LECTURE 5.
VECTOR-
TENSOR ANALYSIS PART 3
Transport Phenomena
CHEB417-01
2024 Spring Semester / MW 11:00 ~ 12:15
Summary of the previous lecture
• Index notation
• Index notation concept
• Summation convention (Einstein convention)
• Preliminary notions
• Vector operations with index notation
3
(1) Dot product of two vectors 𝐚𝐚 � 𝐛𝐛 = 𝑎𝑎𝑖𝑖 𝑏𝑏𝑖𝑖 → � 𝑎𝑎𝑖𝑖 𝑏𝑏𝑖𝑖
𝑖𝑖=1
(2) Cross product of two vectors 𝐚𝐚 × 𝐛𝐛 = 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝐞𝐞𝑘𝑘
(3) Triple product of three vectors 𝐚𝐚 � 𝐛𝐛 × 𝐜𝐜 = 𝐛𝐛 � 𝐜𝐜 × 𝐚𝐚 = 𝐜𝐜 � 𝐚𝐚 × 𝐛𝐛 = 𝜖𝜖𝑖𝑖𝑗𝑗𝑗𝑗 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘
(4) One important formula for the product of two permutation symbols
𝜖𝜖𝑖𝑖𝑖𝑖𝑘𝑘 𝜖𝜖𝑙𝑙𝑙𝑙𝑘𝑘 = 𝛿𝛿𝑖𝑖𝑖𝑖 𝛿𝛿𝑗𝑗𝑗𝑗 − 𝛿𝛿𝑖𝑖𝑚𝑚 𝛿𝛿𝑗𝑗𝑙𝑙
• In LHS, 𝑘𝑘 is the dummy index and
3
𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝜖𝜖𝑙𝑙𝑙𝑙𝑙𝑙 → � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝜖𝜖𝑙𝑙𝑙𝑙𝑙𝑙
𝑘𝑘=1
• The rigorous proof of the above is quite difficult and beyond the scope
• Instead, let’s check the formula for the case 𝑖𝑖 = 1, 𝑗𝑗 = 2, 𝑙𝑙 = 2, 𝑚𝑚 = 1.
3
(LHS) � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝜖𝜖𝑙𝑙𝑙𝑙𝑙𝑙 = 𝜖𝜖𝑖𝑖𝑖𝑖1 𝜖𝜖𝑙𝑙𝑙𝑙1 + 𝜖𝜖𝑖𝑖𝑖𝑖2 𝜖𝜖𝑙𝑙𝑙𝑙2 + 𝜖𝜖𝑖𝑖𝑖𝑖3 𝜖𝜖𝑙𝑙𝑙𝑙3
𝑘𝑘=1
= 𝜖𝜖123 𝜖𝜖213 = 1 −1 = −1
(RHS) 𝛿𝛿𝑖𝑖𝑖𝑖 𝛿𝛿𝑗𝑗𝑗𝑗 − 𝛿𝛿𝑖𝑖𝑖𝑖 𝛿𝛿𝑗𝑗𝑗𝑗 = 𝛿𝛿12 𝛿𝛿21 − 𝛿𝛿11 𝛿𝛿22 = −1
(Note) Try other cases. The above formula can be used to prove the
following important vector identity
(5) 𝐚𝐚 × 𝐛𝐛 × 𝐜𝐜 = 𝐚𝐚 � 𝐜𝐜 𝐛𝐛 − 𝐚𝐚 � 𝐛𝐛 𝐜𝐜
𝐚𝐚 × 𝐛𝐛 × 𝐜𝐜 = 𝑎𝑎𝑖𝑖 𝐞𝐞𝑖𝑖 × 𝑏𝑏𝑗𝑗 𝐞𝐞𝑗𝑗 × 𝑐𝑐𝑘𝑘 𝐞𝐞𝑘𝑘
= 𝑎𝑎𝑖𝑖 𝐞𝐞𝑖𝑖 × 𝜖𝜖𝑗𝑗𝑗𝑗𝑗𝑗 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑙𝑙
= 𝜖𝜖𝑗𝑗𝑗𝑗𝑗𝑗 𝜖𝜖𝑖𝑖𝑙𝑙𝑚𝑚 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚
= 𝜖𝜖𝑗𝑗𝑗𝑗𝑙𝑙 𝜖𝜖𝑚𝑚𝑚𝑚𝑙𝑙 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚
= 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 − 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚
= 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚 − 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚
= 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑖𝑖 𝐞𝐞𝑗𝑗 − 𝑎𝑎𝑖𝑖 𝑏𝑏𝑖𝑖 𝑐𝑐𝑘𝑘 𝐞𝐞𝑘𝑘
= 𝑎𝑎𝑖𝑖 𝑐𝑐𝑖𝑖 𝑏𝑏𝑗𝑗 𝐞𝐞𝑗𝑗 − 𝑎𝑎𝑖𝑖 𝑏𝑏𝑖𝑖 𝑐𝑐𝑘𝑘 𝐞𝐞𝑘𝑘
= 𝐚𝐚 � 𝐜𝐜 𝐛𝐛 − 𝐚𝐚 � 𝐛𝐛 𝐜𝐜
• In the previously slide derivation,
𝐚𝐚 × 𝐛𝐛 × 𝐜𝐜 = 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚 − 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚
3 3 3 3
→ � � � � 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚 − 𝛿𝛿𝑗𝑗𝑗𝑗 𝛿𝛿𝑘𝑘𝑘𝑘 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑘𝑘 𝐞𝐞𝑚𝑚
𝑖𝑖=1 𝑗𝑗=1 𝑘𝑘=1 𝑚𝑚=1
3 3 3 3
= � � 𝑎𝑎𝑖𝑖 𝑏𝑏𝑗𝑗 𝑐𝑐𝑖𝑖 𝐞𝐞𝑗𝑗 − � � 𝑎𝑎𝑖𝑖 𝑏𝑏𝑖𝑖 𝑐𝑐𝑘𝑘 𝐞𝐞𝑘𝑘
𝑖𝑖=1 𝑗𝑗=1 𝑖𝑖=1 𝑘𝑘=1
3 3 3 3
= � 𝑎𝑎𝑖𝑖 𝑐𝑐𝑖𝑖 � 𝑏𝑏𝑗𝑗 𝐞𝐞𝑗𝑗 − � 𝑎𝑎𝑖𝑖 𝑏𝑏𝑖𝑖 � 𝑐𝑐𝑘𝑘 𝐞𝐞𝑘𝑘
𝑖𝑖=1 𝑗𝑗=1 𝑖𝑖=1 𝑘𝑘=1
= 𝐚𝐚 � 𝐜𝐜 𝐛𝐛 − 𝐚𝐚 � 𝐛𝐛 𝐜𝐜
Vector operations with index notation (cont’d)
(6) Position vector in Cartesian coordinate system
:Straight line having one end fixed to a body and the other end attached to a
moving point and used to describe the position of the point relative to the body
𝐱𝐱 = 𝑥𝑥𝑖𝑖 𝐞𝐞𝑖𝑖 → � 𝑥𝑥𝑖𝑖 𝐞𝐞𝑖𝑖 = 𝑥𝑥1 𝐞𝐞1 + 𝑥𝑥2 𝐞𝐞2 + 𝑥𝑥3 𝐞𝐞3
𝑖𝑖=1
(7) Gradient of a scalar function (temperature, concentration)
• Gradient vector is a operator defined as,
3
𝜕𝜕 𝜕𝜕 𝜕𝜕 𝜕𝜕 𝜕𝜕
𝛁𝛁 = 𝐞𝐞𝑖𝑖 → � 𝐞𝐞𝑖𝑖 = 𝐞𝐞1 + 𝐞𝐞2 + 𝐞𝐞3
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3
𝑖𝑖=1
• Gradient of a scalar is thus,
3
𝜕𝜕𝜙𝜙 𝜕𝜕𝜙𝜙 𝜕𝜕𝜙𝜙 𝜕𝜕𝜙𝜙 𝜕𝜕𝜙𝜙
𝛁𝛁𝜙𝜙 = 𝐞𝐞𝑖𝑖 → � 𝐞𝐞𝑖𝑖 = 𝐞𝐞1 + 𝐞𝐞2 + 𝐞𝐞3
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3
𝑖𝑖=1
(8) Divergence of a vector
𝜕𝜕 𝜕𝜕 𝑎𝑎𝑗𝑗 𝐞𝐞𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗
𝛁𝛁 � 𝐚𝐚 = 𝐞𝐞𝑖𝑖 � 𝑎𝑎𝑗𝑗 𝐞𝐞𝑗𝑗 = 𝐞𝐞𝑖𝑖 � = 𝐞𝐞𝑖𝑖 � 𝐞𝐞𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
3
𝜕𝜕𝑎𝑎𝑗𝑗 𝜕𝜕𝑎𝑎𝑖𝑖 𝜕𝜕𝑎𝑎𝑖𝑖 𝜕𝜕𝑎𝑎1 𝜕𝜕𝑎𝑎2 𝜕𝜕𝑎𝑎3
= 𝛿𝛿𝑖𝑖𝑖𝑖 = →� = + +
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3
𝑖𝑖=1
𝜕𝜕
• In the above, the component of a gradient vector is a scalar operator
𝜕𝜕𝑥𝑥𝑖𝑖
and it can cross over the dot product operator (�) but it must act on the
whole thing (𝑎𝑎𝑗𝑗 𝐞𝐞𝑗𝑗 )
𝜕𝜕 𝑎𝑎𝑗𝑗 𝐞𝐞𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗
= 𝐞𝐞𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
• We also used the relation above because 𝐞𝐞𝑗𝑗 are the independent of the
position in the Cartesian coordinate system (the magnitude and the
direction are not changed)
• However, if we consider a curvilinear coordinate system such as
cylindrical and spherical systems, the unit vectors are not independent of
the position because the unit vectors change their direction as shown
below.
𝜕𝜕 𝑎𝑎𝑗𝑗 𝐞𝐞𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗
≠ 𝐞𝐞𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
in curvilinear coordinate system
• The divergence of a vector field is the extent to which the vector field flux
behaves like a source at a given point. It is a local measure of its
"outgoingness" – the extent to which there is more of the field vectors
exiting an infinitesimal region of space than entering it.
a vector field is an assignment of a vector to each
point in a subset of space
𝐯𝐯 𝑥𝑥, 𝑦𝑦
𝜕𝜕𝑣𝑣𝑥𝑥 𝑥𝑥, 𝑦𝑦 𝜕𝜕𝑣𝑣𝑦𝑦 𝑥𝑥, 𝑦𝑦
𝛁𝛁 � 𝐯𝐯 𝑥𝑥, 𝑦𝑦 = +
𝜕𝜕𝑥𝑥 𝜕𝜕𝑦𝑦
𝜕𝜕𝑣𝑣1 𝜕𝜕𝑣𝑣2 𝜕𝜕𝑣𝑣3
𝛁𝛁 � 𝐯𝐯 = + +
𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3
en.wikipedia.org
(9) Curl of a vector
𝜕𝜕
𝛁𝛁 × 𝐚𝐚 = 𝐞𝐞𝑖𝑖 × 𝑎𝑎𝑗𝑗 𝐞𝐞𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖
𝜕𝜕 𝑎𝑎𝑗𝑗 𝐞𝐞𝑗𝑗
= 𝐞𝐞𝑖𝑖 ×
𝜕𝜕𝑥𝑥𝑖𝑖
www.dvo.com
𝜕𝜕𝑎𝑎𝑗𝑗
= 𝐞𝐞𝑖𝑖 × 𝐞𝐞𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖
𝜕𝜕𝑎𝑎𝑗𝑗
= 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞𝑘𝑘
𝜕𝜕𝑥𝑥𝑖𝑖
3 3 3
𝜕𝜕𝑎𝑎𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗
= 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞𝑘𝑘 → � � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝑘𝑘
𝑖𝑖=1 𝑗𝑗=1 𝑘𝑘=1
3 3
𝜕𝜕𝑎𝑎𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗 𝜕𝜕𝑎𝑎𝑗𝑗
= 𝜖𝜖𝑖𝑖𝑖𝑖1 𝐞𝐞1 + 𝜖𝜖𝑖𝑖𝑖𝑖2 𝐞𝐞2 + 𝜖𝜖𝑖𝑖𝑖𝑖3 𝐞𝐞3 → � � 𝜖𝜖 𝑖𝑖𝑖𝑖𝑖 𝐞𝐞1 + 𝜖𝜖 𝑖𝑖𝑖𝑖𝑖 𝐞𝐞 2 + 𝜖𝜖 𝑖𝑖𝑖𝑖𝑖 𝐞𝐞3
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
𝑖𝑖=1 𝑗𝑗=1
𝜕𝜕𝑎𝑎3 𝜕𝜕𝑎𝑎2 𝜕𝜕𝑎𝑎1 𝜕𝜕𝑎𝑎3 𝜕𝜕𝑎𝑎2 𝜕𝜕𝑎𝑎1
= − 𝐞𝐞 + − 𝐞𝐞 + − 𝐞𝐞
𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3 1 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥1 2 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2 3
(10) Laplacian operator (𝛁𝛁2 = 𝛁𝛁 � 𝛁𝛁)
𝜕𝜕 𝜕𝜕 𝜕𝜕 𝜕𝜕
𝛁𝛁 2 = 𝛁𝛁 � 𝛁𝛁 = 𝐞𝐞𝑖𝑖 � 𝐞𝐞𝑗𝑗 = 𝐞𝐞𝑖𝑖 � 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖 𝑗𝑗 𝜕𝜕𝑥𝑥𝑗𝑗
𝜕𝜕 2 𝜕𝜕 2 𝜕𝜕 2
= (𝐞𝐞𝑖𝑖 � 𝐞𝐞𝑗𝑗 ) = 𝛿𝛿𝑖𝑖𝑖𝑖 =
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
3
𝜕𝜕 2 𝜕𝜕 2 𝜕𝜕 2 𝜕𝜕 2 𝜕𝜕 2
= →� = + +
𝜕𝜕𝑥𝑥𝑖𝑖 2 𝜕𝜕𝑥𝑥𝑖𝑖 2 𝜕𝜕𝑥𝑥1 2 𝜕𝜕𝑥𝑥2 2 𝜕𝜕𝑥𝑥3 2
𝑖𝑖=1
𝜕𝜕2
• In the above, we should note that dummy index 𝑖𝑖 appears twice in
𝜕𝜕𝑥𝑥𝑖𝑖 2
𝜕𝜕2 𝜕𝜕2
because =
𝜕𝜕𝑥𝑥𝑖𝑖 2 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
• The Laplacian of a scalar function is given as,
2 𝜙𝜙 2 𝜙𝜙 2 𝜙𝜙 2 𝜙𝜙
𝜕𝜕 𝜕𝜕 𝜕𝜕 𝜕𝜕
𝛁𝛁 2 𝜙𝜙 = = + +
𝜕𝜕𝑥𝑥𝑖𝑖 2 𝜕𝜕𝑥𝑥1 2 𝜕𝜕𝑥𝑥2 2 𝜕𝜕𝑥𝑥3 2
(11) 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 = 0 if 𝐴𝐴𝑖𝑖𝑖𝑖 = 𝐴𝐴𝑗𝑗𝑖𝑖
(Proof) Note first that 𝑖𝑖 and 𝑗𝑗 are the dummy indices and summation is
applied and a dummy index can be changed to any other index
3 3
𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 → � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 = 𝜖𝜖𝑗𝑗𝑖𝑖𝑘𝑘 𝐴𝐴𝑗𝑗𝑗𝑗 (𝑖𝑖 → 𝑗𝑗, 𝑗𝑗 → 𝑖𝑖)
𝑖𝑖=1 𝑗𝑗=1
= 𝜖𝜖𝑗𝑗𝑖𝑖𝑘𝑘 𝐴𝐴𝑖𝑖𝑖𝑖 (𝐴𝐴𝑖𝑖𝑖𝑖 = 𝐴𝐴𝑗𝑗𝑗𝑗 )
= −𝜖𝜖𝑖𝑖𝑗𝑗𝑘𝑘 𝐴𝐴𝑖𝑖𝑖𝑖
𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 = −𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 2𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 = 0
• Therefore 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 = 0 if 𝐴𝐴𝑖𝑖𝑖𝑖 = 𝐴𝐴𝑗𝑗𝑗𝑗
(12) Curl of any gradient of scalar is zero
𝛁𝛁 × 𝛁𝛁𝜙𝜙 = 0
𝜕𝜕 𝜕𝜕𝜙𝜙 𝜕𝜕 𝜕𝜕𝜙𝜙
𝛁𝛁 × 𝛁𝛁𝜙𝜙 = 𝐞𝐞𝑖𝑖 × 𝐞𝐞𝑗𝑗 = 𝐞𝐞𝑖𝑖 × 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖 𝑗𝑗 𝜕𝜕𝑥𝑥𝑗𝑗
𝜕𝜕 2 𝜙𝜙
= 𝐞𝐞𝑖𝑖 × 𝐞𝐞𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗
3 3 3 3 3
𝜕𝜕 2 𝜙𝜙 𝜕𝜕 2 𝜙𝜙 𝜕𝜕 2 𝜙𝜙
→ � � 𝜖𝜖𝑖𝑖𝑖𝑖1 𝐞𝐞1 = 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞 → � � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝑘𝑘 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝑘𝑘
𝑖𝑖=1 𝑗𝑗=1 𝑖𝑖=1 𝑗𝑗=1 𝑘𝑘=1
𝜕𝜕 2 𝜙𝜙 𝜕𝜕 2 𝜙𝜙 𝜕𝜕 2 𝜙𝜙 𝜕𝜕 2 𝜙𝜙 𝜕𝜕 2 𝜙𝜙
= (𝜖𝜖231 + 𝜖𝜖321 ) 𝐞𝐞 = 𝜖𝜖𝑖𝑖𝑖𝑖1 𝐞𝐞 + 𝜖𝜖𝑖𝑖𝑖𝑖2 𝐞𝐞 + 𝜖𝜖𝑖𝑖𝑖𝑖3 𝐞𝐞
𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥2 1 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 1 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 2 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 3
𝜕𝜕 2 𝜙𝜙 𝜕𝜕 2 𝜙𝜙
=( − )𝐞𝐞
𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3 1 𝜕𝜕2 𝜙𝜙 𝜕𝜕2 𝜙𝜙
Since = , 𝛁𝛁 × 𝛁𝛁𝜙𝜙 = 0
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖
(13) Divergence of any curl of a vector is zero
𝛁𝛁 � 𝛁𝛁 × 𝐚𝐚 = 0
(9) Curl of a vector
𝜕𝜕 𝜕𝜕𝑎𝑎𝑘𝑘
𝛁𝛁 � 𝛁𝛁 × 𝐚𝐚 = 𝐞𝐞𝑖𝑖 � 𝜖𝜖𝑗𝑗𝑗𝑗𝑙𝑙 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝑙𝑙
𝜕𝜕 2 𝑎𝑎𝑘𝑘
= (𝐞𝐞𝑖𝑖 � 𝐞𝐞𝑙𝑙 )𝜖𝜖𝑗𝑗𝑗𝑗𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗
𝜕𝜕 2 𝑎𝑎𝑘𝑘
= 𝛿𝛿𝑖𝑖𝑙𝑙 𝜖𝜖𝑗𝑗𝑗𝑗𝑗𝑗
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗
3 3 3
𝜕𝜕 2 𝑎𝑎 𝑘𝑘 𝜕𝜕 2 𝑎𝑎 𝑘𝑘 𝜕𝜕 2 𝑎𝑎𝑘𝑘
= 𝜖𝜖𝑗𝑗𝑗𝑗𝑖𝑖 = 𝜖𝜖𝑖𝑖𝑗𝑗𝑗𝑗 → � � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗
𝑖𝑖=1 𝑗𝑗=1 𝑘𝑘=1
𝜕𝜕2 𝑎𝑎𝑘𝑘 𝜕𝜕2 𝑎𝑎𝑘𝑘
Since = , 𝛁𝛁 � 𝛁𝛁 × 𝐚𝐚 = 0
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖
• Let‘s see the significance of the two relations (12) and (13):
𝛁𝛁 × 𝛁𝛁𝜙𝜙 = 0
𝛁𝛁 � 𝛁𝛁 × 𝐚𝐚 = 0
(i) Potential flow: Ideal fluid flow without viscosity effect
• A flow whose velocity 𝐮𝐮 is given by a gradient of a scalar function as,
𝐮𝐮 = 𝛁𝛁𝜙𝜙
(ii) Vorticity vector
• Vorticity (𝛚𝛚) is defined by, 𝛚𝛚 = 𝛁𝛁 × 𝐮𝐮 (9) Curl of a vector
3 3 3
𝜕𝜕𝑢𝑢𝑗𝑗 𝜕𝜕𝑢𝑢
= 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞𝑘𝑘 → � � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝑗𝑗 𝐞𝐞𝑘𝑘
𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
𝑖𝑖=1 𝑗𝑗=1 𝑘𝑘=1
𝜕𝜕𝑢𝑢𝑗𝑗 𝜕𝜕𝑢𝑢𝑗𝑗 𝜕𝜕𝑢𝑢𝑗𝑗
= 𝜖𝜖𝑖𝑖𝑖𝑖1 𝐞𝐞 + 𝜖𝜖𝑖𝑖𝑖𝑖2 𝐞𝐞 + 𝜖𝜖𝑖𝑖𝑖𝑖3 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 1 𝜕𝜕𝑥𝑥𝑖𝑖 2 𝜕𝜕𝑥𝑥𝑖𝑖 3
𝛚𝛚 = 𝛁𝛁 × 𝐮𝐮
𝜕𝜕𝑢𝑢𝑗𝑗 𝜕𝜕𝑢𝑢𝑗𝑗 𝜕𝜕𝑢𝑢𝑗𝑗
= 𝜖𝜖𝑖𝑖𝑖𝑖1 𝐞𝐞 + 𝜖𝜖𝑖𝑖𝑖𝑖2 𝐞𝐞 + 𝜖𝜖𝑖𝑖𝑖𝑖3 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 1 𝜕𝜕𝑥𝑥𝑖𝑖 2 𝜕𝜕𝑥𝑥𝑖𝑖 3
3 3 3 3 3 3
𝜕𝜕𝑢𝑢𝑗𝑗 𝜕𝜕𝑢𝑢𝑗𝑗 𝜕𝜕𝑢𝑢𝑗𝑗
→ � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞 + � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞 + � � 𝜖𝜖𝑖𝑖𝑖𝑖𝑖 𝐞𝐞
𝜕𝜕𝑥𝑥𝑖𝑖 1 𝜕𝜕𝑥𝑥𝑖𝑖 2 𝜕𝜕𝑥𝑥𝑖𝑖 3
𝑖𝑖=1 𝑗𝑗=1 𝑖𝑖=1 𝑗𝑗=1 𝑖𝑖=1 𝑗𝑗=1
= 𝜔𝜔1 𝐞𝐞1 + 𝜔𝜔2 𝐞𝐞2 + 𝜔𝜔3 𝐞𝐞3
𝜕𝜕𝑢𝑢3 𝜕𝜕𝑢𝑢2 𝜕𝜕𝑢𝑢1 𝜕𝜕𝑢𝑢3 𝜕𝜕𝑢𝑢2 𝜕𝜕𝑢𝑢1
𝜔𝜔1 = − 𝜔𝜔2 = − 𝜔𝜔3 = −
𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2
𝛚𝛚 = 𝛁𝛁 × 𝐮𝐮 = 𝜔𝜔1 𝐞𝐞1 + 𝜔𝜔2 𝐞𝐞2 + 𝜔𝜔3 𝐞𝐞3
𝜕𝜕𝑢𝑢3 𝜕𝜕𝑢𝑢2 𝜕𝜕𝑢𝑢1 𝜕𝜕𝑢𝑢3 𝜕𝜕𝑢𝑢2 𝜕𝜕𝑢𝑢1
𝜔𝜔1 = − 𝜔𝜔2 = − 𝜔𝜔3 = −
𝜕𝜕𝑥𝑥2 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥3 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2
• In order to see the meaning, let us see the velocity distribution
𝜕𝜕𝑢𝑢2 𝜕𝜕𝑢𝑢1
• Therefore 𝜔𝜔3 = − means the net rotation in the counter-clockwise
𝜕𝜕𝑥𝑥1 𝜕𝜕𝑥𝑥2
direction
• This direction of counter-clockwise rotation is denoted by the unit
vector 𝐞𝐞3
• Now let us state several things on the potential flow and vorticity:
(i) Potential flow does not have vorticity (irrotational)
𝐮𝐮 = 𝛁𝛁𝜙𝜙 𝛚𝛚 = 𝛁𝛁 × 𝐮𝐮 = 𝛁𝛁 × (𝛁𝛁𝜙𝜙) = 0
→ (12) Curl of any gradient of scalar is zero 𝛁𝛁 × 𝛁𝛁𝜙𝜙 = 0
(ii) Divergence of vorticity is zero (solenoidal)
→ (13) Divergence of any curl of a vector is zero 𝛁𝛁 � 𝛁𝛁 × 𝐚𝐚 = 0
𝛁𝛁 � (𝛚𝛚) = 𝛁𝛁 � (𝛁𝛁 × 𝐮𝐮) = 0
www.mathworks.com
