Section 1.6 1.6 Vector Calculus 1 - Differentiation Calculus involving vectors is discussed in this section, rather intuitively at first and more formally toward the end of this section. 1.6.1 The Ordinary Calculus Consider a scalar-valued function of a scalar, for example the time-dependent density of a material (t) . The calculus of scalar valued functions of scalars is just the ordinary calculus. Some of the important concepts of the ordinary calculus are reviewed in Appendix B to this Chapter, §1.B.2. 1.6.2 Vector-valued Functions of a scalar Consider a vector-valued function of a scalar, for example the time-dependent displacement of a particle u u(t) . In this case, the derivative is defined in the usual way, du u(t t) u(t) lim , dt t0 t which turns out to be simply the derivative of the coefficients1, du du du du du 1 e 2 e 3 e i e dt dt 1 dt 2 dt 3 dt i Partial derivatives can also be defined in the usual way. For example, if u is a function of the coordinates, u(x1 , x2 , x3 ) , then u u(x x , x , x ) u(x , x , x ) lim 1 1 2 3 1 2 3 x1 0 x1 x1 Differentials of vectors are also defined in the usual way, so that when u1 , u2 , u3 undergo increments du1 u1, du2 u2 , du3 u3 , the differential of u is du du1e1 du2e 2 du3e3 and the differential and actual increment u approach one another as u1, u2 , u3 0 . 1 assuming that the base vectors do not depend on t Solid Mechanics Part III 30 Kelly Section 1.6 Space Curves The derivative of a vector can be interpreted geometrically as shown in Fig. 1.6.1: u is the increment in u consequent upon an increment t in t. As t changes, the end-point of the vector u(t) traces out the dotted curve shown – it is clear that as t 0 , u approaches the tangent to , so that du / dt is tangential to . The unit vector tangent to the curve is denoted by τ : du / dt τ (1.6.1) du / dt τ s s x u 2 ds du2 u(t) u(t t) du1 x1 (a) (b) Figure 1.6.1: a space curve; (a) the tangent vector, (b) increment in arc length Let s be a measure of the length of the curve , measured from some fixed point on . Let s be the increment in arc-length corresponding to increments in the coordinates, T u u1 , u2 , u3 , Fig. 1.6.1b. Then, from the ordinary calculus (see Appendix 1.B), 2 2 2 2 ds du1 du2 du3 so that 2 2 2 ds du du du 1 2 3 dt dt dt dt But du du du du 1 e 2 e 3 e dt dt 1 dt 2 dt 3 so that du ds (1.6.2) dt dt Solid Mechanics Part III 31 Kelly Section 1.6 Thus the unit vector tangent to the curve can be written as du / dt du τ (1.6.3) ds / dt ds If u is interpreted as the position vector of a particle and t is interpreted as time, then v du / dt is the velocity vector of the particle as it moves with speed ds / dt along . Example (of particle motion) 2 2 A particle moves along a curve whose parametric equations are x1 2t , x2 t 4t , x3 3t 5 where t is time. Find the component of the velocity at time t 1 in the direction a e1 3e 2 2e3 . Solution The velocity is dr d v 2t 2e t 2 4t e 3t 5 e dt dt 1 2 3 4e1 2e 2 3e3 at t 1 The component in the given direction is v aˆ , where aˆ is a unit vector in the direction of a, giving 8 14 / 7 . ■ Curvature The scalar curvature (s) of a space curve is defined to be the magnitude of the rate of change of the unit tangent vector: dτ d 2u (s) (1.6.4) ds ds 2 Note that τ is in a direction perpendicular to τ , Fig. 1.6.2. In fact, this can be proved as follows: since τ is a unit vector, τ τ is a constant ( 1), and so dτ τ/ ds 0 , but also, d dτ τ τ 2τ ds ds and so τ and dτ / ds are perpendicular. The unit vector defined in this way is called the principal normal vector: Solid Mechanics Part III 32 Kelly Section 1.6 1 dτ ν (1.6.5) ds 1 s R (s) ν(s) ν(s ds) s τ(s) τ τ(s ds) Figure 1.6.2: the curvature This can be seen geometrically in Fig. 1.6.2: from Eqn. 1.6.5, τ is a vector of magnitude s in the direction of the vector normal to τ . The radius of curvature R is defined as the reciprocal of the curvature; it is the radius of the circle which just touches the curve at s, Fig. 1.6.2. Finally, the unit vector perpendicular to both the tangent vector and the principal normal vector is called the unit binormal vector: b τ ν (1.6.6) The planes defined by these vectors are shown in Fig. 1.6.3; they are called the rectifying plane, the normal plane and the osculating plane. ν Osculating Normal plane plane τ Rectifying plane b Figure 1.6.3: the unit tangent, principal normal and binormal vectors and associated planes Solid Mechanics Part III 33 Kelly Section 1.6 Rules of Differentiation The derivative of a vector is also a vector and the usual rules of differentiation apply, d du dv u v dt dt dt (1.6.7) d dv d (t)v v dt dt dt Also, it is straight forward to show that {▲Problem 2} d da dv d da dv v a v a v a v a (1.6.8) dt dt dt dt dt dt (The order of the terms in the cross-product expression is important here.) 1.6.3 Fields In many applications of vector calculus, a scalar or vector can be associated with each point in space x. In this case they are called scalar or vector fields. For example (x) temperature a scalar field (a scalar-valued function of position) v(x) velocity a vector field (a vector valued function of position) These quantities will in general depend also on time, so that one writes (x,t) or v(x,t) . Partial differentiation of scalar and vector fields with respect to the variable t is symbolised by / t . On the other hand, partial differentiation with respect to the coordinates is symbolised by / xi . The notation can be made more compact by introducing the subscript comma to denote partial differentiation with respect to the 2 coordinate variables, in which case ,i / xi , ui, jk ui / x j xk , and so on. 1.6.4 The Gradient of a Scalar Field Let (x) be a scalar field. The gradient of is a vector field defined by (see Fig. 1.6.4) e1 e 2 e3 x1 x2 x3 ei Gradient of a Scalar Field (1.6.9) xi x The gradient is of considerable importance because if one takes the dot product of with dx , it gives the increment in : Solid Mechanics Part III 34 Kelly Section 1.6 dx ei dx j e j xi dxi (1.6.10) xi d (x dx) (dx) dx x Figure 1.6.4: the gradient of a vector If one writes dx as dxe dxe , where e is a unit vector in the direction of dx, then dd e (1.6.11) dx in e direction dn This quantity is called the directional derivative of , in the direction of e, and will be discussed further in §1.6.11. The gradient of a scalar field is also called the scalar gradient, to distinguish it from the vector gradient (see later)2, and is also denoted by grad (1.6.12) Example (of the Gradient of a Scalar Field) 2 2 Consider a two-dimensional temperature field x1 x2 . Then 2x1e1 2x2e 2 For example, at (1,0) , 1, 2e1 and at (1,1) , 2 , 2e1 2e2 , Fig. 1.6.5. Note the following: (i) points in the direction normal to the curve const. (ii) the direction of maximum rate of change of is in the direction of 2 in this context, a gradient is a derivative with respect to a position vector, but the term gradient is used more generally than this, e.g. see §1.14 Solid Mechanics Part III 35 Kelly Section 1.6 (iii) the direction of zero d is in the direction perpendicular to (1,1) 1 (1,0) 2 Figure 1.6.5: gradient of a temperature field The curves x1, x2 const. are called isotherms (curves of constant temperature). In general, they are called iso-curves (or iso-surfaces in three dimensions). ■ Many physical laws are given in terms of the gradient of a scalar field. For example, Fourier’s law of heat conduction relates the heat flux q (the rate at which heat flows through a surface of unit area3) to the temperature gradient through q k (1.6.13) where k is the thermal conductivity of the material, so that heat flows along the direction normal to the isotherms.