Perturbation Methods for a Continuum Metamodel

  • Chapter
  • First Online:
Perturbation Methods and Nonlinear Phenomena

Abstract

Perturbation methods, already used in the first part of the book to analyze the sample system, are here applied to general structures. The metamodel introduced in Sect. 5.2 is considered, for which several nonlinear problems, of different nature, are investigated. The static perturbation method is used: (i) in elastostatics (in which the stiffness operator is non-singular), and (ii) in buckling (in which the stiffness operator is instead singular). The Multiple Scale Method is employed to tackle dynamic problems, i.e.: (i) the external resonance phenomena (primary, sub-harmonic and super-harmonic); (ii) the parametric excitation of conservative systems with periodically time-variant properties; (iii) the dynamic bifurcation of nonconservative systems. Here, only the bifurcation equations are derived, since their study requires performing the same steps illustrated in Chaps. 24.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Chapter
USD 29.95
Price excludes VAT (Canada)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 34.99
Price excludes VAT (Canada)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Hardcover Book
USD 44.99
Price excludes VAT (Canada)
  • Durable hardcover edition
  • Dispatched in 3 to 5 business days
  • Free ship** worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Notes

  1. 1.

    The multiplier \(\alpha \) itself could be used as perturbation parameter, but the procedure illustrated here is preferred, since systematic.

  2. 2.

    It has been exploited, that \(\boldsymbol{n}_{2}\left( \textbf{u}+\textbf{v},\textbf{u}+\textbf{v}\right) =\boldsymbol{n}_{2}\left( \textbf{u},\textbf{u}\right) +2\boldsymbol{n}_{2}\left( \textbf{u},\textbf{v}\right) +\boldsymbol{n}_{2}\left( \textbf{v},\textbf{v}\right) \), where \(\boldsymbol{n}_{2}\left( \textbf{u},\textbf{v}\right) =\boldsymbol{n}_{2}\left( \textbf{v},\textbf{u}\right) \) (see also Note 3 in the Appendix B). As a scalar example, if \(n_{2}(u,u):=uu'\), it follows that \(n_{2}(u+v,u+v)=(u+v)(u+v)'=uu'+(u'v+uv')+vv'\), from which:

    $$ n_{2}(u,v)=\frac{1}{2}(u'v+uv'). $$

    Similarly, for cubic operators, it is \(\boldsymbol{n}_{3}\left( \textbf{u}+\textbf{v},\textbf{u}+\textbf{v},\textbf{u}+\textbf{v}\right) =\boldsymbol{n}_{3}\left( \textbf{u},\textbf{u},\textbf{u}\right) +3\boldsymbol{n}_{3}\left( \textbf{u},\textbf{u},\textbf{v}\right) +3\boldsymbol{n}_{3}\left( \textbf{u},\textbf{v},\textbf{v}\right) +\boldsymbol{n}_{3}\left( \textbf{v},\textbf{v},\textbf{v}\right) \). For example, if \(n_{3}(u,u,u)=:uu'u''\), by the same reasoning, it follows:

    $$ \begin{aligned}n_{3}(u,u,v) & =\frac{1}{3}(uu'v''+uu''v'+u'u''v),\\ n_{3}(u,v,v) & =\frac{1}{3}(uv'v''+u'vv''+u''vv'). \end{aligned} $$

    .

  3. 3.

    If this is not the case, a change of variable \(\textbf{u}=\textbf{u}^{f}\left( \textbf{s};\mu \right) +\textbf{v}\left( \textbf{s}\right) \), with \(\textbf{u}^{f}\) the known non-trivial path, permits to reduce the problem to the simpler case, although the dependence on \(\mu \) of the total stiffness matrix \(\boldsymbol{\mathcal {K}}\) is no longer linear (see, e.g., [10]).

  4. 4.

    The vector \(\boldsymbol{n}_{2}\) has been expanded with respect to the load as \(\boldsymbol{n}_{2}\left( \textbf{u},\textbf{u};\mu \right) =\boldsymbol{n}_{2}\left( \textbf{u},\textbf{u};\mu _{0}\right) +\epsilon \mu _{1}\boldsymbol{n}_{2,\mu }\left( \textbf{u},\textbf{u};\mu _{0}\right) +\cdots \), where \(\boldsymbol{n}_{2,\mu }:=\frac{\partial \boldsymbol{n}_{2}}{\partial \mu }\).

  5. 5.

    It is assumed here that the operator possesses a one-dimensional kernel.

  6. 6.

    Usually, the interest is on \(\mu >0\), for example, on compression loads acting on the Euler beam.

  7. 7.

    The structure, therefore, behaves as a single degree of freedom system, but its configurations does not span a subspace (i.e., the kernel of the stiffness operator), but rather a manifold, having the same dimension and tangent to it. The analogy with the Center Manifold Theory [4, 5, 18, 19] (mentioned in the Remark 4.3), relevant to dynamical systems, should be noticed.

  8. 8.

    An operator \(\boldsymbol{\mathcal {L}}\), with its boundary conditions \(\boldsymbol{\mathcal {B}}\), is self-adjoint when, for any vectors \(\boldsymbol{\phi }\), \(\boldsymbol{\psi }\) defined in the function space on which \(\boldsymbol{\mathcal {L}}\) and \(\boldsymbol{\mathcal {B}}\) operate, the following Extended Green Identity holds:

    $$ \intop _{\Omega }\!\!\boldsymbol{\psi }^{T}\boldsymbol{\mathcal {L}}\boldsymbol{\phi }\,\textrm{d}\Omega +\intop _{\Gamma }\!\!\boldsymbol{\psi }^{T}\boldsymbol{\mathcal {B}}\boldsymbol{\phi }\,\textrm{d}\Gamma =\intop _{\Omega }\!\!\boldsymbol{\phi }^{T}\boldsymbol{\mathcal {L}}\boldsymbol{\psi }\,\textrm{d}\Omega +\intop _{\Gamma }\!\!\boldsymbol{\phi }^{T}\boldsymbol{\mathcal {B}}\boldsymbol{\psi }\,\textrm{d}\Gamma . $$

    .

  9. 9.

    Namely, for any hk:

    $$ \intop _{\Omega }\!\!\boldsymbol{\phi }_{h}^{T}\boldsymbol{\mathcal {K}}\boldsymbol{\phi }_{k}\,\textrm{d}\Omega +\intop _{\Gamma }\!\!\boldsymbol{\phi }_{h}^{T}\boldsymbol{\mathcal {B}}\boldsymbol{\phi }_{k}\,\textrm{d}\Gamma =\delta _{hk}\omega _{k}^{2}, $$

    where \(\delta _{hk}\) is the Kronecker operator, and the (arbitrary) normalization \(\intop _{\varOmega }\boldsymbol{\phi }_{k}^{T}\boldsymbol{\mathcal {M}}\boldsymbol{\phi }_{k}\,\textrm{d}\Omega =1\) has been used.

  10. 10.

    Rescaling is performed in such a way the load and the dam** forces are made of the same order of \(\left| \textbf{u}\right| ^{3}\), for the reasons which will appear clear soon.

  11. 11.

    Internal resonances [11] are excluded, i.e., the natural frequencies are assumed to be incommensurable.

  12. 12.

    If \(\textbf{x}=\textbf{a}+\bar{\textbf{a}},\textbf{y}=\textbf{b}+\bar{\textbf{b}}\), then \(\boldsymbol{n}_{2}\left( \textbf{x},\textbf{x}\right) =\boldsymbol{n}_{2}\left( \textbf{a},\textbf{a}\right) +2\boldsymbol{n}_{2}\left( \textbf{a},\bar{\textbf{a}}\right) +\boldsymbol{n}_{2}\left( \bar{\textbf{a}},\bar{\textbf{a}}\right) =\boldsymbol{n}_{2}\left( \textbf{a},\textbf{a}\right) +\boldsymbol{n}_{2}\left( \textbf{a},\bar{\textbf{a}}\right) +\mathrm {c.c.}\), the overbar indicating complex conjugate. Similarly, \(\boldsymbol{n}_{2}\left( \textbf{x,y}\right) =\boldsymbol{n}_{2}\left( \textbf{a,b}\right) +\boldsymbol{n}_{2}\left( \textbf{a},\bar{\textbf{b}}\right) +\mathrm {c.c.}\) and \(\boldsymbol{n}_{3}\left( \textbf{x,x},\textbf{x}\right) =\boldsymbol{n}_{3}\left( \textbf{a,a},\textbf{a}\right) +3\boldsymbol{n}_{3}\left( \textbf{a,a},\bar{\textbf{a}}\right) +\mathrm {c.c.}\), to be used later.

  13. 13.

    Indeed, considered the problem:

    $$ \begin{aligned}\boldsymbol{\mathcal {M}}\ddot{\textbf{u}}+\boldsymbol{\mathcal {K}}\textbf{u} & =\textbf{p}e^{i\omega _{r}t},\qquad \textrm{in}\,\Omega ,\\ \boldsymbol{\mathcal {B}}\textbf{u} & =\textbf{q}e^{i\omega _{r}t},\qquad \textrm{on}\,\Gamma , \end{aligned} $$

    and taken the solution as \(\textbf{u}=\hat{\textbf{u}}e^{i\omega _{r}t}\), a space-differential problem follows:

    $$ \begin{aligned}\left( \boldsymbol{\mathcal {K}}-\omega _{r}^{2}\boldsymbol{\mathcal {M}}\right) \hat{\textbf{u}} & =\textbf{p},\\ \boldsymbol{\mathcal {B}}\hat{\textbf{u}} & =\textbf{q}. \end{aligned} $$

    Since the operator \(\boldsymbol{\mathcal {K}}-\omega _{r}^{2}\boldsymbol{\mathcal {M}}\) is singular, whose kernel is spanned by \(\boldsymbol{\phi }_{r}\), the known terms \(\textbf{p},\textbf{q}\) must be orthogonal to \(\boldsymbol{\phi }_{r}\) (i.e., the forces must spend zero virtual work on the ‘floppy’ mode admitted by the operator), namely:

    $$ \intop _{\Omega }\!\!\boldsymbol{\phi }_{r}^{T}\textbf{p}\,\textrm{d}\Omega +\intop _{\Gamma }\!\!\boldsymbol{\phi }_{r}^{T}\textbf{q}\,\textrm{d}\Gamma =0. $$

    .

  14. 14.

    The motion, therefore, takes place on a family of two dimensional manifolds, parametrized by \(\alpha \) and \(\sigma _{r}\).

  15. 15.

    This rescaling is useful for symmetric systems, to make \(\left| \textbf{u}\right| ^{3}=O\left( \epsilon \left| \textbf{u}\right| \right) \).

  16. 16.

    The adjective ‘principal’ refers to the fact that the resonance is the first one encountered in the perturbation scheme, and therefore more important, since of lowest order. ‘Secondary’ resonances would appear to higher orders, not investigated here (see [11]).

  17. 17.

    The motion, therefore, develops on a family of two dimensional manifolds, parametrized by \(\mu \) and \(\sigma _{r}\).

  18. 18.

    They are typical of the aeroelastic forces [10], not addressed in this book.

  19. 19.

    They have no qualitative influence on dynamic bifurcations [10].

  20. 20.

    This operation is called splitting of the parameter, which is something different from expansion (which contains an infinite number of terms).

  21. 21.

    Terms as \(\boldsymbol{n}_{2,\mu }\left( \textbf{u},\textbf{u};\mu _{0}\right) \), mentioned in the Note 4 of this chapter by dealing with buckling, are here of higher order, since the neighborhood of the bifurcation point explored is now smaller, of order \(\epsilon ^{2}\).

  22. 22.

    Equation 6.67 is formally identical to Eq. 6.29, relevant to the external resonance of a conservative system, but now the mode \(\boldsymbol{\phi }_{c}\) is complex, entailing:

    $$ \textbf{u}_{0}=a\left( t_{2}\right) \left\{ \mathrm {Re\left( \boldsymbol{\phi }_{c}\right) }\cos \left[ \omega _{c}t_{0}+\varphi \left( t_{2}\right) \right] -\mathrm {Im\left( \boldsymbol{\phi }_{c}\right) }\sin \left[ \omega _{c}t_{0}+\varphi \left( t_{2}\right) \right] \right\} , $$

    as \(A(t_{2})=\frac{1}{2}a(t_{2})e^{i\varphi (t_{2})}\).

  23. 23.

    The left eigenvalue satisfies the adjoint of Eqs. 6.66, i.e.:

    $$ \begin{aligned}\left( \boldsymbol{\mathcal {K}}_{e}+\mu _{c}\boldsymbol{\mathcal {K}}_{g}^{*}-i\omega _{c}\mathcal {\boldsymbol{\mathcal {C}}}_{s}-\omega _{c}^{2}\boldsymbol{\mathcal {M}}\right) \boldsymbol{\psi }_{c} & =\boldsymbol{0},\\ \left( \boldsymbol{\mathcal {B}}_{s}+\mu _{0}\boldsymbol{\mathcal {B}}_{g}^{*}-i\omega _{c}\boldsymbol{\mathcal {D}}_{s}\right) \boldsymbol{\psi }_{c} & =\textbf{0}, \end{aligned} $$

    where \(\boldsymbol{\mathcal {K}}_{g}^{*}\) is the adjoint stiffness operator and \(\boldsymbol{\mathcal {B}}_{g}^{*}\) the adjoint boundary conditions, which are drawn from the Extended Green Identity:

    $$ \intop _{\Omega }\!\!\bar{\boldsymbol{\psi }}^{T}\boldsymbol{\mathcal {L}}\boldsymbol{\phi }\,\textrm{d}\Omega +\intop _{\Gamma }\!\!\bar{\boldsymbol{\psi }}^{T}\boldsymbol{\mathcal {B}}\boldsymbol{\phi }\,\textrm{d}\Gamma =\intop _{\Omega }\!\!\bar{\boldsymbol{\phi }}^{T}\mathcal {L^{*}}\boldsymbol{\psi }\,\textrm{d}\Omega +\intop _{\Gamma }\!\!\bar{\boldsymbol{\phi }}^{T}\boldsymbol{\mathcal {B}}^{*}\boldsymbol{\psi }\,\textrm{d}\Gamma . $$

    From the same identity, the compatibility condition is derived (see the Appendix A, Sect. A.4).

  24. 24.

    The motion, therefore, takes place on an (invariant) two dimensional manifold, according to the Center Manifold Theory.

References

  1. Andrianov, I., Awrejcewicz, J., Danishevs’kyy, V., Ivankov, A.: Asymptotic Methods in the Theory of Plates with Mixed Boundary Conditions. Wiley, Chichester (2014)

    Google Scholar 

  2. Bolotin, V.V.: Nonconservative Problems of the Theory of Elastic Stability. Macmillan, London (1963)

    Google Scholar 

  3. Bolotin, V.V.: The Dynamic Stability of Elastic Systems. Holden Day, San Francisco (1964)

    Google Scholar 

  4. Carr, J.: Applications of Centre Manifold Theory. Springer, New York (1981)

    Google Scholar 

  5. Guckenheimer, J., Holmes, P.: Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields. Springer, New York (1983)

    Google Scholar 

  6. Hinch, E.J.: Perturbation Methods. Cambridge University Press, Cambridge (1991)

    Google Scholar 

  7. Holmes, M.: Introduction to Perturbation Methods. Springer, New York (2015)

    Google Scholar 

  8. Kevorkian, J., Cole, J.D.: Multiple Scale and Singular Perturbation Methods. Springer, New York (2011)

    Google Scholar 

  9. Leipholz, H.H.: Stability of Elastic Systems. Sijthoff & Noordhoff, Alphen aan den Rijn (1980)

    Google Scholar 

  10. Luongo, A., Ferretti, M., Di Nino, S.: Stability and Bifurcation of Structures: Statical and Dynamical Systems. Springer, Cham (2023)

    Google Scholar 

  11. Nayfeh, A.H., Mook, D.T.: Nonlinear Oscillations. Wiley, New York (1995)

    Google Scholar 

  12. Nayfeh, A.H.: Perturbation Methods. Wiley, New York (1973)

    Google Scholar 

  13. Pignataro, M., Rizzi, N., Luongo, A.: Stability, Bifurcation and Postcritical Behaviour of Elastic Structures. Elsevier, Amsterdam (1990)

    Google Scholar 

  14. Rand, R.: Perturbation Methods. Bifurcation Theory and Computer Algebra. Springer, New York (1987)

    Google Scholar 

  15. Seyranian, A.P., Mailybaev, A.A.: Multiparameter Stability Theory with Mechanical Applications. World Scientific, Singapore (2003)

    Google Scholar 

  16. Thompson, J.M.T., Hunt, G.W.: A General Theory of Elastic Stability. Wiley, London (1973)

    Google Scholar 

  17. Thompson, J.M.T., Hunt, G.W.: Elastic Instability Phenomena. Wiley, Chichester (1984)

    Google Scholar 

  18. Troger, H., Steindl, A.: Nonlinear Stability and Bifurcation Theory: An Introduction for Engineers and Applied Scientists. Springer, Wien (1991)

    Google Scholar 

  19. Wiggins, S.: Introduction to Applied Nonlinear Dynamical Systems and Chaos. Springer, New York (1990)

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Angelo Luongo .

Rights and permissions

Reprints and permissions

Copyright information

© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Check for updates. Verify currency and authenticity via CrossMark

Cite this chapter

Luongo, A., Zulli, D., Ferretti, M., D’Annibale, F. (2024). Perturbation Methods for a Continuum Metamodel. In: Perturbation Methods and Nonlinear Phenomena. Synthesis Lectures on Engineering, Science, and Technology. Springer, Cham. https://doi.org/10.1007/978-3-031-49397-3_6

Download citation

  • DOI: https://doi.org/10.1007/978-3-031-49397-3_6

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-031-49396-6

  • Online ISBN: 978-3-031-49397-3

  • eBook Packages: Synthesis Collection of Technology (R0)

Publish with us

Policies and ethics

Navigation