Numerical Study on Welding Structure of Connecting Fin Used in Thermal Power Plant

Abstract

The background of deep peak-shaving poses demanding requirements for the performance of metal components in thermal power plants. The cracking of the connecting fins on the side wall of the flue seriously endangers the safe operation of the system. Herein, welding simulations were carried out using the finite element method to understand the cracking process of the connecting fins. By changing the welding process and fin size, their effect on stress and deformation was explored. The results showed that increasing the welding rate would decrease the flexural deformation of the flue-side wall. Additionally, the yield range of the connecting fin depended more on width than thickness, and increased with increasing width. As for the relationship between fin size and flexural deformation, the maximum deformation decreased with the increase in thickness, while it first decreased and then increased with increasing width. Overall, the post-welding stress and deformation of the boiler flue side wall exhibited more sensitivity to the fin width compared with the fin thickness and welding rate. This article clarifies the stress distribution status of the connecting fins in the flue side wall under different welding conditions, providing a basis for analyzing its cracking phenomenon, and further providing theoretical guidance for optimizing the structural parameter design of the side wall.

1. Introduction

New energy policies have remarkably promoted the rapid increase in the proportion of renewable clean energy, such as wind power, solar energy, and hydropower [1,2,3,4,5]. As a result, ultra-supercritical thermal power plants (USC-TPPs) are now frequently expected to participate in deep peak shaving. However, this action will produce constant start-up and shut-down for devices, causing a wider range of temperature fluctuation for structural components [6,7,8,9,10]. As such, higher requirements have been put forward for the structural integrity of key components, particularly for the flue-side water-cooled wall. This component plays a crucial role in cooling flue gas and recovering heat, which directly relates to the safety and economy of USC-TPPs. Unfortunately, in recent years, the large temperature gradient and load changes triggered by deep peak shaving have led to frequent cracking failures of structural components, for example, the weld seam and connecting fins on the flue-side wall. Figure 1 shows a typical failure diagram of a flue-side wall. Overall, this kind of accident seriously hinders the safe operation and economic benefit of USC-TPPs [11,12,13,14]. Additionally, due to the practical constraints on the installation of the flue-side wall during welding, the size of the connecting fins varies accordingly. To reduce the cracking probability of welds and connecting fins and ensure the safe operation of thermal power plants, it is necessary to clarify the impact of connecting fin dimensions on residual welding stress.

Due to the large structural dimensions and demanding working conditions associated with flue-side walls or membrane-type water-cooled walls (structures sharing similar designs, but differing in nomenclature), researchers have typically relied on numerical simulations to study temperature and stress distribution. A.H. Assefinejad et al. [15] investigated the critical conditions in a sub-critical drum boiler and found that water wall tubes of all lengths had a serious risk of combustion and overheating if the fluid flow in the tubes was disturbed. C.A. Duarte et al. [16] analyzed the failure locations in the water-cooled wall tubes of a specific power plant. Following extensive experimental analysis, numerical simulation methods were employed to calculate the residual stresses and equivalent stresses of water-cooled wall tubes under bending conditions. J. Fu et al. [17] investigated the impact of combustion condition symmetry on temperature, heat flux density, strain, and stress distribution. This study provided, for the first time, a comprehensive overview of the entire operational state of a boiler’s water-cooled wall, particularly focusing on strain and stress distribution. J.Y. Qiu et al. [18] explored the heat transfer characteristics between the membrane wall and syngas under different working conditions, and found that when the membrane water wall was not protected, the inner surface part of the fin would overheat and the maximum stress would exceed the allowable stress. Q. Fan et al. [19] established a numerical framework that coupled fluid dynamics and solid mechanics to obtain temperature profiles within a furnace, thermal stress, and strain fields on a furnace wall. This framework was employed to predict critical regions essential for mechanical reliability. L. Wei et al. [20] conducted experimental research and numerical simulations for 17 typical flexible operating conditions to investigate the safety of water-cooled walls for combustion boilers under flexible load conditions. Furthermore, research has revealed that the failure and cracking of furnace side walls or membrane-type water-cooled wall structures result from the synergy of many factors, including high-temperature creep, corrosion, and scaling [21,22,23,24,25,26]. However, no previous studies have considered the impact of welding structure dimensions on the residual stress and deformation of water-cooled walls. The actual failure cases have highlighted that the weld seam (especially at the connecting fin) poses a high risk for cracking and failure of water-cooled walls. Moreover, the post-welding residual stress is often a crucial factor triggering structural failure. Therefore, it is of great significance to explore the evolution of post-welding residual stress under different welding parameters and structural sizes. Such research provides guidance for obtaining appropriate welding parameters and structural dimensions, which could ensure the safety of the flue-side wall during service.

At present, the methods for detecting residual stress are mainly divided into two categories: destructive and nondestructive testing. Among them, destructive testing methods include the blind hole, indentation, and cutting methods [27,28,29,30] and nondestructive testing methods include ultrasonic, magnetic, X-ray diffraction, and neutron diffraction methods [31,32,33,34,35,36].

The measurement of residual stress in both destructive and nondestructive testing has its disadvantages, as shown in

Table 1. The disadvantages of destructive and nondestructive testing.

Residual Stress Test Method		Disadvantage
Destructive testing	Blind-hole method	Causes damage
	Indentation method	Causes damage
	Cutting method	Causes damage
Nondestructive Testing	Ultrasonic method	Weak detection depth
	Magnetic method	Imperfect development
	X-ray diffraction method	Only magnetic materials
	Eutron diffraction method	Poor test sensitivity

. Most of the existing test instruments need to frequently change samples, rotate sample angles, manually calculate data, and formulate calibration curves, which are often limited by various actual engineering conditions. Thus, the finite element simulation method is widely used for predicting residual welding stresses. In this study, the finite element method was employed to establish a one-to-one model according to the size of the actual flue-side wall. Then, the effect of fin size on the stress and deformation of the welding structure was explored. Our findings offer a theoretical basis for optimizing the welding structure and ensuring the safe operation of the flue-side wall used in USC-TPPs.

2. Welding Structural Parameter

Figure 2 displays a structural diagram of the welding structure of the flue-side wall in USC-TPPs. Cracking commonly occurs on the connecting fins between different side walls. Figure 3 presents the failure position of the flue-side wall, which is located at the horizontal flue-side wall, with the connecting fin joined with the water-cooled wall. The relevant structural parameters can be found in

Table 2. Structural parameters of water-cooled tubes and connecting fins in a membrane water-cooled wall.

Species	Value
Outer diameter of wall tube D (mm)	32
Thickness of water-cooled wall δ (mm)	8
Pitch x (mm)	49.5
Horizontal flue-side wall cladding wall pipe length L1 (mm)	11,180
Tail flue-side wall package wall pipe length L2 (mm)	19,410
Connecting fin length L3 (mm)	10,520
Weld fin angle α (°)	31
Weld radian radius R (mm)	5
Size of welding toe at connecting fin k1 (mm)	3.33
Size of welding toe at wall pipe k2 (mm)	2.86

.

3. Method

3.1. Model Size and Material Parameter

ABAQUS 6.14 and MARC 2022 were used for finite element simulation in our work. As reported in our previous work [37], the effect of welding structure on the structural stress distribution of the flue-side wall was carried out by ABAQUS 6.14. Meanwhile, the dependence of residual welding stress and deformation on the fin size and welding parameters was investigated using MARC 2022. The finite element model, as shown in Figure 4, was established based on the actual working conditions and structural parameters provided in Table 1. The left side of the model represented the horizontal flue-side wall with a pipe length of 11,180 mm, while the right side represented the tail flue-side wall with a pipe length of 19,410 mm. The connecting fins for both pipe groups had a length of 10,520 mm. The entire model was made of 12Cr1MoV steel, and the material parameters such as heat capacity C (J/Kg∙K), thermal conductivity λ (W/m$·$k), linear expansion coefficient α (10⁻⁶/°C), elastic modulus E (10³ MPa), shear modulus G (10³ MPa), Poisson’s ratio ν, and yield strength σ_s (MPa), which vary with temperature T (°C), are listed in

Table 3. Material parameters of 12Cr1MoV [38].

T	C	λ	α	E	G	ν	σs
20	560	45.2	10.80	214	83.5	0.28	393
100	569	45.2	13.00	211	81.8	0.28	374
200	586	45.2	13.36	206	79.2	0.30	359
300	611	42.7	13.55	195	74.0	0.31	308
400	653	40.5	13.83	187	72.2	0.29	285
500	682	37.7	14.15	179	68.6	0.30	266
600	729	35.5	14.38	167	63.1	0.32	251
700	—	33.4	14.62	—	—	0.32	—

. The meshing strategy adopted in this study was manual meshing, and the meshing type was hexahedral meshing. In the welding simulation process, the encrypted mesh was used in the weld seam and heat-affected zone, where the temperature gradient greatly changed. The temperature gradient in the area far from the weld seam was much smaller, so a relatively sparse grid was used. The final number of simulated elements reached 195,727.

3.2. Calculation Process

3.2.1. Thermal Model

The simulation utilized a double ellipsoidal heat source [39]. As the welding heat source passed through the weld seam, the molten metal filling process in the weld seam was taken into account using the life and death unit technology to fill the gap. The welding unit was activated and filled the gap to the welding seam once the heating temperature of the component reached the melting point of the solder. The welding process involved first welding the two connecting welds between the horizontal flue side and the fin from top to bottom, followed by welding the connecting welds between the tail flue side and the fin.

The ambient temperature and initial temperature of the workpiece were defined as 20 °C. The heat transfer coefficient was set as a constant of 20 W/(m²·°C). During the welding process, heat loss occurred primarily through radiation and convection from the workpiece’s surface to the surrounding environment, with radiation being the primary form of loss. The radiation effect became stronger as the temperature increased. Typically, at temperatures above 1200 °C, radiation loss exceeds convective heat dissipation loss, while convection is the dominant form at lower temperatures. For calculation convenience, the radiation and convection coefficients were combined into a total heat transfer coefficient for simulation purposes. This allowed the heat energy lost due to boundary heat transfer to be expressed as follows:

Q_S = β(T − T₀)

(1)

β = β_e + β_c

(2)

where T represents the surface temperature of the welding piece (°C), T₀ represents the temperature of the surrounding medium, β is the surface heat transfer coefficient (W/(m²·°C)), β_e is the convective heat transfer coefficient, and β_c is the radiation heat transfer coefficient.

3.2.2. Residual Stress Model

During arc welding, the electrical power source produces a voltage (U) between the electrode and base metal, resulting in the flow current. In this process, energy loss is due to the convection and radiation within the electric arc. Consequently, only a tiny fraction of energy is employed for melting materials. To account for this, a variable, namely power efficiency (η), was introduced. The effective welding heat input can be calculated as:

Q = ηUI

(3)

where Q is the heat input, η is the power efficiency, U is the voltage, and I is the current.

Considering the thermal model, the thermal gradient can be evaluated by establishing an energy balance [40]:

$$\rho (T)c(T)\frac{\partial T}{\partial t}=Q+\frac{\partial }{\partial x}[Kx\left(T\right)\frac{\partial T}{\partial x}]+\frac{\partial }{\partial y}[Ky\left(T\right)\frac{\partial T}{\partial y}]+\frac{\partial }{\partial z}[Kz\left(T\right)\frac{\partial T}{\partial z}]$$

(4)

where ρ is the material density, c is the specific heat, Q is the heat input, T is the temperature, t is the time, and K_x, K_y, and K_z are the thermal conductivity coefficients in their respective directions. The heat flow involved is non-linear, and the material’s thermophysical properties are strongly temperature-dependent.

Using the thermal gradients from Equation (4), residual stress can be calculated by evaluating the gradient deformation:

$${\sigma }_v={\delta }_v{\lambda }_{\epsilon kk}+2\mu {\epsilon }_v-{\delta }_v(3\lambda +2\mu )\alpha T$$

(5)

where λ and µ represent the Lame constants, which are associated with the elastic modulus (E) and Poisson’s ratio (v). Variable ε_v is further related to the deformation and displacement:

$${\epsilon }_v = \frac{1}{2}\left(\frac{\partial u_i}{\partial x_j}+\frac{\partial u_j}{\partial x_i}\right)$$

(6)

Accordingly, the residual stress calculation method involves elastoplastic behavior, including isotropic hardening, and the relevant data were extracted from the strain during the welding process. As strain can be categorized by elastic, plastic, and thermal properties, the formula of the total strain can be given as:

$${\epsilon }_{ij}= {\epsilon }_{ij}^e+ {\epsilon }_{ij}^p+ {\epsilon }_{ij}^{th}$$

(7)

where ${\epsilon }_{ij}$ is the total deformation, ${\epsilon }_{ij}^e$ is the elastic deformation, ${\epsilon }_{ij}^p$ is the plastic deformation, and ${\epsilon }_{ij}^{th}$ is the thermal deformation.

Furthermore, according to Hooke’s law, the elastic deformation can be expressed as:

$${\epsilon }_{kl}^e={\sigma }_{ij}^e·{E\left(T\right)}^{-1}$$

(8)

The thermal-induced deformations are as follows:

$${\epsilon }_{kl}^{th}={\alpha }_{ij}·(T-T_{\infty })$$

(9)

where ${\alpha }_{ij}$ is the linear thermal coefficient and $T_{\infty }$ is the reference temperature. Plasticity theory characterizes the elastic–plastic response of materials using mathematical relationships that rely on specific restrictive assumptions. Among these assumptions, plastic deformation is the result of transient stress and is independent of time [41].

Finally, according to the flow rule model that determines the plastic direction of metals under minor displacements, the plastic potential function (g) is equivalent to the flow area capability (f). This relationship is known as the associated flow rule, where the plastic direction is considered to be perpendicular to the flow surface. The plastic strain ratio is determined as:

$${\mathrm{d}\epsilon }_{kl}^{pl}=\mathrm{d}\mathsf{\lambda }\frac{\partial f}{\partial {\sigma }_{ij}}$$

(10)

where λ is a positive constant dependent on the properties of the material. For a perfect elastic–plastic material and for the cases in which a Von Mises surface is used as the flow area capability criterion, parameter λ can be described as:

$$\lambda =\frac{3G{s}_{ij}{s}_{kl}}{{{\sigma }_e}^2}$$

(11)

3.2.3. Boundary Conditions and Parameter Design

Under practical working conditions, the flue-side wall is usually fixed. To limit the displacement of the structure and simultaneously achieve the free expansion deformation of the structure, the boundary constraint conditions applied to the structure are illustrated in Figure 5 [42]: a symmetric constraint about the X-axis plane was applied at the center (X = RY = RZ = 0), and displacement constraints were applied in the X, Y, and Z directions at the upper-left corner of the structure, in the Y and Z directions at the upper-right corner, and in the Y direction at the lower-right corner.

Owing to the significant impact of different welding efficiencies on the welding quality of the flue side wall structure when the linear energy remains constant during the actual welding process, this study aims to investigate the influence of different welding efficiencies on the welding stress and deformation of the structure. To achieve this, the welding speed was varied while maintaining a constant linear energy of q = 346 J/mm, based on the actual working conditions. Furthermore, to examine the effect of structural size on residual stress and deformation after welding, two comparative groups were established: one based on the connecting fin width and another based on the connecting fin thickness. The specific parameters for these groups are provided in

Table 4. Structural parameters of the connecting fin.

Fin Thickness b (mm)	Fin Width s (mm)
8	10, 17.5, 30, 50, 70
6, 8, 10, 12	17.5

.

4. Results

4.1. Sensitivity Analysis of the Number of Water-Cooled Tubes

Owing to the large size of the overall structure, the calculation was simplified by initially studying the stress distribution in models with 8, 16, and 32 tubes. The dimensions of these models are presented in Table 1. Figure 6 displays the calculation results, which indicate a significant decrease in stress in the surrounding tube bundles as the number of tube bundles increased. The distribution curves of equivalent residual stress and longitudinal stress (parallel to the weld axis) on the transverse path (i.e., the path from left to right in Figure 6) are shown in Figure 7. It can be seen from Figure 7b that the longitudinal residual stress in the weld seam and connecting fin area after welding was tensile stress. This was due to the rapid cooling of the weld seam area from high temperatures during the cooling process, resulting in volume shrinkage. However, it was constrained by the surrounding metal and fixed end, resulting in tensile stress. At the same time, the wall cladding pipes on both sides were subjected to compressive stress due to the effect of cooling shrinkage welds. From Figure 7a, it can be seen that the yielding intervals in the 8-, 16-, and 32-tube models were 77.46, 84.15, and 80.81 mm, respectively, and the connecting fin parts all yielded. Comparing the stress curves along the transverse path obtained from the 8/16/32-tube models, it becomes evident that the residual stress gradually decreased at the edge of the model as the number of tubes increased from 8 to 16 and then to 32. The distribution of residual stress also tended to stabilize (Figure 7). Consequently, subsequent calculations were simulated using the 32-tube model.

4.2. Effect of Welding Rate on Residual Stress and Deformation

The influence of different welding rates on welding residual stress and deformation was studied using a 32-tube model under the condition that the actual working condition line energy q was 346 J/mm. The model size is shown in Table 1 and the stress and deformation curves distributed along the path are shown in Figure 8. The welding speed had little effect on the distribution of residual stress and the amplitude of the maximum longitudinal stress. Different welding speeds v = 11, 12, and 13 mm/s corresponded to yield ranges of 78.86, 80.81, and 80.81 mm for the flue side-wall structure (Figure 8a), respectively, while the maximum longitudinal stresses were 409.11, 413.96, and 415.64 MPa (Figure 8b). The welding speed had a significant impact on the distribution of bending deformation in the longitudinal path (i.e., the path from top to bottom in the middle of the structural connecting fins). The bending deformation gradually decreased with the increase in welding speed (Figure 8c), and the maximum bending deformation decreased from 2.11 mm at v = 11 mm/s to 1.14 mm at v = 13 mm/s.

4.3. Effect of Fin Size on Residual Stress and Deformation

Owing to the significant temperature difference on both sides of the connecting fins, they were subjected to significant thermal stress. Owing to the geometric discontinuity caused by structural factors and material discontinuity caused by welding seam factors, significant stress concentration occurred at the connecting fins. In practical engineering, due to various factors, such as size and construction, the size of the connecting fins varied, making it difficult to estimate the distribution of welding residual stress and prone to cracking and failure during operation. Therefore, finite element simulations were conducted on connection fin models with different widths and thicknesses to elucidate the influence of connection fin size on welding residual stress. The models used were consistent with the previous text, except for the width and thickness dimensions of the connecting fins. The stress and deformation curves distributed along the path under different connecting fin thicknesses are shown in Figure 9. As the thickness of the fins gradually increased, the ranges of stress reaching yield on the welded structure were 78.66, 77.55, 75.72, and 75.22 mm, respectively (Figure 9a), which means that, as the thickness of the connecting fins increased, the yield range decreased. The thickness of the connecting fins had little effect on the distribution and amplitude of longitudinal stress. Different fin thicknesses of 6, 8, 10, and 12 mm corresponded to maximum longitudinal stress of 417.74, 413.96, 408.21, and 406.31 MPa (Figure 9b). When the thicknesses of the fins were 6, 8, and 10 mm, the distribution of deflection deformation in the longitudinal path showed a trend of high in the middle and low on both sides. However, when the thickness further increased to 12 mm, the deflection deformation of the flue-side wall structure significantly decreased, and the distribution in the longitudinal path tended to be flat. In addition, the maximum value of deflection deformation on the structure showed a significant downward trend with the increase in fin thickness (Figure 9c).

The stress and deformation curves along the path under different connection fin widths are shown in Figure 10. When the widths were 10, 17.5, 30, 50, and 70 mm, the ranges of welding residual stress on the structure reaching material yield were 66.643, 77.55, 91.293, 110.225, and 120.582 mm, respectively (Figure 10a). As the width of the connecting fins increased, the range within which the structure reached yield also increased. When the fin width was narrow, all the connecting fins yielded after welding, and as the fin width increased to ~70 mm, the middle part of the connecting fins away from the weld no longer yielded. After welding, all the connecting fins reached yield. Therefore, as the width of the connecting fins increased, the range of the structure reaching yield also increased. The maximum longitudinal stresses on welded structures with different fin widths of 10, 17.5, 30, 50, and 70 mm were 413.511, 413.96, 414.441, 417.206, and 412.77 MPa, respectively (Figure 10b). As the fin width increased, the maximum longitudinal stress did not change much and was slightly higher than the yield strength of the material. When the fin widths were 10, 17.5, 30, 50, and 70 mm, the maximum deflection deformations on the structure were 3.832, 1.624, 2.011, 3.521, and 26.92 mm, respectively (Figure 10c). It can be seen that, under the condition of a fin thickness of 8 mm, the structural deflection deformation was relatively small when the fin width was between 10 and 50 mm. When the fin width increased to 70 mm, the deflection deformation at the structural weld seam significantly increased, leading to instability.

5. Discussion

5.1. Relationship between Welding Rate and Stress/Deformation

Welding finite element simulations were conducted on the boiler flue-side wall under different welding speeds (keeping line energy constant). As such, the response of the maximum deflection after welding to the welding rate is shown in Figure 11, in which the welding speed coefficient α = v_i/v_min (v_i denotes the different welding speed and v_min denotes the minimum welding speed) is used as the abscissa and maximum deflection as the ordinate. It can be seen that, with the increase in welding rate, the maximum deflection of the structure welding gradually decreased. However, the change in welding rate had no obvious effect on the maximum deflection. With the enhancement of the welding rate to 18%, the reduction in maximum deflection was only ~1 mm. In addition, the change in welding rate had a negligible impact on the yield range and the longitudinal stress after welding.

5.2. Relationship between Fin Size and Stress/Deformation

According to the results of the finite element calculation, fin thickness had almost no impact on the yield range, while increasing the fin width enabled an increase in the yield range. In addition, the change in the fin thickness or width showed a minor effect on the longitudinal stress, but had a great influence on the flexural deformation. Figure 12 shows the effect of fin size on maximum deflection, where s_i and b_i are different fin widths and thicknesses, s₀ and b₀ are the typical fin width and thickness under actual working conditions (Table 1), y_i is the maximum deflection under different fin size conditions, and y₀ is the maximum deflection under typical fin size conditions. One can notice that the maximum deflection deformation of the structure gradually decreased as the fin thickness increased (Figure 12a), while it first decreased and then increased with the increased fin width (Figure 12b). When the fin thickness b₀ was 8 mm and the fin width s₀ was 17.5 mm, the maximum deflection was 1.62 mm. When the fin thickness was reduced by 25%, the maximum deflection decreased by ~8%. However, when the fin thickness was increased by 50%, the maximum deflection decreased by about 72%. When the fin width was 0.57s₀, the maximum deflection increased by 140%. When the fin width exceeded s₀, the maximum deflection increased with an increase in fin width. When the fin width was 2.8s₀, the maximum deflection increased by nearly 40%. Furthermore, when the width was four times the typical fin width in the structure, the structure became notably unstable, and the maximum deflection was ~17 times that of the typical fin width. Therefore, compared with the welding rate and fin thickness, the maximum deflection of the flue-side wall structure was more sensitive to the changes in fin width.

The thickness of the connecting fins had a visible effect on the flexural deformation, which was attributed to the varying mechanical properties in the welding area. Thicker fins had higher strength and rigidity, and possessed more resistance to thermal expansion and flexural deformation. Furthermore, the increased fin thickness implied deepened constraint conditions near the weld seam.

Increasing the width of the connecting fins weakened the constraint effect, which induced more prone thermal shrinkage and deformation in the welding area, suggesting greater deflection deformation during the cooling process. Additionally, an upward temperature gradient in the welding area arose from the increase in the width of the connecting fins. Accordingly, a larger temperature gradient meant that the temperature difference on both sides of the weld increased, causing a larger thermal stress difference and ultimately greater deflection deformation. It also should be noted that, if the width of the connecting fins is too narrow, the distance between different weld seams will be close enough to produce the superimposed effect of the flexural deformation. Consequently, the maximum flexural deformation on the welding structure becomes larger. In short, the width of the flue-side wall fins should be comprehensively considered in the design process.

The effect of the tube number on the post-welding residual stress was investigated. As shown in Figure 7, it was found that, when the tube number was eight, the peak of longitudinal stress appeared in the middle of the connecting fins. However, as the tube number increased to 32, the peak value of the residual stress was inspected in the vicinity of the weld toe. Thus, the increase in the tube number significantly affected the distribution of longitudinal stress. Additionally, the distribution of longitudinal stress was also influenced by the constraints on both sides of the structure. Furthermore, it can be observed from Figure 8, Figure 9 and Figure 10 that the peak longitudinal stress always arose from the welding toe under different welding rates and fin sizes. It can be concluded that the distribution of the peak longitudinal stress mainly depended on the geometric discontinuity caused by the welding structure, while the welding rate and fin size were not the key factors affecting the distribution of the post-welding stress.

Notably, the weld seam of the flue-side wall had a length of ~10 m. When the width-to-thickness ratio of the connecting fin was not appropriate; the larger the width, the more unstable deformation of the structure will occur, resulting in excessive flexural deformation (Figure 12a).

5.3. Brief Summary

In summary, the higher the welding efficiency in the welding process, the smaller the flexural deformation of the structure after welding. Changing the fin size had an obvious effect on the flexion deformation after welding, but had little effect on the magnitude and distribution of the longitudinal stress and equivalent stress after welding. The weld structure had a great effect on the distribution of stresses, and the distribution of the maximum longitudinal stress mainly depended on the geometric discontinuity caused by the welding structure. Some limitations still exist in this study, when the finite element simulation was utilized to analyze post-weld residual stress. For instance, it did not consider the coupling of structural stress generated after the start-up of the thermal power plant with the welding residual stress. However, due to the large size of the flue-side wall in this study, compared with other residual stress testing methods, the simulation method had the advantage of obtaining more comprehensive residual stress distribution after welding. Moreover, it had stronger adaptability and it was more convenient and economical to study the effects of fin size and the welding process on the distribution of the residual stress.

6. Conclusions

In this work, the residual stress and post-welding deformation of connecting fins of the flue-side wall (horizontal and tail) in USC-TPPs were analyzed through the finite element method. The effect of the line energy and size parameters (width and thickness) of connecting fins on the thermal stress and strain distribution of the welding structure were discussed. The main conclusions are summarized as follows:

(1)

When the welding line energy was fixed, the deflection deformation of the flue-side wall decreased with the increase in the welding rate.

(2)

The connecting fin yielded in all cases. The yield range increased with the increasing width of the connecting fin. The maximum deformation decreased with the increase in thickness, while it decreased first and then increased with the increasing width.

(3)

The post-welding stress and deformation of the boiler flue-side wall exhibited more sensitivity to the width of the connecting fin.