Integrated Composite Analyzer (ICAN/JAVA)
Particle Slicing Computation

Particle Slicing

ICAN/JAVA may be used to analyze a composite material where particles are embedded in a binder/matrix. If the particles are not sliced, the analysis for a spherical particle in an equivalent cube is used. Further information is available about computing volume ratios for a spherical particle. However, if the particles are sliced, there was no analysis that exactly fit that situation.

Because the particles are assumed to be evenly spaced, it is sufficient to consider one particle within a volume that is repeated as required.
Particle Sliced D
The size of the cube containing the particle may be computed from the diameter of the particle, dp, and the particle volume ratio, kp. The particle volume ratio may be expressed as the ratio of the volume of the particle to the volume of the cube that may be repeated.
(1.1)   kp = (pi/6) (dp)³ / D³
where D is length of the side of the cube over which the particles are evenly spaced.

Solving for D, one obtains:
(1.2)   D = dp [(pi/6) / kp](1/3)

Consequently, one need consider a horizontal slice, possibly containing a piece of the particle, only within a cube over which the particles are evenly spaced. One such slice is shown below:
Particle Sliced D
The particle slice volume ratio may be expressed as the ratio of the volume of the particle within the horizontal slice to the volume of the horizontal slice. Further information is available about computing volume ratios for a slice of a spherical particle. The analysis below uses the volumes and volume ratios for the slice.

To analyze further, the horizontal slices may be sliced in the vertical direction; for example, along the indicated lines above. The round edges are approximated by square edges, as shown below:
Particle Subsliced D

Particle Subslicing

In particle subslicing, the horizontal slices are also sliced in the vertical direction to form boxes. To perform the calculations, one must compute or approximate the amount of each material within the subslice. The amount of the particle within the subslice is the volume of the intersection of the particle and the subslice. The particle is considered to be a sphere of diameter dp. The amount of the interphase within the subslice is the intersection of the interphase and the subslice. The interphase is considered to be the space between two concentric spheres of diameter dp (particle) and larger diameter di (interphase).
Intersection of Particle + Subslice D
Let the origin of the spherical particle be denoted by O. The radius, R, is ½ dp, then ½ di , to compute the required intersections. The center point of the subslice, C, and the size of the subslice, (dX,dY,dZ), are specified. The farthest point of the subslice, F, may be computed as F = C + ½ D. The nearest point of the subslice, N, may be computed as N = C - ½ D. Where the size vector, D, is computed as: D = [ dX sign([C - O]X), dY sign([C - O]Y), dZ sign([C - O]Z) ].
  1. If the subslice is entirely inside the sphere, i.e. the sphere contains the entire subslice, there is obviously an intersection. The entire subslice is inside the sphere if the farthest point of the subslice is within a radius from the center of the sphere, i.e. ||F - O|| < R.

  2. If the subslice contains the entire sphere, i.e. the sphere is entirely inside the subslice, there is obviously an intersection.
    Intersection of Particle + Subslice D
    For each axis, let T be a unit vector defining the direction to test. The sphere is within the subslice if, for all such directions, T, |(C-O)·T| < ½ |D·T| - R.

  3. If the subslice is entirely outside the sphere, i.e. the sphere is entirely outside the subslice, there is obviously no intersection.
    Intersection of Particle + Subslice D
    For each axis, let T be unit vector defining the direction to test. The sphere is outside the subslice if, for any such direction, T, |(C-O)·T| > ½ |D·T| + R. This test determines whether the sphere is outside of the box enlarged by the radius of the sphere on each side. The "green" areas may be described as the area outside the locus of points made by the center of the sphere rolling around on the outside of the box and inside a box enlarged by the radius of the sphere on each side. The "green" areas in the figure, near each corner or edge, are left for the following analysis.

  4. If the center of the sphere is outside the subslice by an amount greater than R, the sphere does not intersect the subslice. For each axis, let T be unit vector defining the direction to test. The distance that the center of the sphere is outside the subslice in the direction specified by T, is given by:
    dT = |(C-O)·T| - ½ |D·T|
    If dT < 0, use zero for this term in the next equation, because one only needs the distance outside the corresponding face of the subslice.
    Let d = SQRT( dTx² + dTy² + dTz² ).
    There is no intersection when d > R.

  5. Otherwise, there is an intersection.

The above analysis may be used as follows:

The volume of the intersection may be computed by integration or approximation. The problem with integration is computing the limits of integration, which occur at either the spherical interface or some edge of the subslice, and integrating differently within the different regions (difficult). Approximation allows evaluating the limits along a single line, computing the length of the line within the intersection, and using the weighted sum of many such lengths to represent the integral (easy).

To approximate the integral, sum the vertical lengths over all x (1) and y (2) values. The z location(s) where the vertical line at (x,y) intersects the sphere may be computed as follows:
z = OZ ± SQRT[ R² - (x - OX)²  - (y - OY)² ]
If the computation gives imaginary results, there is no intersection. If real z value(s) can be computed, there will be two solutions z1 and z2, corresponding to the negative and positive signs, respectively. The length inside the sphere at (x,y) is the length of the line from (x,y,z1) to (x,y,z2). The amount of this line to use, i.e. the amount inside the subslice, is computed by limiting the (maximum) z2 value to the maximum for the slice and limiting the (minimum) z1 value to the minimum for the slice. If the resulting (z2 - z1) is positive, sum the corresponding volume: (z2 - z1) * dA.

Normal Moduli

Assuming the elements are connected in series, from the force equilibrium,
(2.1)   sigmac11 = sigmam11 = sigmai11 = sigmap11
From the geometric compatibility, the total displacement equals the sum of the displacements in the component materials. This may be written:
(2.2)   epsilonc11 Lc = epsilonm11 Lm + epsiloni11 Li + epsilonp11 Lp
where L is the distance over which the strain is applicable. Using a simplified Hooke's Law:
(2.3)   epsilon = sigma / E
Substituting, one may write:
(2.4)   Lc sigmac11 / Ec11 = Lm sigmam11 / Em11 + Li sigmai11 / Ei11 + Lp sigmap11 / Ep11
Using (2.1), one obtains:
(2.5)   Lc / Ec11 = Lm / Em11 + Li / Ei11 + Lp / Ep11
The moduli are constant for a given material type over the subslices. The lengths of each type of material may vary over the subslices. The integral of the lengths, over the volume of the slice, will be the corresponding volumes of the material type. Thus, one obtains:
(2.6)   Vc / Ec11 = Vm / Em11 + Vi / Ei11 + Vp / Ep11
Dividing by the total volume of the slice, Vc, the equation may be expressed in terms of volume ratios for the slice:
(2.7)   1 / Ec11 = km / Em11 + ki / Ei11 + kp / Ep11
For a spherical particle, there is no difference in the equations for longitudinal or transverse directions within the slice. Since the 11 and 22 directions are equivalent because of symmetry,
(2.8)   Ec22 = Ec11
The 33 direction does not work the same way because it is sliced. Also, it may not be exactly symmetric because of the MFIM and loads that change through the thickness. Consider that the regions are connected in parallel. From the force equilibrium:
(2.9)   sigmac33 Lc = sigmam33 Lm + sigmai33 Li + sigmap33 Lp
where L is the distance over which the strain is applicable. From the geometric compatibility:
(2.10)   epsilonc33 = epsilonm33 = epsiloni33 = epsilonp33
Using a simplified Hooke's Law (2.3) and substituting:
(2.11)   Ec33 Lc = Em33 Lm + Ei33 Li + Ep33 Lp
As previously noted, one may integrate the lengths to obtain volumes and divide to obtain the equation in terms of volume ratios:
(2.12)   Ec33 = km Em33 + ki Ei33 + kp Ep33

Shear Moduli

Assuming the elements are connected in series, from the force equilibrium,
(3.1)   sigmaSc = sigmaSm = sigmaSi = sigmaSp
From the geometric compatibility, the total displacement equals the sum of the displacements in the component materials. This may be written:
(3.2)   epsilonSc Lc = epsilonSm Lm + epsilonSi Li + epsilonSp Lp
where L is the distance over which the strain is applicable. Using a simplified Hooke's Law:
(3.3)   epsilonS = sigmaS / G
Substituting, one may write:
(3.4)   Lc sigmac / Gc = Lm sigmam / Gm + Li sigmai / Gi + Lp sigmap / Gp
Using (3.1), one obtains:
(3.5)   Lc / Gc = Lm / Gm + Li / Gi + Lp / Gp
The moduli are constant for a given material type over the subslices. The lengths of each type of material may vary over the subslices. The integral of the lengths, over the volume of the slice, will be the corresponding volumes of the material type. Thus, one obtains:
(3.6)   Vc / Gc = Vm / Gm + Vi / Gi + Vp / Gp
Dividing by the total volume, Vc, the equation may be expressed in terms of the volume ratios for the slice:
(3.7)   1 / Gc = km / Gm + ki / Gi + kp / Gp

Poisson's Ratio

Using the rule of mixtures, Poisson's ratio may be written as:
(4.1)   nuc = km num + ki nui + kp nup

Thermal Expansion Coefficients

From the thermal expansion displacement compatibility, the total displacement equals the sum of the displacements in the component materials. This may be written:
(5.1)   alphac Lc = alpham Lm + alphai Li + alphap Lp
This is similar to the Normal moduli. The result being:
(5.2)   alphac = km alpham + ki alphai + kp alphap

Thermal Heat Conductivities

Similar to the derivation for the normal moduli, one obtains:
(6.1)   1 / Kc = km / Km + ki / Ki + kp / Kp

Heat Capacity

Heat capacities are related as follows:
(7.1)   Vc rhoc Cc = Vm rhom Cm + Vi rhoi Ci + Vp rhop Cp
Dividing by Vc and using the appropriate volume ratios, one obtains:
(7.2)   rhoc Cc = km rhom Cm + ki rhoi Ci + kp rhop Cp
The masses add as follows:
(7.3)   Vc rhoc = Vm rhom + Vi rhoi + Vp rhop
Dividing by Vc and using the appropriate volume ratios, one obtains:
(7.4)   rhoc = km rhom + ki rhoi + kp rhop
Substituting 7.4 int 7.2 and solving for Cc, one obtains:
                              km rhom Cm + ki rhoi Ci + kp rhop Cp
(7.5)   Cc = ------------------------------------------------------------
                              km rhom + ki rhoi + kp rhop

Moisture Expansion Coefficient

Similar to the derivation for the Thermal Expansion Coefficient, one obtains:
(8.1)   betac = km betam + ki betai + kp betap

Moisture Diffusivities

Similar to the derivation for the Thermal Heat Conductivities, one obtains:
(9.1)   1 / Dc = km / Dm + ki / Di + kp / Dp


Return to ICAN/JAVA home page.
Responsible Official: Chris Chamis
Last updated on Mon Nov 8 11:34:28 EST 2004
by Louis Handler of the NASA Glenn Flight Software Engineering branch.
Please send notice of any problems to: Louis.M.Handler@nasa.gov

(NASA Privacy Policy and Important Notices)