Jtam-A4.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 54, 1, pp. 239-249, Warsaw 2016 DOI: 10.15632/jtam-pl.54.1.239 MAGNETO-THERMO-MECHANICAL CREEP BEHAVIOR OF NANO-COMPOSITE ROTATING CYLINDER MADE OF POLYPROPYLENE REINFORCED BY MWCNTS Abbas Loghman, Hossein Shayestemoghadam Faculty of Mechanical Engineering, Department of Solid Mechanic, University of Kashan, Kashan, Iran e-mail: aloghman@kashanu.ac.ir; h.shayestemoghadam@gmail.com History of strains, stresses and displacements of a rotating cylinder made of polypropylene reinforced by multi-walled carbon nanotubes (MWCNTs) subjected to magneto-thermo- -mechanical loading is investigated using Burgers viscoelastic creep model. By making use of equations of equilibrium, stress-strain and strain-displacement, a constitutive differential equation containing creep strains is obtained which is solved semi analytically. It has been found that radial displacement, tangential strain and absolute values of radial strain are increasing with time at a decreasing rate so that they finally approach the steady state condition. Effective stresses are decreasing at the inner and increasing at the outer surface of the cylinder. Keywords: nano-composite cylinder, creep analysis, MWCNTs, polypropylene, Burgers model 1. Introduction Composite cylinders subjected to thermo-mechanical load are extensively used in aerospace, pressure vessels and petrochemical industries. Carbon nanotubes reinforced polymer matrix composites have shown high strength properties and are widely used in manufacturing of com- ponents exposed to high pressure environment. Even at room temperature, significant creep deformation is observed for polypropylene tubes. Therefore, creep analysis and creep life as- sessment of such components are very important.When such vessels are loaded, thermo-elastic stresseswill bedeveloped in the cylinder at zero time.However, because of creep evolution, stress redistribution occur during life of the component, which can affect its long-time performance. The analysis of an internally pressurized, homogeneous, orthotropic rotating cylinder sub- jected to a steady state creep condition was investigated by Bhatnagar et al. (1984). In ano- ther work, they considered an orthotropic thick-walled cylinder under primary creep conditions (Bhatnagar et al., 1986). In their second work, the authors presented the analysis of an ortho- tropic, thick-walled cylinder undergoing creep due to combined action of internal and external pressures and rotary inertia. As a result, they found that the material which is strong in the radial directionmaybebeneficial for design of the cylinder as it gives lower values of the effective stress. Evaluation of creep compliances of unidirectional fiber-reinforced compositeswas done by Moal andPerreux (1994). In that paper, amethod based on themodel of Laws andMcLaughlin was proposed for the determination of viscoelastic behavior of unidirectional fiber-reinforced composites. Moreover, the interface problemwas taken into account by using anisotropic elastic coefficients for the reinforcement fiber. The suggested procedure allowed the viscoelasticity of the resin to be characterized. Avariationalmethodwas developed byOhno et al. (2002) for ana- lyzing the matrix creep induced time-dependent change in fiber stress profiles in unidirectional composites. They verified the solutions on the basis of an energy balance equation and a finite 240 A. Loghman, H. Shayestemoghadam difference computation. They also showed that the solution for the fiber pull-out model agreed well with an experiment on a single carbon fiber/acrylic model composite if the initial slip at fiber/matrix interface was taken into account. Singh and Ray (2002) modeled anisotropy and creep in an orthotropic aluminum-silicon carbide composite rotating disc. They observed that the anisotropy helped reduction of the tangential strain rate significantly, more near the inner radius. It was also found the strain rate distribution in the orthotropic disc was lower than that of isotropic disc following von Misses criterion. Creep deformations and stresses in thick-walled cylindrical vessels of functionally gradedmaterials subjected to internal pressurewere investiga- ted byYou et al. (2007). They examined how variations of material parameters along the radial direction affect the stresses in the vessels. Yang et al. (2006) carried out characterization of ten- sile creep resistance of polyamide 66 nano-composites. To develop their works, they presented both a viscoelastic creep model named Burgers model and an empirical method called Findley power law. They revealed that the simulation results from both models agreed quite well with the experimental data. Jia et al. (2011) studied the creep and recovery of polypropylene/multi- -walled carbon nanotube composites. They showed that the creep strain reduceswith a decrease in temperature and an increase in the content of carbon nanotubes.Magneto-thermo-elastic cre- ep analysis of functionally graded cylinders was presented by Loghman et al. (2010). The paper describes time-dependent creep stress redistribution analysis of a thick-walled FGM cylinder subjected to a uniformmagnetic field, temperature field and internal pressure. They calculated stress redistributions iteratively usingmagneto-thermo-elastic stresses as initial values for stress redistributions. The result indicated that the radial stress redistributions were not significant for different material properties, while major redistributions occurred for circumferential and effective stresses. A semi analytical solution of magneto-thermo-elastic stresses was suggested for functionally graded variable thickness rotating disks by Ghorbanpour Arani et al. (2010). In the paper, stresses and perturbation of magnetic field vector in FG rotating disks were de- termined using infinitesimal theory of magneto-thermo elasticity under plane stress conditions. It was found that imposing amagnetic field significantly decreases tensile circumferential stres- ses. Thus, the fatigue life of the disk would be significantly improved by applying themagnetic field. They suggested that the results of that investigation could be applied for optimum design of FG hollow rotating disks with variable thickness. Theory of plasticity for carbon nanotube reinforced composites was mentioned by Barai and Weng (2011). It was found that, with per- fect interface contact, the decreasing of the CNT radius would improve the overall stiffness and plastic strength, but with an imperfect interface the size effect was reversed. Time-dependent thermo-elastic creep analysis of a rotating disk made of Al–SiC composite using Mendelson’s method of successive elastic solution was presented by Loghman et al. (2011). They found that the stresses, displacement, and creep strains were changing with time at a decreasing rate so that after almost 50 years the solution approached the steady-state condition. Themain objective of this paper is to obtain history of creep stresses and deformations of a nano-composite rotating cylindermade of polypropylene reinforced byMWCNTs usingBurgers viscoelastic creep model undermagneto-thermo-mechanical loadings. 2. Geometry, loading condition and material properties 2.1. Geometry and loading condition A long rotating thick-walled nano-composite cylinder made of polypropylene reinforced by MWCNTs with inner radius ri and outer radius ro is considered (Fig. 1). The cylinder is sub- jected to a uniform magnetic field in the axial direction and a uniform temperature field. The following data for geometry and loading conditions are used in this paper: ro/ri = 2, Magneto-thermo-mechanical creep behavior of nano-composite... 241 MWCNTs content = 4.5%, ν = 0.45, T = 80◦C, ω = 52.35rad/s, µ0 = 4π · 10−7H/m, HZ =1 ·108A/m. Fig. 1. Schematic of the rotating thick-walled composite cylinder subjected to uniformmagnetic and thermal fields 2.2. Material properties Young’s modulus of the polymeric composite cylinder reinforced with different MWCNT contents is given in Table 1 based on experimental results reported by Jia et al. (2011). Table 1. Young’s modulus of propylene nano-composite with different MWCNT contents (Jia et al., 2011) MWCNTs content [vol.%] Young’s modulus [GPa] 0 1.83±0.11 0.3 2.10±0.11 0.6 2.12±0.03 2.8 2.33±0.14 4.5 2.42±0.13 3. Burgers viscoelastic creep model Burgers four-element model can be used to predict the viscoelastic creep behavior of polymer based nano-composites. This model is shown in Fig. 2 with Maxwell and Kelvin rheological elements connected in series. Fig. 2. Schematic diagram of Burgers four-elementmodel (Yang et al., 2006) 242 A. Loghman, H. Shayestemoghadam The creep constitutive model based on the Burgers law is written as follows (Yang et al., 2006) ε= σ0 EM + σ0 EK ( 1− e −t τ ) + σ0 ηM t τ = ηK EK (3.1) in which σ0 is the initially applied stress, τ is the retardation time taken to produce 63.2% or (1−e−1) of the total deformation in theKelvin element,EM and ηM are the elasticmodulus and viscosity of theMaxwell spring and dashpotmodel,Ek and ηk are the elastic modulus and the viscosity of theKelvin springanddashpotmodel.TheBurgersmodelwhich includes the essential elements can be satisfactorily applied to describe behavior of viscoelastic materials practically. Differentiating Eq. (3.1) with respect to time gives the strain rate constitutive equation of the Burgers model as ε̇= σ0 ηM + σ0 ηK e −t τ (3.2) The material parameters, ηm,EK and ηk are simulated from the experimental data (Jia et al., 2011). Table 2. The simulated parameters of the Burger model with different MWCNT contents for long term prediction (Jia et al., 2011) MWCNTs [vol.%] Ek [MPa] ηk [MPas] ηm [s] τ [s] 0 5.7 9.00E+07 1.50E+10 1.49E+07 0.3 8.5 1.00E+08 2.10E+10 1.53E+07 0.6 9.2 1.50E+08 2.60E+10 1.63E+07 2.8 9.6 1.70E+08 2.70E+10 1.77E+07 4.5 10.4 2.00E+08 2.80E+10 1.92E+07 4. Theoretical analysis Thestrain-displacement relationship for a long cylinderunderanaxisymmetric loading condition is written as εr = ∂ur ∂r εθ = ur r (4.1) where εr and εθ are the radial and circumferential total strains, andur is the radial displacement. Considering the total strains to be the sum of elastic, thermal and creep strains, the stress- -strain relations may be written as follows σr =C11εr +C12εθ −λrTr− (C11εcr +C12ε c θ ) σθ =C21εr +C22εθ−λθTr− (C21εcr +C22ε c θ ) (4.2) where σr and σθ are the radial and circumferential stresses, ε c r and εc θ are the radial and cir- cumferential creep strains, Tr is the temperature field. and the other coefficients are defined as C11 = E(1−ν) (1+ν)(1−2ν) C12 = Ev (1+ν)(1−2ν) C21 =C12 C22 =C11 λr = Eα 1−2ν λθ = E 1+ν (4.3) Magneto-thermo-mechanical creep behavior of nano-composite... 243 inwhichE, ν andα areYoung’smodulus,Poison’s ratio and the coefficient of thermal expansion of nano-composite, respectively. The equilibrium equation of a thick-walled composite hollow cylinder subjected to a uniform magnetic field is written as(Loghman et al., 2010) ∂σr ∂r + σr−σθ r +fr+ρrω 2 =0 (4.4) in which ρrω2 is the centrifugal body force per unit volume, and fr is the Lorentz force written as fr =µ(r)H 2 z ∂ ∂r (∂ur ∂r + ur r ) (4.5) where µ(r) is the magnetic permeability and Hz is the magnetic field intensity in the axial direction. Substituting the strains from Eqs. (4.1) into Eqs. (4.2), and then substituting the radial and circumferential stresses into equilibriumEq. (4.4) and substituting fr fromEq. (4.5), the following constitutive differential equation for displacement is obtained (C11+µ(r)H 2 z ) ∂2ur ∂r2 + (C11 r + µ(r) r H2 z )∂ur ∂r + (C22 r2 − H2 z r2 µ(r) ) ur = λr−λθ r Tr +λr ∂Tr ∂r + εc r (C11−C21) r + εc θ (C12−C22) r + ∂ ∂r (C11ε c r +C22ε c θ )−ρrω2 (4.6) The above differential equation is summarized as D11 ∂2ur ∂r2 +D12 ∂ur ∂r +D13ur +D14 =0 (4.7) in which D11 =C11+µ(r)H 2 z D12 = C11 r + µ(r) r H2 z D13 = C22 r2 − H2 z r2 µ(r) D14 = λr−λθ r Tr+λr ∂Tr ∂r + εc r (C11−C21) r + εc θ (C12−C22) r + ∂ ∂r (C11ε c r +C22ε c θ )−ρrω2 (4.8) D14 contains creep strains which are time, temperature and stress dependent. Ifwe ignore the time-dependent creep strains in the coefficientD14, thendifferential Eq. (4.7) becomes Navier’s equation, a non-homogenous second-order ordinary differential equation with variable coefficients the solution to which can be found from magneto-thermo-elastic analysis. This analysis is done by making use of the division method (Hosseini Kordkheili and Naghda- bai, 2007). In this method, the cylinder thickness is divided into a finite number of divisions. ThenNavier’s equation for k-th division yields the following differential equation with constant coefficients ( D (k) 11 ∂2 ∂r2 +D (k) 12 ∂ ∂r +D (k) 13 ) uk r +Dk14 =0 (4.9) The coefficients of Eq. (4.9) are evaluated in each division in terms of constants and the radius of kth division. The exact solution to Eq. (4.9) can be written in the form of u(k) r =X (k) 1 exp(η (k) 1 r (k))+X (k) 2 exp(η (k) 2 r (k))− D (k) 14 D (k) 13 (4.10) where η (k) 1 ,η (k) 2 = D (k) 12 ± √ (D (k) 12 ) 2−4D(k)13 D (k) 11 2D (k) 11 (4.11) 244 A. Loghman, H. Shayestemoghadam It is noted that this solution to Eq. (4.9) is valid in the following sub-domain r(k)− t(k) 2 ¬ r¬ r(k)+ t(k) 2 (4.12) where t(k) is the thickness of kthdivision andX (k) 1 ,X (k) 2 are unknownconstants fork-th division. The unknowns X (k) 1 and X (k) 2 are determined by applying the necessary boundary conditions between two adjacent sub-domains. For this purpose, the continuity of the radial displacementu as well as the radial stress σr is imposed at the interfaces of the adjacent sub-domains. These continuity conditions at the interfaces are written as u(k) r ∣ ∣ ∣ ∣ ∣ r=r(k) + t(k) 2 =u(k+1) r ∣ ∣ ∣ ∣ ∣ r=r(k+1) − t(k+1) 2 σ(k) r ∣ ∣ ∣ ∣ ∣ r=r(k) + t(k) 2 =σ(k+1) r ∣ ∣ ∣ ∣ ∣ r=r(k+1) − t(k+1) 2 (4.13) and the global boundary conditions are σr =0 at r= ri σr =0 at r= ro (4.14) Continuity conditions Eqs. (4.13) together with global boundary conditions Eqs. (4.14) yield a set of linear algebraic equations in terms ofX (k) 1 andX (k) 2 . Solving the resultant linear algebraic equations for X (k) 1 and X (k) 2 , the unknown coefficients of Eq. (4.10) are calculated. Then, the displacement component ur and the stresses are determined in each radial sub-domain.Accuracy of the results will be improved by increasing number of divisions. 5. Time-dependent creep analysis For time-dependent creep analysis, the creep strains in the coefficient D14 must be considered. The creep strains are time, temperature, and stress dependent.Creep strain increments are rela- ted to the current stresses and thematerial uni-axial creep behavior byPrandtl-Reuss relations. For problems of a rotating thick-walled composite cylinder with axial symmetry, these relations (Loghman et al., 2010) are written as follows ε̇c r = ε̇c 2σe [2σr − (σθ +σz)] ε̇cθ = ε̇c 2σe [2σθ − (σr+σz)] ε̇c z = ε̇c 2σe [2σz − (σr+σθ)] (5.1) where ε̇c r , ε̇c θ and ε̇c z are the radial, circumferential, and axial creep strain rates, ε̇c and σc are the equivalent creep strain rate and equivalent stress, respectively. These equivalent or effective variables are defined as follows ε̇c = 2 √ 3 √ (ε̇c r )2+(ε̇c θ )2+(ε̇c z )2 σe = √ 1 2 [(σr −σθ)2+(σr−σz)2+(σθ−σz)2] ε̇c z =0 → σz = 1 2 (σr+σθ) (5.2) The material creep constitutive model can be rewritten in terms of the equivalent creep strain rateand equivalent stress as ε̇c = σ0 ηM + σ0 ηK e −t τ (5.3) Magneto-thermo-mechanical creep behavior of nano-composite... 245 Equations (5.1), (5.2), and (5.3) together with differential Eq. (4.9) are used in a numerical procedure based on Mendelson’s method of successive elastic solution (Mendelson, 1968) to obtain history of stresses and deformations during creep process. The numerical procedure is explained by Loghman et al. (2011). 6. Numerical results and discussion The results presented in this study are based on the data presented in Section 2 for geometry, material properties and loading conditions. The elastic properties are dependent on the volume percent of MWCNTs content and are given in Table 1. In this research, a 4.5% volume content of MWCNTs is considered. The history of radial displacement, radial and circumferential stresses, effective stress, radial and circumferential strains, radial and circumferential creep strains are plottedwith andwithout the effect of magnetic field in Figs. 3 to 10. Figures 3a and 3b show the radial displacement histories. Generally, the radial displacements are increasing with time at a decreasing rate during life of the cylinder so that finally approach the steady state condition.However, in the presence of amagnetic field, the radial displacements are lower in magnitude. Fig. 3. Radial displacement of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) Fig. 4. Radial stress of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) 246 A. Loghman, H. Shayestemoghadam Histories of the radial stress are illustrated in Figs. 4a and 4b in which the boundary con- ditions at the inner and outer surfaces of the cylinder are satisfied. The radial stresses are decreasing with time during life of the cylinder. The radial stresses with the effect of magnetic field are of lower magnitudes. Histories 9f the circumferential stress are illustrated in Figs. 5a and 5b. The circumferential stresses are decreasing at the inner surface of the cylinder and are increasing at the outer surface so that the reference point can be identifiedwhere the circumferential stress is not changingwith time. Fig. 5. Circumferential stress of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) Histories of the effective stress are demonstrated inFigs. 6a and 6b.The effective stresses are very similar to circumferential stresses. This is because the circumferential stresses are almost ten times greater than the radial stresses and therefore are dominant. Fig. 6. Effective stress of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) Histories of the radial strain are shown inFigs. 7a and 7b.The radial strains are compressive because of highly tensile circumferential stresses. Histories of the circumferential strain are shown in Figs. 8a and 8b. The circumferential strains are positive due to highly tensile circumferential stresses. Histories of the radial creep strain are shown in Figs. 9a and 9b. It is clear that the radial creep strains at zero time are zero, however, their absolute values are increasing with time due to creep. The creep strains are negative because of highly tensile circumferential stresses. Magneto-thermo-mechanical creep behavior of nano-composite... 247 Fig. 7. Radial strain of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) Fig. 8. Circumferential strain of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) Fig. 9. Radial creep strain of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) Histories of the circumferential creep strain are shown in Figs. 10a and 10b. Due to the incompressibility condition of the material, the circumferential strains are positive. They are also increasing with time due to creep deformation. Generally, the stresses, strains anddisplacements are changingwith time at a decreasing rate during life of the cylinder so that they finally approach the steady state condition. However, in the presence of a magnetic field they are all lower in magnitude. 248 A. Loghman, H. Shayestemoghadam Fig. 10. Circumferential creep strain of the nano-composite cylinder (a) without and (b) with the effect of magnetic field (H =1E+8) 7. Conclusion Time-dependent creep stress, strain and displacement analysis of a rotating thick-walled nano-composite cylinder made of polypropylene reinforced by multi-walled carbon nanotubes (MWCNTs)subjected to magnetic, thermal and mechanical load is investigated using Burgers viscoelastic creepmodel. The results are presentedwith andwithout the effect ofmagnetic field. It has been found that the radial displacement, tangential strain and absolute values of the ra- dial strain are increasing with time at a decreasing rate so that they finally approach the steady state condition. The effective stresses are decreasing at the inner surface and increasing at the outer surface of the cylinder and approach their steady state condition after 30 years. In the presence of magnetic field stresses, the strains and radial displacement are lower inmagnitude. Acknowledgement The authors are grateful to University of Kashan for supporting this work by giving research grant. References 1. Barai P., Weng G.J., 2011, A theory of plasticity for carbon nanotube reinforced composites, International Journal of Plasticity, 27, 539-559 2. Bhatnagar N.S., KulkarniP.S., Arya V.K., 1984, Creep analysis of an internally pressurized orthotropic rotating cylinder,Nuclear Engineering and Design, 83, 379-388 3. Bhatnagar N.S., Kulkarni P.S., Arya V.K., 1986, Analysis of an orthotropic thick-walled cylinder underprimary creep conditions, International Journal of Pressure Vessels and Piping, 23, 165-185 4. GhorbanpourAraniA., LoghmanA., ShajariA.R.,AmirS., 2010,Semi-analytical solution ofmagneto-thermo-elastic stresses for functionally gradedvariable thickness rotatingdisks,Journal of Mechanical Science and Technology, 24, 2107-2117 5. Hosseini Kordkheili S.A., Naghdabadi R., 2007, Thermo-elastic analysis of a functionally graded rotating disk,Composite Structures, 79, 508-516 6. Jia Y., Peng K., Gong X.L., Zhang Z., 2011, Creep and recovery of polypropylene/carbon nanotube composites, International Journal of Plasticity, 27, 1239-1251 7. Loghman A., Ghorbanpour Arani A., Amir S., Vajedi A., 2010, Magneto-thermo-elastic creep analysis of functionally graded cylinders, International Journal of Pressure Vessels and Pi- ping, 87, 389-395 Magneto-thermo-mechanical creep behavior of nano-composite... 249 8. Loghman A., Ghorbanpour Arani A.,Shajari A.R., and Amir S., 2011, Time-dependent thermo-elastic creep analysis of rotating disk made of Al-SiC composite, Archive of Applied Me- chanics, 81, 1853-1864 9. Mendelson A., 1968,Plasticity Theory and Applications, TheMacmillan Company, NewYork 10. MoalP.L., PerreuxD., 1994,Evaluation of creep compliances of unidirectional fiber-reinforced composites,Composites Science and Technology, 51, 469-477 11. Ohno N., Ando T., Miyake T., Biwa S., 2002, A variational method for unidirectional fiber- reinforced compositeswithmatrix creep, International Journal of Solid and Structures,39, 159-174 12. Singh S.B., Ray S., 2002, Modeling the anisotropy and creep in orthotropic aluminum-silicon carbide composite rotating disc,Mechanics of Materials, 34, 363-372 13. Yang J.L., Zhang Z., Schlarb A.K., Friedrich K., 2006, On the characterization of tensile creep resistance of polyamide 66,Polymer, 47, 6745-6758 14. YouL.H.,Ou Z.Y., ZhengH., 2007,Creep deformations and stresses in thick-walled cylindrical vessels,Composite Structures, 78, 285-291 Manuscript received November 17, 2014; accepted for print August 13, 2015