A Comparison Study of Fatigue Behavior of S355J2+N, S690QL and X37CrMoV5-1 Steel

: Steel of the mild-strength S355J2+N steel grade is the most often used steel for manufac-turing carrying sections of constructions exposed to fatigue loads. The use of high-strength steels, such as S690QL, allows for the creation of structures that are light and simple to construct. However, increasing the yield strength of high-strength steels does not result in a corresponding increase in fatigue resistance. As a result, using high-strength steels for constructions subjected to fatigue loading can be a major design concern, raising the question of whether high-strength steels should be used at all. Most of the experimental investigations regarding the hot work tool steel X37CrMoV5-1 found in the literature are focused on its machining and wear resistance, with insufﬁcient attention paid to the cyclic loads. This article evaluates the fatigue properties of mild-strength S355J2+N, high-strength S690QL, and X37CrMoV5-1 steel grades. A SHIMADZU servo-hydraulic testing machine is used to perform uniaxial tensile tests under uniaxial fatigue stress-controlled, fully reversed conditions (tensile–compression testing with R = − 1 stress ratio) in accordance with EN ISO and ASTM standards. The aim of this paper is to highlight the fatigue characteristics of these three steels that are among the most used in their respective groups. Steel S355J2+N belongs to the group of hot-rolled normalized steels, S690QL belongs to the group of improved (quenched + tempered) steels with increased strength, and X37CrMoV5-1 belongs to the group of high-alloyed tool steels for hot work. This choice was made as the tested steels can be considered typical representatives of their groups. Based on the test results of these three steels, which are organized in S – N curves, the fatigue behavior of the entire mentioned groups of steels can be foreseen.


Introduction
Despite the development of new alloys [1] and composite materials [2], steels are still the most widely utilized materials in mechanical [3] and civil engineering [4]. Steel structural elements and constructions are frequently subjected to varying loads over their service (fatigue) lives [5,6]. Due to the nature of their sufficient qualities and inexpensive prices, structural steels S235 [7], S275 [8], and S355 [9,10] are the most extensively used steels for structural elements and constructions exposed to fatigue stress [11]. High-strength steel [12] is used in order to meet the requirements for light constructions with simple designs while having excellent structural performance.
The S-N curves proposed in the design regulations EN 1993-1-9 [13] and EN 13001-3-1 [14] do not reveal the exact material dependence; hence, the exact fatigue properties of different steel types must be determined. In general, the fatigue behavior of structural steels is well-understood and investigated by many researchers [9,10,15], but high-strength steels are still understudied, and determining their mechanical properties (fatigue) has become the particular focus of several researchers [16][17][18][19].
Experimental testing of steel is unavoidable and crucial in fatigue life prediction as different specimens made of the same material can have different impurities, non-metallic inclusions, and slightly different compositions of alloying elements, which, depending on the steel grade, can scatter results significantly.
Fatigue life prediction is a field gaining considerable attention in material science, and multiaxial fatigue models [20], which can be based on stress, strain, energy, or fracture mechanics [21], are utilized. Each of these approaches is suitable for specific loading conditions; for instance, stress-based models are the most appropriate for high cycle fatigue (such as the one presented in this paper), while strain-based models are better suited for low cycle fatigue in which plastic deformation is notable [21]. Energy-based models consider both contributions from stress and strain and can predict the reduction in fatigue life due to the out-of-phase hardening [21], while the Strain Energy Density-based (SED) fatigue model can predict ratcheting phenomena [22]. Based on the experimental analysis of crack formation and growth studied in the fracture mechanics field, the critical plane method is used to predict ductile or brittle failure [21]. The maximum normal stress range plane is used to predict brittle failure, and the maximum shear stress range plane is applied to estimate ductile failure [21]. This analysis requires notched specimens and a high-resolution camera with an optical device with variable magnification [23]. Another similar issue is the requirement of servo-hydraulic machines that can perform multiaxial loading (including torsion); hence, researchers often propose the extrapolation of data obtained using a uniaxial test [20][21][22][23]. In fact, in [20], the authors determined most of the coefficients using a monotonic tensile test, while the authors in [23] used two uniaxial fully reversed strain-controlled tests, one with a higher strain amplitude and another with a lower strain amplitude, to evaluate the strain-life relationship and two elastic-plastic numerical models, which they used to complement the experimental data. In this paper, we present our contribution to the field of fatigue damage assessment of high-strength steels, which complements existing research and provides engineers with valuable data that will facilitate new lightweight construction designs. The fatigue characteristics of the highstrength steel grade S690QL, also known as STRENX 700 (according to manufacturer SSAB Corporation, Stockholm, Sweden), in the gigacycle region of loading are presented in [24]. However, the experiments presented in [24] are conducted at the high-frequency loading of f ≈ 20 kHz, which is unrealistic for the most real-life steel constructions. The ultrasonic impact treatment of welded joints made of S690QL steel is studied in [25]. This treatment increases the fatigue limit at N = 10 7 cycles and at the testing frequency of f = 35 Hz by about 12% [25]. In our research, we use even lower frequencies of 10 and 15 Hz for up to N = 2 × 10 6 cycles on untreated material, filling the gap between [24] and [25].
The hot work tool steel X37CrMoV5-1, also known as H11 [26], according to the American Iron and Steel Institute (AISI), is used primarily for making tools for the die casting of aluminum alloys, copper, plastic, and other materials. It has high thermal fatigue [26] and wear resistance [27], and it is suitable for the machining processing [28]. In the literature, hot work tool steels are usually tested to withstand low cycle fatigue [29], which is, of course, in accordance with their purpose [26]. However, in this paper, we present the experimental fatigue testing of X37CrMoV5-1 with N = 2 × 10 6 cycles, the same as the two other steel grades, in order to carry out a corresponding comparison between them. This addresses the feasibility of expanding X37CrMoV5-1 implementation for the construction of parts and constructions subjected to long-lasting fatigue loads and provides engineers with the appropriate data for designing such elements.
The remainder of the paper is organized as follows: in the Materials and Methods Section, a theoretical background of fatigue behavior analysis is given, followed by an experimental setup description. This includes a description of the chemical composition of the tested materials and specimen dimensions, as well as a description of the testing machine and equipment. References to related standards are also given. The Results Section is dedicated to the presentation of the S-N curve charts and fatigue strengths given in tables. The Discussion Section features a comparison between the tested steels and the possible implications of the obtained results on mechanical parts and construction design. In the Conclusion, we point out that although S355J2+N has a significant strain hardening, while S690QL and X37CrMoV5-1 steel have only minor strain hardening, the reduction in the dimension of structural elements made using high-strength steels can lead to the elevation of the fatigue sensitivity in these elements.

Fatigue Behavior Assessment
Fatigue tests evaluate a material's resistance to damage, loss of strength, and failure when subjected to cyclical loading. The fatigue test approaches are classified into S-N and ε-N linear fracture mechanics approaches as classical approaches, and the increasingly popular energy-based approaches. Equation (1) [30,31], Equation (2) [31,32], and Equation (3) [31,33] shown below are the most well-known relationships for describing fatigue behavior: In Equations (1)-(3) σ a is the true stress amplitude, and 2N f is the number of reversals to failure. The fatigue strength coefficient is denoted with σ f and the fatigue strength exponent with b. The fatigue ductility coefficient is given with ε f and the fatigue ductility exponent with c. ε a , ε a,e , ε a,p are the total, elastic, and plastic strain amplitudes, respectively. E is the Young's modulus. The commonly known stress-life (S-N) approach, based on the Basquin model [30] specified in Equation (1), was used to determine the fatigue properties of the studied materials (S355J2+N, S690QL, and X37CrMoV5-1). The S-N approach is a global strategy that is related to the stress range, and it is the foundation of several fatigue life standards, such as Eurocode 3, part 1-9 [32]. Given the fact that there is only one stress component in most fatigue life calculations, the math is relatively simple. The results of the S-N approach show a clear relationship between a global definition of the stress range (stress amplitude) and the total number of reversals to failure. The Basquin equation (Equation (1)) is often adopted for representing the Wöhler curve [34] as a straight line in a double logarithmic plot.

Basic Properties of Studied Steels
Mild-strength S355J2+N steel, high-strength S690QL steel, and hot work tool steel grade X37CrMoV5-1 steel were used for this experimental study. A comparison of the chemical composition (in terms of weight %) of these steels [35][36][37][38] was obtained using spark emission spectrometry and is shown in Table 1. The S355J2+N and S690QL steels are suitable for welding; however, due to the higher amount of alloy elements, the weldability of high-strength steels is often less than that of mild-strength steel. On the other hand, X37CrMoV5-1 has poor weldability but good machinability. - The specimens used to determine and validate the static strength properties of the three steel grades used in the experimental program were prepared according to the standards EN ISO 6892-1 [39] and ASTM E8M-01 [40]. Uniaxial tensile tests were performed on three representative flat specimens for each steel, with the same thickness in all cross-sections to investigate the static strength properties.
The technical drawing and real shape of one of the investigated specimens, before testing, is shown in Figure 1. All nominal dimensions of the specimen shown in Figure 1 are in millimeters (mm). The uniaxial tensile tests on the specimens to determine their static strength properties (mechanical characteristics) were performed using a SHIMADZU type EHF EV101K3-070-0A servo-hydraulic testing machine (Shimadzu Corporation, Tokyo, Japan) with a force of ±100 kN and a stroke of ±100 mm ( Figure 2).
The uniaxial tensile tests were carried out in accordance with EN ISO 6892-1 [39] and ASTM E8M-01 [40] at room temperature (23 ± 5 • C) with a constant stroke control rate of 4 mm/min (strain rate 10 −3 s −1 ) without a change in the speed of testing. One of the investigated specimens (S355J2+N steel grade) at the end of the uniaxial tensile test is presented in Figure 3a. An MFA25 extensometer (MF Mess & Feinwerktechnik GmbH, Velbert, Germany) with a gauge length of 50 mm was used to determine the Young's modulus and elongation, and it is shown in Figure 3b.  As an outcome of the uniaxial tensile tests on the three flat specimens, for each steel, we determined the mechanical characteristics of the S355J2+N, S690QL, and X37CrMoV5-1 steel grades. The average yield strength, tensile strength, and Young's modulus for each steel are shown in Table 2. The force-displacement responses were recorded for all tested specimens, and the response of "Specimens with results close to average values of mechanical characteristics" was selected as the representative for the determination of the tensile strain-hardening exponent, n, and strength coefficient, K, according to ASTM E646-00 [41], for each tested steel grade.
Another result of the conducted uniaxial tensile testing is the stress-strain curves for the S355J2+N, S690QL, and X37CrMoV5-1 steel grades that are shown in Figure 4 for one of the representative specimens (a specimen with results close to the average values of the mechanical characteristics) for each investigated steel.

Fatigue Analysis Using the Experimental Method
This section describes a complete fatigue characterization of mild-strength S355J2+N steel, high-strength S690QL steel, and hot work tool steel grade X37CrMoV5-1 steel, carried out according to the internal procedures of the Centre for Engineering Software and Dynamic Testing at the Faculty of Engineering University of Kragujevac using a SHIMADZU type EHF EV101K3-070-0A servo-hydraulic testing machine (Shimadzu Corporation, Tokyo, Japan) with a force of ±100 kN and a stroke of ±100 mm, based on ASTM E468-90 [42].
All specimens utilized to determine the fatigue properties of each steel were prepared according to the standard E468-90 [42]. The technical drawing and real shape of one of the investigated specimens for fatigue testing, before testing, is shown in Figure 5. All nominal dimensions of the specimen shown in Figure 5 are in millimeters (mm). All specimens were finely polished to minimize surface roughness effects. The mean roughness level achieved on the surface of the gauge length of the specimens was in the range of 1-5 µm after polishing. To calculate the roughness measurement the SJ-210 Portable Surface Roughness Tester (Mitutoyo America Corporation, Aurora, IL, USA) was used. One series of 15 specimens was prepared for each steel. Uniaxial tension-compression fatigue tests were performed by applying a sinusoidal wave on the SHIMADZU servohydraulic testing machine (Shimadzu Corporation, Tokyo, Japan). The specimens were exposed to high cycle fatigue under stress-controlled, fully reversed (tensile-compression) testing conditions in accordance with ASTM E468-90 [42]. The stress ratio for all tested specimens was R = −1. The stress levels used to control the fatigue tests were chosen from the previously performed monotonic uniaxial tensile test described in Section 2.2.
For mild-strength S355J2+N steel and high-strength S690QL steel, 6 levels of stress amplitude were used for fatigue testing. We had three repetitions per level with a high range of stress amplitude and two repetitions per level with a low range of stress amplitude.
Given the scatter in the data obtained during the testing of the specimens of the hot work tool steel grade X37CrMoV5-1, 4 levels of stress amplitude were used for fatigue testing. We conducted three and four repetitions per two level stress amplitude in order to obtain a better repetition of results.
During the testing of fatigue properties, the frequency was 10 or 15 Hz, and complete failure of the specimen was used as a criterion for the test stop.

Results
The initial uniaxial tensile testing gave us the stress-strain curves shown in Figure 4. Based on these curves, we could compare the yield region of all steels and observe the initial strain hardening behavior of the S355J2+N steel. It is clear that the S355J2+N steel shows a yield plateau, after which a very significant strain hardening can be verified. The S690QL steel does not show that yield plateau, and a relatively small strain hardening can be observed. The X37CrMoV5-1 steel has a significantly higher limit of proportionality, as can be seen in Figure 4, but it yields at less than 3% strain. Therefore, we chose different stress amplitudes for each steel grade in order to cover a wide range of cyclic load conditions. For S355J2+N, we used amplitudes between 200 and 350 MPa; for S690QL, we used amplitudes between 300 and 540 MPa; and for X37CrMoV5-1, we used amplitudes between 500 and 1000 MPa. The common amplitude for S355J2+N and S690QL is 350 MPa, and for S690QL and X37CrMoV5-1, it is 500 MPa.  1  1000  10  10,613  2  1000  10  4025  3  1000  10  12,839  4  1000  10  12,280  5  800  10  87,089  6  800  10  12,554  7  800  10  95,546  8  600  15  38,035  9  600  15  58,871  10  600  15  25,192  11  600  15  70,825  12  600  15  51,055  13  500  15  164,237  14  500  15  2,110,777  15  500  15 2,583,827 The results of the experimental fatigue tests on smooth cylindrical specimens shown in Figure 5 for mild-strength S355J2+N, high-strength S690QL, and hot work tool steel grade X37CrMoV5-1 are displayed in Table 3. Table 3 summarizes the results of the fatigue tests carried out with smooth specimens under stress-controlled conditions. This table includes the controlled stress range; testing frequency; and the resulting number of cycles to failure, N f , for each specimen. During the testing of specimens for S355J2+N and S690QL, the number of repeats was limited to two million cycles, and specimens with the designation of fifteen did not fail. For the testing of specimens for X37CrMoV5-1, the number of repeats was limited to five million cycles.
According to the experimental data shown in Table 3, the Basquin model described in Equation (1) and statistical analysis (linear model Y = A + BX, log-normal fatigue life distribution with constant variance along the entire interval of X used in testing) in accordance with standard ASTM E739-91 [43], the fatigue properties of the mild-strength S355J2+N steel, high-strength S690QL steel, and hot work tool steel grade X37CrMoV5-1 steel were determined and are shown in Table 4. Table 4. Fatigue properties under uniaxial stress-controlled fully reversed testing, stress ratio R = −1.

Steel Grade
Fatigue Strength Coefficient Based on uniaxial tension-compression stress-controlled experiments, the S-N curves (semi-log representation) for S355J2+N, S690QL, and X37CrMoV5-1 steel grades were determined and are shown in Figures 6-8, respectively, and combined in Figure 9.

Discussion
The stress-strain curves depicted in Figure 4 allow us to perform a comparison of the yield regions of all steel grades as well as the initial strain hardening behavior. It is clear that the S355J2+N steel shows a yield plateau, after which a very significant strain hardening can be verified. The S690QL steel does not show a yield plateau, and a relatively small strain hardening can be observed. When it comes to the X37CrMoV5-1 steel grade, from the diagram shown in Figure 4, it can be seen that it has extremely high tensile strength but very little maximum strain <2% due to its brittle nature caused by its martensitic structure. Such material behavior is common for hot work tool steels that need to be able to withstand high impact forces during operation without the occurrence of permanent deformations. However, if a critical force is reached and permanent deformations occur, the fracture of such steels occurs relatively quickly because the strain hardening is insignificant.
The reference value of the fatigue strength at N C = 2 million cycles is denoted as ∆σ c , [13], and for the S355J2+N steel grade, its calculated value is about the same as class 160 of EN 1993-1-9 [13]. The results obtained for the S690QL steel grade show a fatigue behavior similar to that of the S355J2+N steel grade; however, the calculated ∆σ c level of the S690QL steel grade is much higher than that of the S355J2+N value.
The best dynamic properties were certainly observed for the X37CrMoV5-1 steel grade. This steel grade has a high fatigue (endurance) limit, which was attained at about 500 MPa. The two tested samples achieved more than two million cycles, which is significantly more than the other two tested steel grades. For the X37CrMoV5-1 steel grade, the scattering of the results is somewhat higher than that in the other cases. The scattering of the results is most likely due to the high strength of the steel, the martensitic-carbide microstructure, and the large number of carbide inclusions of high hardness. However, based on the S-N curve, shown in Figure 8, the results can be clearly read. Such conclusions were somewhat expected given the tensile strength of the steel grade and its purpose.
Based on the visual examination of the fracture surfaces, it was determined that the steels S355J2+N and S690QL have a larger fatigue zone; i.e., for them, the time from the appearance of the initial crack to failure is longer than with the X37CrMoV5-1 steel.
Bearing in mind the mechanical characteristics of steel-above all, the ratio of strength to plasticity-such a conclusion was expected.
The visual examination of the fracture determined that the fracture that occurred in the S355J2+N and S690QL steels was assessed as predominantly ductile and the fracture of the X37CrMoV5-1 steel as predominantly brittle, which was expected considering the estimated structure of the steel.

Conclusions
The fatigue properties of mild-strength S355J2+N, high-strength S690QL, and hot work tool steel grade X37CrMoV5-1 were evaluated in this work. Some key points are listed below:

•
The mild-strength S355J2+N steel was used as a reference point and was compared with S690QL and X37CrMoV5-1 steel.

•
The fatigue behavior was investigated using uniaxial tension-compression stresscontrolled experiments, with stress ratio R = −1, on smooth cylindrical specimens.

•
Considering the high cycle fatigue regime and obtained S-N curves for all materials, the X37CrMoV5-1 steel grade showed superior fatigue behavior, and due to its yield strength, the X37CrMoV5-1 steel grade showed a higher endurance limit than the S355J2+N and S690QL steel grades at the same stress amplitudes.

•
For the high cycle load, if we observe Basquin curves, the fatigue limit for the X37CrMoV5-1 steel grade tends to be similar to that of S690QL, even though these materials have very different stress-strain curves for the uniaxial tensile test.

•
The X37CrMoV5-1 steel has a huge scattering of fatigue results, and it is very important to produce this steel under conditions as controlled as possible so that the structure is as uniform as possible.

•
The utilization of high-strength steels increases the fatigue sensitivity of the construction components in comparison to components made of structural (mild-strength) steels because of the reduced cross-section and, consequently, increased stress. • Therefore, such a new design would require verification using a numerical method, such as finite element analysis, and extensive experimental testing of the prototype [7], but the final product would have better quality and superior features.

•
As a concluding remark, the design of construction components with high-strength steels (S690QL steel grade) should take advantage of the superior resistance of these steels to static and service (fatigue) loads.

•
The S690QL steel grade has inferior fatigue properties to those of X37CrMoV5-1; however, it has much better weldability [12], and, therefore, it represents the optimum solution for designing lightweight, highly loaded constructions.  Data Availability Statement: Not applicable.

Conflicts of Interest:
The authors declare no conflict of interest.