What youβll learn
Theorems and identities : Get familiar with theorems and identities which are often used in derivations.
Strong and weak formulation : Understand what is the strong formulation, how to formulate the weak formulation and why it is useful.
This section contains an overview of general approach to formulating the physics that is solved in Allsolve.
Vector calculus identities
Divergence of a curl
β β
( β Γ A ) = 0 \begin{align}
\nabla \cdot (\nabla \times \boldsymbol{A})
= 0
\end{align} β β
( β Γ A ) = 0 β β
Leibniz rule for nabla operator
β β
( a b ) = β a β
b + a ( β β
b ) \begin{align}
\nabla \cdot (a \boldsymbol{b})
= \nabla a \cdot \boldsymbol{b}
+ a(\nabla \cdot \boldsymbol{b})
\end{align} β β
( a b ) = β a β
b + a ( β β
b ) β β
Per partes theorem for scalars
Let U U U and V V V be a scalar functions, Ξ© \Omega Ξ© a set with boundary Ξ \Gamma Ξ . Then let n \boldsymbol{n} n be a unit normal vector pointing outwards. Then
β« Ξ© β
β U β V β
β d Ξ© = β« Ξ β
β U V n β
β d Ξ β β« Ξ© β
β V β U β
β d Ξ© \begin{align}
\int_{\Omega}\; U\nabla V\; d \Omega
= \int_{\Gamma}\; UV\boldsymbol{n}\; d \Gamma
- \int_{\Omega}\; V\nabla U\; d \Omega
\end{align} β« Ξ© β U β V d Ξ© = β« Ξ β U V n d Ξ β β« Ξ© β V β U d Ξ© β β
Per partes theorem for vectors
Let A \boldsymbol{A} A be a vector field and f f f a scalar function defined on a domain Ξ© \Omega Ξ© with boundary Ξ \Gamma Ξ snd let n \boldsymbol{n} n be a unit normal vector pointing outwards. Then
β« Ξ© β
β ( A β
β f ) β
β d Ξ© = β« Ξ β
β f β
β ( A β
n ) β
β d Ξ β β« Ξ© β
β f β
β ( β β
A ) β
β d Ξ© \begin{align}
\int_{\Omega}\; (\boldsymbol{A} \cdot \nabla f)\; d \Omega
= \int_{\Gamma}\; f\; (\boldsymbol{A} \cdot \boldsymbol{n})\; d \Gamma
- \int_{\Omega}\; f\; (\nabla \cdot \boldsymbol{A})\; d \Omega
\end{align} β« Ξ© β ( A β
β f ) d Ξ© = β« Ξ β f ( A β
n ) d Ξ β β« Ξ© β f ( β β
A ) d Ξ© β β
Divergence theorem
Let Ξ© \Omega Ξ© be a set with boundary Ξ \Gamma Ξ , n \boldsymbol{n} n be a unit normal vector pointing outwards and T \boldsymbol{T} T be a vector field or a second-order tensor. Then
β« Ξ© β β
T β
β d Ξ© = β« Ξ β
β T β
n β
β d Ξ \begin{align}
\int_{\Omega} \nabla \cdot \boldsymbol{T}\; d \Omega
= \int_{\Gamma}\; \boldsymbol{T} \cdot \boldsymbol{n}\; d \Gamma
\end{align} β« Ξ© β β β
T d Ξ© = β« Ξ β T β
n d Ξ β β
Common example of PDE (partial differential equation) is the Poissonβs equation.
Ξ u = f β
β on Ξ© \begin{align}
\Delta u
&= f\; \text{on}\ \Omega
\end{align} Ξ u β = f on Ξ© β β
This is the strong formulation of Poissonβs equation.
The strong formulation refers to a precise mathematical representation of a problem with very strict requirements for a solution.
The weak formulation is an alternative mathematical representation that relaxes the regularity requirements for solutions.
The weak formulation is computationally preferable as it provides the foundation for numerical methods such as the Finite Element Method (FEM).
Both Divergence theorem and Per partes theorem (Integration by parts) can be applied on the strong formulation.
Using these theorems we derive the weak formulation from the strong formulation.
To begin, we choose a test function u β² u^{\prime} u β² .
We then multiply both sides of the differential equations by u β² u^{\prime} u β² and integrate over Ξ© \Omega Ξ© .
β« Ξ© β
β ( Ξ u u β² β f u β² ) β
β d Ξ© = 0 \begin{align}
\int_{\Omega}\; (\Delta uu^{\prime} - fu^{\prime})\; d \Omega
= 0
\end{align} β« Ξ© β ( Ξ u u β² β f u β² ) d Ξ© = 0 β β
For the first term of the left side we apply Per partes theorem.
In this case U = u β² U = u^{\prime} U = u β² , V = β u V = \nabla u V = β u and n \boldsymbol{n} n is the unit normal vector.
β« Ξ© β
β Ξ u β
β u β² β
β d Ξ© = β« Ξ© β
β ( β β
β u ) β
β u β² β
β d Ξ© = β« Ξ β
β ( β u β
n ) β
β u β² β
β d Ξ β β« Ξ© β
β β u β
β u β² β
β d Ξ© \begin{align}
\int_{\Omega}\; \Delta u\; u^{\prime}\; d \Omega
= \int_{\Omega}\; (\nabla \cdot \nabla u)\; u^{\prime}\; d \Omega
= \int_{\Gamma}\; (\nabla u \cdot \boldsymbol{n})\; u^{\prime}\; d \Gamma - \int_{\Omega}\; \nabla u \cdot \nabla u^{\prime}\; d \Omega
\end{align} β« Ξ© β Ξ u u β² d Ξ© = β« Ξ© β ( β β
β u ) u β² d Ξ© = β« Ξ β ( β u β
n ) u β² d Ξ β β« Ξ© β β u β
β u β² d Ξ© β β
Overall we get a weak formulation of (3).
β« Ξ β
β ( β u β
n ) β
β u β² β
β d Ξ β β« Ξ© β
β ( β u β
β u β² + f β
β u β² ) β
β d Ξ© = 0 \begin{align}
\int_{\Gamma}\; (\nabla u\ \cdot \boldsymbol{n})\; u^{\prime}\; d \Gamma
- \int_{\Omega}\; (\nabla u \cdot \nabla u^{\prime} + f\; u^{\prime})\; d \Omega
= 0
\end{align} β« Ξ β ( β u β
n ) u β² d Ξ β β« Ξ© β ( β u β
β u β² + f u β² ) d Ξ© = 0 β β
Now, we have less strict regularity requirements for u u u and f f f .
Certainly we know that solution to (3) is also a solution to (7) and every weak solution smooth enough is a solution to (3).