Jtam.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 41, 3, pp. 537-544, Warsaw 2003 NON-LINEAR STABILITY PROBLEM OF SPHERICAL SHELL LOADED WITH TORQUE Stefan Joniak Institute of Applied Mechanics, Poznań University of Technology e-mail: Stefan.Joniak@put.poznan.pl A thin-walled spherical shell is pivoted at both edges. One of the edges may rotate around the shell axis. Moreover, it is loaded with a torque. Theproblemof shell stability is considered.The systemof equations cha- racterizing the problem consists of a non-linear equation of equilibrium and non-linear compatibility equation. Both equations are solved with Bubnov-Galerkin’s method, assuming beforehand the form of deflection and force-functions. As a result of the solution, an algebraic equation is obtained, with respect to a dimensionless load parameter. The critical loadparameter corresponding to theminimal critical load value is deter- mined from this equation. The number m at which the load parameter has theminimumvalue determines themode of stability loss. The paper is supplied with a numerical example. Key words: shells, non-linear stability 1. Introduction A thin-walled spherical shell, being a subject of the analysis, is shown in Figure 1. Its bottom edge is fixed and pivoted. The upper edge is also pivoted, butmay rotate around the vertical axis of the shell. The upper edge is loaded with a rotationalmoment.Theproblemof the loss of stability is considered. In order to solve the problem, non-linear stability equations are applied, defined in themonographbyMushtari andGalimov (1957). The systemof equations is solvedwithmakinguseofBubnov-Galerkin’smethodwithdeflection functions and force functions assumed in advance. The final goal is the determination of the critical load. As the problem is very complex, only obtaining a numerical solution is possible. 538 S.Joniak Fig. 1. 2. Equations of the stability problem The system of stability equations takes the following form ∇ 2 ∇ 2Ψ =Eh(κ212 −κ11κ22−κ11k22−κ22k11) (2.1) D∇2∇2w+2Sκ12+T1(k11+κ11)+2Sκ12+T2(k22+κ22)= 0 where Ψ – force-function w – deflection function upon the loss stability kii – main curvatures κii,κ12 – curvature variations and surface torsion D – plate stiffness, D=Eh3/[12(1−ν2)] S – tangent force in pre-critical state Ti,S – critical state forces, and ∇ 2 = 1 R2 ( cotθ ∂ ∂θ + ∂2 ∂θ2 + 1 sin2θ ∂2 ∂ϕ2 ) Changes in the curvatures and torsions of the spherical shell surface are the following functions of the deflection Non-linear stability problem of spherical shell... 539 κ11 =− 1 R2 ∂2w ∂θ2 κ22 =− 1 R2 ( 1 sin2θ ∂2w ∂ϕ2 +cotθ ∂w ∂θ ) (2.2) κ12 = 1 R2 sinθ ( cotθ ∂w ∂ϕ − ∂2w ∂ϕ∂θ ) The forces corresponding to the critical state depend on the force functions as in the following T1 = 1 R2 ( 1 sin2θ ∂2Ψ ∂ϕ2 +cotθ ∂Ψ ∂θ ) T2 = 1 R2 ∂2Ψ ∂θ2 (2.3) S= 1 R2 sinθ ( cotθ ∂Ψ ∂ϕ − ∂2Ψ ∂ϕ∂θ ) On the other hand, the tangent force of the pre-critical state takes the form, see Łukasiewicz (1976) S= M0 2πR2 sin2θ (2.4) Changes in curvatures (2.2) and cross-section forces (2.3), (2.4) should be introduced into equations (2.1). This gives a system of nonlinear partial differential equations with respect to w and Ψ. 3. Solution to the stability equations The stability equations are solved with the help of Bubnov-Galerkin’s me- thod. This causes the need to adopt a form of the deflection function and the force-functions that possibly satisfies all the boundary conditions of the problem. The edges of the shell are characterized by the following boundary conditions θ= θ1 w=0 Mθ =0 S= M0 2πR2 sin2θ1 T1 =0 (3.1) and θ=π−θ1 w=0 Mθ =0 S= M0 2πR2 sin2θ1 T1 =0 (3.2) 540 S.Joniak The force functions and deflection functions are assumed in the form Ψ = [bϕ+csin(mϕ)]sin2θ (3.3) w= asin π(θ−θ1) π−2θ1 sin [π(θ−θ1) π−2θ1 +mϕ ] sin2θ where a,b,c are constants and m – an integer number. Deflection function (3.3)2 meets accurately the first of conditions (3.1), while the second one is met in an integral sense. The force function satisfies the third conditions at both edges, with the accuracy to a constant, while the condition for force T1 remains unsatisfied. Inseparability equation (2.1)1 is solved with Bubnov-Galerkin’s method. The function subject to the orthogonalization has of the form F(θ,ϕ)=∇2∇2Ψ−Eh(κ212−κ11κ22−κ11k22−κ22k11) while the orthogonalization conditions are ∫ A F(θ,ϕ)fi(θ,ϕ) dA=0 (3.4) where fi(θ,ϕ) – orthogonalization factors, i.e. the components of the Ψ function A – middle surface of the shell. The orthogonalization conditions are as follows π−θ1 ∫ θ1 2π ∫ 0 F(θ,ϕ)ϕsin3θ dθdϕ=0 (3.5) π−θ1 ∫ θ1 2π ∫ 0 F(θ,ϕ ) sin(mϕ)sin3θ dθdϕ=0 An expansion of conditions (3.5) provides two algebraic equations, serving as a basis for the determination of factors of the Ψ force function. They are presented by the expressions b=EhH1a 2 c=EhH2a 2 (3.6) where: H1,H2 are the constants including θ1 and the number m. Non-linear stability problem of spherical shell... 541 Finally, the solution to equation (2.1)1 takes the form Ψ =Eha2[H1ϕ+H2 sin(mϕ)]sin 2θ (3.7) In order to apply theBubnov-Galerkinmethod to equilibriu equation (2.1)2 its left-hand side should be considered as the function subject to orthogonaliza- tion, with the orthogonalization condition duly defined. The function subject to the orthogonalization is G(θ,ϕ)=D∇2∇2w+2Sκ12+T1(k11+κ11)+2Sκ12+T2 ( k22+κ22) while the orthogonalization condition has the form ∫ A G(θ,ϕ)g(θ,ϕ) dA=0 with g(θ,ϕ) being the orthogonalization factor corresponding to the right- hand side of deflection function (3.3)2. Finally, the orthogonalization condition takes the following form π−θ1 ∫ θ1 2π ∫ 0 G(θ,ϕ)sin π(θ−θ1) π−2θ1 sin [π(θ−θ1) π−2θ1 +mϕ]sin3θ dθdϕ=0 (3.8) The mathematical program Derive was used to solve orthogonalization con- ditions (3.5) and (3.8). Differentiation and integration procedures, and also procedures for transforming algebraical expressions from Derive were used. Finally, an algebraical equation was obtained describing the torque M0. The result of the solution to this equation was an expression for the dimensionless torque M = M0 Eh3 = 1 12(1−ν2) C1+C2 (a h )2 (3.9) where Ci are constants depending on the angle θ1 and the number m. Their form is very complex. Equation (3.9)was thenconverted to the formwithadimensionless tangent stress t= τ0 E = M 2π sin2θ1 (h R )2 = ( h R )2 2π sin2θ1 [ 1 12(1−ν2) C1+C2 (a h )2] (3.10) where τ0 = M0 2πR2hsin2θ1 542 S.Joniak The dimensionless stress written by equation (3.13) depends on the num- ber m describing the form of the loss of stability. The equation enables one to find the minimal value tmin with respect to m. The minimal value of the stress is equal to the critical stress tcr, while the number m determining the stress is equal to mcr. 4. Example of calculation The search for the critical load is practically only possible with a nume- rical method. For this purpose the Ci constants must be calculated for a pre-determined value of the θ1 angle and some numbers m. The constants Ci were calculated by the procedures included in the Derive. Table 1 presents examples of the constants for θ1 =π/12 and θ1 =π/4. Table 1 θ1 =π/12 θ1 =π/4 M C1 C2 C1 C2 1 64.1419 0.4631 226.1292 0.1896 2 62.2741 3.7563 154.5125 1.9270 3 113.4468 12.4487 170.8259 6.1066 4 221.0470 29.2795 233.6061 14.0478 The mathematical program Derive for Windows was applied to draw the diagrams shown in Fig.2. Figure 2 shows the plots against the co-ordinates t−a/h for θ1 = π/12, h/R = 0.005 and ν = 0.3, with various numbers m. As shown, the minimal value of t corresponds to m=2. Table 2 θ1 mcr t(R/h) 2 π/12 2 13.5497 π/10 2 10.0269 π/8 2 7.4554 π/7 2 5.0503 π/5 2 4.4628 π/4 2 4.5260 Non-linear stability problem of spherical shell... 543 Fig. 2. Table 2 specifies the dimensionless critical stresses of the shells for some angles θ1. It is evident that for any θ1 the number determining the formof the stability loss is equal to m=mcr =2. It should be noticed that for θ1 =π/6, θ1 =3π/10, and θ1 =5π/14 the problem remains unsolved. This is certainly a result of the assumed forms of the deflection function and the force-function. References 1. Avdonin A.S., 1969, Prikladnye metody razchyota obolochek i tonkostennykh konstrukčıi, Izd. ”Mashynostroenie”,Moskva 2. Łukasiewicz S., 1976, Obciążenia skupione w płytach, tarczach i powłokach, PWN,Warszawa 3. MushtariG.M.,GalimovK.Z., 1957,Neliněınaya teoria uprugikh obolochek, Tatknigizdat, Kazań 4. Nowacki W., 1979,Dźwigary powierzchniowe, PWN,Warszawa 5. Volmir A.S., 1967, Ustojchivost deformiruemykh system, Izd. ”Nauka”, Mo- skva 544 S.Joniak Nieliniowe zagadnienie stateczności powłoki kulistej obciążonej momentem obrotowym Streszczenie Cienkościenna powłoka kulista jest podparta przegubowo na obu brzegach. Je- den z brzegów ma możliwość obrotu wokół osi powłoki; do tego brzegu przyłożony jest moment obrotowy. Rozpatruje się zagadnienie stateczności powłoki. Układ rów- nań zagadnienia tworzą nieliniowe równanie równowagi oraz nieliniowe równanie nie- rozdzielności. Oba równania rozwiązuje się metodą Bubnowa-Galerkina, przyjmując uprzednio postać funkcji ugięcia i funkcji sił. Efektem rozwiązania jest równanie al- gebraiczne na bezwymiarowyparametr obciążenia. Z tego równaniawyznacza się pa- rametr obciążenia krytycznego, odpowiadającyminimalnej jego wartości. Liczba m, przy której parametr obciążenia osiąga minimum wyznacza postać utraty stateczno- ści. Praca kończy się przykładem liczbowym. Manuscript received December 2, 2002; accepted for print April 2, 2003