本文主要内容如下:
1. 不变性微分算子
Hamilton 算子(Nabla 算子)又称作不变性微分算子,这是因为它对张量场所作微分运算的形式不随坐标系的改变而改变,如:在不同坐标系
{
X
i
′
−
G
⃗
i
′
}
\{\mathscr{X}^{i'}-\vec{\mathscr{G}}_{i'}\}
{Xi′−Gi′}与
{
x
i
−
g
⃗
i
}
\{x^i-\vec{g}_i\}
{xi−gi} 中计算张量的梯度
▽
T
=
G
⃗
i
′
∂
T
∂
X
i
′
=
G
⃗
i
′
(
∂
T
∂
x
j
∂
x
j
∂
X
i
′
)
=
(
∂
x
j
∂
X
i
′
G
⃗
i
′
)
∂
T
∂
x
j
=
(
β
i
′
j
G
⃗
i
′
)
∂
T
∂
x
j
=
g
⃗
j
∂
T
∂
x
j
\bigtriangledown\bold T =\vec{\mathscr{G}}^{i'}\dfrac{\partial\bold T}{\partial\mathscr{X}^{i'}} =\vec{\mathscr{G}}^{i'}\left(\dfrac{\partial\bold T}{\partial x^j}\frac{{\partial x^j}}{\partial\mathscr{X}^{i'}}\right) =\left(\frac{{\partial x^j}}{\partial\mathscr{X}^{i'}}\vec{\mathscr{G}}^{i'}\right)\dfrac{\partial\bold T}{\partial x^j} =(\beta^{j}_{i'}\vec{\mathscr{G}}^{i'})\dfrac{\partial\bold T}{\partial x^j} =\vec{g}^j\dfrac{\partial\bold T}{\partial x^j}
▽T=Gi′∂Xi′∂T=Gi′(∂xj∂T∂Xi′∂xj)=(∂Xi′∂xjGi′)∂xj∂T=(βi′jGi′)∂xj∂T=gj∂xj∂T
2. 散度
对于
r
(
r
≥
1
)
r(r\ge 1)
r(r≥1) 阶张量场
T
\bold T
T,定义:
左散度:
▽
⋅
T
≜
g
⃗
i
⋅
∂
T
∂
x
i
≜
d
i
v
T
右散度:
T
⋅
▽
≜
∂
T
∂
x
i
⋅
g
⃗
i
左散度:\bigtriangledown\cdot\bold{T}\triangleq\vec{g}^i\cdot\frac{\partial \bold T}{\partial x^i}\triangleq div\bold T\\\ \\ 右散度:\bold{T}\cdot\bigtriangledown\triangleq\frac{\partial \bold T}{\partial x^i}\cdot\vec{g}^i
左散度:▽⋅T≜gi⋅∂xi∂T≜divT 右散度:T⋅▽≜∂xi∂T⋅gi
显然,一般
▽
⋅
T
≠
T
⋅
▽
\bigtriangledown\cdot\bold{T}\ne\bold{T}\cdot\bigtriangledown
▽⋅T=T⋅▽
举例:
- 向量场的散度:
▽ ⋅ v ⃗ = v ⃗ ⋅ ▽ = v ; i i = v , i i + v j Γ i j i = v , i i + 1 g ∂ g ∂ x j v j = 1 g ∂ ( g v j ) ∂ x j \bigtriangledown\cdot\vec{v}=\vec{v}\cdot\bigtriangledown=v^i_{;i} =v^i_{,i}+v^j\Gamma_{ij}^{i} =v^i_{,i}+\dfrac{1}{\sqrt{g}}\dfrac{\partial\sqrt{g}}{\partial x^j}v^j =\dfrac{1}{\sqrt{g}}\dfrac{\partial(\sqrt{g}v^j)}{\partial x^j} ▽⋅v=v⋅▽=v;ii=v,ii+vjΓiji=v,ii+g1∂xj∂gvj=g1∂xj∂(gvj) - 二阶张量场的散度:
▽ ⋅ A = A ; i i j g ⃗ j = A ∙ j i ∣ ; i g ⃗ j A ⋅ ▽ = A ; j i j g ⃗ i = A i ∙ j ∣ ; j g ⃗ i \bigtriangledown\cdot\bold{A}=A^{ij}_{;i}\vec{g}_j=A_{\bullet j}^{i}|_{;i}\vec{g}^{j}\\\ \\ \bold{A}\cdot\bigtriangledown=A^{ij}_{;j}\vec{g}_i=A^{\bullet j}_{i}|_{;j}\vec{g}^{i} ▽⋅A=A;iijgj=A∙ji∣;igj A⋅▽=A;jijgi=Ai∙j∣;jgi
通过上式可知:对称二阶张量的左右散度相等。 - 三阶张量场的散度:
▽ ⋅ T = A ; i i j k g ⃗ j g ⃗ k A ⋅ ▽ = A ; j i k j g ⃗ i g ⃗ k \bigtriangledown\cdot\bold{T}=A^{ijk}_{;i}\vec{g}_j\vec{g}_k\\\ \\ \bold{A}\cdot\bigtriangledown=A^{ikj}_{;j}\vec{g}_i\vec{g}_k ▽⋅T=A;iijkgjgk A⋅▽=A;jikjgigk
书写规则:
- 由于梯度点乘时总是自带逆变基,因此为方便点积,将张量分量靠近Nabla 算子的指标取为逆变指标,从而省去度量张量的分量;
- 张量分量靠近Nabla 算子的指标与协变导数的坐标指标相同,其余指标与基向量的指标形成哑指标。
3. 旋度
对于
r
(
r
≥
1
)
r(r\ge 1)
r(r≥1) 阶张量场
T
\bold T
T,定义:
左散度:
▽
×
T
≜
g
⃗
i
×
∂
T
∂
x
i
≜
c
u
r
l
T
=
ϵ
:
(
▽
T
)
右散度:
T
×
▽
≜
∂
T
∂
x
i
×
g
⃗
i
=
(
T
▽
)
:
ϵ
左散度:\bigtriangledown\times\bold{T}\triangleq\vec{g}^i\times\frac{\partial \bold T}{\partial x^i}\triangleq curl\bold T=\epsilon:(\bigtriangledown\bold{T})\\\ \\ 右散度:\bold{T}\times\bigtriangledown\triangleq\frac{\partial \bold T}{\partial x^i}\times\vec{g}^i=(\bold{T}\bigtriangledown):\epsilon
左散度:▽×T≜gi×∂xi∂T≜curlT=ϵ:(▽T) 右散度:T×▽≜∂xi∂T×gi=(T▽):ϵ
显然,一般
▽
×
T
≠
T
×
▽
\bigtriangledown\times\bold{T}\ne\bold{T}\times\bigtriangledown
▽×T=T×▽
举例:矢量场的旋度
▽
×
v
⃗
=
v
j
;
i
ϵ
i
j
k
g
⃗
k
=
(
v
j
,
i
−
v
m
Γ
j
i
m
)
ϵ
i
j
k
g
⃗
k
=
v
j
,
i
ϵ
i
j
k
g
⃗
k
=
1
g
∣
g
⃗
1
g
⃗
2
g
⃗
3
∂
∂
x
1
∂
∂
x
2
∂
∂
x
3
v
1
v
2
v
3
∣
v
⃗
×
▽
=
v
j
;
i
ϵ
j
i
k
g
⃗
k
=
−
▽
×
v
⃗
\bigtriangledown\times\vec{v}=v_{j;i}\epsilon^{ijk}\vec{g}_k =(v_{j,i}-v_m\Gamma^m_{ji})\epsilon^{ijk}\vec{g}_k =v_{j,i}\epsilon^{ijk}\vec{g}_k =\frac{1}{\sqrt{g}}\begin{vmatrix} \vec{g}_1 & \vec{g}_2 & \vec{g}_3 \\\\ \dfrac{\partial}{\partial x^1} & \dfrac{\partial}{\partial x^2} & \dfrac{\partial}{\partial x^3} \\\\ v_1 & v_2 & v_3 \end{vmatrix}\\\ \\ \vec{v}\times\bigtriangledown=v_{j;i}\epsilon^{jik}\vec{g}_k=-\bigtriangledown\times\vec{v}
▽×v=vj;iϵijkgk=(vj,i−vmΓjim)ϵijkgk=vj,iϵijkgk=g1
g1∂x1∂v1g2∂x2∂v2g3∂x3∂v3
v×▽=vj;iϵjikgk=−▽×v
书写规则:
- 由于梯度点乘时总是自带逆变基,为方便叉积,将张量分量靠近Nabla 算子的指标取为协变指标,从而省去度量张量的分量;
- 旋度的分量由协变导数与置换张量的逆变分量组成。
命题 给定向量场
v
⃗
\vec{v}
v,则反对称张量场
1
2
(
v
⃗
▽
−
▽
v
⃗
)
,
1
2
(
−
v
⃗
▽
+
▽
v
⃗
)
\dfrac{1}{2}(\vec{v}\bigtriangledown-\bigtriangledown\vec{v}),\dfrac{1}{2}(-\vec{v}\bigtriangledown+\bigtriangledown\vec{v})
21(v▽−▽v),21(−v▽+▽v)
的对偶矢量分别为:
ω
⃗
1
=
1
2
(
▽
×
v
⃗
)
,
ω
⃗
2
=
1
2
(
v
⃗
×
▽
)
\vec{\omega}_1=\dfrac{1}{2}(\bigtriangledown\times\vec{v}),\vec{\omega}_2=\dfrac{1}{2}(\vec{v}\times\bigtriangledown)
ω1=21(▽×v),ω2=21(v×▽)
证明如下:
ω
⃗
=
−
1
4
ϵ
:
(
v
⃗
▽
−
▽
v
⃗
)
=
−
1
4
(
v
⃗
×
▽
−
▽
×
v
⃗
)
=
1
2
▽
×
v
⃗
\begin{aligned} &\vec{\omega}=-\dfrac{1}{4}\epsilon:(\vec{v}\bigtriangledown-\bigtriangledown\vec{v})\\\\ &\ \ =-\dfrac{1}{4}(\vec{v}\times\bigtriangledown-\bigtriangledown\times\vec{v})\\\\ &\ \ =\dfrac{1}{2}\bigtriangledown\times\vec{v} \end{aligned}
ω=−41ϵ:(v▽−▽v) =−41(v×▽−▽×v) =21▽×v
4. Laplace 算子
定义
r
(
r
≥
1
)
r(r\ge 1)
r(r≥1) 阶张量场
T
\bold T
T 的Laplace 算子:
▽
2
T
=
▽
⋅
(
▽
T
)
=
d
i
v
(
g
r
a
d
T
)
\bigtriangledown^2\bold T=\bigtriangledown\cdot(\bigtriangledown \bold T)=div(grad\bold T)
▽2T=▽⋅(▽T)=div(gradT)
若
▽
2
T
=
0
\bigtriangledown^2\bold T=0
▽2T=0 则称
T
\bold T
T 是调和的。
举例:标量场的Laplace 算子
▽
2
ϕ
=
▽
⋅
(
▽
ϕ
)
=
▽
⋅
(
ϕ
,
i
g
⃗
i
)
=
g
i
j
(
ϕ
,
i
)
;
j
=
g
i
j
(
ϕ
,
i
j
−
ϕ
,
m
Γ
i
j
m
)
=
(
g
i
j
ϕ
,
i
)
;
j
=
(
g
i
j
ϕ
,
i
)
,
j
+
g
i
m
ϕ
,
i
Γ
m
j
j
=
(
g
i
m
ϕ
,
i
)
,
m
+
g
i
m
ϕ
,
i
1
g
∂
g
∂
x
m
=
1
g
∂
(
g
g
i
m
ϕ
,
i
)
∂
x
m
\begin{aligned} &\bigtriangledown^2\phi=\bigtriangledown\cdot(\bigtriangledown\phi)=\bigtriangledown\cdot(\phi_{,i}\vec{g}^i)\\\\ &\quad\quad\ =g^{ij}(\phi_{,i})_{;j}=g^{ij}(\phi_{,ij}-\phi_{,m}\Gamma^m_{ij})\\\\ &\quad\quad\ =(g^{ij}\phi_{,i})_{;j}=(g^{ij}\phi_{,i})_{,j}+g^{im}\phi_{,i}\Gamma^j_{mj}\\\\ &\quad\quad\ =(g^{im}\phi_{,i})_{,m}+g^{im}\phi_{,i}\frac{1}{\sqrt{g}}\frac{\partial \sqrt{g}}{\partial x^m}=\frac{1}{\sqrt{g}}\frac{\partial (\sqrt{g}g^{im}\phi_{,i})}{\partial x^m} \end{aligned}
▽2ϕ=▽⋅(▽ϕ)=▽⋅(ϕ,igi) =gij(ϕ,i);j=gij(ϕ,ij−ϕ,mΓijm) =(gijϕ,i);j=(gijϕ,i),j+gimϕ,iΓmjj =(gimϕ,i),m+gimϕ,ig1∂xm∂g=g1∂xm∂(ggimϕ,i)