The conventional V groove volume is comparatively higher, and it necessitates a more considerable amount of weld filler material to be deposited. It causes increased driving force employed by the base material against the excessive amount of shrinkages caused by the higher number of weld thermal cycles and finally leads to a larger amount of locked-in strains. In the transverse direction, both the residual stress patterns in different groove geometries followed the same trend, but the peak tensile residual stress in narrow V groove was 40% lower than that of the conventional V groove. In both the groove welds, the transverse residual stress in the middle of the cross-section was found to be compressive in nature with almost identical magnitudes. The magnitude of residual stresses in conventional and narrow V groove welds significantly differed at the upper half of the weld cross-section. In contrast, at the bottom portion, nearly similar magnitudes were observed.

Figure 12.

The effect of groove geometry on the through-thickness residual stress distribution at the weld zone: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

When the residual stresses at the HAZ of different weld groove geometries were compared (Figure 13), the obtained results were quite different compared to the trends of the weld fusion zone. At the HAZ, the peak magnitude of tensile longitudinal residual stress acquired in the narrow groove weld specimen was higher than the conventional groove weld. However, the transverse residual stress at the HAZ of the narrow groove weld was significantly low compared to the conventional groove weld.

Figure 13.

The effect of groove geometry on the through-thickness residual stress distribution at the HAZ: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

Effect of number of weld passes

Figure 14 compares the through-thickness residual stress distribution in conventional groove welds having a different number of weld passes (10 and 17 passes). It was observed that in both the longitudinal and transverse directions, the magnitudes of residual stresses were almost equal at the top surface of the weld zone in both the welds. But, the peak longitudinal stress significantly increased in the CG-26 (17 passes) at a depth of 5mm from the top surface. CG-26 (17 passes) weld encountered a 17% hike in peak longitudinal residual stress compared to CG-26 (10 passes) weld. However, both the samples experienced an almost similar trend of longitudinal residual stress in the middle section. At the bottom of the weld samples, the peak compressive longitudinal residual stress increased in CG-26 (17 passes). In the transverse direction, it was observed that the number of weld passes has a minor effect on the through-thickness residual stress distribution. Although the weld sample with an increased number of weld passes encountered a slightly higher magnitude of transverse residual stresses of both the tensile and compressive nature.

Figure 14.

The effect of the number of weld passes on the through-thickness residual stress distribution at the weld zone: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

In the HAZ of CG-26 (17 passes), due to a higher number of weld beads, the peak magnitude of longitudinal and transverse residual stress significantly increased around 70% compared to the CG-26 (10 passes) weld sample, as shown in Figure 15.

Figure 15.

The effect of the number of weld passes on the through-thickness residual stress distribution at the HAZ: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

Effect of constraints

The FE model of CG-26 (10 passes) was used to study the effect of constraints on the distribution of through-thickness residual stresses. The displacement and rotation against lateral shrinkage were made zero to replicate the probable restrained behavior under clamped condition. Figure 16 depicts the longitudinal and transverse through-thickness residual stress fields under both the restraining conditions. It was observed that under the influence of external constraints, the magnitude of residual stresses in both the longitudinal and transverse directions significantly increased at the top surface of the weld. At the bottom portion of the weld, residual stresses were near to the same magnitude. Jones et al.⁴³ also investigated the same for surface residual stress distribution using the analytical and blind hole drilling method and concluded that transverse residual stress increased with imposed restraint on the weld joint.

Figure 16.

The effect of external constraints on the through-thickness residual stress distribution at the weld zone: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

The effect of constraints was also significant on the residual stress trends at the HAZ. At the HAZ, due to the constrained condition, around 26% and 72 % hike in the magnitudes of longitudinal and transverse residual stresses were observed, respectively, as shown in Figure 17.

Figure 17.

The effect of external constraints on the through-thickness residual stress distribution at the HAZ: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

Effect of preheat

Residual stress occurs due to localized heating during welding, which causes nonuniform expansion of the weld zone and the adjacent region under thermal loads followed by different cooling rates. The yielding of weld material occurs after solidification to account for the local strain gradients caused by thermal contraction against relatively cooler base material. The cooling rate of the weld zone and its adjacent region can be altered by introducing a preheat temperature to minimize the induced residual stresses. Adedayo and Adeyemi⁴⁴ observed around 65% reduction in residual stress magnitudes at the surface of the weld zone by the application of 200°C preheat temperature. For the present investigation, three different homogeneous preheat temperatures (202°C, 308°C, and 408°C) were obtained in the FE model of CG-26 (10 passes) at the initial modeling step. Figures 18 and 19 show the variation in longitudinal and transverse residual stress patterns under the influence of preheat temperatures at the weld zone and HAZ, respectively. It was observed that an increase in preheating temperature substantially reduced the longitudinal residual stress magnitude. At the preheat temperature of 408°C, the longitudinal residual stress got decreased around 30%. However, on the other hand, considerable rise or fall in peak transverse residual stress was not observed with the change in preheat temperature at the top surface of the weldment, but the through-thickness transverse residual stress distribution got altered significantly with increasing preheat temperature. The compressive transverse residual stress fields present in the middle of the plate thickness transformed into tensile in nature and kept intensifying with increasing preheat temperature. The increasing preheat temperature influences the equilibrium state of temperature dispersed across the weldment and the total amount of heat input. So, their cumulative effect may have suppressed the effect of preheating on the reduction of transverse residual stress distribution. Lin and Lee found preheating increased the residual stress magnitude in SS304 weldment. It was suggested that the application of preheat temperature may lead to a broader HAZ and reduce the strength of the base material.⁴⁵

Figure 18.

The effect of preheating temperature on the through-thickness residual stress distribution at the weld zone: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

Figure 19.

The effect of preheating temperature on the through-thickness residual stress distribution at the HAZ: (a) longitudinal residual stress (S33) and (b) transverse residual stress (S11).

As of the weld zone, a similar phenomenon was also observed at the HAZ under the influence of preheat temperature (Figure 19).

Effect of welding conditions on distortion

Figure 20 shows the comparison of angular distortion between the weld samples of two conventional groove joints having different number of weld passes and one narrow groove joint. Both the experimental measurement of deflection and the numerical prediction was done at the center of the plate and in a plane normal to the weld axis, starting at a point 10 mm away from the plate edge. From Figure 20, the effects of the number of weld passes and different groove geometries on the angular distortion can be distinguished. It was observed that the major angular distortion occurred in the case of the conventional weld sample with a higher number of weld passes. In contrast, the conventional groove weld with 10 number of passes got distorted comparatively less. Murugan and Gunaraj⁴⁶ made the same conclusion in a study of gas metal arc welded structural steel welded plates. On the other hand, in a comparison between conventional and narrow groove welds, it was observed that the narrow groove weld joint exhibited significantly lower angular deflection.

Figure 20.

Effect of welding conditions on angular distortion.

Evolution of through-thickness residual stresses

The weld induced residual stresses develop due to nonuniform volumetric expansion of the weld zone and the HAZ.⁴⁷ Consequently, locked up strains increase due to plastic deformation caused by thermal softening and phase transformation. In order to understand the final residual stress distribution across the weld zone, the evolution of through-thickness residual stress fields with corresponding subsequent weld passes has been represented in Figure 21. At the initial stage of weld deposition, it was observed that due to a higher temperature gradient at the weld center, volumetric expansion of the weld zone was accommodated in terms of plastic flow by the thermally softened HAZ. So, upon cooling, the HAZ endured higher tensile residual stresses in the longitudinal direction. In the transverse direction, the weld metal endured resistance against lateral shrinkage exerted by the parent material leading to higher tensile residual stresses at the weld zone. It can be observed from the pass no. 4th and 7th that with subsequent deposition of weld passes, the severe effect of the last active weld pass on the residual stress distribution got distributed among the previously deposited weld passes and the HAZ. Consequently, volumetric expansion and contraction on the HAZs significantly got reduced with succeeding weld passes. Along with that, when the solid-state phase transformation of the HAZ occurs from austenite, the strains associated with the phase transformation opposes the volumetric contraction and reduces tension at the HAZ.⁴⁸ Thus finally, the concentration of a higher magnitude of tensile residual stresses was developed at the weld center, and gradually it decreased and transformed into compressive stresses across the weld cross-section to maintain the self-equilibrium state.

Figure 21.

Evolution of through-thickness residual stresses in CG-26 (10 passes).

Conclusion

Both the FE numerical modeling and experimental approach were employed to investigate the residual stress fields distributed through the thickness of butt-welded SA516 Grade 70 pressure vessel steel welded joints. Using the deep hole drilling technique and with the help of numerical models, the effect of welding groove design, external mechanical constraints, number of weld passes, and preheating of base plates on the trend of through-thickness residual stresses and distortion were studied. The outcomes of the present investigation can be concluded as below:

(1) Heterogeneous microstructural appearance was observed across the weldments with corresponding variations in the microhardness values. The CGHAZ appeared to be the hardest region across the weld cross-section.

(2) The peak magnitude of residual stresses was found around 5 mm below the top layer instead of at the surface.

(3) The narrow groove weld joint exhibited reductions in the longitudinal and transverse peak residual stress magnitudes by 25% and 40%, respectively, compared to the conventional groove weld of 10 passes.

(4) The peak magnitude of tensile residual stresses increased in both the longitudinal and transverse directions by 17% with the increase in the number of weld passes.

(5) According to the results drawn from the numerical analysis, the peak longitudinal and transverse residual stresses comparatively increased by 18% and 43%, respectively, by the influence of external constraints on the conventional groove weld of 10 passes.

(6) The addition of a preheating temperature (408°C) before welding effectively reduced the peak longitudinal residual stress by 30% and significantly altered the through-thickness trend of transverse residual stress.

(7) A substantial difference in angular distortion was noticed due to the influence of the number of weld passes and weld groove geometry.

(8) Previously deposited weld passes had a significant influence on the final residual stress distribution across the weldments.

Footnotes

Acknowledgements

The authors are highly thankful to BRNS, BARC-India, and IIT Bhubaneswar for providing experimental facilities along with financial assistance for the present research work.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study received financial support for the present research work from BRNS, BARC India.

ORCID iD

Pradeeptta Kumar Taraphdar

References

Dong

Brust

FW.

Welding residual stresses and effects on fracture in pressure vessel and piping components: a millennium review and beyond. J Press Vessel Technol 2000; 122: 329–338.

Lee

Chang

Lee

CY.

Comparative study of welding residual stresses in carbon and stainless steel butt welds. Proc IMechE, Part B: J Engineering Manufacture 2008; 222: 1685–1694.

Lee

Chang

KH.

Effect of the welding sequence in the circumferential direction on residual stress distribution in a thin-walled pipe weld. Proc IMechE, Part B: J Engineering Manufacture 2009; 223: 723–735.

Denlinger

Heigel

Michaleris

Residual stress and distortion modeling of electron beam direct manufacturing Ti-6Al-4V. Proc IMechE, Part B: J Engineering Manufacture 2015; 229: 1803–1813.

Wang

Effective thermal elastic plastic finite element computation for welding distortion investigation of pozidriv-type welded structure with rectangular pipes and its mitigation. Proc IMechE, Part B: J Engineering Manufacture 2020; 234: 1729–1741.

Liu

Zhang

JX.

Investigation of external restraining force effects on welding residual stresses using three-dimensional thermal elastic-plastic multi-body coupling finite element model. Proc IMechE, Part B: J Engineering Manufacture 2009; 223: 1591–1600.

Aalami-Aleagha

Eslampanah

AH.

Mechanical constraint effect on residual stress and distortion in T-fillet welds by three-dimensional finite element analysis. Proc IMechE, Part B: J Engineering Manufacture 2013; 227: 315–323.

Zargar

Farahani

Givi

MKB

. Numerical and experimental investigation on the effects of submerged arc welding sequence on the residual distortion of the fillet welded plates. Proc IMechE, Part B: J Engineering Manufacture 2016; 230: 654–661.

Park

GB.

Effect of welding sequence to minimize fillet welding distortion in a ship’s small component fabrication using joint rigidity method. Proc IMechE, Part B: J Engineering Manufacture 2016; 230: 643–653.

10.

Leggatt

RH.

Residual stresses in welded structures. Int J Press Vessel Pip 2008; 85: 144–151.

11.

Mousavi

SAAA

Miresmaeili

. Experimental and numerical analyses of residual stress distributions in TIG welding process for 304L stainless steel. J Mater Process Technol 2008; 208: 383–394.

12.

Sattari-Far

Farahani

MR.

Effect of the weld groove shape and pass number on residual stresses in butt-welded pipes. Int J Press Vessel Pip 2009; 86: 723–731.

13.

Chen

Shenoi

RA.

Influence of welding sequence on welding deformation and residual stress of a stiffened plate structure. Ocean Eng 2015; 106: 271–280.

14.

Javadi

Investigation of clamping effect on the welding residual stress and deformation of monel plates by using the ultrasonic stress measurement and finite element method. J Press Vessel Technol 2014; 137: 011501.

15.

Teng

Lin

CC.

Effect of welding conditions on residual stresses due to butt welds. Int J Press Vessel Pip 1998; 75: 857–864.

16.

Mitra

Prasad

Ram

GDJ

. Estimation of residual stresses in an 800 mm thick steel submerged arc weldment. J Mater Process Technol 2016; 229: 181–190.

17.

Egea

AJS

Rodríguez

Celentano

, et al. Joining metrics enhancement when combining FSW and ball-burnishing in a 2050 aluminium alloy. Surf Coat Technol 2019; 367: 327–335.

18.

Colegrove

Ikeagu

Thistlethwaite

, et al. Welding process impact on residual stress and distortion. Sci Technol Weld Join 2009; 14: 717-725.

19.

Balasubramanian

Guha

Effect of welding processes on toe cracking behaviour of pressure vessel grade steel. Eng Fail Anal 2004; 11: 575–587.

20.

Prajapati

Badheka

Mehta

KP.

Hybridization of filler wire in multi-pass gas metal arc welding of SA516 Gr70 carbon steel. Mater Manuf Process 2018; 33: 315–322.

21.

Prajapati

Badheka

Mehta

An outlook on comparison of hybrid welds of different root pass and filler pass of FCAW and GMAW with classical welds of similar root pass and filler pass. Sadhana: Acad Proc Eng Sci 2018; 43: 75.

22.

Mondal

Lohit

Biswas

, et al. Prediction of weld-induced distortion of large structure using equivalent load technique. Proc IMechE, Part B: J Engineering Manufacture 2018; 232: 499–512.

23.

Liu

Zhang

Xue

CB.

Numerical investigation on residual stress distribution and evolution during multipass narrow gap welding of thick-walled stainless steel pipes. Fusion Eng Des 2011; 86: 288–295.

24.

Brickstad

Josefson

BL.

A parametric study of residual stresses in multi-pass butt-welded stainless steel pipes. Int J Press Vessel Pip 1998; 75: 11–25.

25.

Taraphdar

Pandey

Mahapatra

MM.

Finite element investigation of IGSCC-prone zone in AISI 304L multipass groove welds. Arch Civ Mech Eng 2020; 20: 54.

26.

Kant

Joshi

SN.

Numerical modeling and experimental validation of curvilinear laser bending of magnesium alloy sheets. Proc IMechE, Part B: J Engineering Manufacture 2014; 228: 1036–1047.

27.

Haddag

Atlati

Nouari

, et al. Finite element formulation effect in three-dimensional modeling of a chip formation during machining. Int J Mater Form 2010; 3: 527–530.

28.

Deng

Kiyoshima

. FEM prediction of welding residual stresses in a SUS304 girth-welded pipe with emphasis on stress distribution near weld start/end location. Comput Mater Sci 2010; 50: 612–621.

29.

Xia

Jin

Analysis of residual stresses and variation mechanism in dissimilar girth welded joints between tubular structures and steel castings. Int J Press Vessel Pip 2018; 165: 104–113.

30.

Lee

Chang

Temperature fields and residual stress distributions in dissimilar steel butt welds between carbon and stainless steels. Appl Therm Eng 2012; 45–46: 33–41.

31.

Biswas

Mahapatra

Mandal

NR.

Numerical and experimental study on prediction of thermal history and residual deformation of double-sided fillet welding. Proc IMechE, Part B: J Engineering Manufacture 2010; 224: 125–134.

32.

Taraphdar

Mahapatra

Pradhan

, et al. Evaluation of through-thickness residual stresses in conventional and narrow grooved stainless steel welds. Proc IMechE, Part L: J Materials Design and Applications 2020; 234: 1165–1179.

33.

Giri

Mahapatra

Sharma

, et al. A study on the effect of weld groove designs on residual stresses in SS 304LN thick multipass pipe welds. Int J Steel Struct 2017; 17: 65–75.

34.

Taraphdar

Thakare

Pandey

, et al. Novel residual stress measurement technique to evaluate through thickness residual stress fields. Mater Lett 2020; 277: 128347.

35.

Zhdanov

Gonchar

Determining the residual welding stresses at a depth in metals. Autom Weld 1978; 319: 22–24.

36.

Beaney

Measurement of sub-surface stress. Report Rd/B/N4325, Central Electricity Generating Board, 1978.

37.

Mahmoudi

Hossain

Truman

, et al. A new procedure to measure near yield residual stresses using the deep hole drilling technique. Exp Mech 2009; 49: 595–604.

38.

Leggatt

Smith

, et al. Development and experimental validation of the deep hole method for residual stress measurement. J Strain Anal Eng Des 1996; 31: 177–186.

39.

Amanie

Oguocha

INA

Yannacopoulos

Effect of submerged arc welding parameters on microstructure of SA516 steel weld metal. Can Metall Q 2012; 51: 48–57.

40.

Francis

Bhadeshia

HKDH

Withers

PJ.

Welding residual stresses in ferritic power plant steels. Mater Sci Technol 2007; 23: 1009–1020.

41.

Aurrekoetxea

De Lacalle

LNL

Llanos

Machining stresses and initial geometry on bulk residual stresses characterization by on-machine layer removal. Materials (Basel). Epub ahead of print 22 March 2020. DOI: 10.3390/ma13061445.

42.

Suárez

Veiga

De Lacalle

LNL

, et al. Effects of ultrasonics-assisted face milling on surface integrity and fatigue life of Ni-Alloy 718. J Mater Eng Perform 2016; 25: 5076–5086.

43.

Jones

Emery

Marburger

An analytical and experimental study of the effects of welding parameters on fusion welds. Weld J 1993; 72: 51s–59s.

44.

Adedayo

Adeyemi

MB.

Effect of preheat on residual stress distributions in arc-welded mild steel plates. J Mater Eng Perform 2000; 9: 7–11.

45.

Lin

Lee

KH.

Effect of preheating on the residual stress in type 304 stainless steel weldment. J Mater Process Technol 1997; 63: 797–801.

46.

Murugan

Gunaraj

Effects of process parameters on angular distortion of gas metal arc welded structural steel plates. Weld J 2005; 84: 165–171.

47.

Iltaf

Junaid

Khan

, et al. Influence of electron beam oscillation patterns on the microstructure, texture, residual stress and mechanical properties of Ti-5Al-2.5Sn alloy weldments. Proc IMechE, Part B: J Engineering Manufacture. Epub ahead of print 20 July 2020. DOI: 10.1177/0954405420937537.

48.

Ishigami

Roy

Walsh

, et al. The effect of the weld fusion zone shape on residual stress in submerged arc welding. Int J Adv Manuf Technol 2017; 90: 3451–3464.