METAL 2008 13.-15.5.2008 Hradec nad Moravicí COMPARISON BETWEEN THE METHODS OF DETERMINING THE CRITICAL STRESS FOR INITIATION OF DYNAMIC RECRYSTALLIZATION IN 316 STAINLESS STEEL M. Jafari, A. Najafizadeh Department of Materials Engineering, Isfahan University of Technology, Isfahan, Iran. [email protected] E-mail: [email protected] Abstract The critical stress for initiation of dynamic recrystallization (σc) could be identified by the inflection point on the strain hardening rate (θ=dσ/dε) versus flow stress ( σ ) curve. Several methods have been proposed to calculate this value on the basis of mathematical methods. One of them was proposed by Stewart, Jonas and Montheillet in which this critical point appears as a distinct minimum in the (-dθ/dσ versus σ ) through differentiating from θ versus σ . Another one was presented by Najafizadeh and Jonas which obtain by modifying the Poliak and Jonas method. According to this method, the strain hardening rate was plotted against flow stress, and the value of σc was obtained numerically from the coefficients of the third-order equation that was the best fit from the experimental θ - σ data. In this study, hot compression tests were used in the range of 1000 to 1100 °C and strain rates of 0.01-1s-1 and strain of 1 for the above mentioned methods to compare the value of the σc for 316 stainless steel. The result shows that the method presented by Najafizadeh and Jonas not only simpler than the previous one but also, it has a good agreement with microstructures. Furthermore, the normalized critical stress for this steel was obtained uc = σc/σp = 0.92. Keywords: Dynamic Recrystallization; Critical Stress; Strain hardening rate; . 1. Introduction Dynamic recrystallization (DRX) is a powerful tool for controlling microstructural evolution during industrial hot deformation processing [1-5]. When this type of softening mechanism is operating, both nucleation as well as growth take place while the strain is being applied [1-2]. Calculation of the critical condition for the initiation of DRX is of considerable interest for modeling of industrial processes [6-9]. It depends on the chemical composition of the material, the grain size prior to deformation, and the deformation condition (T and ε& ) [7-10]. Several researchers have proposed mathematical relations to predict the initiation of DRX. For example, M. R. Barnett, G. L. Kelly and P. D. Hodgson [11] have identified the critical strain for initiation of DRX using the kinetics of static recrystallization (SRX) modified to allow SRX to begin before the end of deformation. This approach defines the critical stress for the initiation of DRX. On the basis of a dislocation density work hardening model, G. METAL 2008 13.-15.5.2008 Hradec nad Moravicí Gottstein, M. Frommert, M. Goerdeeler, and N. Schafer [12] have also predicted the critical strain for initiation of DRX. The method developed by Ryan and McQueen [10] defines the critical strain as the strain at which the experimental flow curve deviates from the idealized dynamic recovery curve, based upon differences in the strain hardening behaviors associated with DRX and dynamic recovery. An alternative method based on the thermodynamics of irreversible processes was developed by Poliak and Jonas [6-9]. This method eliminates the need for extrapolated flow stress data; the onset of DRX can be identified by an inflection point in the θ - σ curve corresponding to the appearance of an additional thermodynamic degree of freedom in the system. Another one is Najafizadeh and Jonas [13] method which obtains by modifying the Poliak and Jonas method. According to this method, the strain hardening rate was plotted against flow stress, and the third-order equation that the best fit the experimental θ - σ data from yield to the peak stress was found for each of the deformation conditions. Then the value of εc was obtained numerically from the some coefficients of the third-order equation. In this method the critical stress for the initiation of DRX is given by the simple relationship σc=B/3A, where A and B are two of the four coefficients of the third order equation. In the present work, the two latter methods were compared according to microstructures which obtained from quenched samples at critical points with different deformation conditions. The value of σc depends on the experimental conditions, if instead the normalized stressstrain curve is employed, according to the Najafizadeh and Jonas [13] method the normalized stress is given by uc= σc/σp=-B'/3A', where A' and B' are two of four coefficients of normalized third order equation. The value of uc is constant and independent of the deformation conditions The validity of these relationships was examined for 316 stainless steel. 2. Experimental procedure The chemical composition of the 316 stainless steel employed in this work is given in Table.1. This material was supplied with the form of a hot rolled bar with a diameter of 10 mm. Cylindrical samples 5 mm in diameter and 7.5 mm in height were prepares with their axes aligned along the rolling direction. For minimize the coefficient of friction during hot compression testing, glass plates were put into graves (Rastegave) as lubricant. One-hit hot compression tests were carried out on a dilatometer machine (BAHR Thermoanalysen GmbH) equipped with induction furnace and a vacuum chamber. The samples totally were deformed after preheated at 1150 °C for 15 min in a vacuum chamber. Then they cooled to test temperature. Tests were performed at the temperature of 950-1100 °C with strain rates of 0.01-1s-1 to the strain of 1. Samples were cut along compression direction and were polished electrochemically in a 65% nitric acid solution. In order to minimize the coefficient of friction during hot compression testing, glass plates were into graves as lubricant. In order to minimize the effect of friction on the stress-strain curves the method of Ebrahimi and Najafizadeh [14] were used. 3. Results and Discussion Fig. 1 present the typical DRX stress-strain curves of 316 stainless steel. In order to calculate the critical stress for initiation of DRX (σc) the Poliak [6-9] and Najafizadeh [13] methods were used. The σc could be identified by the inflection point on the strain hardening rate (θ=dσ/dε) versus flow stress ( σ ) curve. The Poliak method requires differentiation of the stress- strain curve, but short- range noise can render such differentiation calculus impossible. The METAL 2008 13.-15.5.2008 Hradec nad Moravicí variation in the derivate of the true stress can be several orders of magnitude larger than the mean value. In order to solve this problem, a smoothed stress-strain curve was obtained by modeling the region of interest (the region of plastic deformation that encompasses the stress peak) as a 9th order polynomial. These polynomials are presented as common form of below. σ = a + bε + cε 2 + dε 3 + eε 4 + fε 5 + gε 6 + hε 7 + iε 8 + jε 9 (2) The 9th order polynomials were obtained for the curves of Fig. 1 is presented in the table 2. As differentiation of the polynominal is simple and free of noise, the work hardening profile can be readily derived. The θ-σ plot corresponding to the stress-strain curve of 1100 °C with the strain rate of 0.1 s-1 is shown in Fig. 2. The θ-σ curve is directly related to concave-downward shape of the stress-strain curve. In the absence of dynamic recrystallization, the work hardening rate will taper off to zero as the strain goes to arbitrarily large values. However, dynamic recrystallization causes a downward inflection in the θ-σ curve, leading to zero and negative work hardening rates (corresponding to the peak and post peak softening, respectively). The method involves differentiating the above θ-σ curve. The critical point then appears as a distinct minimum represents the initiation of dynamic recrystallization (Fig.3). 2- Najafizadeh and Jonas method In the approach of Poliak and Jonas [6-9, 15] and N. D. Ryan and H. J. McQueen [10], the initiation of DRX is associated with the point of inflection in the θ-σ curve. To plot this kind of curve, it is necessary to find an equation that fits the experimental θ-σ data from yield until the peak stress. The simplest equation that has an inflection point is: θ = Aσ 3 + Bσ 2 + C σ + D (3) Where: A, B, C, D is constants for a given set of deformation conditions. All the third order equations obtained from this study are presented in Table.3. Differentiation of equation 3 with respect to σ results in: dθ = 3 Aσ 2 + 2 Bσ + C dσ d 2θ = 0 ⇒ 6 Aσ + 2 B = 0 dσ 2 σc = − B 3A (4) Fig.4 shows for example a θ-σ curve from yield until the peak stress at 1100 °C and strain rates of 0.1 s-1. In the table 4 the compressions of these two methods are presented and the fraction's values σc/σP in a certain range show approximately independence to various deformation conditions. As can be seen in Fig.5 those obtained from Najafizadeh and Jonas method are in good agreement with microstructures from quenched samples at critical points in different condition. Thus, this method is simpler than another one. So we suggest usage of this method. METAL 2008 13.-15.5.2008 Hradec nad Moravicí The normalized stress-strain curve Once normalized stress (u=σ/σp) - stain (w=ε/εp) plot have been established, the flow stress at any strain, strain rate and temperature can be calculated from this single curve. Some normalized stress-strain curves of the 316L stainless steel deformed under different conditions of deformation are illustrated in Fig.6. The normalized strain hardening rate (θN=du/dw) versus normalized stress (u) plots obtained from such curves have inflection points that present the onset of initiation of DRX. For a given steel, the inflection points of θN-u curve correspond to the approximately constant normalized critical stress uc= σc/ σp = 0.92. Consequently for a given steel deformed under different conditions of deformation the ratio of the critical strain to the peak strain, wc=εc/εp, is also approximately constant. To obtain the critical stress for initiation of DRX under different conditions of deformation, the normalized stress-strain curve was used. The best fit equation corresponding to these data had the form: θ N = A′u 3 + B ′u 2 + C ′u + D ′ (5) The coefficients of this equation do not depend on the deformation conditions. The critical stress for initiation of DRX can be calculated: dθ N = 3 A′u 2 + 2 B ′u + C ′ du (6) d 2θ N = 0 ⇒ 6 A′u c + 2 B ′ = 0 du 2 uc = − B′ 3 A′ (7) Using equation 7, the critical stress for any deformation conditions can be calculated by using σp for such condition. This approach also helps to explain the constancy of the σc/σp. As can be seen in Fig. 6 there are small discrepancies that arise from a weak dependence of the curves on the value of Z. This source of error can be reduced by deriving 4 normalized curves pertaining to four different sets of deformation conditions and then averaging the third order of equation 6 (Fig.7): θ = −375.55σ 3 + 1038 .2σ 2 − 959.52σ + 296.83 (8) Thus the uc will be; uc = − 1038.2 = 0.92 (9) 3 × 375.55 The uc value is expected to be constant under all conditions of deformation. This hypothesis can be assessed by plotting the measured values of σc versus those calculated from equation 9. Such a comparison is presented in Fig. 8. As can be seen in this figure there is a reasonably good agreement between these two sets of values. METAL 2008 13.-15.5.2008 Hradec nad Moravicí 4. Conclusion The principal conclusions that can be drawn from the present work are the following: 1. The Najafizadeh et.al. method for calculated the onset of DRX is in good agreement with microstructures obtained from quenched samples at this point for different condition of deformation. 2. The method of Najafizadeh et al. is simpler than Poliak et.al. because he used the third in place of nine order equation. 3. The normalized critical stress for initiation of DRX (uc) that calculated from the each method is approximately constant and independent of the deformation conditions. 4. The critical value of uc for 316L stainless steel obtained from Najafizadeh et.al. method is 0.92. Acknowledgments The author expresses theirs thanks to the Center of Excellence for Steel for supporting of this work References 1. J. J. Jonas, Materials science and Engineering, 1994, A184, pp.155. 2. R. D. Doherty, D. A. Hughes, F.J. Humphreys, J. J. Jonas, D. Juul Jensen, M. E. Kassner, W.E. King, T. R. McNelley, H. J. McQueen, A. D. Rollet, Materials Science and Engineering, 1997, A238, pp.219. 3. H. J. McQueen, S. Yue, N. D. Ryan and E. Fry, Materials processing Technology, 1995, A53, pp.293. 4. G. R. Stewart, J. J. Jonas and F. Montheillet, ISIJ International, 2004, Vol. 44, No.9, pp.1581. 5. F. Montheillet, J. J. Jonas, Encyclopedia of Applied Physics, 1996, Vol. 16, pp.205. 6. E. I. Poliak, J. J. Jonas, Acta mater, 1996, Vol. 44, No.1, pp.127. 7. E. I. Poliak, J. J. Jonas, ISIJ International, 2003, Vol. 43, No.5 , pp.684. 8. E. I. Poliak, J. J. Jonas, ISIJ International, 2003, Vol. 43, No.5 , pp.692. 9. J. J. Jonas, E. I. Poliak, Mater. Sci. Forum, Acta mater, 2003, Vols. 426-432, pp.57. 10. N.D. Ryan and H. J. McQueen, Can, Metall, Q., 1990, Vol.29, pp.147 11. M. R. Barnett, G. L. Kelly, P. D. Hodgson, Scripta mater, 2000, Vol. 43, pp. 365. 12. G. Gottstein. M. Frommert, M. Goerdeeler, N. Schafer, science and Engineering, 2004, A387-389, pp.604. 13. A. Najafizadeh , J. J. Jonas, ISIJ International, 2006, Vol. 46, No. 11, pp. 1679. 14. R. Ebrahimi and A. Najafizadeh, "A new method for evaluation of friction in bulk metal forming", Materials processing Technology, Vol. 152, Issue 2, 2004, PP.136-143. 15. E. I. Poliak, J. J. Jonas, ISIJ International, 2004, Vol. 44, No.11 , pp.1874. Table.1. Chemical composition of 316 stainless steel %Cr %Mo %Cu %C %Si %S %P %Mn %Ni 0.043 0.285 0.022< 0.044 1.635 10.225 16.679 2.207 0.489 METAL 2008 13.-15.5.2008 Hradec nad Moravicí Table.2. Coefficients for stress (σ) versus strain (ε) equations T=1000-1100°C ε& =0.01-1 s-1. Strain rate σ −ε T°°C (s-1) σ = 22.95 + 1.2 × 103 ε − 2.3 × 10 4 ε 2 + 1.4 × 105 ε 3 − 5.1 × 105 ε 4 + 1 1100 1.14 × 106 ε 5 − 1.6 × 106 ε 6 + 1.38 × 106 ε 7 − 6.57 × 105 ε 8 + 1.33 × 105 ε 9 σ = 24 + 1.84 × 10 3 ε − 2.07 × 10 4 ε 2 + 1.29 × 10 5 ε 3 − 4.71 × 10 5 ε 4 + 0.1 1.07 × 10 6 ε 5 − 1.5 × 10 6 ε 6 + 1.3 × 10 6 ε 7 − 6.2 × 10 5 ε 8 + 12.6 × 10 4 ε 9 σ = 18.23 + 1.34 × 103 ε − 1.6 × 104 ε 2 + 1.1 × 105 ε 3 − 4.3 × 105 ε 4 + 0.01 1000 and 1.03 × 106 ε 5 − 1.51× 106 ε 6 + 1.34 × 106 ε 7 − 6.54 × 105 ε 8 + 1.36 × 105 ε 9 σ = 22 + 2.4 × 103 ε − 2.3 × 104 ε 2 + 1.3 × 105 ε 3 − 4.56 × 105 ε 4 + 0.1 9.88 × 105 ε 5 − 1.35 × 106 ε 6 + 1.13 × 106 ε 7 − 5.28 × 105 ε 8 + 1.05 × 105 ε 9 Table.3. Coefficients for stress (θ) versus strain (σ) equations T=1000-1100°C and ε& =0.01-1 s-1 obtained from yield until the peak stress. T °C Strain rate (S-1) θ/σrelation 1 θ =-0.21σ3 +61.63σ2 -5991σ +194146 1100 1000 0.1 θ =-0.018σ3 +6.9σ2 -874.4σ +37071 0.01 θ =-0.32σ3 +63.67σ2 -4168.7σ +91045 0.1 θ =-0.012σ3 +5.17σ2 -765.99σ +38068 Table.4. Two sets of results obtained from two kinds of methods mentioned in this article. Poliak-Jonas method Najafizadeh-Jonas method Deformation condition 1100 °C، 0.1 s-1 σc σc/σp σc σc/σp 98.42 0.95 97.82 0.94 124.23 0.88 127.77 0.90 65.84 0.95 66.32 0.96 147.9 0.91 143.61 0.88 -1 1100 °C، 1 s 1100 °C، 0.01 s-1 -1 1000 °C01، 0.1 s METAL 2008 13.-15.5.2008 Hradec nad Moravicí Fig.1. Stress- strain curves obtained from one-hit compression tests 2000.00 θ=dσ/dε 1500.00 1000.00 500.00 0.00 -500.00 0 20 40 60 80 100 120 σ Fig. 2. The θ-σ plot corresponding to the stress-strain curve at 1100 °C and strain rate 0.1 s-1 METAL 2008 13.-15.5.2008 Hradec nad Moravicí Fig.3. The dθ/dσ versus σ curve at 1100 °C and strain rates of 0.1 s-1, these graphs must be centered on the inflection point to determine the critical strain for onset of DRX. Fig.4. The θ-σ plot corresponding to the stress-strain curve at 1100°C and strain rate 0.1 s-1 obtained from yield until the peak stress. Fig.5. Microstructures obtained from quenched samples at critical strian for initiation of DRX in different condition. ; Arrow showed bulging and serration produce by the initiation of DRX. METAL 2008 13.-15.5.2008 Hradec nad Moravicí Fig.6. The normalized stress -stain curves obtained under different conditions of deformation. Fig.7. Normalized strain hardening rate versus normalized stress at 1100°C and strain rates of 0.1 s-1. Fig.8. Critical stress for the initiation of DRX (σc). Here the measured values are compared with those estimated using the relation σ= 0.92σP.
© Copyright 2024