Front. Mater. Frontiers in Materials Front. Mater. 2296-8016 Frontiers Media S.A. 1778608 10.3389/fmats.2026.1778608 Original Research Fatigue safety estimation of eh690 high-strength steel and welded joints in extreme marine environments Lu et al. 10.3389/fmats.2026.1778608 Lu Chengming 1 * Writing - original draft Writing - review and editing Liu Zhengyuan 1 Writing - original draft Writing - review and editing Zhao Yang 1 Writing - original draft Writing - review and editing Xue Bingqing 1 Writing - review and editing Writing - original draft Chen Cheng 2 Writing - review and editing Writing - original draft Peng Zhuo 2 Writing - original draft Writing - review and editing Shanghai Branch of CNOOC (China) Co., Ltd., Shanghai, China School of Civil Engineering and Geomatics, Southwest Petroleum University, Chengdu, China Correspondence: Chengming Lu, chengminglucnooc@163.com 25 02 2026 2026 13 1778608 31 12 2025 01 02 2026 10 02 2026 Copyright © 2026 Lu, Liu, Zhao, Xue, Chen and Peng. 2026 Lu, Liu, Zhao, Xue, Chen and Peng https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

This study investigates the EH690 high-strength steel base metal and its welded joints under neutral salt spray tests simulating extreme marine atmospheric conditions. The research systematically examines the evolution of surface morphology under different corrosion durations and its influence on fatigue performance, with the stress concentration factor applied to quantify the effect of corrosion pits. A total of 18 base metal and welded joint specimens were subjected to corrosion exposures of 240, 480, and 960 h. The corrosion behavior was evaluated using mass loss measurement, 3D morphology scanning, and fatigue tests. Results show that with prolonged corrosion time, the surface roughness of the base metal increases significantly, and the fatigue life decreases drastically under high stress ratio. Moreover, at the weld toe, the corrosion notches deepen and the notch radius decreases, leading to aggravated stress concentration and more severe fatigue degradation. Based on experimental data, an S-N prediction model incorporating surface roughness and the stress concentration factor was developed, and a fatigue life assessment method applicable to corroded EH690 base metal and welded joints was proposed. This study provides a theoretical basis for the anti-corrosion fatigue design of such materials in marine environments.

 corrosion fatigue corrosion morphology scanning EH690 high-strength steel extreme marine environments salt spray test National Natural Science Foundation of China 10.13039/501100001809 52208211 The author(s) declared that financial support was received for this work and/or its publication. This research was funded by the National Natural Science Foundation of China (No. 52208211). The authors also gratefully acknowledge the support provided by the Structural Engineering Laboratory of the School of Civil Engineering and Geomatics at Southwest Petroleum University. section-at-acceptance Mechanics of Materials Introduction

With the rapid advancement of global infrastructure construction and the swift development of marine engineering, the demand for high-performance steel has become increasingly urgent in fields such as bridges, offshore platforms, transmission towers, and aerospace (Fu et al., 2018; Yang et al., 2025; Ban and Shi, 2018). EH690 high-strength steel, renowned for its excellent strength-to-weight ratio, good toughness and plasticity, and outstanding corrosion fatigue resistance, has become a key material in extreme marine engineering structures (Shao et al., 2025; Shen et al., 2025). It is widely used in critical facilities including deep-sea platforms, cross-sea bridges, and offshore wind turbine foundations. However, when these critical structures serve in extreme marine environments, they are perennially exposed to harsh corrosive conditions–such as high salt spray, chloride ion erosion, acidic seawater, and industrial pollutants–while simultaneously bearing extreme cyclic or impact loads like wave shocks, seismically induced vibrations, and tsunami loads. This dual exposure places high-strength steel and its welded structures at serious risk of corrosion fatigue failure (Guo H. C. et al., 2021). Corrosion fatigue, as a complex coupled environmental-mechanical damage mechanism, is particularly pronounced under extreme marine conditions. It significantly reduces the material’s load-bearing capacity and service life, posing a severe threat to the reliability and safety of engineering structures.

Compared to dry air environments, corrosive media readily induce stress concentration on the steel surface, forming corrosion pits that accelerate fatigue crack initiation (Guo HC. et al., 2021; Chen et al., 2025; Wu et al., 2021). During the stable crack growth stage, the interaction between the corrosive medium and the material further accelerates the crack growth rate, with the crack growth threshold decreasing and the growth rate increasing as the stress ratio rises (Guan et al., 2025; Li et al., 2023). As the degree of corrosion intensifies, material mass loss increases and surface integrity is compromised, leading to a significant reduction in fatigue life (Li et al., 2023; Heng et al., 2025; Yan et al., 2025). Compared to the base metal, welded joints are further degraded in fatigue performance due to microstructural inhomogeneity induced by the welding thermal cycle, residual stress concentration, and geometric discontinuities (especially in the weld toe region) (Shojai et al., 2025a). This makes them more susceptible to crack initiation and unstable propagation in extreme marine corrosive environments, particularly under high-amplitude dynamic loads induced by earthquakes or tsunamis (Liu et al., 2025). Existing research indicates that corrosion shortens the fatigue crack initiation period and accelerates the propagation rate, with stress concentration induced by corrosion pits being a key mechanism, effectively explainable through stress concentration models (Wu et al., 2025).

In the extreme marine environment, high-strength steel components are prone to corrosion and form corrosion pits and other corrosion morphology, which causes local geometric defects in steel components and produces stress concentration phenomena. Currently, the Stress Concentration Factor (K SCF) is a key parameter for quantifying the local stress amplification effect caused by geometric defects in structures. Its core definition is the ratio of the local actual maximum stress to the nominal stress in a component, serving as a critical bridge for revealing damage mechanisms and establishing life prediction models. Zhang HP. et al. (2025), studying high-strength steel wires, found that the geometric parameters and spatial distribution of corrosion pits significantly influence K SCF. Yang et al. (2022), investigating finger-type expansion joints, confirmed a positive correlation between corrosion depth and K SCF. Garbatov et al. (2014) demonstrated, using small-scale specimens, that corrosion morphology degrades fatigue performance through K SCF. Li et al. (2021), exploring the corrosion fatigue mechanism of bolted steel plates, found that grain refinement indirectly affects K SCF. Jiang et al. (2020) pointed out that the maximum corrosion pit in stay cable wires causes a sharp increase in K SCF. These studies quantify stress concentration coefficients by describing corrosion morphology geometry parameters to study the effect on fatigue performance. However, existing studies mostly focus on the relationship between a single corrosion parameter (e.g., pit depth) and K SCF, lacking systematic quantification of multi-parameter coupling (pit depth, radius).

Based on this background, this study focuses on EH690 high-strength steel base metal and its welded joints. Neutral Salt Spray (NSS) tests are conducted to simulate extreme marine atmospheric corrosion environments (high salt spray + chloride ion erosion), followed by fatigue performance testing. The aim is to systematically reveal the influence of corrosion morphology on the fatigue performance of both base metal and welded joints, obtaining fatigue life data under different corrosion durations and loading parameters. For the corrosion fatigue of the base metal, relational functions are fitted based on scanned roughness parameter values, yielding S-N life prediction curves controlled by roughness. For the welded joints, the study introduces the stress concentration factor as a core quantitative tool to reveal the evolution law of stress concentration in EH690 high-strength steel welded joints (focusing on the weld toe region) under extreme marine environments and its influence mechanism on fatigue crack initiation/propagation. Ultimately, a unified “corrosion degree - stress concentration factor - fatigue life” prediction model is established. This research provides a theoretical basis and data support for the anti-corrosion fatigue design, life assessment, and safe service of engineering structures in earthquake/tsunami-prone sea areas.

 Experimental materials and methods Material properties

The base material used in this study was 16 mm thick EH690 high-strength steel manufactured by Nanjing Iron & Steel Co., Ltd. (China). The steel plates were delivered in the quenched and tempered (Q&T) condition. The manufacturing process involves quenching and tempering. The welding wire used was OK AristoRod 79 (Ø 1.2 mm), manufactured by ESAB AB (Gothenburg, Sweden). The chemical compositions and material properties of the EH690 high-strength steel and the welding wire are provided in Table 1 and 2, respectively.

Chemical composition of EH690 steel and welding materials (%).

Material C Si Mn P S Cr Ni V Mo Cu B Al Nb Ti Fe
EH690 0.13 0.20 1.36 0.008 0.002 0.22 0.31 0.034 0.235 0.02 0.0017 0.035 0.015 0.014 Bal.
OKAristoRod79 0.09 0.89 1.82 - - 0.25 2.03 - 0.64 - - - - - Bal.

Mechanical properties of EH690 steel and welding materials.

Material Elastic modulus E/MPa Yield strength R P0.2/MPa Tensile strength R m /MPa Elongation after fracture δ(%)
EH690 2.10 × 105 794 834 15.4
OKAristoRod79 - 825 900 17.0
Specimen preparation

Base metal corrosion specimens were machined from EH690 steel plates using wire electrical discharge machining (WEDM), with dimensions of 300 mm × 30 mm × 16 mm. The specimen surfaces were polished to ensure consistent surface quality across all specimens, thereby minimizing the influence of initial surface defects. After completing the cleaning, weighing, and scanning of the corroded specimens, they were further processed to the dimensions shown in Figure 1a. A separate set of uncorroded specimens was prepared as a control group.

Dimensions of corrosion-fatigue specimens: (a) base material, (b) welded joint.

Technical drawings labeled (a) and (b) compare a rectangular specimen’s dimensions with millimeter units, each showing a length of three hundred, thickness of sixteen, width of thirty, and a central notch with radius eighty and depth ten. Drawing (b) includes a dark crack-like feature at the notch center.

Welded joints were produced using gas metal arc welding, followed by heat preservation treatment to prevent cold cracking. The joints were subjected to non-destructive testing to confirm the absence of welding defects before being accepted for experimental use. Using the weld center as a reference, corrosion specimens measuring 300 mm × 30 mm × 10 mm were extracted by WEDM from steel plates measuring 1,000 mm × 402 mm × 10 mm. The structural design, process parameters, and quality of the welded joints complied with the requirements of the current Chinese standard GB/T 50661-2011 (Ministry of Housing and Urban-Rural Development of the People’s Republic of China, 2011). After corrosion testing, fatigue test specimens were finally machined by WEDM from the corroded steel plates (300 mm × 30 mm × 10 mm), again using the weld center as the reference point. The specimen design dimensions are shown in Figure 1b, and the welding process parameters are listed in Table 3.

Welding process parameters.

Welding pass Voltage U/V Current I/A Travel speed v/(cm/min) Heat input Q/(kJ/cm)
Root 20–23 170–190 28–32 ≤9.36
Fill 26–30 240–280 36–40 ≤14.00
Neutral salt spray test

This study employed a salt spray test chamber (Model YWX-750) that complies with the Chinese standard GB/T 10125-2021 (China Iron and Steel Association, 2021). The effective dimensions of the test chamber are 1,100 mm × 750 mm × 500 mm. The detailed test parameters are listed in Table 4.

Salt spray test conditions.

Test method Corrosive medium Temperature Average settlement rate (per 80 cm2 horizontal area) NaCl concentration pH
Neutral salt spray 5% NaCl solution 35 °C ± 2 °C 1.5 mL/h ± 0.5 mL/h 50 g/L ± 5 g/L 6.5∼7.2

The test included three corrosion durations: 240 h, 480 h, and 960 h, with three parallel specimens set for each duration. The specimens were placed horizontally in the salt spray chamber, and their positions were rotated every 24 h to ensure uniform corrosion progression. Figure 2 shows the salt spray test chamber and the arrangement of EH690 high-strength steel specimens. During the corrosion test, specimens were removed at predetermined intervals according to the experimental design. Following the standard GB/T 16545-2015 (China Iron and Steel Association, 2015), the specimens were derusted and subsequently used for corrosion morphology scanning tests.

(a) Salt spray corrosion test chamber. (b) Layout of base material specimens. (c) Layout of welded joint specimens.

(a) Laboratory corrosion testing chamber with transparent cover partially open, containing multiple metallic specimens arranged inside, surrounded by safety barriers and control panel visible on the right. (b) Close-up of metallic test samples aligned horizontally in the chamber under a spraying apparatus, showing surface residue. (c) Angled overhead view of metallic specimens in the chamber beneath an open protective lid, with spray nozzles positioned above for corrosion testing. Scanning system configuration

The scanning system used in this study primarily consists of the following components: the corroded specimen, a fixed support, a HonorScanII high-precision blue-light 3D scanner (with an equipment accuracy of 0.008 mm), power and data cables, and a computer, as shown in Figure 3a. The scanner was positioned near the specimen to be scanned, ensuring an optimal scanning distance, and directly captured the point cloud data of the parallel areas on the specimen. The specific scanning area is illustrated in Figure 3b and the specific welded joint scanning area is illustrated in Figure 3c.

(a) 3D scanning system; (b) Schematic diagram of the base material scanning area; (c) Schematic diagram of the scan area of the welded joint.

Figure with three labeled panels. Panel (a) shows a 3D scanner mounted on a tripod above a hold down support and specimen. Panel (b) displays a microscopic view of a material surface with a 100 micrometer by 100 micrometer scanning area marked. Panel (c) presents a similar surface with a weld joint in the center and two 100 micrometer by 120 micrometer scanning areas marked on either side of the joint. Fatigue testing methodology

The fatigue tests were conducted on a SUNS-890 electro-hydraulic servo fatigue testing machine. The loading device and data acquisition system are shown in Figure 4. Loading was applied under force control using a constant-amplitude sinusoidal waveform. Throughout the experiment, a frequency of 10 Hz and a stress ratio of R = 0.1 were used. The maximum test stress for intact specimens was set at 0.6 times the yield strength. The testing proceeded from higher stress levels to lower ones. High-cycle fatigue tests were carried out at three different stress levels, and the number of fatigue cycles at each level was recorded to obtain the S-N curve. To account for the potential of infinite life at low stress levels and considering time constraints, a run-out threshold of 2 million cycles was set.

SUNS-890 electro-hydraulic servo fatigue testing machine.

Universal testing machine applying a tensile load to a metallic specimen held between two cylindrical grips, with a control panel featuring dials, switches, and a remote on the front below the testing area. Results and discussion: corrosion behavior Mass loss and thickness degradation

Figure 5 shows the comparison of specimen conditions after different corrosion durations. The cleaning solution used was a 50% hydrochloric acid solution. The degree of corrosion degradation (DOD) is represented by the weight loss rate, while the equivalent thickness loss, Δh eq , serves as a parameter for characterizing corrosion thickness. This value can be calculated from the mass loss using the following Equations 1, 2: D O D = m 0 − m 1 m 0 Δ h e q = m 0 − m 1 2 ρ A where:

m 0 = initial mass of the uncorroded specimen (g)

m 1 = mass of the specimen after derusting (g)

ρ = density of the EH690 steel plate (7.85 g/cm3)

A = surface area of the uncorroded specimen (cm2)

Corrosion condition of specimens at different corrosion times. (a) Specimen after 240 h of corrosion. (b) Specimen after 480 h of corrosion. (c) Specimen after 960 h of corrosion.

Three rectangular metallic specimens, each with a narrowed center section, display increasing surface discoloration and corrosion after 240, 480, and 960 hours of exposure, respectively, labeled as figure a, b, and c.

After the salt spray corrosion test, the changes in corrosion weight are presented in Table 5, in which BM represents the base material and WJ represents the welded joint, and the relationship between the degree of corrosion degradation and thickness loss is shown in Figure 6. The condition of the specimens under different corrosion durations is described as follows:

Data sheet for corroded specimens.

Specimen Corrosion time T/h Initial weight m 0/g Weight after corrosion m 1/g Mass loss Δ m/g Degree of deterioration DOD/% Thickness loss Δh eq /μm Average thickness loss Δ h e q ¯ /μm
BM-1-1 240 h 1079.57 1067.7 11.87 1.11 52.94 45.6
BM-1-2 240 h 1076.47 1068 8.47 0.80 37.78
BM-1-3 240 h 1083.67 1073.9 9.77 0.91 43.58
BM-2-1 480 h 1077.37 1062 15.37 1.44 68.56 69.7
BM-2-2 480 h 1078.27 1063.1 15.17 1.42 67.66
BM-2-3 480 h 1081.37 1064.5 16.87 1.58 75.25
BM-3-1 960 h 1078.37 1056.9 21.47 2.02 95.76 111.4
BM-3-2 960 h 1078.47 1055.6 22.87 2.15 102.01
BM-3-3 960 h 1079.47 1051.2 28.27 2.65 126.1
WJ-1-1 240 h 1093.0 1082.3 10.7 0.98 47.73 50.7
WJ-1-2 240 h 1092.2 1081.3 10.9 1.00 48.62
WJ-1-3 240 h 1092.2 1079.7 12.5 1.14 55.75
WJ-2-1 480 h 1094.6 1076.4 18.2 1.66 81.18 69.6
WJ-2-2 480 h 1091.0 1074.4 16.6 1.52 74.04
WJ-2-3 480 h 1092.9 1080.9 12.0 1.10 53.52
WJ-3-1 960 h 1093.8 1068.4 25.4 2.32 113.29 110.8
WJ-3-2 960 h 1090.9 1066.4 24.5 2.25 109.28
WJ-3-3 960 h 1092.5 1067.9 24.6 2.25 109.73

Relationship between: (a) corrosion depth and corrosion extent, (b) average corrosion depth and corrosion time.

Two adjacent scientific line charts compare corrosion characteristics of base material and welded joint. Left chart plots thickness loss versus degree of corrosion, both showing linear increases with distinct fitting curves. Right chart plots average thickness loss against corrosion time, also with linear trends; base material and welded joint are clearly differentiated by markers and fitted lines.

At 240 h, slight thickness thinning occurred, with localized fine pitting corrosion and loss of metallic luster.

At 480 h, pitting corrosion increased and connected into small corrosion zones, accompanied by noticeable darkening in color.

At 960 h, the corrosion depth increased significantly, with evident corrosion pits and loss of surface flatness.

The data indicate a significant positive correlation between the degree of corrosion degradation and the equivalent thickness loss. The average corrosion depth increased over time, though the rate of increase slowed, which is attributed to the accumulation of corrosion products within the pits (Jia et al., 2018; Chat et al., 2015; Na et al., 2025).

 Corrosion morphology parameters

The surface roughness of the base material was calculated according to the formulae recommended in ISO 25178-2:2021 (International Organization for Standardization, 2021) (Equations 3–5): S a = 1 A ∬ Z x , y d x d y S q = 1 A ∬ Z 2 x , y d x d y 1 A ∬ Z x , y d x d y = 0

Where Z (x,y) denotes the height coordinate of each point and A represents the effective scanning area. The maximum surface height S z is primarily employed to characterize the maximum height difference of the corroded surface, while the arithmetic mean height S a and root mean square height S q are utilized to assess the average roughness of the corroded surface. The corrosion morphology of the base material is illustrated in Figure 7, and the calculated parameters are tabulated in Table 6. The data demonstrate that all three roughness parameters (S a, S q, and S z) within the base material region exhibit a positive correlation with corrosion time, as shown in Figure 8.

SEM images of base material corrosion morphology: (a) 240 h; (b) 480 h; (c) 960 h.

Three 3D surface plots labeled a, b, and c show varying surface roughness in micrometers over a nine millimeter square area, with elevation indicated by color gradients from blue (lowest) to red (highest) alongside vertical color bars indicating height scales specific to each plot.

Roughness parameters of the corroded material.

Roughness parameters Arithmetic mean height S a/μm Root mean square height S q/μm Maximum surface height S z/μm
Location Base material Left weld Right weld Base material Left weld Right weld Base material Left weld Right weld
BM-1-1 15.68 - - 20.09 - - 149.7 - -
BM-1-2 13.92 - - 22.94 - - 331.9 - -
BM-1-3 15.60 - - 19.90 - - 249.8 - -
BM-2-1 20.66 - - 26.12 - - 333.9 - -
BM-2-2 31.71 - - 39.92 - - 322.9 - -
BM-2-3 22.93 - - 29.21 - - 353.7 - -
BM-3-1 40.91 - - 50.46 - - 362.6 - -
BM-3-2 65.59 - - 80.29 - - 524.3 - -
BM-3-3 52.76 - - 68.12 - - 547.5 - -
WJ-1-1 - 107.2 145.3 - 125.7 166.8 - 678.1 713.8
WJ-1-2 - 164.0 128.2 - 187.9 150.5 - 863.6 725.1
WJ-1-3 - 107.9 107.6 - 128.4 126.4 - 770.6 726.8
WJ-2-1 - 109.7 90.2 - 129.1 107.6 - 728.9 611.2
WJ-2-2 - 131.8 121.0 - 166.8 141.5 - 867.4 788.2
WJ-2-3 - 116.3 123.0 - 138.7 144.4 - 933.7 907.5
WJ-3-1 - 119.1 135.9 - 148.8 171.1 - 753.2 887.7
WJ-3-2 - 165.0 126.7 - 188.4 151.2 - 772.1 819.3
WJ-3-3 - 76.01 113.6 - 95.87 140.7 - 875.3 710.1

Relationship between corrosion time and roughness parameters: (a) Arithmetic mean height; (b) Root mean square height; (c) Maximum surface height.

Three line graphs display the increase in surface roughness parameters against corrosion time in hours. Graph (a) uses a blue line for arithmetic mean height, graph (b) a green line for root mean square height, and graph (c) a red line for a different root mean square height, all showing upward trends.

No clear correlation was observed between the roughness parameters of the weld zone and corrosion time, which is attributed to the accelerated notch development at the weld due to corrosion (Shojai et al., 2025b). In this study, the focus was placed on the evolution of corrosion-induced morphological changes at the weld. However, the experimental results indicated that surface roughness alone could not adequately quantify the morphological changes or determine the geometric parameters at the weld. Therefore, geometric parameters such as the relative depth of corrosion notches at the weld toe were obtained by processing the scanned data. Figure 9 shows the morphology of the weld zone. As illustrated, at 240 h of corrosion, isolated pitting pits appeared at the weld and gradually connected to form linear corrosion grooves along the weld. By 480 h, the corrosion grooves expanded laterally, and the notch radius decreased. After 960 h, the corrosion grooves developed into continuous slots parallel to the weld, with the notch bottom becoming increasingly sharp. In the initial stage of corrosion, the preferential corrosion at the weld was driven by residual stresses and microstructural inhomogeneity. However, in the later stage, the accumulation of corrosion products inhibited further expansion of the corrosion notch, leading to the formation of sheet-like corrosion pits around the original pits (Zhang QH. et al., 2025).

Scanning results of weld seams under different corrosion times: (a) Specimen after 240 h; (b) Specimen after 480 h; (c) Specimen after 960 h.

Three 3D surface plots labeled a, b, and c show colored height maps of textured surfaces, each with an inset vertical view and corresponding color bar ranging from blue at lowest elevations to red at highest.

The relative notch depth, h q, is defined as the distance from the surface of the corroded specimen to the bottom of the deepest corrosion pit within the corrosion notch at the weld toe. The equivalent notch radius, r, is derived by approximating the irregular pit bottom as an arc-shaped surface and fitting it using the minimum circumscribed circle method. A schematic illustration is provided in Figure 10, with specific data summarized in Table 7. The relationship between corrosion time and the corresponding parameters is presented in Figure 11.

Schematic diagram for the calculation of relative notch depth and equivalent notch radius.

Diagram illustrating a cross-section of a welded joint with a corroded notch at the interface of base material and weld joint. The upper segment shows the full weld cross-section with a red highlighted zoom area. The lower left inset details the measurement of average height after corrosion and relative notch depth, while the lower right inset shows radius and minimum circumscribed circle at the notch tip, all with labeled base material and weld joint regions.

Notch parameters at weld toe.

Specimen Notch location Relative corrosion notch depth h q/mm Standard deviation/mm Mean value/mm Corrosion notch radius r/μm Standard deviation/μm Mean value/μm
WJ-1-1 L 0.729 0.068 0.649 375 22.5 394.0
R 0.596 424
WJ-1-2 L 0.686 385
R 0.634 375
WJ-1-3 L 0.551 384
R 0.698 421
WJ-2-1 L 0.691 0.067 0.716 185.6 27.1 206.5
R 0.766 164.8
WJ-2-2 L 0.629 216.8
R 0.718 211.2
WJ-2-3 L 0.816 240.8
R 0.675 220
WJ-3-1 L 0.809 0.075 0.755 127.2 26.0 127.3
R 0.662 98.4
WJ-3-2 L 0.769 116
R 0.75 112.8
WJ-3-3 L 0.68 136
R 0.858 173.6

Relationship between corrosion time and notch parameters: (a) Relative notch depth; (b) Notch radius.

Panel (a) presents a line graph showing the average value of relative notch depth in millimeters increasing with corrosion time in hours, with data points and a blue curve. Panel (b) displays a line graph illustrating the average notch radius in micrometers decreasing as corrosion time increases, with data points and a green curve.

As shown in the figure, the average notch depth measures 0.645 mm at 240 h, increases to 0.708 mm at 480 h, and reaches 0.757 mm at 960 h. The depth at 960 h shows an increase of 17.4% compared to that at 240 h, demonstrating a significant upward trend with prolonged corrosion time, though the rate of increase slows. This indicates continuous attack by the corrosive medium at the weld toe region, leading to progressive deepening of the notch and enhanced geometric discontinuity of the cross-section. Meanwhile, the average notch radius is 397.96 μm at 240 h, decreases to 199.04 μm at 480 h, and further reduces to 134.94 μm at 960 h. The radius at 960 h exhibits a reduction of 66.1% compared to that at 240 h, showing a notable downward trend over time, with a decelerating rate of decrease. This suggests that corrosion preferentially attacks the bottom and sidewalls of the notch, resulting in a progressively sharper notch geometry.

 Fatigue performance and life prediction

Fatigue tests were conducted on a total of 15 base metal specimens and 15 welded joint specimens. The fatigue results for uncorroded specimens and those subjected to different corrosion durations are summarized in Table 8 and 9, in which BM-0 denotes the uncorroded base metal specimens (Table 8) and WJ-0 denotes the uncorroded welded joint specimens (Table 9). All fractures in the corroded welded joint specimens occurred in the weld region.

Fatigue test results of base material.

Specimen Corrosion time T/h Load ratio R Max. Stress S max/MPa Test stress amplitude Δ σ T /MPa Surface roughness S z/μm Stress concentration factor K C Modified stress amplitude Δ σ /MPa Fatigue life N/cycles
BM-0-1 0 0.7 554.48 304.96 - 1 304.96 133,805
BM-0-2 0 0.6 475.27 261.4 - 1 261.4 285,044
BM-0-3 0 0.55 435.66 239.61 - 1 239.61 438,674
BM-0-4 0 0.5 396.06 217.83 - 1 217.83 683,154
BM-0-5 0 0.4 316.84 174.26 - 1 174.26 2,000,000
BM-0-6 0 0.3 237.63 130.7 - 1 130.7 2,000,000
BM-1-1 240 0.6 475.35 261.44 149.7 1.146 299.65 157,754
BM-1-2 240 0.5 396.23 217.93 331.9 1.327 293.01 182,131
BM-1-3 240 0.4 316.91 174.3 249.8 1.286 223.73 506,052
BM-2-1 480 0.6 475.53 261.54 333.9 1.379 351.96 74,936
BM-2-2 480 0.5 396.22 217.92 322.9 1.298 291.81 198,893
BM-2-3 480 0.4 316.97 174.33 353.7 1.387 236.54 374,055
BM-3-1 960 0.6 475.51 261.53 362.6 1.443 356.08 62,421
BM-3-2 960 0.5 396.09 217.85 524.3 1.401 310.13 146,624
BM-3-3 960 0.4 316.86 174.27 547.5 1.446 249.21 316,628

Fatigue test results.

Specimen Corrosion time T/h Load ratio R Max. test force P max / kN Min. test force P min / kN Test stress amplitude Δ σ T /MPa Stress concentration factor K c Modified stress amplitude Δ σ /MPa Ratio of relative notch depth to notch radius (h q/r) Fatigue life N/cycles
WJ-0-1 0 0.7 77.32 7.73 434.94 - 434.94 - 24,527
WJ-0-2 0 0.6 66.27 6.63 372.75 - 372.75 - 55,471
WJ-0-3 0 0.55 60.75 6.07 341.75 - 341.75 - 63,829
WJ-0-4 0 0.5 55.22 5.52 310.63 - 310.63 - 79,282
WJ-0-5 0 0.4 44.18 4.42 248.50 - 248.50 - 481,872
WJ-0-6 0 0.3 33.14 3.31 186.44 - 186.44 - 863,735
WJ-1-1 240 0.6 66.27 6.63 373.99 1.229 459.63 1.944 17,699
WJ-1-2 240 0.5 55.22 5.52 311.55 1.195 372.31 1.782 46,973
WJ-1-3 240 0.4 44.18 4.42 249.26 1.185 295.37 1.658 137,261
WJ-2-1 480 0.6 66.27 6.63 374.60 1.341 502.34 4.648 11,728
WJ-2-2 480 0.5 55.22 5.52 312.21 1.319 411.80 3.400 29,445
WJ-2-3 480 0.4 44.18 4.42 249.66 1.334 333.04 3.389 78,717
WJ-3-1 960 0.6 66.27 6.63 375.08 1.359 509.74 6.728 Test
WJ-3-2 960 0.5 55.22 5.52 312.84 1.372 429.22 6.649 24,305
WJ-3-3 960 0.4 44.18 4.42 250.21 1.359 340.03 5.000 71,492

The S-N curve method is commonly used for fatigue life assessment. This method is based on the linear cumulative damage hypothesis, where each cycle of loading causes a certain degree of damage to the specimen. Failure occurs when the accumulated damage reaches a critical value. The S-N curve is expressed by a power function relationship (Equation 6): S max m N = C where:

S max: maximum applied stress (MPa),

N: fatigue life (number of cycles),

m and C: material-dependent constants.

For the high-cycle fatigue tests discussed in this paper, a functional relationship expressed in terms of stress amplitude is commonly used (Equation 7): Δ σ m N = C

Where Δ σ is the stress amplitude in MPa.

The baseline S-N curve for uncorroded EH690 base metal was obtained by fitting the experimental data. The fitted stress amplitude values were derived by substituting the fatigue life of corroded specimens into this curve. The stress concentration factor, K c, was then determined by dividing these fitted stress amplitudes by the experimental stress amplitudes. By fitting the relationship between the stress concentration factor K c and surface roughness parameters using the least squares method, it was found that S a and S q exhibited poor correlation with K c. However, the surface roughness parameter S z showed a positive correlation with K c, increasing consistently, as illustrated in Figure 12. Therefore, the fatigue life of the base metal under different roughness conditions can be evaluated by modifying the experimental stress amplitude using the following relationships (Equations 8–10): K c = 5.6261 ln ⁡ S z S z − 0.8 + 1.6587 Δ σ M = K c .   Δ σ T N = 5892.612 Δ σ M − 0.24888 where:

K c: stress concentration factor,

S z: maximum surface corrosion height, with valid values ranging from 149.7 to 547.5 μm,

Δ σ T : experimental stress amplitude in MPa,

Δ σ M : modified stress amplitude in MPa.

Fitted curve: relationship between S z and K C .

Scatter plot with black squares showing data for Stress Concentration Factor Kc versus Sz in micrometers, displaying a red curve indicating a positive correlation as Sz increases from 100 to 600 micrometers.

After correction of the experimental stress amplitude, the resulting data are presented in Figure 13. In practical engineering applications, accurate prediction of the fatigue life of EH690 steel can be achieved simply by obtaining the actual stress amplitude and the maximum surface corrosion height, and applying this functional relationship.

S-N curves for base material: (a) Unmodified stress amplitude; (b) Modified stress amplitude.

Two side-by-side scatter plots show the relationship between stress amplitude and fatigue life for different specimens. Plot (a) features colored points representing samples exposed to 0, 240, 480, and 900 hours, along with a red S-N curve for the base material. Plot (b) displays blue squares for corrosion specimens, red squares for base materials, and a black S-N curve for the base material, each indicating a negative correlation between stress amplitude and fatigue life.

The test results indicate that with increasing corrosion time, the combined effects of material loss and increased surface roughness lead to a significant and progressive reduction in the fatigue life of the specimens. At higher load levels (0.6–0.5 times the yield strength), surface roughness exerts a major influence on fatigue life, as the associated stress concentration becomes the primary driver for crack initiation. In contrast, the effect of roughness is relatively weaker at lower load levels (0.4 times the yield strength).

The fracture positions of the 15 corrosion fatigue specimens (Contains 6 uncorroded specimens and 9 corroded specimens) tested in this test all occur in the corrosion notch of the weld toe, as shown in Figure 14. By observing the macroscopic section of the fatigue specimen, the cracks in the butt fatigue specimen always start and expand in the weld area. After fatigue loading of the butt joint, cracks occur in the corrosion pit at the bottom of the notch. As shown in Figure 15.

Schematic of failure location and macro fracture: (a) Uncorroded fatigue specimens; (b) Corrosion fatigue specimens.

Four metallurgical sample pieces are shown in a vertical row in panel a, with an arrow pointing to a close-up of a single fractured cross-section. Panel b displays three similar samples with visible surface cracks and an arrow pointing to another close-up cross-section showing a different fracture surface.

The location of the corrosion notch crack at the weld toe after fatigue loading of the butt joint.

Microscopy image showing a dark region at the top and a lighter textured area at the bottom, with a red dashed rectangle highlighting the crack origination location, indicated by an arrow and label; scale bar denotes one hundred micrometers.

The maximum stress was modified by calculating the stress concentration factor. Corrosion notches on both the left and right sides of the weld are shown in Figure 16.

Cross-sectional view of the deepest corrosion notch at the weld.

Cross-sectional diagram illustrating a corroded region around a weld joint, with labeled measurements including corrosion test specimen thickness h, minimum notch thickness d, relative notch depth hn, and sections for base material and weld joint.

In the course of the previous research, the S-N curve of the EH690 high-strength steel butt joint without corrosion has been established (ΔσT 4.6322⋅N = 3.8088 × 1016). The test results of fatigue life of EH690 high-strength steel butt joints after corrosion are shown in Table 9 and Figure 17a. The test results show that under the same corrosion conditions, the fatigue life of the specimen decreases significantly with the increase of stress amplitude, and the fatigue life index m value is basically the same as that of the uncorroded specimen. Under the same fatigue load, the fatigue life of the specimen gradually decreases with the extension of corrosion time, and the longer the corrosion time, the more obvious the decreasing trend of life is. Through micromorphological observation and stress analysis, it can be seen that the corrosion environment will cause the weld toe of the specimen to form a corrosion notch, and form a stress concentration in the local corrosion pit at the bottom of the corrosion notch, so that cracks will occur under the action of fatigue load. The deeper the relative corrosion notch depth h q, the smaller the notch radius h q of the local corrosion pit, the more obvious the stress concentration phenomenon and the more the fatigue life decreases. Based on the fact that the fatigue life index m of EH690 high-strength steel before and after corrosion is basically unchanged, it is assumed that the actual stress amplitude Δσ of the corroded specimen is the same as that of the uncorroded specimen under the same fatigue life, and the stress concentration coefficient K c is calculated based on linear regression. The actual stress amplitude and fatigue life of the corrosion fatigue specimen corrected by the stress concentration coefficient are shown in Figure 17b.

S-N curves for welded joint: (a) Unmodified stress amplitude; (b) Modified stress amplitude.

Left panel (a) shows a scatter plot of test stress amplitude versus fatigue life for welded joints at four corrosion durations, with a downward-sloping red S-N curve; right panel (b) displays a similar plot using modified stress amplitude, with blue and red points representing corrosion specimens and welded joints, along with a black S-N curve for the welded joint.

According to the existing research results (Singh, 2025) and the test results of this test, the stress concentration degree of the corrosion notch is affected by the relative notch depth h q and the notch radius r, which is directly proportional to the relative corrosion notch depth h q and inversely proportional to the local corrosion pit notch radius r. Based on the experimentally measured h q, r (Table 7), stress amplitude ΔσT and fatigue life N, the functional relationship between the stress concentration coefficient K c and the corrosion notch geometric parameters (such as the ratio of relative notch depth h q to the notch radius r) is derived by the least squares method, as shown in Figure 18 and Equation 11. K c = 1 + 0.15482 h q r

Fitted curves: relationship between h q / r and K c .

Scatter plot with a red trend line showing Stress Concentration Factor Kc on the y-axis versus Ratio of relative notch depth to notch radius hq/r on the x-axis, indicating Kc increases as hq/r rises.

In practical engineering applications, by measuring the corrosion morphology parameters of EH690 butt-welded joints (relative notch depth h q and notch radius r), the stress concentration factor can be obtained directly through this equation, enabling accurate prediction of their fatigue life. This approach significantly reduces the workload associated with fatigue life assessment in engineering practice.

 Conclusion

Based on the corrosion and fatigue tests conducted on EH690 high-strength steel base metal and its welded joints, the following conclusions are drawn:

With prolonged corrosion time (240 h–960 h), the mass loss and equivalent thickness reduction of the base metal intensified continuously, accompanied by increased surface roughness and localized deep pitting. The mass loss increased by 75.74% from 240 h to 960 h, and the surface roughness (S a) value rose to 203.37 μm at 960 h (compared to 47.15 μm at 240 h). At the weld toe, the corrosion notch depth increased significantly while the notch radius decreased consistently, resulting in a progressively sharper geometry. By 960 h, the notch depth had increased by 17.4%, and the radius had decreased by 66.1% relative to the values at 240 h.

With increasing exposure time in the salt spray environment, the fatigue life of corroded specimens decreased significantly compared to uncorroded specimens under the same stress level. The influence of surface roughness on fatigue life was more pronounced under higher load levels. The evolution of corrosion notch morphology led to a monotonic increase in the stress concentration factor, which consequently resulted in a notable reduction in the fatigue life of welded joints. The actual stress-based S-N curve more accurately reflects the fatigue behavior under service conditions.

A correlation model between the geometric parameters of corrosion notch morphology and the stress concentration factor was established. A method for modifying the stress amplitude of corroded EH690 welded joints was proposed, enabling accurate fatigue life prediction through the measurement of corrosion notch geometric parameters. This provides a theoretical basis for the anti-fatigue design, life assessment, and maintenance cycle planning of steel structures.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

 Author contributions

CL: Writing – original draft, Writing – review and editing. ZL: Writing – original draft, Writing – review and editing. YZ: Writing – original draft, Writing – review and editing. BX: Writing – review and editing, Writing – original draft. CC: Writing – review and editing, Writing – original draft. ZP: Writing – original draft, Writing – review and editing.

 Conflict of interest

Author(s) CL, ZL, YZ, and BX were employed by Shanghai Branch of CNOOC (China) Co., Ltd.

The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

 Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References Ban H. Shi G. (2018). A review of research on high-strength steel structures. Proc. Institution Civ. Engineers-Structures Build. 171 (8), 625–641. 10.1680/jstbu.16.00197 Chatisathien C. Phetchakul T. Po-yai A. (2015). Atmospheric corrosion behavior assessment of carbon steel pipes using cyclic salt spray test. J. Mater. Eng. Perform. 24 (10), 3833–3841. Chen M. T. Pan Y. Ren F. M. Cao H. Lai M. Zhang J. (2025). Experimental investigation and predictive model on tensile behavior of corroded high-strength steel and butt-welded connections. Thin-Walled Struct. 210, 112982. 10.1016/j.tws.2025.112982 China Iron and Steel Association (2015). Corrosion of metals and alloys-removal of corrosion products from corrosion test specimens GB/T 16545-2015. Beijing: China Standard Press. China Iron and Steel Association (2021). Artificial atmosphere corrosion tests-salt spray tests GB/T 10125-2021. Beijing: China Standard Press. Fu L. Xu G. B. Yan Y. Yang J. Xie J. (2018). The application and research progress of high strength and high performance steel in building structure. IOP Conf. Ser. Mater. Sci. Eng. 392 (2), 022008. 10.1088/1757-899x/392/2/022008 Garbatov Y. Guedes S. Parunov J. (2014). Fatigue strength experiments of corroded small scale steel specimens. Int. J. Fatigue 59, 137–144. 10.1016/j.ijfatigue.2013.09.005 Guan X. D. Guo H. C. Li G. Q. Wang Y. Pan Y. (2025). Study on fatigue crack growth behaviour of Q690D high strength steel with corrosion damage. Eng. Fail. Anal. 174, 109495. 10.1016/j.engfailanal.2025.109495 Guo H. C. Wei H. H. Li G. Q. Wang Y. (2021a). Experimental research on fatigue performance of corroded Q690 high strength steel. J. Mater. Civ. Eng. 33 (11), 0899–1561. 10.1061/(asce)mt.1943-5533.0003929 Guo H. C. Wei H. H. Li G. Q. Sun F. (2021b). Experimental research on fatigue performance of butt welds of corroded Q690 high strength steel. J. Constr. Steel Res. 184, 106801. 10.1016/j.jcsr.2021.106801 Heng L. Liang Z. Ke L. X. (2025). Experimental investigation on fatigue performance of corroded Q690D high-strength steel. J. Constr. Steel Res. 207, 109405. 10.1016/j.jcsr.2025.109405 International Organization for Standardization (2021). Geometrical product specifications (GPS)—Surface texture: areal—part 2: terms, definitions and surface texture parameters ISO 25178-2:2021. Geneva: ISO. Jia C. Shao Y. Guo L. Liu Y. (2018). Incipient corrosion behavior and mechanical properties of low-alloy steel in simulated industrial atmosphere. Constr. Build. Mater. 187, 1242–1252. 10.1016/j.conbuildmat.2018.08.082 Jiang C. Wu C. Cai C. Jiang X. Xiong W. (2020). Corrosion fatigue analysis of stay cables under combined loads of random traffic and wind. Eng. Struct. 206, 110153. 10.1016/j.engstruct.2019.110153 Li L. Li Y. Shi W. Li C. Q. (2021). Deterioration of fatigue strength of bolted connection plates under combined corrosion and fatigue. J. Constr. Steel Res. 179, 106559. 10.1016/j.jcsr.2021.106559 Li Q. Yao Q. Sun L. Y. Ma H. Zhang C. Wang N. (2023). Effect of micro-galvanic corrosion on corrosion fatigue cracking of the weld joint of high strength bridge steel. Int. J. Fatigue 170 (5), 107568. 10.1016/j.ijfatigue.2023.107568 Liu P. C. Chen C. Shao Y. B. (2025). Fatigue properties of microstructure in coarse-grained heat-affected zone of EH690 high strength steel welded joint. Trans. Mater. Heat Treat. 46 (10), 200–212. 10.13289/j.issn.1009-6264.2024-0539 Ministry of Housing and Urban-Rural Development of the People's Republic of China (2011). Code for welding of steel structures GB 50661-2011. Beijing: China Standard Press. Na Y. S. Pradhan S. K. Kim D. G. Yun Y. S. Oh M. S. (2025). Electrochemical evaluation of corrosion behavior and product characterization of Zn–Mg–Al alloy-coated steels in a laboratory-simulated full immersion marine environment. Mater. Today Commun. 49, 114036. 10.1016/j.mtcomm.2025.114036 Shao Y. Wang M. Gao X. Chen C. Zhu H. (2025). Fatigue crack growth of EH36/EH690 dissimilar steel welded joints in corrosive environments. J. Constr. Steel Res. 235, 109926. 10.1016/j.jcsr.2025.109926 Shen X. K. Chen C. Shao Y. B. Gao X. D. Zhu H. J. Yu Z. B. (2025). Investigation of welded residual stress in EH690 high-strength steel butt joints. J. Constr. Steel Res. 232, 109613. 10.1016/j.jcsr.2025.109613 Shojai S. Kabha K. Woitzik C. Braun M. Ghafoori E. (2025a). Fatigue behaviour of 12-month corroded offshore steel joints under accelerated salt spray exposure: an experimental and numerical analysis. Weld. World 69 (8), 2351–2369. 10.1007/s40194-025-02043-0 Shojai S. Schönamsgruber F. Köhler M. Ghafoori E. (2025b). Impact of accelerated corrosion on weld geometry, hardness and residual stresses of offshore steel joints over time. Mater. Des. 251, 113578. 10.1016/j.matdes.2024.113578 Singh K. L. (2025). Pitting corrosion as a failure mechanism in aircraft panels: a modelling and simulation approach. Adv. Aerosp. Sci. Technol. 10 (2), 64–79. 10.4236/aast.2025.102005 Wu H. Lei H. Frank C. Y. (2021). Comparison on mechanisms of high-cycle fatigue performance of structural steel exposed to urban industrial atmosphere and laboratory simulated corrosive environment based on infrared thermography. Int. J. Fatigue 145, 106098. 10.1016/j.ijfatigue.2020.106098 Wu J. Xu S. Li A. Wang B. Wang Y. (2025). Prediction model of fatigue crack propagation life of corroded steel plate considering pit characteristics. Int. J. Fatigue 190, 108591. 10.1016/j.ijfatigue.2024.108591 Yan L. Liu Z. X. Wang P. P. (2025). Corrosion failure mechanism and electrolytic accelerated fatigue performance of butt-welded joints. J. Constr. Steel Res. 227, 109354. 10.1016/j.jcsr.2025.109354 Yang M. Y. Kainuma S. Nabeshima W. Yamauchi M. Ji B. (2022). Synergistic effect of corrosion and residual stress on fatigue cracks of finger-type bridge expansion joints. Eng. Fail. Anal. 131, 105894. 10.1016/j.engfailanal.2021.105894 Yang L. Guo Y. Y. Chen L. (2025). Research progress on mechanical properties of ultra-high strength (fy>690 MPa) steel structures. Steel Construction (Chinese and English) 40 (3), 1–19. 10.13206/j.gjgS24042201 Zhang H. P. Liu H. J. Deng Y. Cao Y. He Y. Liu Y. (2025a). Fatigue behavior of high-strength steel wires considering coupled effect of multiple corrosion-pitting. Corros. Sci. 244, 112633. 10.1016/j.corsci.2024.112633 Zhang Q. H. Li X. B. Liu X. H. (2025b). Corrosion evolution and intelligent performance prediction method for welded joints of high-strength weathering steel. China Civ. Eng. J. 58 (3), 1–12.

Edited by: Ehsan Ghassemali, Jönköping University, Sweden

Reviewed by: Qing Zhang, Jönköping University, Sweden

Katerina Chantziara, Karlstad University, Sweden