本文主要内容如下:
1. 张量的代数运算
1.1. 张量的相等
若两个张量 T 与 S \bold{T}与\bold{S} T与S在同一坐标系中的协变分量(逆变分量/混变分量)均一 一对应相等,则称张量 T \bold{T} T等于张量 S \bold{S} S,记作 T = S \bold{T}=\bold{S} T=S。
1.2. 张量的加(减)法
两个张量
T
\bold{T}
T与
S
\bold{S}
S在同一坐标系中的协变分量(逆变分量/混变分量)一 一对应相加(减),如
U
i
j
k
=
T
i
j
k
+
S
i
j
k
U_{ijk}=T_{ijk}+S_{ijk}
Uijk=Tijk+Sijk得到的新有序数组满足坐标变换关系:
U
i
′
j
′
k
′
=
T
i
′
j
′
k
′
+
S
i
′
j
′
k
′
=
β
i
′
p
β
j
′
q
β
k
′
l
(
T
p
q
l
+
S
p
q
l
)
=
β
i
′
p
β
j
′
q
β
k
′
l
U
p
q
l
U_{i'j'k'} =T_{i'j'k'}+S_{i'j'k'} =\beta_{i'}^{p}\beta_{j'}^{q}\beta_{k'}^{l}(T_{pql}+S_{pql}) =\beta_{i'}^{p}\beta_{j'}^{q}\beta_{k'}^{l}U_{pql}
Ui′j′k′=Ti′j′k′+Si′j′k′=βi′pβj′qβk′l(Tpql+Spql)=βi′pβj′qβk′lUpql将这种得到新张量的运算方式称之为张量
T
\bold{T}
T与张量
S
\bold{S}
S的和(差),记作:
U
=
T
±
S
\bold{U}=\bold{T}\pm\bold{S}
U=T±S即,
U
i
j
k
g
⃗
i
g
⃗
j
g
⃗
k
=
T
i
j
k
g
⃗
i
g
⃗
j
g
⃗
k
+
S
i
j
k
g
⃗
i
g
⃗
j
g
⃗
k
U_{ijk}\vec{g}^i\vec{g}^j\vec{g}^k=T_{ijk}\vec{g}^i\vec{g}^j\vec{g}^k+S_{ijk}\vec{g}^i\vec{g}^j\vec{g}^k
Uijkgigjgk=Tijkgigjgk+Sijkgigjgk
1.3. 张量的数乘
将张量
T
\bold{T}
T在一坐标系中的协变分量(逆变分量/混变分量)均乘以同一标量
k
k
k(
k
k
k不随坐标变化),如
U
i
j
=
k
T
i
j
U_{ij}=kT_{ij}
Uij=kTij得到的新有序数组满足坐标变换关系:
U
i
′
j
′
=
k
T
i
′
j
′
=
β
i
′
p
β
j
′
q
k
T
p
q
=
β
i
′
p
β
j
′
q
U
p
q
U_{i'j'} =kT_{i'j'} =\beta_{i'}^{p}\beta_{j'}^{q}kT_{pq} =\beta_{i'}^{p}\beta_{j'}^{q}U_{pq}
Ui′j′=kTi′j′=βi′pβj′qkTpq=βi′pβj′qUpq将这种得到新张量的运算方式称之为张量
T
\bold{T}
T与标量
k
k
k的数乘,记作:
U
=
k
T
\bold{U}=k\bold{T}
U=kT即,
U
i
j
g
⃗
i
g
⃗
j
=
k
T
i
j
g
⃗
i
g
⃗
j
U_{ij}\vec{g}^i\vec{g}^j=kT_{ij}\vec{g}^i\vec{g}^j
Uijgigj=kTijgigj
1.4. 张量的并乘
在同一坐标系中,将张量
T
\bold{T}
T任意形式的分量与张量
S
\bold{S}
S任意形式的分量两两相乘也可得到新的有序数组,如
U
∙
∙
∙
p
q
l
m
n
i
j
k
=
T
i
j
k
S
p
q
l
m
n
U^{ijk}_{\bullet\bullet\bullet pqlmn}=T^{ijk}S_{pqlmn}
U∙∙∙pqlmnijk=TijkSpqlmn显然,新有序数组也满足坐标变换关系:
U
∙
∙
∙
p
′
q
′
l
′
m
′
n
′
i
′
j
′
k
′
=
T
i
′
j
′
k
′
S
p
′
q
′
l
′
m
′
n
′
=
β
a
i
′
β
b
j
′
β
c
k
′
T
a
b
c
β
p
′
d
β
q
′
e
β
l
′
f
β
m
′
h
β
n
′
r
S
d
e
f
h
r
=
β
a
i
′
β
b
j
′
β
c
k
′
β
p
′
d
β
q
′
e
β
l
′
f
β
m
′
h
β
n
′
r
U
∙
∙
∙
d
e
f
h
r
a
b
c
U^{i'j'k'}_{\bullet\bullet\bullet p'q'l'm'n'} =T^{i'j'k'}S_{p'q'l'm'n'} =\beta^{i'}_{a}\beta^{j'}_{b}\beta^{k'}_{c}T^{abc}\beta_{p'}^{d}\beta_{q'}^{e}\beta_{l'}^{f}\beta_{m'}^{h}\beta_{n'}^{r}S_{defhr} =\beta^{i'}_{a}\beta^{j'}_{b}\beta^{k'}_{c}\beta_{p'}^{d}\beta_{q'}^{e}\beta_{l'}^{f}\beta_{m'}^{h}\beta_{n'}^{r}U^{abc}_{\bullet\bullet\bullet defhr}
U∙∙∙p′q′l′m′n′i′j′k′=Ti′j′k′Sp′q′l′m′n′=βai′βbj′βck′Tabcβp′dβq′eβl′fβm′hβn′rSdefhr=βai′βbj′βck′βp′dβq′eβl′fβm′hβn′rU∙∙∙defhrabc将这种得到新张量的运算方式称之为张量
T
\bold{T}
T与张量
S
\bold{S}
S的并乘/张量乘,记作:
U
=
T
S
\bold{U}=\bold{T}\bold{S}
U=TS即
U
=
U
∙
∙
∙
p
q
l
m
n
i
j
k
g
⃗
i
g
⃗
j
g
⃗
k
g
⃗
p
g
⃗
q
g
⃗
l
g
⃗
m
g
⃗
n
=
(
T
i
j
k
g
⃗
i
g
⃗
j
g
⃗
k
)
(
S
p
q
l
m
n
g
⃗
p
g
⃗
q
g
⃗
l
g
⃗
m
g
⃗
n
)
=
T
S
\bold{U} =U^{ijk}_{\bullet\bullet\bullet pqlmn}\vec{g}_i\vec{g}_j\vec{g}_k\vec{g}^p\vec{g}^q\vec{g}^l\vec{g}^m\vec{g}^n =(T^{ijk}\vec{g}_i\vec{g}_j\vec{g}_k)(S_{pqlmn}\vec{g}^p\vec{g}^q\vec{g}^l\vec{g}^m\vec{g}^n) =\bold{T}\bold{S}
U=U∙∙∙pqlmnijkgigjgkgpgqglgmgn=(Tijkgigjgk)(Spqlmngpgqglgmgn)=TS注意到并乘所得新张量的阶数为参与并乘的张量阶数之和,新张量指标的前后次序上下位置与参与并乘的张量分别保持一致且张量并乘时次序不可随意调换,即
T
S
≠
S
T
\bold{TS}\ne\bold{ST}
TS=ST
1.5. 张量的缩并
将张量的基张量中任意两个基矢量进行点积,如对于张量
T
\bold{T}
T:
T
=
T
∙
∙
m
n
i
j
g
⃗
i
g
⃗
j
g
⃗
m
g
⃗
n
\bold{T}=T^{ij}_{\bullet\bullet mn}\vec{g}_i\vec{g}_j\vec{g}^m\vec{g}^n
T=T∙∙mnijgigjgmgn(1) 若第一、二基矢量进行点积:
S
=
S
m
n
g
⃗
m
g
⃗
n
=
T
∙
∙
m
n
i
j
g
i
j
g
⃗
m
g
⃗
n
\bold{S}=S_{mn}\vec{g}^m\vec{g}^n=T^{ij}_{\bullet\bullet mn}g_{ij}\vec{g}^m\vec{g}^n
S=Smngmgn=T∙∙mnijgijgmgn(2) 若第二、三基矢量进行点积:
P
=
P
∙
n
i
g
⃗
i
g
⃗
n
=
T
∙
∙
m
n
i
j
δ
j
m
g
⃗
i
g
⃗
n
=
T
∙
∙
j
n
i
j
g
⃗
i
g
⃗
n
\bold{P}=P_{\bullet n}^i\vec{g}_i\vec{g}^n=T^{ij}_{\bullet\bullet mn}\delta^{m}_{j}\vec{g}_i\vec{g}^n=T^{ij}_{\bullet\bullet jn}\vec{g}_i\vec{g}^n
P=P∙nigign=T∙∙mnijδjmgign=T∙∙jnijgign(3) 若第三、四基矢量进行点积:
Q
=
Q
i
j
g
⃗
i
g
⃗
j
=
T
∙
∙
m
n
i
j
g
m
n
g
⃗
i
g
⃗
j
\bold{Q}=Q^{ij}\vec{g}_i\vec{g}_j=T^{ij}_{\bullet\bullet mn}g^{mn}\vec{g}_i\vec{g}_j
Q=Qijgigj=T∙∙mnijgmngigj得到的新有序数组满足坐标变换关系:
(
1
)
S
m
′
n
′
=
T
∙
∙
m
′
n
′
i
′
j
′
g
i
′
j
′
=
β
a
i
′
β
b
j
′
β
m
′
c
β
n
′
d
T
∙
∙
c
d
a
b
g
i
′
j
′
=
β
m
′
c
β
n
′
d
T
∙
∙
c
d
a
b
g
a
b
=
β
m
′
c
β
n
′
d
S
c
d
(
2
)
P
∙
n
′
i
′
=
T
∙
∙
m
′
n
′
i
′
j
′
δ
m
′
j
′
=
β
a
i
′
β
b
j
′
β
m
′
c
β
n
′
d
T
∙
∙
c
d
a
b
δ
m
′
j
′
=
β
a
i
′
β
b
j
′
β
j
′
c
β
n
′
d
T
∙
∙
c
d
a
b
=
β
a
i
′
β
n
′
d
T
∙
∙
c
d
a
b
δ
b
c
=
β
a
i
′
β
n
′
d
T
∙
∙
b
d
a
b
=
β
a
i
′
β
n
′
d
P
∙
d
a
(
3
)
Q
i
′
j
′
=
T
∙
∙
m
′
n
′
i
′
j
′
g
m
′
n
′
=
β
a
i
′
β
b
j
′
β
m
′
c
β
n
′
d
T
∙
∙
c
d
a
b
g
m
′
n
′
=
β
a
i
′
β
b
j
′
T
∙
∙
c
d
a
b
g
c
d
=
β
a
i
′
β
b
j
′
Q
a
b
(1) S_{m'n'} =T^{i'j'}_{\bullet\bullet m'n'}g_{i'j'} =\beta^{i'}_{a}\beta^{j'}_{b}\beta_{m'}^{c}\beta_{n'}^{d}T^{ab}_{\bullet\bullet cd}g_{i'j'} =\beta_{m'}^{c}\beta_{n'}^{d}T^{ab}_{\bullet\bullet cd}g_{ab} =\beta_{m'}^{c}\beta_{n'}^{d}S_{cd}\\\ \\ (2)P_{\bullet n'}^{i'} =T^{i'j'}_{\bullet\bullet m'n'}\delta^{j'}_{m'} =\beta^{i'}_{a}\beta^{j'}_{b}\beta_{m'}^{c}\beta_{n'}^{d}T^{ab}_{\bullet\bullet cd}\delta^{j'}_{m'} =\beta^{i'}_{a}\beta^{j'}_{b}\beta_{j'}^{c}\beta_{n'}^{d}T^{ab}_{\bullet\bullet cd} =\beta^{i'}_{a}\beta_{n'}^{d}T^{ab}_{\bullet\bullet cd}\delta^{c}_{b} =\beta^{i'}_{a}\beta_{n'}^{d}T^{ab}_{\bullet\bullet bd} =\beta^{i'}_{a}\beta_{n'}^{d}P_{\bullet d}^{a}\\\ \\ (3)Q^{i'j'} =T^{i'j'}_{\bullet\bullet m'n'}g^{m'n'} =\beta^{i'}_{a}\beta^{j'}_{b}\beta_{m'}^{c}\beta_{n'}^{d}T^{ab}_{\bullet\bullet cd}g^{m'n'} =\beta^{i'}_{a}\beta^{j'}_{b}T^{ab}_{\bullet\bullet cd}g^{cd} =\beta^{i'}_{a}\beta^{j'}_{b}Q^{ab}
(1)Sm′n′=T∙∙m′n′i′j′gi′j′=βai′βbj′βm′cβn′dT∙∙cdabgi′j′=βm′cβn′dT∙∙cdabgab=βm′cβn′dScd (2)P∙n′i′=T∙∙m′n′i′j′δm′j′=βai′βbj′βm′cβn′dT∙∙cdabδm′j′=βai′βbj′βj′cβn′dT∙∙cdab=βai′βn′dT∙∙cdabδbc=βai′βn′dT∙∙bdab=βai′βn′dP∙da (3)Qi′j′=T∙∙m′n′i′j′gm′n′=βai′βbj′βm′cβn′dT∙∙cdabgm′n′=βai′βbj′T∙∙cdabgcd=βai′βbj′Qab将这种得到新张量的运算方式称之为张量
T
\bold{T}
T的缩并。
注意到:每缩并一次,张量阶数降低两阶。
1.6. 张量的点积/内积
前张量的最后的基矢量与后矢量的第一个基矢量进行点积,该运算称作张量的点积/内积,显然该运算得到的结果仍然是张量,因张量的点积等价于张量并乘后进行相应地缩并运算。记作:
U
=
T
∙
S
\bold{U}=\bold{T}\bullet\bold{S}
U=T∙S如,
T
∙
S
=
T
i
j
g
⃗
i
g
⃗
j
∙
S
m
n
g
⃗
m
g
⃗
n
=
T
i
j
g
⃗
i
S
m
n
g
⃗
n
δ
j
m
=
T
i
j
S
j
n
g
⃗
i
g
⃗
n
=
U
∙
n
i
g
⃗
i
g
⃗
n
\bold{T}\bullet\bold{S} =T^{ij}\vec{g}_i\vec{g}_j\bullet S_{mn}\vec{g}^m\vec{g}^n =T^{ij}\vec{g}_iS_{mn}\vec{g}^n\delta_j^m =T^{ij}S_{jn}\vec{g}_i\vec{g}^n =U^i_{\bullet n}\vec{g}_i\vec{g}^n
T∙S=Tijgigj∙Smngmgn=TijgiSmngnδjm=TijSjngign=U∙nigign若将张量并乘后连续进行两次点积则称该运算为双点积,包括两种情形,举例进行说明:
(1)并联式:
T
:
S
=
T
i
j
S
m
n
δ
m
i
δ
n
j
=
T
i
j
S
i
j
(前前后后)
\bold{T}:\bold{S}=T^{ij}S_{mn}\delta^i_m\delta^j_n=T^{ij}S_{ij}\qquad(前前后后)
T:S=TijSmnδmiδnj=TijSij(前前后后)(2)串联式:
T
∙
∙
S
=
T
i
j
S
m
n
δ
m
j
δ
n
i
=
T
i
j
S
j
i
(内内外外)
\bold{T}\bullet\bullet\bold{S}=T^{ij}S_{mn}\delta^j_m\delta^i_n=T^{ij}S_{ji}\qquad(内内外外)
T∙∙S=TijSmnδmjδni=TijSji(内内外外)显然,同样张量的串、并联式的计算结果不同且张量点积与点乘次序相关。
1.7. 张量的叉积/矢积
1.7.1. 矢量的叉积
根据基矢量的叉积可得出任意矢量的叉积计算:
c
⃗
=
a
⃗
×
b
⃗
=
a
i
g
⃗
i
×
b
j
g
⃗
j
=
a
i
b
j
(
g
⃗
i
×
g
⃗
j
)
=
a
i
b
j
ϵ
i
j
k
g
⃗
k
=
c
k
g
⃗
k
⟹
c
k
=
a
i
b
j
ϵ
i
j
k
c
⃗
=
a
⃗
×
b
⃗
=
a
i
g
⃗
i
×
b
j
g
⃗
j
=
a
i
b
j
(
g
⃗
i
×
g
⃗
j
)
=
a
i
b
j
ϵ
i
j
k
g
⃗
k
=
c
k
g
⃗
k
⟹
c
k
=
a
i
b
j
ϵ
i
j
k
\vec{c}=\vec{a}\times\vec{b} =a^i\vec{g}_i\times b^j\vec{g}_j =a^ib^j(\vec{g}_i\times\vec{g}_j) =a^ib^j\epsilon_{ijk}\vec{g}^k =c_k\vec{g}^k \Longrightarrow c_k=a^ib^j\epsilon_{ijk} \\\ \\ \vec{c}=\vec{a}\times\vec{b} =a_i\vec{g}^i\times b_j\vec{g}^j =a_ib_j(\vec{g}^i\times\vec{g}^j) =a_ib_j\epsilon^{ijk}\vec{g}_k =c^k\vec{g}_k \Longrightarrow c^k=a_ib_j\epsilon^{ijk}
c=a×b=aigi×bjgj=aibj(gi×gj)=aibjϵijkgk=ckgk⟹ck=aibjϵijk c=a×b=aigi×bjgj=aibj(gi×gj)=aibjϵijkgk=ckgk⟹ck=aibjϵijk此外,可通过并联式双点积进行向量叉积的表示:
ϵ
:
a
⃗
b
⃗
=
(
ϵ
p
q
l
g
⃗
p
g
⃗
q
g
⃗
l
)
:
(
a
i
b
j
g
⃗
i
g
⃗
j
)
=
ϵ
p
q
l
a
i
b
j
δ
i
q
δ
j
l
g
⃗
p
=
a
i
b
j
ϵ
i
j
p
g
⃗
p
=
a
⃗
×
b
⃗
=
−
a
⃗
⋅
ϵ
⋅
b
⃗
=
b
⃗
⋅
ϵ
⋅
a
⃗
a
⃗
b
⃗
:
ϵ
=
(
a
i
b
j
g
⃗
i
g
⃗
j
)
:
(
ϵ
p
q
l
g
⃗
p
g
⃗
q
g
⃗
l
)
=
ϵ
p
q
l
a
i
b
j
δ
i
p
δ
j
q
g
⃗
l
=
a
i
b
j
ϵ
i
j
l
g
⃗
l
=
a
⃗
×
b
⃗
\epsilon:\vec{a}\vec{b} =(\epsilon_{pql}\vec{g}^p\vec{g}^q\vec{g}^l):(a^ib^j\vec{g}_i\vec{g}_j) =\epsilon_{pql}a^ib^j\delta^q_i\delta^l_j\vec{g}^p =a^ib^j\epsilon_{ijp}\vec{g}^p =\vec{a}\times\vec{b} =-\vec{a}\cdot\epsilon\cdot\vec{b} =\vec{b}\cdot\epsilon\cdot\vec{a}\\\ \\ \vec{a}\vec{b}:\epsilon =(a^ib^j\vec{g}_i\vec{g}_j):(\epsilon_{pql}\vec{g}^p\vec{g}^q\vec{g}^l) =\epsilon_{pql}a^ib^j\delta^p_i\delta^q_j\vec{g}^l =a^ib^j\epsilon_{ijl}\vec{g}^l =\vec{a}\times\vec{b}
ϵ:ab=(ϵpqlgpgqgl):(aibjgigj)=ϵpqlaibjδiqδjlgp=aibjϵijpgp=a×b=−a⋅ϵ⋅b=b⋅ϵ⋅a ab:ϵ=(aibjgigj):(ϵpqlgpgqgl)=ϵpqlaibjδipδjqgl=aibjϵijlgl=a×b
1.7.2. 矢量的混合积
(
a
⃗
×
b
⃗
)
∙
c
⃗
=
(
a
i
b
j
ϵ
i
j
k
g
⃗
k
)
∙
(
c
h
g
⃗
h
)
=
a
i
b
j
c
k
ϵ
i
j
k
=
a
⃗
b
⃗
c
⃗
⋮
ϵ
=
ϵ
⋮
a
⃗
b
⃗
c
⃗
(
a
⃗
×
b
⃗
)
∙
c
⃗
=
(
a
i
b
j
ϵ
i
j
k
g
⃗
k
)
∙
(
c
h
g
⃗
h
)
=
a
i
b
j
c
k
ϵ
i
j
k
=
a
⃗
b
⃗
c
⃗
⋮
ϵ
=
ϵ
⋮
a
⃗
b
⃗
c
⃗
(\vec{a}\times\vec{b})\bullet\vec{c} =(a^ib^j\epsilon_{ijk}\vec{g}^k)\bullet(c^h\vec{g}_h)=a^ib^jc^k\epsilon_{ijk} =\vec{a}\vec{b}\vec{c}\ \vdots\ \epsilon =\epsilon\ \vdots\ \vec{a}\vec{b}\vec{c}\\\ \\ (\vec{a}\times\vec{b})\bullet\vec{c} =(a_ib_j\epsilon^{ijk}\vec{g}_k)\bullet(c_h\vec{g}^h)=a_ib_jc_k\epsilon^{ijk} =\vec{a}\vec{b}\vec{c}\ \vdots\ \epsilon =\epsilon\ \vdots\ \vec{a}\vec{b}\vec{c}
(a×b)∙c=(aibjϵijkgk)∙(chgh)=aibjckϵijk=abc ⋮ ϵ=ϵ ⋮ abc (a×b)∙c=(aibjϵijkgk)∙(chgh)=aibjckϵijk=abc ⋮ ϵ=ϵ ⋮ abc另外推知:
(1)混合积中点积与叉积的位置可互换;
(2)参与混合积的向量可依次交换位置而不改变混合积的值。即:
a
⃗
∙
(
b
⃗
×
c
⃗
)
=
(
b
⃗
×
c
⃗
)
∙
a
⃗
=
ϵ
⋮
b
⃗
c
⃗
a
⃗
=
a
k
b
i
c
j
ϵ
k
i
j
=
ϵ
⋮
a
⃗
b
⃗
c
⃗
=
(
a
⃗
×
b
⃗
)
∙
c
⃗
⟹
a
⃗
∙
(
b
⃗
×
c
⃗
)
=
(
a
⃗
×
b
⃗
)
∙
c
⃗
(
a
⃗
×
b
⃗
)
∙
c
⃗
=
(
b
⃗
×
c
⃗
)
∙
a
⃗
=
(
c
⃗
×
a
⃗
)
∙
b
⃗
=
a
i
b
j
c
k
ϵ
i
j
k
\vec{a}\bullet(\vec{b}\times\vec{c}) =(\vec{b}\times\vec{c})\bullet\vec{a} =\epsilon\ \vdots\ \vec{b}\vec{c}\vec{a} =a^kb^ic^j\epsilon_{kij} =\epsilon\ \vdots\ \vec{a}\vec{b}\vec{c} =(\vec{a}\times\vec{b})\bullet\vec{c} \Longrightarrow \vec{a}\bullet(\vec{b}\times\vec{c})=(\vec{a}\times\vec{b})\bullet\vec{c}\\\ \\ (\vec{a}\times\vec{b})\bullet\vec{c} =(\vec{b}\times\vec{c})\bullet\vec{a} =(\vec{c}\times\vec{a})\bullet\vec{b} =a_ib_jc_k\epsilon^{ijk}
a∙(b×c)=(b×c)∙a=ϵ ⋮ bca=akbicjϵkij=ϵ ⋮ abc=(a×b)∙c⟹a∙(b×c)=(a×b)∙c (a×b)∙c=(b×c)∙a=(c×a)∙b=aibjckϵijk
1.7.3. 矢量的三重叉积
f
⃗
=
(
a
⃗
×
b
⃗
)
×
c
⃗
=
(
a
i
b
j
ϵ
i
j
k
g
⃗
k
)
×
(
c
p
g
⃗
p
)
=
a
i
b
j
c
p
ϵ
i
j
k
ϵ
p
q
k
g
⃗
q
=
f
q
g
⃗
q
⟹
f
q
=
a
i
b
j
c
p
ϵ
i
j
k
ϵ
p
q
k
f
⃗
=
(
a
⃗
×
b
⃗
)
×
c
⃗
=
(
a
i
b
j
ϵ
i
j
k
g
⃗
k
)
×
(
c
p
g
⃗
p
)
=
a
i
b
j
c
p
ϵ
i
j
k
ϵ
p
q
k
g
⃗
q
=
f
q
g
⃗
q
⟹
f
q
=
a
i
b
j
c
p
ϵ
i
j
k
ϵ
p
q
k
f
⃗
=
a
⃗
×
(
b
⃗
×
c
⃗
)
=
(
a
i
g
⃗
i
)
×
(
b
j
c
p
ϵ
j
p
k
g
⃗
k
)
=
a
i
b
j
c
p
ϵ
j
p
k
ϵ
i
k
q
g
⃗
q
=
f
q
g
⃗
q
⟹
f
q
=
a
i
b
j
c
p
ϵ
j
p
k
ϵ
q
i
k
f
⃗
=
a
⃗
×
(
b
⃗
×
c
⃗
)
=
(
a
i
g
⃗
i
)
×
(
b
j
c
p
ϵ
j
p
k
g
⃗
k
)
=
a
i
b
j
c
p
ϵ
j
p
k
ϵ
i
k
q
g
⃗
q
=
f
q
g
⃗
q
⟹
f
q
=
a
i
b
j
c
p
ϵ
j
p
k
ϵ
q
i
k
\vec{f} =(\vec{a}\times\vec{b})\times\vec{c} =(a^ib^j\epsilon_{ijk}\vec{g}^k)\times(c_p\vec{g}^p) =a^ib^jc_p\epsilon_{ijk}\epsilon^{pqk}\vec{g}_q =f^q\vec{g}_{q} \Longrightarrow f^q=a^ib^jc_p\epsilon_{ijk}\epsilon^{pqk}\\\ \\ \vec{f} =(\vec{a}\times\vec{b})\times\vec{c} =(a_ib_j\epsilon^{ijk}\vec{g}_k)\times(c^p\vec{g}_p) =a_ib_jc^p\epsilon^{ijk}\epsilon_{pqk}\vec{g}^q =f_q\vec{g}^{q} \Longrightarrow f_q=a_ib_jc^p\epsilon^{ijk}\epsilon_{pqk}\\\ \\ \vec{f} =\vec{a}\times(\vec{b}\times\vec{c}) =(a^i\vec{g}_i)\times(b_jc_p\epsilon^{jpk}\vec{g}_k) =a^ib_jc_p\epsilon^{jpk}\epsilon_{ikq}\vec{g}^q =f_q\vec{g}^{q} \Longrightarrow f_q=a^ib_jc_p\epsilon^{jpk}\epsilon_{qik}\\\ \\ \vec{f} =\vec{a}\times(\vec{b}\times\vec{c}) =(a_i\vec{g}^i)\times(b^jc^p\epsilon_{jpk}\vec{g}^k) =a_ib^jc^p\epsilon_{jpk}\epsilon^{ikq}\vec{g}_q =f^q\vec{g}_{q} \Longrightarrow f^q=a_ib^jc^p\epsilon_{jpk}\epsilon^{qik}
f=(a×b)×c=(aibjϵijkgk)×(cpgp)=aibjcpϵijkϵpqkgq=fqgq⟹fq=aibjcpϵijkϵpqk f=(a×b)×c=(aibjϵijkgk)×(cpgp)=aibjcpϵijkϵpqkgq=fqgq⟹fq=aibjcpϵijkϵpqk f=a×(b×c)=(aigi)×(bjcpϵjpkgk)=aibjcpϵjpkϵikqgq=fqgq⟹fq=aibjcpϵjpkϵqik f=a×(b×c)=(aigi)×(bjcpϵjpkgk)=aibjcpϵjpkϵikqgq=fqgq⟹fq=aibjcpϵjpkϵqik由此可知:三重叉积不具有结合律,即:
(
a
⃗
×
b
⃗
)
×
c
⃗
≠
a
⃗
×
(
b
⃗
×
c
⃗
)
(\vec{a}\times\vec{b})\times\vec{c}\ne\vec{a}\times(\vec{b}\times\vec{c})
(a×b)×c=a×(b×c)另外,由
ϵ
∼
δ
\epsilon\sim\delta
ϵ∼δ 关系式:
(
a
⃗
×
b
⃗
)
×
c
⃗
=
a
i
b
j
c
p
ϵ
i
j
k
ϵ
p
q
k
g
⃗
q
=
a
i
b
j
c
p
(
δ
p
i
δ
q
j
−
δ
q
i
δ
p
j
)
g
⃗
q
=
(
a
i
c
i
)
(
b
j
g
⃗
j
)
−
(
b
j
c
j
)
a
i
g
⃗
i
=
(
a
⃗
∙
c
⃗
)
b
⃗
−
(
b
⃗
∙
c
⃗
)
a
⃗
a
⃗
×
(
b
⃗
×
c
⃗
)
=
a
i
b
j
c
p
ϵ
j
p
k
ϵ
q
i
k
g
⃗
q
=
a
i
b
j
c
p
(
δ
j
q
δ
p
i
−
δ
p
q
δ
j
i
)
g
⃗
q
=
(
a
i
c
i
)
(
b
j
g
⃗
j
)
−
(
a
i
b
i
)
(
c
q
g
⃗
q
)
=
(
a
⃗
∙
c
⃗
)
b
⃗
−
(
a
⃗
∙
b
⃗
)
c
⃗
(\vec{a}\times\vec{b})\times\vec{c} =a^ib^jc_p\epsilon_{ijk}\epsilon^{pqk}\vec{g}_q =a^ib^jc_p(\delta^i_p\delta^j_q-\delta^i_q\delta^j_p)\vec{g}_q =(a^ic_i)(b^j\vec{g}_j)-(b^jc_j)a^i\vec{g}_i =(\vec{a}\bullet\vec{c})\vec{b}-(\vec{b}\bullet\vec{c})\vec{a}\\\ \\ \vec{a}\times(\vec{b}\times\vec{c}) =a_ib^jc^p\epsilon_{jpk}\epsilon^{qik}\vec{g}_q =a_ib^jc^p(\delta^q_j\delta^i_p-\delta^q_p\delta^i_j)\vec{g}_q =(a_ic^i)(b^j\vec{g}_j)-(a_ib^i)(c^q\vec{g}_q) =(\vec{a}\bullet\vec{c})\vec{b}-(\vec{a}\bullet\vec{b})\vec{c}
(a×b)×c=aibjcpϵijkϵpqkgq=aibjcp(δpiδqj−δqiδpj)gq=(aici)(bjgj)−(bjcj)aigi=(a∙c)b−(b∙c)a a×(b×c)=aibjcpϵjpkϵqikgq=aibjcp(δjqδpi−δpqδji)gq=(aici)(bjgj)−(aibi)(cqgq)=(a∙c)b−(a∙b)c
1.7.4. 张量的叉积
类似于矢量的叉积同样可以定义张量的叉积,以二阶张量为例进行演算来说明问题:
U
=
T
×
S
=
T
i
j
g
⃗
i
g
⃗
j
×
S
p
q
g
⃗
p
g
⃗
q
=
T
i
j
S
p
q
ϵ
j
p
k
g
⃗
i
g
⃗
k
g
⃗
q
=
U
∙
k
i
∙
q
g
⃗
i
g
⃗
k
g
⃗
q
\bold{U} =\bold{T}\times\bold{S} =T^{ij}\vec{g}_i\vec{g}_j\times S^{pq}\vec{g}_p\vec{g}_q =T^{ij}S^{pq}\epsilon_{jpk}\vec{g}_i\vec{g}^k\vec{g}_q =U^{i\bullet q}_{\bullet k}\vec{g}_i\vec{g}^k\vec{g}_q
U=T×S=Tijgigj×Spqgpgq=TijSpqϵjpkgigkgq=U∙ki∙qgigkgq叉积得到的物理量仍为张量,因其满足坐标转换关系:
U
∙
k
′
i
′
∙
q
′
=
T
i
′
j
′
S
p
′
q
′
ϵ
j
′
p
′
k
′
=
(
β
a
i
′
β
b
j
′
T
a
b
)
(
β
c
p
′
β
d
q
′
S
c
d
)
(
β
j
′
e
β
p
′
f
β
k
′
h
ϵ
e
f
h
)
=
β
a
i
′
β
d
q
′
β
k
′
h
δ
b
e
δ
c
f
T
a
b
S
c
d
ϵ
e
f
h
=
β
a
i
′
β
d
q
′
β
k
′
h
T
a
b
S
c
d
ϵ
b
c
h
=
β
a
i
′
β
d
q
′
β
k
′
h
U
∙
h
a
∙
d
U^{i'\bullet q'}_{\bullet k'} =T^{i'j'}S^{p'q'}\epsilon_{j'p'k'} =(\beta^{i'}_{a}\beta^{j'}_{b}T^{ab})(\beta^{p'}_{c}\beta^{q'}_{d}S^{cd})(\beta_{j'}^{e}\beta_{p'}^{f}\beta_{k'}^{h}\epsilon_{efh}) =\beta^{i'}_{a}\beta^{q'}_{d}\beta_{k'}^{h}\delta_{b}^{e}\delta_{c}^{f}T^{ab}S^{cd}\epsilon_{efh} =\beta^{i'}_{a}\beta^{q'}_{d}\beta_{k'}^{h}T^{ab}S^{cd}\epsilon_{bch} =\beta^{i'}_{a}\beta^{q'}_{d}\beta_{k'}^{h}U^{a\bullet d}_{\bullet h}
U∙k′i′∙q′=Ti′j′Sp′q′ϵj′p′k′=(βai′βbj′Tab)(βcp′βdq′Scd)(βj′eβp′fβk′hϵefh)=βai′βdq′βk′hδbeδcfTabScdϵefh=βai′βdq′βk′hTabScdϵbch=βai′βdq′βk′hU∙ha∙d显然,m阶张量与n阶张量的叉积结果为(m+n-1)阶张量。
1.8. 张量的商法则
若某有序数组与任意p阶张量的“内积”均为q阶张量则该有序数组应为(p+q)阶张量。以下面的情形来说明该法则的正确性:
设一个五指标的数组与任意二阶张量点积均可得到三阶张量,即:
T
(
i
,
j
,
k
,
m
,
n
)
S
m
n
=
P
i
j
k
T(i,j,k,m,n)S^{mn}=P^{ijk}
T(i,j,k,m,n)Smn=Pijk则
T
(
i
′
,
j
′
,
k
′
,
m
′
,
n
′
)
S
m
′
n
′
=
P
i
′
j
′
k
′
=
β
p
i
′
β
q
j
′
β
l
k
′
P
p
q
l
=
β
p
i
′
β
q
j
′
β
l
k
′
T
(
p
,
q
,
l
,
u
,
v
)
S
u
v
=
β
p
i
′
β
q
j
′
β
l
k
′
β
m
′
u
β
n
′
v
T
(
p
,
q
,
l
,
u
,
v
)
S
m
′
n
′
T(i',j',k',m',n')S^{m'n'} =P^{i'j'k'} =\beta^{i'}_{p}\beta^{j'}_q\beta^{k'}_{l}P^{pql} =\beta^{i'}_{p}\beta^{j'}_q\beta^{k'}_{l}T(p,q,l,u,v)S^{uv} =\beta^{i'}_{p}\beta^{j'}_q\beta^{k'}_{l}\beta_{m'}^{u}\beta_{n'}^{v}T(p,q,l,u,v)S^{m'n'}
T(i′,j′,k′,m′,n′)Sm′n′=Pi′j′k′=βpi′βqj′βlk′Ppql=βpi′βqj′βlk′T(p,q,l,u,v)Suv=βpi′βqj′βlk′βm′uβn′vT(p,q,l,u,v)Sm′n′则
[
T
(
i
′
,
j
′
,
k
′
,
m
′
,
n
′
)
−
β
p
i
′
β
q
j
′
β
l
k
′
β
m
′
u
β
n
′
v
T
(
p
,
q
,
l
,
u
,
v
)
]
S
m
′
n
′
=
0
[T(i',j',k',m',n')-\beta^{i'}_{p}\beta^{j'}_q\beta^{k'}_{l}\beta_{m'}^{u}\beta_{n'}^{v}T(p,q,l,u,v)]S^{m'n'}=0
[T(i′,j′,k′,m′,n′)−βpi′βqj′βlk′βm′uβn′vT(p,q,l,u,v)]Sm′n′=0由
S
m
′
n
′
S^{m'n'}
Sm′n′的任意性可得:
T
(
i
′
,
j
′
,
k
′
,
m
′
,
n
′
)
=
β
p
i
′
β
q
j
′
β
l
k
′
β
m
′
u
β
n
′
v
T
(
p
,
q
,
l
,
u
,
v
)
T(i',j',k',m',n')=\beta^{i'}_{p}\beta^{j'}_q\beta^{k'}_{l}\beta_{m'}^{u}\beta_{n'}^{v}T(p,q,l,u,v)
T(i′,j′,k′,m′,n′)=βpi′βqj′βlk′βm′uβn′vT(p,q,l,u,v)说明
T
\bold T
T 满足坐标变换关系,故其应为五阶张量。
2. 张量的转置
2.1. 张量的转置运算
保持某张量的基矢量排列次序不变而调换张量分量的指标前后次序且不改变指标的协逆变性质可得到一个新的张量,该新张量称作原张量的转置张量。显然,对于高阶张量,对不同指标进行转置结果不同。
2.2. 张量的对称化与反对称化
- 对某张量的两指标 i , j i,j i,j进行转置运算,若原张量与转置张量相等,则称原张量对指标 i , j i,j i,j具有对称性,或者说关于指标 i , j i,j i,j,原张量为对称张量;
- 对某张量的两指标 i , j i,j i,j进行转置运算,若原张量与转置张量相差一负号,则称原张量对指标 i , j i,j i,j具有反对称性,或者说关于指标 i , j i,j i,j,原张量为反对称张量;
根据张量的反对称性,可推知:反对称张量的对角元素为零,所谓张量的对角元素是指两个协逆变性质相同的指标取值相同时所对应的一组张量分量。
任取张量 T \bold{T} T,对于指标 i , j i,j i,j进行转置运算得到张量 S \bold{S} S,那么张量 P = T + S 2 ( 张量的对称化运算 ) \bold{P}=\frac{\bold{T}+\bold{S}}{2}\qquad(张量的对称化运算) P=2T+S(张量的对称化运算)关于指标 i , j i,j i,j为一对称张量。而张量 Q = T − S 2 ( 张量的反对称化运算 ) \bold{Q}=\frac{\bold{T}-\bold{S}}{2}\qquad(张量的反对称化运算) Q=2T−S(张量的反对称化运算)关于指标 i , j i,j i,j为一反对称张量。
证明:因
S
i
j
=
T
j
i
S^{ij}=T^{ji}
Sij=Tji故
(
P
i
j
)
T
=
[
T
i
j
+
S
i
j
2
]
T
=
T
j
i
+
S
j
i
2
=
T
j
i
+
T
i
j
2
=
S
i
j
+
T
i
j
2
=
P
i
j
(P^{ij})^T =\left[\frac{T^{ij}+S^{ij}}{2}\right]^T =\frac{T^{ji}+S^{ji}}{2} =\frac{T^{ji}+T^{ij}}{2} =\frac{S^{ij}+T^{ij}}{2} =P^{ij}
(Pij)T=[2Tij+Sij]T=2Tji+Sji=2Tji+Tij=2Sij+Tij=Pij同理
(
Q
i
j
)
T
=
[
T
i
j
−
S
i
j
2
]
T
=
T
j
i
−
S
j
i
2
=
T
j
i
−
T
i
j
2
=
S
i
j
−
T
i
j
2
=
−
T
i
j
−
S
i
j
2
=
−
Q
i
j
(Q^{ij})^T =\left[\frac{T^{ij}-S^{ij}}{2}\right]^T =\frac{T^{ji}-S^{ji}}{2} =\frac{T^{ji}-T^{ij}}{2} =\frac{S^{ij}-T^{ij}}{2} =-\frac{T^{ij}-S^{ij}}{2} =-Q^{ij}
(Qij)T=[2Tij−Sij]T=2Tji−Sji=2Tji−Tij=2Sij−Tij=−2Tij−Sij=−Qij另外
T
=
P
+
Q
\bold{T=P+Q}
T=P+Q即任意张量关于任意两指标可分解为对称部分与反对称部分。