4-Eq-Dim-Reduction
4-Eq-Dim-Reduction
Coordinates
Manus AI
March 27, 2025
1 Introduction
This document presents the dimension reduction of four important equations in both cylindrical coordinates
(axisymmetry) and spherical coordinates (radial symmetry). The equations considered are:
1. Allen-Cahn Equation
2. Cahn-Hilliard Equation
3. Convection-Diffusion Equation
4. Navier-Stokes Equations
Dimension reduction exploits symmetry to reduce the computational complexity of partial differential
equations. By assuming axisymmetry (cylindrical symmetry) or radial symmetry (spherical symmetry), we
can reduce 3D problems to 2D or 1D problems, respectively.
For axisymmetric problems, where there is no dependence on the angular coordinate ϕ, we have ∂
∂ϕ = 0,
and the operators simplify to:
Gradient (axisymmetric):
∂f ∂f
∇f = er + ez (4)
∂r ∂z
Divergence (axisymmetric):
1 ∂ ∂vz
∇·v= (rvr ) + (5)
r ∂r ∂z
1
Laplacian (axisymmetric):
1 ∂ ∂2f
∂f
∇ f=
2
r + (6)
r ∂r ∂r ∂z 2
1 ∂ 2 1 ∂ 1 ∂vϕ
∇·v= (r vr ) + (sin θvθ ) + (8)
r2 ∂r r sin θ ∂θ r sin θ ∂ϕ
Laplacian of a scalar field f :
1 ∂ 1 1 ∂2f
∂f ∂ ∂f
∇2 f = r2 + sin θ + (9)
r2 ∂r ∂r r sin θ ∂θ
2 ∂θ r2 sin θ
2 ∂ϕ2
For radially symmetric problems, where there is no dependence on the angular coordinates θ and ϕ, we
have ∂θ
∂
= ∂ϕ
∂
= 0, and the operators simplify to:
Gradient (radially symmetric):
∂f
∇f = er (10)
∂r
Divergence (radially symmetric):
1 ∂
∇ · v = 2 (r2 vr ) (11)
r ∂r
Laplacian (radially symmetric):
1 ∂ ∂2f 2 ∂f
2 ∂f
∇ f= 2
2
r = + (12)
r ∂r ∂r ∂r2 r ∂r
1 ′
∂ϕ M ′
= −M W (ϕ) − ϵ∇2 ϕ =− W (ϕ) + M ϵ∇2 ϕ (14)
∂t ϵ ϵ
1 ∂ ∂2ϕ
∂ϕ
∇2 ϕ = r + 2 (15)
r ∂r ∂r ∂z
2
3.1.2 Step 2: Substitute the axisymmetric Laplacian into the Allen-Cahn equation
1 ∂ ∂2ϕ
∂ϕ M ∂ϕ
= − W ′ (ϕ) + M ϵ r + 2 (16)
∂t ϵ r ∂r ∂r ∂z
3.1.4 Step 4: Final form of the Allen-Cahn equation in cylindrical coordinates with axisym-
metry
∂ ϕ 1 ∂ϕ ∂ 2 ϕ
2
∂ϕ M ′
= − W (ϕ) + M ϵ + + 2 (18)
∂t ϵ ∂r2 r ∂r ∂z
This is a 2D problem in the (r, z) plane, reduced from the original 3D problem.
∂ϕ
(t, 0, z) = 0 (19)
∂r
2. At the outer boundary (r = R or r → ∞), depending on the specific problem:
• For a finite domain (r = R): Typically Dirichlet or Neumann conditions
• For an infinite domain (r → ∞): ϕ approaches a constant value
3.2.2 Step 2: Substitute the radially symmetric Laplacian into the Allen-Cahn equation
∂ ϕ 2 ∂ϕ
2
∂ϕ M
= − W ′ (ϕ) + M ϵ + (21)
∂t ϵ ∂r2 r ∂r
3.2.3 Step 3: Final form of the Allen-Cahn equation in spherical coordinates with radial
symmetry
∂ ϕ 2 ∂ϕ
2
∂ϕ M
= − W ′ (ϕ) + M ϵ + (22)
∂t ϵ ∂r2 r ∂r
This is a 1D problem in the radial coordinate r, reduced from the original 3D problem.
3
3.2.4 Step 4: Boundary conditions
For a complete description, we need appropriate boundary conditions:
1. At the center (r = 0), we need regularity conditions:
∂ϕ
(t, 0) = 0 (23)
∂r
2. At the outer boundary (r = R or r → ∞), depending on the specific problem:
• For a finite domain (r = R): Typically Dirichlet or Neumann conditions
• For an infinite domain (r → ∞): ϕ approaches a constant value
1 ′
∂ϕ M 2 ′
= M ∇2 W (ϕ) − ϵ∇2 ϕ = ∇ W (ϕ) − M ϵ∇4 ϕ (25)
∂t ϵ ϵ
1 ∂ ∂2ϕ ∂ 2 ϕ 1 ∂ϕ ∂ 2 ϕ
∂ϕ
∇2 ϕ = r + 2 = + + 2 (26)
r ∂r ∂r ∂z ∂r2 r ∂r ∂z
∂ ϕ 1 ∂ϕ ∂ 2 ϕ
2
∇ ϕ=∇
4 2
+ + 2 (27)
∂r2 r ∂r ∂z
Expanding this expression is complex, but for our purposes, we can leave it in this form.
4.1.4 Step 4: Final form of the Cahn-Hilliard equation in cylindrical coordinates with ax-
isymmetry
∂ϕ M 2 ′
= ∇ W (ϕ) − M ϵ∇4 ϕ (29)
∂t ϵ
where ∇2 and ∇4 are the axisymmetric Laplacian and biharmonic operators in cylindrical coordinates.
This is a 2D problem in the (r, z) plane, reduced from the original 3D problem.
4
4.1.5 Step 5: Boundary conditions
For a complete description, we need appropriate boundary conditions:
1. At the symmetry axis (r = 0), we need regularity conditions:
∂ϕ
(t, 0, z) = 0 (30)
∂r
∂ 1 ′
W (ϕ) − ϵ∇ ϕ (t, 0, z) = 0
2
(31)
∂r ϵ
2. At the outer boundary (r = R or r → ∞), depending on the specific problem:
• For a finite domain (r = R): Typically no-flux conditions
• For an infinite domain (r → ∞): ϕ approaches a constant value
4.2.4 Step 4: Final form of the Cahn-Hilliard equation in spherical coordinates with radial
symmetry
∂ϕ M 2 ′
= ∇ W (ϕ) − M ϵ∇4 ϕ (35)
∂t ϵ
where ∇2 and ∇4 are the radially symmetric Laplacian and biharmonic operators in spherical coordinates.
This is a 1D problem in the radial coordinate r, reduced from the original 3D problem.
5
5 Dimension Reduction of Convection-Diffusion Equation
The convection-diffusion equation is given by:
∂u
+ ∇ · (vu) − D∇2 u = q (38)
∂t
where u is the concentration or temperature, v is the velocity field, D is the diffusion coefficient, and q
is the source/sink term.
5.1.3 Step 3: Final form of the convection-diffusion equation in cylindrical coordinates with
axisymmetry
∂u 1 ∂ ∂ u 1 ∂u ∂ 2 u
2
∂
+ (rvr u) + (vz u) − D + + =q (42)
∂t r ∂r ∂z ∂r2 r ∂r ∂z 2
This is a 2D problem in the (r, z) plane, reduced from the original 3D problem.
6
5.2.2 Step 2: Substitute into the convection-diffusion equation
1 ∂ ∂ u 2 ∂u
2
∂u
+ 2 (r2 vr u) − D + =q (46)
∂t r ∂r ∂r2 r ∂r
5.2.3 Step 3: Final form of the convection-diffusion equation in spherical coordinates with
radial symmetry
1 ∂ 2 ∂ u 2 ∂u
2
∂u
+ 2 (r vr u) − D + =q (47)
∂t r ∂r ∂r2 r ∂r
This is a 1D problem in the radial coordinate r, reduced from the original 3D problem.
where v is the velocity field, ρ is the density, η is the dynamic viscosity, P is the pressure, and f is the
external force.
7
6.1.2 Step 2: Simplify the viscous terms
The viscous terms can be simplified as follows:
r-component:
∂ 1 ∂ ∂ 2 vr ∂ 2 vr 1 ∂vr ∂ 2 vr
vr vr
(rvr ) + − = + − + (54)
∂r r ∂r ∂z 2 r2 ∂r2 r ∂r r2 ∂z 2
z-component:
1 ∂ ∂ 2 vz ∂ 2 vz 1 ∂vz ∂ 2 vz
∂vz
r + 2
= 2
+ + (55)
r ∂r ∂r ∂z ∂r r ∂r ∂z 2
6.1.3 Step 3: Final form of the Navier-Stokes equations in cylindrical coordinates with ax-
isymmetry
Continuity equation:
1 ∂ ∂vz
(rvr ) + =0 (56)
r ∂r ∂z
r-component of momentum equation:
1 ∂vr
2
∂ 2 vr
∂vr ∂vr ∂vr ∂P ∂ vr vr
ρ + vr + vz =− +η + − 2+ + fr (57)
∂t ∂r ∂z ∂r ∂r2 r ∂r r ∂z 2
1 ∂vz
2
∂ 2 vz
∂vz ∂vz ∂vz ∂P ∂ vz
ρ + vr + vz =− +η + + + fz (58)
∂t ∂r ∂z ∂z ∂r2 r ∂r ∂z 2
This is a 2D problem in the (r, z) plane, reduced from the original 3D problem.
vr (t, 0, z) = 0 (59)
∂vz
(t, 0, z) = 0 (60)
∂r
2. At solid boundaries: no-slip conditions
vr = vz = 0 (61)
1 ∂ 2
∇·v = (r vr ) = 0 (62)
r2 ∂r
This implies:
∂ 2
(r vr ) = 0 (63)
∂r
8
Integrating:
r 2 vr = C (64)
where C is a constant. For a non-singular solution at r = 0, we must have C = 0, which implies vr = 0
everywhere. This means that a non-trivial incompressible radially symmetric flow is not possible.
However, for completeness, we’ll derive the momentum equation for the radial component:
1 ∂ 2vr
∂vr ∂vr ∂P ∂vr
ρ + vr =− +η 2 r2 − 2 + fr (65)
∂t ∂r ∂r r ∂r ∂r r
6.2.3 Step 3: Final form of the Navier-Stokes equations in spherical coordinates with radial
symmetry
Continuity equation:
1 ∂ 2
(r vr ) = 0 (67)
r2 ∂r
Radial momentum equation:
2 ∂vr 2vr
2
∂vr ∂vr ∂P ∂ vr
ρ + vr =− +η + − + fr (68)
∂t ∂r ∂r ∂r2 r ∂r r2
As noted earlier, the continuity equation implies vr = 0 for a non-singular solution, which makes the
momentum equation trivial. This illustrates a fundamental limitation: a non-trivial incompressible radially
symmetric flow is not possible.
For compressible flows or flows with sources/sinks, non-trivial radially symmetric solutions are possible.
7 Conclusion
This document has presented the dimension reduction of four important equations in both cylindrical coor-
dinates (axisymmetry) and spherical coordinates (radial symmetry). The key results are:
1. Allen-Cahn Equation: - Reduced to 2D in cylindrical coordinates with axisymmetry - Reduced to 1D
in spherical coordinates with radial symmetry
2. Cahn-Hilliard Equation: - Reduced to 2D in cylindrical coordinates with axisymmetry - Reduced to
1D in spherical coordinates with radial symmetry
3. Convection-Diffusion Equation: - Reduced to 2D in cylindrical coordinates with axisymmetry - Re-
duced to 1D in spherical coordinates with radial symmetry
4. Navier-Stokes Equations: - Reduced to 2D in cylindrical coordinates with axisymmetry - For spherical
coordinates with radial symmetry, a non-trivial incompressible flow is not possible
These dimension reductions significantly simplify the computational complexity of the original 3D prob-
lems, making them more tractable for numerical simulations and analytical studies.