Академический Документы
Профессиональный Документы
Культура Документы
Thus the transformation matrix A, defined by (18.5), is an orthogonal matrix. For this
reason, the transformation formula (18.4) is called an orthogonal transformation.
is orthogonal.
Solution
⎛ ⎞⎛ ⎞ ⎛ ⎞
4 1 8 4 7 −4 81 0 0
1 ⎝ 1 ⎝
BT · B = 7 4 −4 ⎠ ⎝ 1 4 8 ⎠ = 0 81 0 ⎠ = 1.
81 81
−4 8 1 8 −4 1 0 0 81
Since this example is a special case of the preceding theory, the matrix A in (18.8) must
be orthogonal. [You should check this directly.] A rotation of C about the O x1 or O x2
axis is treated in the same way.
496 Chapter 18 Tensor algebra and the inertia tensor
x3 x 3 n
x 2
x2
M E
ψ
O ψ x1 e
E
x1 O e
Solution
The formula corresponding to (18.8) when C is rotated through an angle ψ about the
axis O x1 is
⎛ ⎞
1 0 0
A = ⎝ 0 cos ψ sin ψ ⎠ .
0 − sin ψ cos ψ
The components of the vector v in C are then given by the elements of the column
vector v , where
⎛ ⎞ ⎛ ⎞⎛ ⎞ ⎛ ⎞
3 1 √ 0 0 3 3√
v = A · ⎝ −4 ⎠ = ⎝ 0 3/2 √1/2 ⎠ ⎝ −4 ⎠ = ⎝ 1 − 2√ 3 ⎠ .
2 0 −1/2 3/2 2 2+ 3
√ √
The components of the vector v in C are therefore {3, 1 − 2 3, 2 + 3 }.
In the general case, the rotation of C can take place about an axis {O, n}, where n is
any unit vector. The argument is messier but we can still find an explicit formula for A.
18.2 Rotated and reflected coordinate systems 497
Figure 18.2 (Right) shows how a typical unit vector e of the coordinate system C is obtained from the
corresponding vector e belonging to C. Let e be resolved into parts parallel and perpendicular to the axis
{O, n} in the form
e = e + e⊥ ,
where
−→ −→
e = O M= (e · n) n, e⊥ = M E= e − (e · n) n.
e = e + e ⊥ ,
If we now take e = e1 , e = e1 and take the scalar product of equation (18.9) successively with e1 , e2 and
e3 , we obtain
where n 1 , n 2 , n 3 are the components of the vector n in the coordinate system C. This gives the elements of
the first row of the transformation matrix A, and the elements of the second and third rows can be found in
a similar manner. The full transformation matrix A is given by
⎛ ⎞
cos ψ + (1 − cos ψ)n 21 (1 − cos ψ)n 1 n 2 + n 3 sin ψ (1 − cos ψ)n 1 n 3 − n 2 sin ψ
⎜ ⎟
⎝ (1 − cos ψ)n 2 n 1 − n 3 sin ψ cos ψ + (1 − cos ψ)n 22 (1 − cos ψ)n 2 n 3 + n 1 sin ψ ⎠ (18.10)
(1 − cos ψ)n 3 n 1 + n 2 sin ψ (1 − cos ψ)n 3 n 2 − n 1 sin ψ cos ψ + (1 − cos ψ)n 23
Solution
−→ √ √ √
The unit vector n in the direction O P has components {1/ 3, 1/ 3, −1/ 3} in C ,
and ψ = 60◦ . It follows from the formula (18.10) that the tranformation matrix from
498 Chapter 18 Tensor algebra and the inertia tensor
C to C is
⎛ ⎞
2 −1 −2
1⎝
A= 2 2 1⎠.
3
1 −2 2
The components of the vector v in C are then given by the elements of the column
vector v , where
⎛ ⎞ ⎛ ⎞⎛ ⎞ ⎛ ⎞
3 2 2 1 3 1
T ⎝ ⎠ 1⎝
v = A · −6 = −1 2 −2 ⎠ ⎝ −6 ⎠ = ⎝ −11 ⎠ .
3
9 −2 1 2 9 2
Reflections
The coordinate system C may also be defined by reflecting C in a plane through O. For
example, if C is reflected in the coordinate plane O x1 x2 , the transformation matrix is
easily found to be
⎛ ⎞
1 0 0
A = ⎝0 1 0⎠.
0 0 −1
In the general case, if C is reflected in the plane through O with unit normal vector n, then
the transformation matrix is
⎛ ⎞
1 − 2n 21 −2n 1 n 2 −2n 1 n 3
A = ⎝ −2n 2 n 1 1 − 2n 22 −2n 2 n 3 ⎠ , (18.11)
−2n 3 n 1 −2n 3 n 2 1 − 2n 32
where n 1 , n 2 , n 3 are the components of the vector n in C . [The method of proof is similar
to that for the general rotation.]
Solution
The equation x3 = 2x1 + 2x2 can be written in the form
It can be proved that the most general orthogonal transformation corresponds either
to (i) a rotation of coordinates, or (ii) a reflection followed by a rotation. Which of these
classes a particular orthogonal transformation belongs to can be decided by examining the
sign of the determinant det A.
If we take determinants of equation (18.7) we obtain
det 1 = det AT · A = det AT × det A = det A × det A = (det A)2 ,
from which it follows that det A can only take the values ±1. Surprisingly, we can work
out the determinant of the general A defined by equation (18.5) quite easily. The first row
of A contains the three components of the vector e1 in the coordinate system C . Likewise,
the second and third rows contain the components of the vectors e2 and e3 . It follows from
the vector formula (1.7) that, provided C is a right-handed coordinate system,
det A = e1 ×e2 · e3 .