Search
Fig.
1.
Spray quenching equipment: (a) photograph; (b) schematic.
| Cite this article as: |
Ning Fan, Zhihui Li, Yanan Li, Xiwu Li, Yongan Zhang, and Baiqing Xiong, Residual stress with asymmetric spray quenching for thick aluminum alloy plates, Int. J. Miner. Metall. Mater., 30(2023), No. 11, pp.2200-2211. https://dx.doi.org/10.1007/s12613-023-2645-2
|
The 7xxx aluminum alloy has been extensively used in the aviation and aerospace industry due to its superior specific strength and stiffness, excellent corrosion, and fatigue resistance [1–2]. To achieve the desired mechanical properties of heat-treatable aluminum alloy, scholars obtained the supersaturated solid solution by solution and quenching treatments [3–5]. Given the highly efficient and controllable heat transfer behavior, spray quenching has been broadly used for industrial large-scale thick aluminum alloy plates to meet the demands of microstructure, mechanical properties, and residual stress simultaneously [6–7]. However, the quenching treatment results in severe temperature gradients between the surface and the interior, which lead to inhomogeneous plastic strain and significant residual stress magnitudes [8]. High residual stress magnitudes have a detrimental effect on mechanical properties, such as stress corrosion and fatigue performance, and cause unexpected distortion during subsequent machining processes [9].
Residual stress distribution and magnitudes have been investigated by several experimental methods [10], such as X-ray diffraction, neutron diffraction, crack compliance, contour, and layer removal methods. Compared with the scarcity and inaccessibility of the neutron diffraction method [11], the X-ray diffraction method is the more common residual stress measurement method because of its non-destructive and convenient nature. Portable X-ray devices based on sin2Ψ [12] and cos α [13] methods are widely used in laboratory and industrial measurement. However, the X-ray diffraction method can only measure the surface residual stress because of its lower penetration depth for metal materials. Crack compliance [14] and contour methods [15] are novel residual stress measuring methods that have been studied and validated by numerous researchers. However, these methods have higher measuring deviations and complex measuring processes that need to combine with finite element (FE) simulation. By contrast, the layer removal method is a proven and accessible method used to measure the residual stress along the thickness direction [16].
The quenching process of aluminum alloy is a short-time, transient, and multi-field coupling. No evidence indicates that the transient strain–stress behavior can be monitored by experimental methods, which is helpful in predicting residual stress with various heat treatment parameters and controlling stress by adjusting parameters. With the development of the finite element method (FEM), the residual stress can be predicted by accurate material properties, suitable mesh elements, realistic boundary conditions, and proper loading and solving methods. The surface heat transfer coefficient (HTC), as the most critical boundary condition, has been widely concerned by researchers. HTCs with various types and temperatures of quenchants were calculated by inverse heat conduction analysis and iterative method [17–20]. Many studies have compared simulated residual stress with experimental results by various measurement methods [21–23]. Tanner and Robinson [24] assumed that the flow stress is strain rate dependent and behaves in a plastic manner. Based on this assumption, an FE model was established to predict the quenched residual stress of 7010 aluminum alloy. Subsequently, they focused on the stress/displacement development during the quenching process of the “Navy C-ring” combined with simulation and experiments [25]. Chobaut et al. [26] identified a thermo-mechanical behavior law of 7xxx aluminum alloys with representative interrupted cooling paths in a Gleeble machine. A model accounting for recovery and precipitation was used to simulate the quench residual stress more accurately. Liu et al. [27] observed that the elastic lattice distortion caused by residual stress can promote nucleation during the artificial aging process of 7085 aluminum alloy. Cai et al. [28] held a similar conclusion and proposed a modified constitutive model of stress relaxation aging behavior for AA7150-T7751.
On the other hand, most researchers only concentrated on residual stresses with the immersion quenching of small-sized samples under laboratory conditions. Few studies have reported the distribution and evolution of residual stresses with spray quenching. Under industrial production conditions, large-scale thick aluminum alloy plates are quenched in a roller hearth furnace. Hundreds of nozzles are distributed in the furnace in the cross-array arrangement. The complexity of the jet beam can lead to different cooling behavior at the upper and lower surfaces of thick plates. The inevitable uneven spraying characteristics can result in an asymmetric residual stress distribution, which is significant in controlling stress redistribution and component deformation in subsequent machining processes.
Thick aluminum alloy plates with symmetric or asymmetric spray quenching were processed on self-designed spray equipment. The quenched residual stresses were measured by the layer removal method. The effect of water flow rates on quenched residual stress was discussed. Moreover, the simulated residual stress distribution and magnitudes by FEM were validated by experimental results. Furthermore, the strain–stress evolution of typical positions by FEM was analyzed to explain the experimental phenomena.
Experiments were conducted on Al–8.11Zn–1.90Mg–2.24Cu alloy. The four side surfaces of the specimens were milled into a dimension of 320 mm × 60 mm. The upper and lower surfaces were ground into a thickness of 26 mm. The surface roughness of the specimen affected the heat transfer capacity in the quenching process, which in turn influenced the uniformity of quenched residual stress distribution. Therefore, the sprayed surface of specimens must be ground to a consistent surface roughness of 0.8 μm.
The spray quenching equipment was consisted of an air circulation furnace (THERMCONCEPT KU 70/06/A), a water circulation system, and a spray chamber (Fig. 1). The temperature deviation of the air circulation furnace was less than 1°C. The water circulation system comprised a water tank, centrifugal pump (PRODN CHM4-2DC), valves, filters, turbine flow meters (LWYC DN10), and stainless-steel hoses. The water temperature was (25 ± 2)°C, and it was kept constant during the quenching process due to the sufficient quenchant in the water tank. The spray pressure was set to 0.2 MPa by a variable frequency centrifugal pump. The water flow rate was adjusted through valves and monitored through flow meters instantly.
The spray chamber is an important component in the spray quenching equipment. Full-cone spray nozzles with a diameter of 5 mm and spray angle of 76° were used. The liquid was broken up into plenty of small droplets and sprayed out evenly with the action of the blade of the spray nozzle, forming a circular spray zone on the plane. The jet axis of the nozzles was set to be perpendicular to the specimen surface. The distance between the nozzle and the specimen was designed to be 70 mm to ensure that the entire surface of the specimen was covered by the spraying jet stream. With comprehensive consideration of the specimen size and jet impact area, one of the longitude-long transverse (L-LT) surfaces of the specimen was impacted by a group of jets via eight nozzles that were fixed on stainless-steel plates in a cross-array arrangement and with 97 mm distance from each other (Fig. 2(a)). The flow rate of one group of nozzles was adjusted by a master valve together. The test using the flow meters confirmed that the water flow rate of each pipe was consistent with the maximum deviation of 0.1 m3/h. Baffles were used to keep the rest of the surface as free as possible from the impact of the jet stream. In addition, the necessary fasteners were added to increase the structural stiffness of the spraying chamber and ensure specimen stabilization during the spray quenching process. The spray chamber photograph of top and side views is shown in Fig. 2(b) and (c), respectively.
The spray parameters were preset first, and then, the specimens were heated in the furnace at 475°C for 2 h to maintain their uniform temperature. Moreover, the initial residual stress can be eliminated with a high temperature. Subsequently, the specimens were transferred to the spray chamber within 10 s. Finally, the pump was switched on to conduct the spray quenching process, which was requested to last more than 1 min.
The residual stresses along the thickness were calculated by deflections, which were measured by the removal of successive layers of the specimen (Fig. 3). The layer removal method was based on the following assumptions:
(1) Young’s modulus was constant in the specimen;
(2) The removal of layers did not disturb the initial residual stresses of specimens (in other words, the machining-induced stresses can be ignored);
(3) Residual stresses varied only throughout the thickness.
b and h0 denote the width and height of the specimen, respectively, and ls is the support distance on the deflection-measured apparatus. hi−1 and hi are the thickness of the specimen before and after the ith removed layer, respectively, and di is the deflection of the specimen measured on the platform. In general, the maximum deflection occurred in the middle of the length direction for the rectangular cross section specimens with uniform stress distribution in the horizontal plane.
When the first layer of the specimen was removed, the force and moment of the specimen were rebalanced through bending deformation. Therefore, the change in the bending moment in the first layer (ΔM1) can be expressed as the following equation:
| ΔM1=−12σ1⋅b⋅(h0−h1)⋅h0 | (1) |
where σ1 is the average stress of the specimen in the first layer.
According to the correspondence between the maximum deflection and the bending moment, ΔM1 can also be expressed as follows:
| ΔM1=8E⋅I1⋅d1l2s | (2) |
where E is Young’s modulus of the specimen. The moment of inertia is computed as I1=b⋅h3112 for specimens with a rectangular cross section. The average stress of the first layer of the specimen can be calculated by combining Eqs. (1) and (2):
| σ1=−43El2sd1⋅h31(h0−h1)⋅h0 | (3) |
The relationship between the change in the bending moment ΔMtot2 after the second layer of the specimen removed, and the deflection d2 can be expressed by the following:
| ΔMtot2=8E⋅I2⋅d2l2s | (4) |
where I2=b⋅h3212. Four factors contributed to the bending moment ΔMtot2. The first factor was the change in the bending moment ΔM12 caused by the stress release in the second layer:
| ΔM12=−12σ2⋅b⋅(h1−h2)⋅h1 | (5) |
The second factor was the change in the bending moment ΔM22 caused by the release of stress in the first layer before the second layer removed. In other words, the second factor is equal to the change in the bending moment in the first layer ΔM1:
| ΔM22=ΔM1=8E⋅I1⋅d1l2s | (6) |
The third change in the bending moment ΔM32 can be considered as the bending moment of the second layer removed based on the deflection d1 from the stress release of the first layer, which has been removed:
| ΔM32=−8E⋅IΔh⋅d1l2s | (7) |
where IΔh=b⋅(h1−h2)312. The stress release in the first layer resulted in a downward shift of the neutral layer by 12(h0−h1). Therefore, the fourth factor of the change of bending moment ΔM42 can be expressed as follows:
| ΔM42=−12σ1⋅b⋅(h1−h2)⋅(h0−h1) | (8) |
According to the right-hand screw rule of bending moment and Eqs. (4)–(8), ΔM12 can also be expressed using the following equation:
| ΔM12=ΔMtot2−ΔM22−ΔM32−ΔM42=2Ed2bh323l2s−2Ed1bh313l2s+2Ed1b(h1−h2)33l2s+12σ1b(h1−h2)(h0−h1) | (9) |
In summary, the average stress in the second layer of the specimen can be calculated as the following equation:
| σ2=−2ΔM12b(h1−h2)h1 | (10) |
The average stress in the nth layer of the specimen σn can be deduced by analogy in the following equation:
| σn=−2ΔM1nb(hn−1−hn)hn−1 | (11) |
where the change in the bending moment ΔM1n can be expressed as follows:
| ΔM1n=2Ednbh3n3l2s−2Edn−1bh3n−13l2s+2Edn−1b(hn−1−hn)33l2s+n−1∑m=1σmb(hn−1−hn)(hm−1−hm)2 | (12) |
The specimen was fixed on the CNC milling machine using hexagonal screws and T-nuts. Successive removals of 1 mm layers were performed through a disk cutter with a diameter of 100 mm. Table 1 shows the specific machining parameters. Screws were released after removing one layer, and the specimen was placed on the deflection-measured apparatus. The dial gauge (Mitutoyo 543-470B) with a resolution of 0.001 mm was located at the midpoint of the specimen to measure the deflections. The actual thickness of the specimen after layer removal was measured using a vernier caliper with a resolution of 0.01 mm. Thickness and deflection measurements were repeated thrice, and the average value was used to calculate the residual stress. The machining and measuring processes were repeated 13 times until the specimen was half of the original thickness.
| Machining sequences | Tool speed / (r·min−1) | Feed rate / (mm·min−1) | Depth of cut / mm |
| Rough machining | 3000 | 500 | 0.8 |
| Finish machining | 3000 | 300 | 0.2 |
DownLoad:
CSV
The quenching simulation was conducted using the commercial FE analysis software ANSYS. The influence of phase transformation on the thermal and stress–strain fields can be negligible because of the rapid cooling during the quenching process of aluminum alloy. The thermal strain caused by plastic strain can also be ignored because of the slight deformation in the quenching process. Therefore, sequential coupling of the thermal field and stress–strain field was used. The thermal field result was substituted into the stress–strain field as the boundary condition.
The constitutive equation for the alloy was described as bilinear kinematic hardening in ANSYS, which means that the effective stress versus effective strain curve is bilinear. The initial slope of the curve was the elastic modulus of the specimen. Once the stress was beyond the yield stress, plastic deformation developed, and the stress versus total strain continued along a line with a slope of the tangent modulus. For uniaxial compression followed by uniaxial tension, the magnitude of the compressive yield stress decreased as the tensile yield stress increased. The thermophysical and strengthened parameters of the alloy are shown in Table 2. The surface HTC of spraying surfaces can be described as the following equation [29]:
| Temperature / °C | Density / (kg·m−3) | Thermal conductivity / (W·m−1·K−1) | Specific heat / (J·kg−1·K−1) | Linear expansion coefficient / (10−6 K−1) | Young’s modulus / GPa | Poisson ratio | Yield strength / MPa | Tangent modulus / GPa |
| 20 | 2842 | 145 | 852.3 | 21.6 | 72.0 | 0.3 | 412 | 1 |
| 100 | 2828 | 152 | 894.5 | 22.7 | 65.2 | 0.3 | 382 | 0.5 |
| 200 | 2807 | 160 | 940.5 | 24.0 | 56.3 | 0.3 | 370 | 0.25 |
| 300 | 2787 | 167 | 982.5 | 24.2 | 38.0 | 0.3 | 120 | 0.1 |
| 400 | 2761 | 171 | 1003.2 | 26.0 | 31.5 | 0.3 | 50 | 0.01 |
| 500 | 2735 | 178 | 1045 | 27.5 | 25.0 | 0.3 | 20 | 0.005 |
DownLoad:
CSV
| h(Ts)=A⋅exp[−(Ts−BC)2] | (13) |
where Ts is the surface temperature of the specimen. A, B, and C are parameters of first-order Gaussian function and can be described as Fourier functions related to the volumetric flux (flow rate per unit area, Q″):
| {A(Q″)=8203+605.4cos(0.06447Q″)−1924sin(0.06447Q″)B(Q″)=104.6+7.503cos(0.07366Q″)−11.46sin(0.07366Q″)C(Q″)=97.62+5.614cos(0.06415Q″)−6.535sin(0.06415Q″) | (14) |
The HTC of other side surfaces, which was heat transfer with air, was set as 50 W/(m2·K).
Considering the symmetry feature of geometry and boundary conditions, a one-four model of the full specimen was used to reduce the simulation time (Fig. 4). The SOLID90 and SOLID186 elements were selected for thermal and stress–strain fields, respectively. Given the drastic temperature gradient near the surface during the quenching process, the element size was 1 mm × 1 mm × 1 mm, and the elements along the thickness direction at the 1 mm area near the upper and lower surfaces were refined to 0.1 mm through the mesh independence test.
The residual stresses of the two specimens under uniform spraying quenching with a water flow rate of 0.60 m3/h were measured to verify the reproducibility and stability of the experiment process. Specimen #A1 was a layer removed from the left surface, which meant that the inner stress of the left half of the specimen was measured, and specimen #A2 was a layer removed from the right surface. As shown in Fig. 5, the deflection and inner residual stress of the two specimens were consistent. The deviation rate of deflection was 0.37%, and the maximum difference was 4.43 MPa. Therefore, consistent experimental results with spray quenching equipment can be obtained by consistent experimental parameters and operating procedures.
To evaluate the accuracy of the equation for calculating residual stress, we used an FE model based on a birth–death element method of the layer removal method. The initial residual stress of the FE model was simulated by appropriate HTCs according to a water flow rate of 0.60 m3/h. Fig. 6 shows the contour map and evolution of deflection during the removal of different layers by the birth–death element method. The deflection in the contour map was magnified five times to display the deformation more clearly after the layer was removed. The linearity of deflections along the length direction is a parabolic type with symmetry at the midpoint. The deflection gradually increased with the increased number of removed layers. The residual stress calculated by deflections is shown as a black line in Fig. 7, and the average quenched residual stress in each layer of the model by FEM is presented as a red line. The calculated stress was the same as the simulated stress. The accuracy and validity of the equation for calculating residual stress can be confirmed. The maximal deviation of the calculated and simulated stress was 3.89 MPa, which was located on the layer near the surface.
The symmetric spray quenching experiments were conducted, and the effect of various water flow rates on residual stress was investigated. The water flow rates of the two sides were set to 0.60, 0.45, 0.30, and 0.15 m3/h. As mentioned above, the deflection increased with the increased number of removed layers, and the maximum deflection appeared when the 13th layer was removed. Fig. 8(a) shows the deflections after successive layers were removed at different water flow rates. The maximum deflections of different water flow rates were 0.536, 0.405, 0.302, and 0.187 mm.
The linearity of the residual stress along the thickness direction, which was calculated by deflections, is a parabolic type (Fig. 8(b)). The maximum compressive stress developed on the layer near the surface, and the maximum tensile stress developed on the central layer of the specimens. The absolute value of the maximum compressive stress was higher than that of the maximum tensile stress. The closer to the core, the smaller the residual stress gradients were. The water flow rates of 0.45, 0.30, and 0.15 m3/h resulted in a reduction in the maximum compressive stresses of 24.73%, 40.65%, and 64.52%, respectively, compared with that at the water flow rate of 0.60 m3/h. In addition, compared to the water flow rate of 0.60 m3/h, the reductions in maximum tensile stresses of 19.80%, 42.46%, and 67.20% were observed at the water flow rate of 0.45, 0.30, and 0.15 m3/h, respectively. Thus, the results on deflections and stresses showed that the residual stress magnitudes decreased with the decrease in the water flow rate.
The solution and quenching heat treatments are generally carried out in roller hearth furnaces with industrial production conditions. However, the inconsistent spraying water flow rate of the upper and lower nozzles may lead to an asymmetric residual stress distribution. Therefore, the effect of different water flow rate ratios on the residual stress magnitude and distribution with asymmetric spray quenching was investigated. The water flow rate ratios were set as 0.60:0.45, 0.60:0.30, and 0.60:0.15. Successive layer removal machining was executed from the high- and low-flow rate sides to the center layer, respectively.
Fig. 9 shows the deflection with asymmetric spray quenching at different flow rate ratios. Fig. 9(a) displays that the maximum deflection with asymmetric spray quenching at a flow rate ratio of 0.60:0.45 was between the maximum deflection with symmetric spray quenching at flow rates of 0.60 and 0.45 m3/h. The maximum deflection from the high- and low-flow rate sides were 0.468 and 0.476 mm, respectively. Fig. 9(b) and (c) reveals the same regularity with asymmetric spray quenching at flow rate ratios of 0.60:0.30 and 0.60:0.15. As shown in Fig. 9(d), with the increase in the asymmetry of flow rates on the two surfaces, the maximum deflection after layer removal gradually decreased, and the deviation of the two sides of the maximum deflection increased.
Fig. 10(a)–(c) shows the residual stress with asymmetric spray quenching at different flow rate ratios. The stress magnitude with asymmetric spray quenching of flow rate ratios of x : y was between stress magnitudes with symmetric spray quenching at the flow rates of x and y m3/h. The maximum compressive stresses and maximum tensile stresses decreased with the increase in the asymmetry of flow rates on the two surfaces. According to Eq. (3), the residual stress of the first layer near the surface was proportional to the deflection after the first layer was removed. Fig. 10(d) shows that the first layer stresses of the two sides decreased with the increase in the water flow rate ratios. The residual stress of the first layer near the lower flow rate surface was higher than that of the first layer near the higher flow rate surface. When the flow rate ratios were 0.60:0.45, 0.60:0.30, and 0.60:0.15, the deviations of the residual stress of the first layer near the higher and lower flow rate surfaces were 3.64, 12.91, and 15.35 MPa, respectively.
Fig. 11 shows the residual stresses along the thickness direction of the experimental and simulated results with symmetric and asymmetric heat transfers. The high and low HTC surfaces were calculated by Eqs. (13)–(14) according to the experimental water flow rate and set at the lower and upper surfaces of the FE model, respectively. The HTC was considered to be proportional to the water flow rate. The accuracy of the FE model can be validated because the simulated residual stress distribution and magnitude showed great agreement with the experimental results. The residual stress was symmetrical along the thickness direction when the HTCs of the two surfaces were the same. When the HTCs of the upper and lower surfaces differed from each other, the compressive stress of the high HTC surface was less than that of the low HTC surface. The difference in residual stresses in the upper and lower surfaces increased with the increase in the difference in HTCs, which was consistent with the experimental results. Therefore, the FE model with asymmetric HTCs can be used to predict the residual stress distribution and magnitude of quenched thick aluminum alloy plates.
Fig. 12 shows the stress contour of the FE model with asymmetric HTCs, which is consistent with the experiment with a water flow rate of 0.60:0.15. The compressive stress developed at the surface, and the tensile stress developed at the interior of the model, which is consistent with the typical quenched residual stress distribution. On the one hand, compared with the upper surface, the color of the lower surface was lighter, which indicated that the compressive stress of the lower surface was less than that of the upper surface. Therefore, the higher the HTC, the smaller the compressive residual stress on the surface under asymmetric heat transfer conditions. On the other hand, the undeformed edge and deformed shape indicated that the model was concave downward. The shrinkage amount of the lower surface was greater than that of the upper surface because the HTC of the former was greater than that of the latter, which led to a higher cooling rate.
To analyze the abnormal phenomenon of residual stress distribution with asymmetric heat transfer, we extracted the evolution of temperature, stress, plastic, and elastic strain of typical points during the quenching process. Points A and B were located in the middle of the lower and upper surface, respectively, and point C was located in the middle of the interior of the FE model.
Fig. 13 shows the cooling curves and stress evolution of points A, B, and C during the whole quenching process. The cooling rate of point A was greater than that of point B in Fig. 13(a) because the heat transfer capacity is proportional to the HTC according to Newton’s law of cooling. The cooling behavior of the metal interior was heat conduction, whose heat transfer efficiency was several orders of magnitude lower than heat transfer with water. Therefore, the cooling rate of point C was the slowest.
As shown in Fig. 13(b), at the initial stage of the quenching process, when the high-temperature materials were exposed to the low-temperature quenchant, the upper and lower surfaces were cooled rapidly and shrunk violently to produce tensile stress. Meanwhile, the interior produced compressive stress under the condition of equilibrium of forces and moments. The maximum temperature gradient between the surface and interior was reached at 2.05 s. The maximum tensile stress of the surface and maximum compressive stress of the interior were produced at the same moment. Under continuous quenching, the surface temperature was relatively low. The interior began to cool and shrink to produce tensile stress. To coordinate with the tensile stress of interior, the tensile stress of surface turns into compressive stress. When the quenching reached 10 s, temperatures at different points were below 100°C, and the temperature gradient between the surface and interior gradually decreased. Meanwhile, the stresses of the surface and interior slightly increased until stabilization. Therefore, the subsequent discussion mainly focused on the stress and strain evolution of the early quenching process (0–10 s).
According to Fig. 14(a), the quenching process can be divided into three stages. From the beginning of quenching to 2.05 s, tensile stresses were produced on surfaces and gradually increased. The tensile stress of point A was higher than that of point B. Given that the HTC of the lower surface was higher than that of the upper surface, the cooling rate of the lower surface was higher, which led to higher shrinkage and higher tensile stress accordingly. As shown in Fig. 14(b), plastic tensile strains were produced at the surface and increased gradually during quenching stage 1. Plastic compressive strains of the interior did not occur until 0.84 s. As the yield strength was relatively low due to the high temperature at this stage, the stress caused by a larger temperature gradient was greater than the yield strength. Moreover, before 0.84 s, the plastic strain of point A was slightly higher than that of point B because of the higher shrinkage deformation, as mentioned above. Afterward, the plastic strain of point A was less than that of point B. Given that the temperature of the lower surface was lower than that of the upper surface, which meant that the yield strength of the lower surface was higher than that of the upper surface, the deformation mode of the lower surface was mainly elastic strain. After 2.05 s, given the lower temperature and higher yield strength of the whole model, the strain caused by quenching shrinkage was insufficient for plastic flow. Moreover, the temperature gradient between the surface and interior gradually decreased. Therefore, the plastic strain of the model did not change in the subsequent quenching process.
The temperature of surfaces was below 200°C from 2.05 s. Therefore, the surfaces showed relatively high yield strength and resistance to deformation. In addition, the interior with higher temperatures began to cool down, and the shrinkage of the interior dominated in this stage. A low-temperature surface, which can be considered a “rigid shell,” limited the shrinkage of the interior. As shown in Fig. 14(c), the elastic compressive strains of the interior decreased, and the elastic tensile strains of surfaces decreased correspondingly from 2.05 s to 3.5 s.
After 3.5 s, with a continuous drop in the temperature of the interior, the compressive stress of point C turned into tensile stress and gradually increased as the quenching proceeded. The tensile stresses of points A and B had successively transformed into compressive stresses accordingly. Given the higher shrinkage of the lower surface in stage 1, the length of the lower surface was less than that of the upper surface during the entire quenching, which led to a less compressive effect. Therefore, the compressive residual stress of the lower surface was less than that of the upper surface in the subsequent quenching process.
According to the analysis above, the residual stress magnitudes and distribution of asymmetric HTC by the FE method showed great agreement with that of the asymmetric spray quenching experiment. Experiment and simulation results are helpful for the prediction and improvement of the residual stress distribution with asymmetric spray quenching of roller quenching equipment under industrial conditions for large-scale thick aluminum alloy plates.
The residual stress of thick aluminum alloy plates with symmetric and asymmetric spray quenching was investigated. The residual stress magnitude and distribution were measured by the layer removal method based on deflections and verified by FEM. With the above results, several conclusions can be made as follows:
(1) The residual stress linearity along the thickness direction was a symmetric parabolic type under the symmetric spray quenching condition. The residual stress magnitude decreased with the decrease in water flow rates. The water flow rates of 0.45, 0.30, and 0.15 m3/h resulted in reductions in the maximum compressive stress of 24.73%, 40.65%, and 64.52%, respectively, compared with that at the water flow rate of 0.60 m3/h.
(2) With the asymmetric spray quenching condition, the maximum compressive stresses and maximum tensile stresses decreased with the increase in the asymmetry of flow rates on the two surfaces. The subsurface stress magnitude of the high-flow rate surface was lower than that of the low-flow-rate surface, and the difference between the two sub-surface stresses increased with an increase in the difference in water flow rates. When the water flow rates on the two surfaces were 0.60 and 0.15 m3/h, respectively, the difference between the two subsurface stresses was 15.38 MPa.
(3) The simulated magnitude and distribution residual stresses by FEM showed great agreement with experimental results. The FE model can accurately predict quenched residual stress. The simulation result is also helpful to analyze stress evolution during the quenching process.
This research was financially supported by the National Key Research and Development Program of China (No. 2020YFF0218200).
The authors declare no conflicts of interest.
| [1] |
T. Dursun and C. Soutis, Recent developments in advanced aircraft aluminium alloys, Mater. Des., 56(2014), p. 862. DOI: 10.1016/j.matdes.2013.12.002
|
| [2] |
A. Heinz, A. Haszler, C. Keidel, S. Moldenhauer, R. Benedictus, and W.S. Miller, Recent development in aluminium alloys for aerospace applications, Mater. Sci. Eng. A, 280(2000), No. 1, p. 102. DOI: 10.1016/S0921-5093(99)00674-7
|
| [3] |
A. Azarniya, A.K. Taheri, and K.K. Taheri, Recent advances in ageing of 7xxx series aluminum alloys: A physical metallurgy perspective, J. Alloys Compd., 781(2019), p. 945. DOI: 10.1016/j.jallcom.2018.11.286
|
| [4] |
P.A. Rometsch, Y. Zhang, and S. Knight, Heat treatment of 7xxx series aluminium alloys—Some recent developments, Trans. Nonferrous Met. Soc. China, 24(2014), No. 7, p. 2003. DOI: 10.1016/S1003-6326(14)63306-9
|
| [5] |
G. Sha and A. Cerezo, Early-stage precipitation in Al–Zn–Mg–Cu alloy (7050), Acta Mater., 52(2004), No. 15, p. 4503. DOI: 10.1016/j.actamat.2004.06.025
|
| [6] |
G.T. Liang and I. Mudawar, Review of spray cooling - Part 1: Single-phase and nucleate boiling regimes, and critical heat flux, Int. J. Heat Mass Transfer, 115(2017), p. 1174.
|
| [7] |
G.T. Liang and I. Mudawar, Review of spray cooling - Part 2: High temperature boiling regimes and quenching applications, Int. J. Heat Mass Transfer, 115(2017), p. 1206.
|
| [8] |
J.S. Robinson, D.A. Tanner, and C.E. Truman, 50th anniversary article: The origin and management of residual stress in heat-treatable aluminium alloys, Strain, 50(2014), No. 3, p. 185. DOI: 10.1111/str.12091
|
| [9] |
J. Guo, H.Y. Fu, B. Pan, and R.K. Kang, Recent progress of residual stress measurement methods: A review, Chin. J. Aeronaut., 34(2021), No. 2, p. 54. DOI: 10.1016/j.cja.2019.10.010
|
| [10] |
N.S. Rossini, M. Dassisti, K.Y. Benyounis, and A.G. Olabi, Methods of measuring residual stresses in components, Mater. Des., 35(2012), p. 572. DOI: 10.1016/j.matdes.2011.08.022
|
| [11] |
R. Pan, T. Pirling, J.H. Zheng, J.G. Lin, and C.M. Davies, Quantification of thermal residual stresses relaxation in AA7xxx aluminium alloy through cold rolling, J. Mater. Process. Technol., 264(2019), p. 454. DOI: 10.1016/j.jmatprotec.2018.09.034
|
| [12] |
M.K. Khan, M.E. Fitzpatrick, S.V. Hainsworth, A.D. Evans, and L. Edwards, Application of synchrotron X-ray diffraction and nanoindentation for the determination of residual stress fields around scratches, Acta Mater., 59(2011), No. 20, p. 7508. DOI: 10.1016/j.actamat.2011.08.034
|
| [13] |
K. Tanaka, The cosα method for X-ray residual stress measurement using two-dimensional detector, Mech. Eng. Rev., 6(2019), No. 1, art. No. 18-00378. DOI: 10.1299/mer.18-00378
|
| [14] |
S. Nervi and B.A. Szabó, On the estimation of residual stresses by the crack compliance method, Comput. Meth. Appl. Mech. Eng., 196(2007), No. 37-40, p. 3577. DOI: 10.1016/j.cma.2006.10.037
|
| [15] |
C. Liu and X. Yi, Residual stress measurement on AA6061-T6 aluminum alloy friction stir butt welds using contour method, Mater. Des., 46(2013), p. 366. DOI: 10.1016/j.matdes.2012.10.030
|
| [16] |
R.G. Treuting and W.T. Read, A mechanical determination of biaxial residual stress in sheet materials, J. Appl. Phys., 22(1951), No. 2, p. 130. DOI: 10.1063/1.1699913
|
| [17] |
R. Kopun, L. Škerget, M. Hriberšek, D.S. Zhang, B. Stauder, and D. Greif, Numerical simulation of immersion quenching process for cast aluminium part at different pool temperatures, Appl. Therm. Eng., 65(2014), No. 1-2, p. 74. DOI: 10.1016/j.applthermaleng.2013.12.058
|
| [18] |
Y.B. Dong, W.Z. Shao, L.X. Lu, J.T. Jiang, and L. Zhen, Numerical simulation of residual stress in an Al–Cu alloy block during quenching and aging, J. Mater. Eng. Perform., 24(2015), No. 12, p. 4928. DOI: 10.1007/s11665-015-1758-9
|
| [19] |
G.S. Zhang, C.H. Mao, J. Wang, N. Fan, and T.T. Guo, Numerical analysis and experimental studies on the residual stress of W/2024Al composites, Materials, 12(2019), No. 17, art. No. 2746. DOI: 10.3390/ma12172746
|
| [20] |
C. Şimşir and C.H. Gür, 3D FEM simulation of steel quenching and investigation of the effect of asymmetric geometry on residual stress distribution, J. Mater. Process. Technol., 207(2008), No. 1-3, p. 211. DOI: 10.1016/j.jmatprotec.2007.12.074
|
| [21] |
W.C. Jiang, W. Woo, G.B. An, and J.U. Park, Neutron diffraction and finite element modeling to study the weld residual stress relaxation induced by cutting, Mater. Des., 51(2013), p. 415. DOI: 10.1016/j.matdes.2013.04.053
|
| [22] |
S.R. Yazdi, D. Retraint, and J. Lu, Study of through-thickness residual stress by numerical and experimental techniques, J. Strain Anal. Eng. Des., 33(1998), No. 6, p. 449. DOI: 10.1243/0309324981513147
|
| [23] |
N. Murugan and R. Narayanan, Finite element simulation of residual stresses and their measurement by contour method, Mater. Des., 30(2009), No. 6, p. 2067. DOI: 10.1016/j.matdes.2008.08.041
|
| [24] |
D.A. Tanner and J.S. Robinson, Residual stress prediction and determination in 7010 aluminum alloy forgings, Exp. Mech., 40(2000), No. 1, p. 75. DOI: 10.1007/BF02327551
|
| [25] |
D.A. Tanner and J.S. Robinson, Time transient validation of residual stress prediction models for aluminium alloy quenching, Mater. Sci. Technol., 32(2016), No. 14, p. 1533. DOI: 10.1080/02670836.2016.1195122
|
| [26] |
N. Chobaut, D. Carron, S. Arsène, P. Schloth, and J.M. Drezet, Quench induced residual stress prediction in heat treatable 7xxx aluminium alloy thick plates using Gleeble interrupted quench tests, J. Mater. Process. Technol., 222(2015), p. 373. DOI: 10.1016/j.jmatprotec.2015.03.029
|
| [27] |
J. Liu, Z.Y. Du, J.L. Su, et al., Effect of quenching residual stress on precipitation behaviour of 7085 aluminium alloy, J. Mater. Sci. Technol., 132(2023), p. 154. DOI: 10.1016/j.jmst.2022.06.010
|
| [28] |
Y.X. Cai, L.H. Zhan, Y.Q. Xu, et al., Stress relaxation aging behavior and constitutive modelling of AA7150-T7751 under different temperatures, initial stress levels and pre-strains, Metals, 9(2019), No. 11, art. No. 1215. DOI: 10.3390/met9111215
|
| [29] |
N. Fan, B.Q. Xiong, Z.H. Li, et al., Heat transfer behavior during water spray quenching of 7xxx aluminum alloy plates, J. Therm. Sci. Eng. Appl., 14(2022), No. 4, art. No. 041013. DOI: 10.1115/1.4051824
|
| [30] |
Y.N. Li, Y.A. Zhang, X.W. Li, et al., Effects of heat transfer coefficients on quenching residual stresses in 7055 aluminum alloy, Mater. Sci. Forum, 877(2016), p. 647. DOI: 10.4028/www.scientific.net/MSF.877.647
|
| [1] | Zheng-jie Shao, Hai-po Liu, Xiao-chun He, Bing Zhou, Yang Li, Shang-zhou Zhang, Meng-jin Li, Shu-jun Li. Microstructure and finite element analysis of hot continuous rolling of doped tungsten rod [J]. International Journal of Minerals, Metallurgy and Materials, 2019, 26(3): 369-376. DOI: 10.1007/s12613-019-1746-4 |
| [2] | Guo-fa Sui, Jin-shan Li, Hong-wei Li, Feng Sun, Tie-bang Zhang, Heng-zhi Fu. Investigation on the explosive welding mechanism of corrosion-resisting aluminum and stainless steel tubes through finite element simulation and experiments [J]. International Journal of Minerals, Metallurgy and Materials, 2012, 19(2): 151-158. DOI: 10.1007/s12613-012-0531-4 |
| [3] | Jian-xun Fu, Jing-she Li, Hui Zhang. Effects of the width and thickness of slabs on broadening in continuous casting [J]. International Journal of Minerals, Metallurgy and Materials, 2010, 17(6): 723-728. DOI: 10.1007/s12613-010-0380-y |
| [4] | Zi-qian Huang, Bo Xiang, Yue-hui He, Bai-yun Huang. Thermal residual stress analysis of coated diamond grits [J]. International Journal of Minerals, Metallurgy and Materials, 2009, 16(2): 215-219. DOI: 10.1016/S1674-4799(09)60036-4 |
| [5] | Qijin Huang, Guoquan Liu, Yong Li, Jin Gao, Zhengqiu Gu, Yuanzheng Ma, Haibin Xue. Pilot biomechanical design of biomaterials for artificial nucleus prosthesis using 3D finite-element modeling [J]. International Journal of Minerals, Metallurgy and Materials, 2004, 11(6): 561-565. |
| [6] | Zhanghua Chen, Jiajian Jin, Jiumei Xiao. Numerical simulation of strain localization and damage evolution in large plastic deformation using mixed finite element method [J]. International Journal of Minerals, Metallurgy and Materials, 2004, 11(3): 273-277. |
| [7] | Renbo Song, Yonglin Kang, Xueping Ren, Hongbo Dong, Jiwen Wang. Two-dimensional simulation on the rolling process of semi-solid 60Si2Mn by finite element method [J]. International Journal of Minerals, Metallurgy and Materials, 2002, 9(4): 273-276. |
| [8] | Junping Pu, Yuanfeng Wang. Contact position controlling for two-dimensional motion bodies by the boundary element method [J]. International Journal of Minerals, Metallurgy and Materials, 2002, 9(2): 138-144. |
| [9] | Xichun Chen, Deguang Zhou, Jie Fu, Weiguo Xu. 3-D Finite-Element Numerical Simulation of Centrifugal Induction Electrosalg Casting Solidification Process [J]. International Journal of Minerals, Metallurgy and Materials, 2001, 8(4): 254-258. |
| [10] | Zhiyuan Shi, Wanjun Wang, Zhigang Kan, Jinglin Wen, Xinhua Wang. Simulation of Temperature and Flow Field by Three Dimension Finite Element Method for Castex Process of AS Wire [J]. International Journal of Minerals, Metallurgy and Materials, 1999, 6(4): 262-267. |
Figures(14) / Tables(2)
Copyright©2019 Editorial Office of International Journal of Minerals, Metallurgy and Materials 京ICP备13030111号
Supported by:
Beijing Renhe Information Technology Co., Ltd.
Email:
info@rhhz.net
| Machining sequences | Tool speed / (r·min−1) | Feed rate / (mm·min−1) | Depth of cut / mm |
| Rough machining | 3000 | 500 | 0.8 |
| Finish machining | 3000 | 300 | 0.2 |
DownLoad:
CSV
| Temperature / °C | Density / (kg·m−3) | Thermal conductivity / (W·m−1·K−1) | Specific heat / (J·kg−1·K−1) | Linear expansion coefficient / (10−6 K−1) | Young’s modulus / GPa | Poisson ratio | Yield strength / MPa | Tangent modulus / GPa |
| 20 | 2842 | 145 | 852.3 | 21.6 | 72.0 | 0.3 | 412 | 1 |
| 100 | 2828 | 152 | 894.5 | 22.7 | 65.2 | 0.3 | 382 | 0.5 |
| 200 | 2807 | 160 | 940.5 | 24.0 | 56.3 | 0.3 | 370 | 0.25 |
| 300 | 2787 | 167 | 982.5 | 24.2 | 38.0 | 0.3 | 120 | 0.1 |
| 400 | 2761 | 171 | 1003.2 | 26.0 | 31.5 | 0.3 | 50 | 0.01 |
| 500 | 2735 | 178 | 1045 | 27.5 | 25.0 | 0.3 | 20 | 0.005 |
DownLoad:
CSV
