Skip to content

Material Derivative

Intuition

Stand on a riverbank with a thermometer: you record how the water temperature at your spot changes. Now jump in and drift: you feel something different — the water around you changes both because the river is cooling overall and because the current carries you into warmer or colder patches. The material derivative \(D/Dt\) is the rate of change experienced by a moving fluid particle: local change plus the change you "run into".

Formal definition

For any field \(f(\mathbf{x}, t)\) (scalar or vector):

\[\boxed{\,\frac{Df}{Dt} = \frac{\partial f}{\partial t} + (\mathbf{v}\cdot\nabla) f\,}\]
  • \(\partial f/\partial t\) — local (unsteady) change at a fixed point
  • \((\mathbf{v}\cdot\nabla)f\) — convective change from being carried through spatial gradients

Applied to velocity itself, it gives the acceleration of a fluid particle:

\[\frac{D\mathbf{v}}{Dt} = \frac{\partial \mathbf{v}}{\partial t} + (\mathbf{v}\cdot\nabla)\mathbf{v}\]

— the left-hand side of both Euler's equation and Navier–Stokes. The convective term is nonlinear in \(\mathbf{v}\), and that single fact is the root of almost all difficulty in fluid mechanics, from instability to turbulence.

Worked example: acceleration in a converging nozzle

Incompressible, inviscid fluid enters a converging nozzle with speed \(u\) through area \(A\) and leaves through area \(\alpha < A\) over a length \(\Delta x\). Find the axial acceleration.

Steady flow ⇒ \(\partial u/\partial t = 0\); on the axis only \(u\,\partial u/\partial x\) survives:

\[a_x = u_\text{avg}\frac{\Delta u}{\Delta x}, \qquad u_\text{avg} = \frac{u + v}{2}, \quad v = \frac{uA}{\alpha}\ \text{(continuity)}\]
\[\boxed{\,a_x = \frac{u^2}{2\Delta x}\left[\left(\frac{A}{\alpha}\right)^2 - 1\right]}\]

Nothing about the flow changes in time, yet particles accelerate hard — a pure convective effect. (The lecture notes verify the same result with a direct \(\Delta u/\Delta t\) argument.)

Physical interpretation

\(D/Dt\) is the Lagrangian rate of change written in Eulerian variables — the infinitesimal version of the Reynolds transport theorem. It lets us apply Newton's laws (which follow matter) while computing with fields (which live on a grid).

Common mistakes

  • Dropping the convective term in steady flow. Steady means \(\partial/\partial t = 0\), not \(D/Dt = 0\).
  • Treating \((\mathbf{v}\cdot\nabla)\mathbf{v}\) as \(\mathbf{v}(\nabla\cdot\mathbf{v})\). The operator \(\mathbf{v}\cdot\nabla = u\partial_x + v\partial_y + w\partial_z\) acts component-wise on the vector that follows.
  • Sign/order confusion: it is \(\mathbf{v}\cdot\nabla f\), the velocity dotted with the gradient — "how fast you move through the field's slope".

Knowledge graph position

Prerequisites: Eulerian vs Lagrangian. Leads to: Reynolds transport theorem, Euler's equation.

Quiz

Q1 (computational). In the 1-D steady field \(u(x) = u_0(1 + x/L)\), what is the acceleration of a particle at \(x = L\)?

Answer

\(a = u\,\partial u/\partial x = u_0(1 + 1)\cdot(u_0/L) = 2u_0^2/L\).

Q2 (conceptual). Can \(Df/Dt = 0\) while \(\partial f/\partial t \neq 0\)? Give an example.

Answer

Yes — a "frozen" pattern advected by the flow: \(f(x,t) = F(x - ut)\) satisfies \(\partial f/\partial t = -u\,\partial f/\partial x \neq 0\) but \(Df/Dt = 0\). Particles carry constant \(f\); the field at a fixed point still changes as the pattern sweeps by.

Q3 (multiple choice). The nonlinearity of the Navier–Stokes equation comes from:

  • (a) the viscous term \(\mu\nabla^2\mathbf{v}\)
  • (b) the pressure gradient
  • (c) the convective term \((\mathbf{v}\cdot\nabla)\mathbf{v}\)
  • (d) gravity
Answer

(c). It is quadratic in velocity; all other terms are linear in \(\mathbf{v}\) or \(P\).