Contents

The Generalized Self-Consistent Scheme (GSCS) predicts the transverse shear modulus $G_{23}$ of a fiber composite by picturing one fiber, wrapped in a matrix shell, sitting inside an “effective medium” that stands in for all the other fibers around it.

Why it beats the alternatives:

  • Mori-Tanaka puts each fiber in pure matrix and ignores its neighbors. At high fiber volume fractions this underpredicts $G_{23}$ — it actually coincides with the Hashin-Shtrikman lower bound.
  • Self-Consistent drops the matrix shell entirely and embeds the fiber straight into the effective medium, which overpredicts stiffness.
  • GSCS keeps the matrix shell and the surrounding effective medium, so it captures the squeeze each matrix pocket feels between neighboring fibers. The result lands between the two extremes and matches experiments and FEA much more closely, especially above $v_f \approx 0.4$.

See also: Eshelby’s inclusion: interactive 3D  ·  Analytical bounds and homogenization models  ·  Tensor visualization (rank 2, 3, 4)  ·  FFT homogenization

GSCS Homogenization

Enter matrix and fiber properties and the fiber volume fraction. The GSCS computes $G_{23}$ from the Christensen-Lo (1979) quadratic, combined with CCA exact results for $E_1, \nu_{12}, k_t, G_{12}$. The 3D views show the deformed three-phase cylinder under transverse shear $\sigma_{23}$, colored by von Mises stress and equivalent strain.

Matrix
Fiber
Microstructure
Presets

Three-phase model: fiber (blue), matrix shell (green), effective medium (gray). Drag to orbit.

Von Mises stress under σ₂₃ shear. Drag to orbit.

Equivalent strain (ε̄) under σ₂₃ shear. Drag to orbit.

Theory

The three-phase model

The GSCS (Christensen & Lo 1979 [1]) models a unidirectional composite as three concentric phases:

  1. Fiber (radius $a$, moduli $E_f, \nu_f$)
  2. Matrix shell (outer radius $b$, moduli $E_m, \nu_m$), with $v_f = (a/b)^2$
  3. Effective medium (extending to infinity, unknown moduli $K^, G^$)

The effective medium replaces all surrounding fibers and matrix with a single homogeneous material. It provides the far-field constraint that, in a real composite, comes from neighboring fibers. The matrix shell preserves the correct fiber-to-matrix ratio, so the local stress transfer at the interface is captured exactly. The effective shear modulus $G^* = G_{23}$ is found by requiring that inserting or removing the coated cylinder leaves the far-field stress unchanged (self-consistency). This condition yields a quadratic in $G_{23}/G_m$.

CCA exact results

Four of the five transversely isotropic constants follow exactly from the Composite Cylinder Assemblage (CCA) model (Hashin 1979 [4]), independently of the GSCS quadratic.

Plane-strain bulk modulus:

\[k_t = k_m^{\rm ps} + \frac{v_f}{\dfrac{1}{k_f^{\rm ps} - k_m^{\rm ps}} + \dfrac{1-v_f}{k_m^{\rm ps} + G_m}} \tag{2}\]

where $k^{\rm ps} = K + G/3$ is the plane-strain (2D) bulk modulus.

Axial shear modulus:

\[G_{12} = G_m\,\frac{G_f(1+v_f) + G_m(1-v_f)}{G_f(1-v_f) + G_m(1+v_f)} \tag{3}\]

Longitudinal modulus (with Poisson coupling correction):

\[E_1 = E_f\,v_f + E_m(1-v_f) + \frac{4(\nu_f - \nu_m)^2\,v_f(1-v_f)}{\dfrac{1-v_f}{k_f^{\rm ps}} + \dfrac{v_f}{k_m^{\rm ps}} + \dfrac{1}{G_m}} \tag{4}\]

Axial Poisson ratio:

\[\nu_{12} = \nu_f\,v_f + \nu_m(1-v_f) + \frac{(\nu_f - \nu_m)\!\left(\dfrac{1}{k_m^{\rm ps}} - \dfrac{1}{k_f^{\rm ps}}\right)\!v_f(1-v_f)}{\dfrac{1-v_f}{k_f^{\rm ps}} + \dfrac{v_f}{k_m^{\rm ps}} + \dfrac{1}{G_m}} \tag{5}\]

The GSCS quadratic for $G_{23}$

Define the modulus ratio and plane-strain parameters:

\[g_r = \frac{G_f}{G_m}, \qquad \eta_f = 3 - 4\nu_f, \qquad \eta_m = 3 - 4\nu_m \tag{6}\]

The transverse shear modulus satisfies (Christensen & Lo 1979, Eq. 4.11 [1]):

\[A\left(\frac{G_{23}}{G_m}\right)^2 + B\left(\frac{G_{23}}{G_m}\right) + D = 0 \tag{7}\]

with coefficients (Eqs. 4.12-4.14, corrected in [2]):

\[\begin{aligned} D &= 3c(1-c)^2(g_r-1)(g_r+\eta_f) \\ &\quad + \bigl(g_r\eta_m + (g_r-1)c + 1\bigr)\bigl(g_r + \eta_f + (g_r\eta_m - \eta_f)\,c^3\bigr) \end{aligned} \tag{8}\] \[\begin{aligned} B &= -6c(1-c)^2(g_r-1)(g_r+\eta_f) \\ &\quad + \bigl(g_r\eta_m + (g_r-1)c + 1\bigr)\bigl((\eta_m-1)(g_r+\eta_f) - 2c^3(g_r\eta_m - \eta_f)\bigr) \\ &\quad + (\eta_m+1)\,c\,(g_r-1)\bigl(g_r + \eta_f + (g_r\eta_m - \eta_f)\,c^3\bigr) \end{aligned} \tag{9}\] \[\begin{aligned} A &= 3c(1-c)^2(g_r-1)(g_r+\eta_f) \\ &\quad + \bigl(g_r\eta_m + \eta_f\eta_m - (g_r\eta_m - \eta_f)\,c^3\bigr)\bigl(\eta_m\,c\,(g_r-1) - (g_r\eta_m + 1)\bigr) \end{aligned} \tag{10}\]

where $c = v_f$. The physical root is the minimum positive solution of the quadratic.

Derived transverse constants

Once $G_{23}$ is known, the remaining transverse constants follow from TI compliance relations:

\[\frac{1}{E_2} = \frac{1}{4k_t} + \frac{1}{4G_{23}} + \frac{\nu_{12}^2}{E_1} \tag{11}\]

and the transverse Poisson ratio:

\[\nu_{23} = \frac{k_t - G_{23}}{k_t + G_{23}} \tag{12}\]

Why GSCS improves on Mori-Tanaka

Mori-Tanaka embeds each fiber in infinite matrix, ignoring fiber-fiber proximity. For $G_{23}$ this coincides exactly with the Hashin-Shtrikman lower bound. The GSCS constrains the matrix shell on its outer boundary by the effective medium, which is stiffer than pure matrix. This confinement effect reflects the reality that each matrix pocket is squeezed between fibers. It restricts shear deformation and raises $G_{23}$:

\[G_{23}^{\rm MT} \;\leq\; G_{23}^{\rm GSCS} \;\leq\; G_{23}^{\rm SC} \tag{13}\]

Dilute limit

As $v_f \to 0$ the quadratic reduces to the dilute (Eshelby) estimate, confirming consistency:

\[\frac{G_{23}}{G_m} \approx 1 + \frac{(G_f - G_m)(1 + \eta_m)}{G_f\,\eta_m + G_m}\,v_f + O(v_f^2) \tag{14}\]

Benveniste’s result

Benveniste (2008) [3] proved that the embedding volume fraction must equal $v_f$. The proof requires that average strains in the fiber core and matrix shell each match their respective phase averages, from which $c_0 = v_f$ follows necessarily. This validates the Christensen-Lo assumption without energy equivalence.

Limitations

  1. Aligned continuous fibers only. Short or randomly oriented fibers require the coated-ellipsoid BVP or orientation averaging.
  2. Isotropic phases. Anisotropic fibers (e.g., carbon) need the transversely isotropic Eshelby tensor.
  3. Linear elasticity. Nonlinear or viscoelastic behavior requires incremental or secant extensions.

References

[1] Christensen, R. M. & Lo, K. H. (1979). Solutions for effective shear properties in three phase sphere and cylinder models. Journal of the Mechanics and Physics of Solids, 27(4), 315-330.

[2] Christensen, R. M. (1986). Erratum to Christensen & Lo (1979). Journal of the Mechanics and Physics of Solids, 34(6), 639.

[3] Benveniste, Y. (2008). Revisiting the generalized self-consistent scheme in composites: Clarification of some aspects and a new formulation. Journal of the Mechanics and Physics of Solids, 56(10), 2984-3002.

[4] Hashin, Z. (1979). Analysis of properties of fiber composites with anisotropic constituents. Journal of Applied Mechanics, 46(3), 543-550.

[5] Mura, T. (1987). Micromechanics of Defects in Solids (2nd ed.). Martinus Nijhoff.

[6] Hill, R. (1965). A self-consistent mechanics of composite materials. Journal of the Mechanics and Physics of Solids, 13(4), 213-222.

[7] Mori, T. & Tanaka, K. (1973). Average stress in matrix and average elastic energy of materials with misfitting inclusions. Acta Metallurgica, 21(5), 571-574.