Mechanical and Metallurgical Characterization of Advance High Strength Steel Q&P1180 Produced by Two Different Suppliers

Abstract

Through mechanical analysis, a comparison of the same type of cold rolled steel produced by two steel manufacturers, supplier 1 and supplier 2, has been carried out. The considered material is a steel that has undergone a quenching and partitioning heat treatment, i.e., a rapid cooling from the austenitizing temperature, followed by a holding treatment at a suitable temperature, so that the residual austenite is stabilized at room temperature. The following tests for mechanical properties were carried out: formability, through Nakajima test, tensile test, bending test, hole expansion test and fatigue strength analysis, through high cycle fatigue and low cycle fatigue test. In addition, to derive useful data for future simulations, tensile and Nakajima tests were analyzed by digital image correlation, which uses a monochrome camera to capture frames during the test, in order to analyze local deformations on investigated samples. Finite elements modeling has been carried out. A suitable calibration of a material card for the Abaqus Finite Element Analysis software has been performed. Through the combination of obtained results, a rational comparison of the two analyzed products has been obtained.

1. Introduction

Dual Phases (DP) steel was developed to produce thin, strong automotive steel sheets to improve fuel economy and to reduce gas emissions. DP, martensite/ferrite, steels are amongst the first advanced high strength steels (AHSS) developed for cold formed sheet applications in the automotive sector [1]. DP steels are high-strength steels with a high-volume fraction of martensite. This steel grade has high strength but low elongation and, consequently, poor formability, which limits its application for automotive sheet metal. As reported in ref. [2], which considers DPs by comparing them with third-generation AHSS steels, DPs have half the elongation (7.8% vs. 18.0%). The third-generation AHSS take advantage of the austenite-to-martensite transformation to improve hardening and retard material instability during deformation. It is also observed that, for the same tensile strength, third-generation steels exhibit higher ductility than DPs, indicating favorable behavior in advanced stamping processes. To avoid the elongation problem, transformation-induced plasticity (TRIP) steels, which show high formability, can be considered. As shown in refs. [3,4], this grade of steel has a microstructure composed of ferrite (soft and ductile), bainite (which confers strength) and retained austenite, which, during deformation, transforms into martensite, further improving strength. This phenomenon is called the TRIP effect, which is the transformation of austenite into martensite under deformation, that generally increases the strength of the material, improving performance during cold forming. TRIP steels offer an exceptional combination of high mechanical strength and good ductility due to their microstructure.

However, the tensile strength of TRIP steels is lower than that of DP steels. To produce metallic materials with suitable mechanical properties for improving automotive safety and energy saving, a new type of steel called quenching and partitioning (Q&P) has been developed in recent years [5]. The idea behind this new family of steels is the so-called “quenching and partitioning” process, which consists of the diffusion of carbon from martensite to retained austenite (RA) for its stabilization. The aim is to produce steels in which significant fractions of appropriately localized RA are present in a martensitic matrix. In Q&P steels, the martensite has the scope to improve the tensile strength, while the RA increases the elongation. The presence of RA between martensite platelets improves the ability of grain boundaries to hinder the movement of dislocations, thus increasing strength and work hardening. Results are steels with tensile strengths above 1200 MPa combined with high ductility, i.e., an elongation higher than 14% [6].

The Q&P steel grade is subjected to a quenching and partitioning heat treatment, where there is initially complete austenization, by an increase in temperature, followed by quenching and isothermal maintenance with a cooling rate greater than 50 °C/s between the martensite start (Ms) and martensite finish (Mf) temperature [7,8]. The consequence of this heat treatment is the formation of a partially hardened martensitic or bainitic matrix (Figure 1) [6]. During isothermal maintenance (partition phase), carbon diffuses into the residual austenite leading to the enrichment of the latter [9]. This step allows the residual austenite to be stabilized at room temperature.

The Q&P steel is normally composed of carbon-depleted martensite and carbon-enriched retained austenite after full austenization. This microstructure allows for a high strength and ductility that can outperform those of DP steels [11]. Q&P steels own an excellent combination of strength and overall deformability, but the high carbon content reduces weldability. This steel grade enjoys properties that allow it to be used as a substitute for DP in the manufacture of molded parts, due to its higher edge strength.

Challenging mechanical properties generally require alloy steels, which are very expensive, but may significantly decrease the vehicle weight. Q&P steels, which contain a small amount of alloying elements, could limit the cost of raw materials, though reducing fuel consumption and thus, carbon dioxide emissions [12]. This study analyzes the mechanical and microstructural properties of a Q&P steel produced by two different suppliers, with the aim of evaluating its potential application in the automotive sector.

2. Materials and Methods

The characterization of materials plays a crucial role in the design and development of automotive components. Through the use of mechanical tests, valuable information can be obtained to characterize and validate the application of materials in the automotive field, such as mechanical strength, sheet metal formability, and behavior under dynamic loads, which are essential elements to meet the increasingly stringent demands of the industry.

2.1. Selected Q&P Steels

Q&P1180 steel is named for the ultimate tensile strength, in this case 1180 MPa. The plates analyzed came from two international steel suppliers, both with global supply capabilities. For production reasons of the two suppliers, obtained samples have two different thicknesses. For supplier 1, the thickness is 1.67 mm, while, for supplier 2, it is 2.0 mm.

Table 1. Chemical composition range (wt%) of Q&P1180 steels provided by supplier 1 and supplier 2.

Supplier	C	Si	Mn	Al	Cu	Ti + Nb	Ni + Cr + Mo	P	S	B
1 min	0.18	1.6	2.6	0.01	0.00	0.01	0.05	0.00	0.000	0.001
1 max	0.20	1.8	3.0	0.06	0.05	0.05	0.20	0.01	0.010	0.005
2 min	0.10	1.2	2.2	0.02	0.00	0.02	0.06	0.00	0.001	0.000
2 max	0.16	1.9	3.3	0.04	0.01	0.03	0.09	0.01	0.003	0.001

shows the range of chemical compositions of Q&P1180 steels produced by the two different suppliers.

The heat treatment is aligned with that already reported in the study covered in ref. [6], for which there is an annealing temperature of about 870 °C, quenching temperature of about 300 °C and partitioning temperature of about 400 °C. Following these annealing temperatures, the carbon content of RA increases [6].

The chemical composition of steel from suppliers 1 and 2 are similar to the Q&P grades studied by Xia et al. [13]. As it is possible to observe, the composition of steel obtained from supplier 2 has less C and Si, if compared with that provided by supplier 1. This difference facilitates the welding operation, because of the lower carbon equivalent value, but it leads to the requirement of faster cooling during production in the steel mill.

2.2. Microstructural and Chemical Analysis

Microstructural analysis was obtained by a Reichert–Jung MeF3 M 1 optical microscope, using UNI 3244:1966 [14] to evaluate nonmetallic inclusions in steels. Nittal 3% reagent was used to reveal the microstructure, as indicated in ASTM E407-07 (reapproved 2015) [15]. For quantitative chemical analysis, OES-Quantometer-ARL3460 was used as the instrument.

2.3. Mechanical Tests

2.3.1. Bending Test

The VDA 238-100 [16] bend test was used to characterize the deformability and fracture limits of metal sheets, subjected to bending under severe plane strain conditions. This test is particularly important for the selection of advanced high-strength steels for energy-absorbing structural components [17]. The bending angle is the principal result obtained from this test. It is calculated by dividing the vertical deformation (difference between initial and final height of the point of load application) by the length of the sample. The universal tensile machine Sun 20 from Galdabini was used to obtain the bending value.

2.3.2. Formability Test

The Formability Test (FLD), according to ISO 12004 [18] allowed the formability of metal sheets to be evaluated. The FLD test determines the material failure limit through the formability limit curve (FLC), which is determined by cupping tests on sheet metal test specimens. Six different geometries (Figure 2) have been used for each strain state, going from a thin useful stretch for uniaxial study to a round one for biaxial deformations. These tests can be performed using hemispherical or flat punches, such as those defined by the Nakajima or Marciniak methods. The maximum deformations of the various types of specimens define the limit formability curve of the material [19]. The Hydraulic Erichseen machine was used to obtain FLD results.

2.3.3. Hole Expansion Test

The Hole Expansion Test (HET) was used to determine the ductility of perforated plates, by quantifying the elongation capacity of the edge. High values of the hole expansion ratio are associated with steel grades with improved local formability. The sample that contains the hole is blocked in place, and a conical punch expands the initial hole. The test stops when a crack is observed through the thickness or when there is a decrease in load above a critical threshold. Hole expansion ratio (HER), also called hole expansion capacity (HEC), is the percent expansion of the initial hole diameter, usually denoted by the Greek letter lambda, λ. The Hydraulic Erichsen machine was used to obtain FLD results.

2.3.4. Fatigue Test

To evaluate the High Cycle Fatigue (HCF) behavior, fatigue tests are conducted in the longitudinal orientation to the rolling direction on hourglass specimens according to ISO 1099 [20]. The sample had a stress concentrator factor kt = 1.06. Cut edges have been smoothed off using abrasive papers to minimize the effect of the sample machining on the fatigue response. Fully reversed axial fatigue tests (stress ratio R = −1) have been carried out under load control, by using a RUMUL resonant fatigue machine, at a nominal frequency of 100 Hz. Antibuckling guides have been used to prevent buckling in the fully reversed fatigue tests.

Data analysis has been performed, combining finite life and endurance limit testing. The endurance limit corresponding to a fatigue life of 5 × 10⁶ cycles is based on staircase testing method following ISO 12107/UNI 3964 [21] std., employing 15 samples and 5 MPa stress increments. Median (R50) and reliability of 90% (R90) values for endurance limit have been evaluated following ISO-12107/UNI3964 std. Reliability of 90% and confidence of 90% (R90C90) values for endurance limit were evaluated by using the Dixon–Mood method. The output data are the following: endurance limit R50 (S_A^SC_R50), endurance limit R90 (S_A^SC_R90), and endurance limit R90C90 (S_A^SC_R90C90).

The medium-to-high-cycle fatigue life in the range between nearly 6 × 10⁴ and 1 × 10⁶ cycles was explored at various load levels, using up to four samples for each of them. The generation of the finite life portion of the median SN curves was obtained by a linear regression (Least Squares Analysis) in logarithmic coordinates of the experimental data according to Equation (1):

$$S_a=S_A{\left({\displaystyle \frac{N_A}{N_f}}\right)}^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$k$}\right.}}$$

(1)

where S_a is the engineering stress amplitude, S_A is the fatigue strength at knee of SN curve, N_A is the fatigue life at knee given by the intersection of the median finite life curve and the median endurance limit, N_f is number of cycles, and k is the slope factor of SN curve. The curve, representing a reliability of 90% (R90) was calculated by using the probability coefficients corresponding to a normal distribution (inverse of standard normal distribution based on reliability level of interest). The design curve, representing a reliability of 90% and a confidence of 90% (R90C90), was calculated using the modified Owen’s approach. It was assumed that the design SN curve (R90C90) has the same slope as the median SN curve (R50). The output data are the following: fatigue strength at knee of SN curve R50 (S_A^SN_R50), fatigue strength at knee of SN curve R90 (S_A^SN_R90), fatigue strength at knee of SN curve R90C90 (S_A^SN_R90C90), fatigue life at knee of SN curve (N_A) and slope factor of SN curve (k).

To evaluate the Low Cycle Fatigue (LCF) behavior, fatigue tests have been conducted in the longitudinal orientation to the rolling direction on constant cross section specimens, according to ISO 12106 [22]. Cut edges have been smoothed off using abrasive papers to minimize the effect of the sample machining on the fatigue response. Fully reversed axial fatigue tests (strain ratio R = −1) were carried out under strain control by using an MTS servo hydraulic fatigue machine at a constant strain rate. Antibuckling guides were used to prevent buckling in the fully reversed fatigue tests. To define the strain–life curve, Morrow’s equation (Equation (2)) was used [11]:

$${\epsilon }_a={\displaystyle \frac{{{\sigma }^{\prime }}_f}{E}}{\left(2N_f\right)}^b+{{\epsilon }^{\prime }}_f{\left(2N_f\right)}^c$$

(2)

where σ′_f is the fatigue strength coefficient, b is the fatigue strength exponent, ε′_f is the fatigue ductility coefficient, c is the fatigue ductility exponent, E is the average cyclic modulus of elasticity and 2N_f is the number of reversals. The σ_f′ and b parameters represent the median stress–life curve (Basquin’s equation) and ε_f′ and c represent the median plastic strain–life curve (Coffin-Manson equation). The design strain–life curves, representing a reliability of 90% and confidence of 90%, have been calculated using the modified Owen’s approach. The output data are the following: fatigue strength coefficient R90C90 (σ′_f,R90C90) and fatigue ductility coefficient R90C90 (ε′_f,R90C90).

2.3.5. Tensile Test

The tensile test on the selected specimens was performed with a Sun 20 equipment from Galdabini (Cardano al Campo, VA, Italy), with a maximum load capacity of 200 kN tensile force, according to ISO 6892-1:2019 [23]. During the tensile tests, characterizations were performed to study the triaxiality of the material. To obtain this information, samples having different geometries, able to experience different stresses in the useful sections, were prepared. Investigated geometries are reported in Figure 3, and they can be defined as follows:

(a)

dog-bone samples at L0 50: a classical tensile specimen subjected to σ load parallel to the movement of the traverse race;

(b)

notched samples: a specimen with specular notches placed in the center of the useful section to simulate a point of greatest concentration of stresses during tension;

(c)

shear samples: specimens having at the center of the useful section a portion of material perpendicular to the traverse race that is subjected to purely shear stress τ during the test.

The output data are the following: yield strength (Rp_0.2), ultimate tensile strength (Rm), and elongation (A).

2.4. Virtual Simulation and Finite Elements Analysis

To collect data regarding the global and local deformation of the specimens during the test, the Digital Image Correlation (DIC) method was applied, using a monochrome camera. This was performed by collecting the various images (frames) taken by the camera while pulling the specimens at a rate of 2 mm/min.

Regarding virtual testing by means of a Finite Elements Analysis (FEA), it has to be considered that, during the design phases of vehicle structures, Computer-Aided Engineering (CAE) tools are normally used to carry out strength checks and damage prediction of components subjected to various load conditions, typical of vehicle operations (static, fatigue, thermo-mechanical, crash, etc.). To develop a FEA, it is necessary to employ analytical models that describe how the material deforms under the action of external loads. Typically, models are chosen to trace the stress and strain patterns that the material exhibits during a standard experimental test. For all metallic and nonmetallic materials, the challenge is to define analytical laws that correlate well with the plastic regime of the material and with the damage and ultimate failure phase. For metallic materials in particular, an appropriate yield function must be defined based on the hardening characteristics of the material. This approach is useful, for example, to calibrate the analytical model to the experimental data. The result of the calibration process is a material card that can be used for strength verification through FEA analysis of vehicle systems and components made from the analyzed material. The FEA analysis was carried out on the three types of specimens, as shown in Figure 4. Considering the small thickness, which leads to the assumption of a plane stress state, a two-dimensional model was realized for each specimen, with an appropriate refining of the computational grid to capture the variation of stresses in the plane, as reported in Figure 4. In terms of boundary conditions, a simulation mimicking the experimental test was performed, namely, controlling the displacement with one end of the sample clamped (Δ) and imposing a quasi-static displacement on the opposite end (D) (Figure 4).

The experimental stress–strain curve obtained from the tensile tests was referred to each geometry for the simulation. From an analytical point of view, two approaches were used simultaneously:

• In the first approach, the hardening behavior was analyzed using the Swift–Voce hardening equation for the tensile specimens, while a Johnson–Cook model was used for the notch and shear specimens.
• In the second approach, considering the last section of the σ–ε curve, for an efficient calibration of the numerical techniques, it was necessary to define the parameters of a progressive damage mechanics model. Figure 5 shows a depiction of the progressive damage model. In this type of modeling, it is necessary to define a damage initiation criterion that defines the point at which the stiffness of the material begins to degrade [24]. The ductile criterion, which is useful for damage initiation due to the nucleation and growth or the coalescence of voids, assumes that the plastic deformation at damage initiation is a function of stress triaxiality and strain rate. The material behavior is described as undamaged up to point d (Figure 5). Therefore, only the Swift–Voice (or Johnson–Cook) model will work in this section. Subsequently, depending on the local triaxial stress, damage starts at point d and progresses to complete failure in section d–e.

The calibration procedure has been carried out according to the following stages:

• convert engineering stress–strain curve to true curve;
• extract only the plastic part of the stress–strain curve true;
• calibration of an analytical model on the plastic part of the true stress–strain curve.
• define the parameters of the damage initiation and evolution starting from experimental data;
• preparation of the FEM model of the sample;
• apply load to specimen, determine stress–strain curve and compare to experimental curve;
• adjust the parameters of the analytical model and damage initiation and evolution, repeat the running of the simulation, and compare again with the experimental data until the desired correlation quality is obtained.

The program employed for virtual testing was Abaqus 2022 FEA software.

The Swift model (Equation (3)) aims to model the plastic behavior of the material in the early part of the strain hardening curve by considering the rapid work hardening that the material achieves, whereas the Voce model (Equation (4)) focuses more on the acquisition of stress saturation as deformation increases in materials that show a clear asymptotic behavior near the maximum stress. To calibrate the analytical model on the plastic section of the stress–strain curves, a combination of two approximation models, namely Swift and Voice (Equation (5)), was used, as implemented for the simulation in the software. The set of equations to be considered is therefore:

$$\mathrm{Swift} \mathrm{model}: {\sigma }_{sw}=A{({\epsilon }_p+{\epsilon }_0)}^n$$

(3)

$$\mathrm{Voce} \mathrm{model}: {\sigma }_V={\sigma }_0+Q\left({1-e}^{-\beta {\epsilon }_p}\right)$$

(4)

$$\mathrm{Swift}\text{–}\mathrm{Voce} \mathrm{model}: {\sigma }_{Sw-V}=\left(1-\alpha \right){\sigma }_V+\alpha {\sigma }_{sw}$$

(5)

where

• σ_sw is the flow stress for the Swift model,
• σ_v is the flow stress for the voce model,
• A is strength coefficient,
• ε_p is the true plastic strain,
• ε₀ is the residual deformation due to the sheet metal forming process,
• n is the Swift hardening exponent, indicating how rapidly the material strengthens with plastic deformation,
• β is the Voce hardening parameter, which controls the saturation behavior of the stress-strain curve.
• Q in approximate way is the differences between yield stress and the stress at saturation
• α is the coefficient that combines the swift and voce model.

Another analytical model used is the Johnson–Cook hardening model. The Johnson–Cook model aims to provide an accurate description of the behavior of materials under dynamic loading conditions, considering variations in temperature and strain rate without considering the thermal effect of the hardening part of the Johnson–Cook model that specifically describes how the flow stress of a material evolves with plastic deformation. It is given by the following formula (Equation (6)):

$$\sigma =A+B{\epsilon }^n$$

(6)

where

• σ is the flow stress.
• A is the initial yield stress.
• B is the strain hardening coefficient.
• ε is the plastic strain.
• n is the strain hardening exponent.

For simulations, an isotropic hardening model was used in which the hardening surface of the material changes size uniformly in all directions. In this way, stress increases or decreases in all directions.

3. Results and Discussion

3.1. Q&P 1180 Microstructure Analysis

Figure 6 shows the optical micrography of Q&P1180 obtained from supplier 1 (a) and supplier 2 (b). In both cases, a matrix of tempered martensite is observed, with traces of residual austenite RA (marked with arrows in the photos). The observed microstructures confirm a proper cooling process during production, although the chemical compositions show some differences, as mentioned earlier. The microstructure is comparable with those observed by Xia et al. [13], Nyyssönen et al. [25], and Kaar et al. [26].

3.2. Formability

Results on HET and bending test are reported in

Table 2. HET and bending test results for Q&P 1180 steels obtained from different suppliers.

	Supplier 1	Supplier 2
HET	Av. strain % = 34.0	Av strain % = 28.0
Bending	α = 152 ÷ 155	α = 152 ÷ 154
FLD0	0.134	0.144

. For the HET test, 30 samples were examined for each supplier and an average strain value slightly higher for the Q&P 1180 obtained from supplier 1 has been obtained. A similar range for the bending angle α can be observed by both suppliers after the testing of five samples each. The HET results of both Q&P1180 steels are higher than DP1180 [27] proving the good flangiability of this new steel family compared to the currents.

Through the Nakajima test, which provides the ability of the plates to deform to the point of failure, FLC was obtained for both materials, using a set of six geometries—five samples for each one–and results are shown in Figure 7. The lowest point of the FLD curve (FLD0) is reported in Table 2. It can be observed that the two materials obtained from different suppliers show similar trends for different sample geometries, with a slightly higher strain value before the cracking of the steel obtained from supplier 2. This difference can be associated with the different thicknesses of the samples. In fact, by increasing the thickness, the necking phase is extended before the crack nucleation and propagation into the sample. It can be concluded that the formability of Q&P1180 steel obtained from supplier 1 is comparable with that of the steel provided by supplier 2. In Figure 7, the comparison with a conventional Dual Phase 980 steel grade (DP980), which has 980 MPa minimum of tensile strength, is also reported. It is possible to observe that QP1180 grades have a similar formability compared to DP980, even if the tensile and yield strengths are higher. This result proves the better forming behavior of the new steel family compared to the conventional ones.

3.3. Fatigue

To study the fatigue behavior of investigated materials, the analysis was carried out for both HCF and LCF. This analysis required 2 HCF and 2 LCF characterizations involving 110 samples. Results of HCF are reported in Figure 8 and Figure 9, that show the S-N curves obtained for samples provided by supplier 1 and supplier 2, respectively. Obtained values are summarized in

Table 3. Woehler’s curves parameters of samples provided by supplier 1 and supplier 2.

Steel	Stair-Case				Constant of S-N Curve
Name	SASC [MPa]			S.D. [MPa]	SASN [MPa]			NA [Cycles]	k
	R50	R90	R90C90		R50	R90	R90C90
Supplier 1	403.2	398.7	395.0	3.5	403.2	388.5	383.0	965,692	10.92
Supplier 2	416.7	401.1	386.3	9.1	416.7	401.4	394.1	783,807	10.13

.

The HCF behavior of the tested steels is quite similar (Table 3). Both S-N curves have the same slope factor (k) and close fatigue life at knee (N_A). The steel produced by supplier 2 has a little higher endurance limit R50 with respect to that provided by supplier 1, but it also has a higher scattering, which leads to a lower endurance limit, considering a reliability of 90% (R90) and reliability and confidence of 90% (R90C90).

Concerning LCF measurements, Figure 10 and Figure 11 show results obtained for samples provided by supplier 1 and supplier 2, respectively. Obtained values are summarized in

Table 4. Morrow’s curves parameters of samples provided by supplier 1 and supplier 2.

Steel	Morrow’s Curve Parameters
Name	E [GPa]	Fatigue Strength Coefficient [MPa]		Fatigue Strength Exponent	Fatigue Ductility Coefficient		Fatigue Ductility
		σ′t	σ′f,R90C90	b	ε′f	ε′f,R90C90	c
Supplier 1	196.4	2881	2649	−0.140	2.145	1.791	−0.831
Supplier 2	196.7	3975	3625	−0.173	11.95	8.734	−1.027

.

Looking to a comparison of LCF results for tested steels (Table 4), it is possible to highlight that, in the elastic field, for high levels of number of cycles, the Q&P1180 obtained from supplier 1 shows better behavior in terms of performance than that provided from supplier 2. However, as the strain amplitude increases until it reaches the value of 0.5%, the difference between the two materials falls to zero, while, for levels of strain amplitude higher than 0.5%, there is a reversed trend, so that the material obtained from supplier 2 shows better performance than that provided by supplier 1. This result may be attributed to the different static properties of the two materials. Specifically, an examination of the static tensile curves reveals that the steel from supplier 2 exhibits higher load values than that provided by supplier 1 near the yield point. This characteristic contributes to a reduced fatigue life under conditions of equivalent imposed deformation. Conversely, with increasing deformation, the more pronounced softening observed in the material from supplier 2, compared to that obtained from supplier 1, accounts for its extended fatigue life. The gap between the R50 and R90C90 of fatigue strength coefficient and fatigue ductility coefficient indicates a similar level in terms of statistical dispersion of experimental results for both materials.

It can be concluded that the experimental data do not highlight significant differences on fatigue properties between materials provided by the two suppliers.

3.4. Tensile Test

The mechanical properties were evaluated for dog-bone samples subjected to uniaxial loading. Specimens obtained in the three directions in relation to the rolling direction were tested to observe possible anisotropy. Five samples were tested for each orientation and then the curve that best approximated the test group was considered. Results are reported in Figure 12, Figure 13 and Figure 14 for tests performed in longitudinal, diagonal, and transversal direction, respectively. Obtained data are summarized in

Table 5. Mechanical properties of samples provided by supplier 1 and supplier 2.

Sample vs. Rolling Direct	Steelmaker	Rp0.2	Rm	A
		MPa	MPa	%
Longitudinal	Supplier 1	936	1204	14.1
	Supplier 2	931	1217	13.9
Diagonal (45°)	Supplier 1	952	1177	16.5
	Supplier 2	980	1219	14.0
Transversal (90°)	Supplier 1	901	1193	15.3
	Supplier 2	924	1218	13.0

.

From obtained results of tensile tests, it is observed that the values of mechanical properties are comparable with the typical values observed for Q&P steels, as reported in ref. [2], i.e., Rp about 900–1000 MPa, Rm about 1150–1230 MPa, and A about 5–19%. The tensile curves, in all examined directions, exhibit higher stress values at low deformation for the steel provided by supplier 2 compared to that obtained from supplier 1. This behavior may be attributed to variations in the deformation applied during the rolling process or differences in the time or temperature conditions during the annealing process. This type of behavior has a positive aspect, as previously observed, for the fatigue properties, although it limits the material’s elongation.

3.5. Simulations

Since the two materials are comparable mechanically, one of the two was selected for virtual simulation in order to observe the most appropriate numerical model to convert the experimental stress and strain values into true values. Five samples were tested for each geometry. In the case of dog-bone and notch samples, for the comparison between the experimental data and the data derived from the FEA simulations of the tensile test, the plastic part of the true stress–strain curve has been modeled with the Swift–Voce equation, accompanied by a progressive damage mechanics model. Parameters used for the models used into FEA software are shown in

Table 6. Parameters for dog-bone tensile simulation.

Swift–Voce Model Parameters for the Tensile Simulation
A = 1800	n = 0.15
s0 = 1035 MPa	Q = 435 MPa
b = 11	a = 0.415
e0 = 0.042
Progress damage mechanics parameter for Abaqus standard
Damage Initiation
Criteria: Ductile
Plastic Strain: 0.125
Stress triaxiality: 450 MPa
Strain rate: 0.003
Damage Evolution
Criteria: Displacement, softening exponential
Total displacement: 0.14
Exponent: 4

. Figure 15 shows the calculated true stress-true strain curve superimposed on the experimental data for the tensile test of the dog-bone sample.

From the simulation point of view, the L11 (true deformation in tensile direction) and the S11 (true stress in tensile direction) expressed during the test have been taken into account, as shown in the following Figure 16. These components are more consistent with the Swift–Voce model and demonstrate rigid behavior of the specimen during the test.

The parameters for the notch sample used in the FEA software are shown in

Table 7. Parameters for notch tensile simulation.

Swift–Voce Model Parameters for the Notch Simulation
A = 1644	n = 0.35
s0 = 1030 MPa	Q = 650 MPa
b = 30	a = 0.635
e0 = 0.35
Progress damage mechanics parameter for Abaqus standard
Damage Initiation
Criteria: Ductile
Plastic Strain: 0.07
Stress triaxiality: 486 MPa
Strain rate: 0.001
Damage Evolution
Criteria: Displacement, softening exponential
Total displacement: 0.095
Exponent: 4

.

Figure 17 shows the calculated true stress–true strain curve superimposed on the experimental data for the tensile test of notch sample.

The L11 and the S11 for the notch sample are expressed in Figure 18.

In the case of the specimen subjected to shear stress, as mentioned, for calibrating the simulation to the experimental data, it was decided to model the plastic part of the stress–strain curve using the Johnson–Cook equation, and results are shown in Figure 19. In fact, in the case of the shear test, a larger deformation state than in the tensile and notch tests was obtained. In fact, as shown in Figure 20, the results of the calculation code must consider the true deformation component L12 and the stresses expressed by the von Mises criterion.

Table 8. Parameters for shear tensile simulation.

Johnson–Cook Model Parameters for the Shear Simulation
A = 1644	n = 0.35
s0 = 1030 MPa	Q = 650 MPa
b = 30	a = 0.635
e0 = 0.35
Progress damage mechanics parameter for Abaqus standard
Damage Initiation
Criteria: Ductile
Plastic Strain: 0.07
Stress triaxiality: 486 MPa
Strain rate: 0.001
Damage Evolution
Criteria: Displacement, softening exponential
Total displacement: 0.095
Exponent: 4

shows the parameters obtained for the Johnson–Cook model and the progressive damage mechanic model.

It is possible to observe that, even if there are slight differences between the experimental and simulated tensile curves, the FEM simulations properly predict the real behavior of different samples, as seen in Figure 20, where the simulated most stressed areas (Figure 20a) correspond to those observed during the experimental tests (Figure 20b).

In summary, microstructural analyses and mechanical tests have shown that the two investigated materials have more in common than the reverse. Optical microscopy confirmed that the two materials have the same constituent phases: tempered martensite matrix with dispersed retained austenite that influences the formability of the steel grade, promoting the TRIP effect during the plastic deformation. Through various mechanical tests, it was observed that both samples exhibit equal bending behavior, with an alpha range of 153°:155°. Both materials have similar yield strength and ultimate tensile strength, while the elongation-at-break values of steel obtained from supplier 1 are higher, albeit slightly, than those of the steel provided by supplier 2. The steel obtained from supplier 2 is slightly more formable, with higher FLC values, than that provided by supplier 1. As already anticipated, this difference can be associated with the different thickness of the samples, which strongly influences the FLC curve.

As already mentioned, Q&P steel use in the automotive field is directed towards components previously manufactured in DP. Observed mechanical properties return an excellent behavior during the automotive crash events, being able to locally absorb the deformation without/delaying the component cracking. For this reason, the Q&P1180 steel could be suitable for windshield reinforcement, side-sill, or A pillar applications. In fact, during a pole lateral crash, there is a local deformation and, in general, the materials have to absorb the energy, limiting the intrusion as much as possible in order to guarantee the safety of the vehicle occupants. If the Q&P1180 is compared to the DP1180, with the same level of tensile strength, it is possible to observe 40% more elongation, which is a crucial point to stamp complex shapes as per the automotive market [28].

To be able to proceed with possible simulations of the Q&P1180 on the vehicle, a FEM method was identified to be able to convert the real data obtained through the mechanical tests into virtual data. An approximation of the true curve of the three analyzed geometries (dog-bone, notched and shear) was obtained. In particular, the dog-bone and carved samples were approximated by means of the Swift–Voce law, while the shear samples were approximated by using the Johnson–Cook law. In all cases, a progressive damage mechanism model was used, i.e., the consideration of the ascending section of the stress–strain curve of a tensile test in the simulation. All the simulations returned to a good correlation with experimental tests, which proves the proper approximation of tensile curves of different samples. So, such models could be used to predict the behavior of components made by Q&P1180.

4. Conclusions

The comparison of the Q&P1180 steels produced by supplier 1 and supplier 2 was carried out by means of micrographic analysis using an optical microscope and mechanical tests, to study mechanical strength, fatigue strength, ductility and formability. The following conclusions can be highlighted:

• Both steels have a microstructure characterized by a tempered martensite matrix, with residual austenite.
• Both exhibit comparable yield strength and ultimate tensile strength, with higher elongation in the case of Q&P1180 obtained from supplier 1.
• Both steels exhibit similar formability, as observed through the Nakajima test.
• The Abaqus program was used to simulate the material for each geometry treated and, through damage and anisotropy calculation methods, the graphs obtained are superimposable on experimental data of the samples tested in the laboratory, proving the good reliability of simulation compared to the physical tests.
• The Q&P1180 steel of both suppliers could be used as alternatives in automotive components, thanks to observed similar mechanical properties.
• The Q&P steel grades will be widely introduced in the automotive application, thanks to the opportunity to reduce the weight without compromising the safety and durability performances of the car.