Lagrangian and Eulerian descriptions of fluid motion
Lagrangian and Eulerian viewpoints, the material derivative with local and convective acceleration, and converting a velocity field into particle paths, with stagnation-flow and unsteady-flow numericals.
Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.
Why it matters
Every equation of fluid dynamics is written in one of two viewpoints: following a fluid particle (Lagrangian) or watching fixed points in space (Eulerian). Newton's second law applies to particles, but wind tunnels, probes and CFD grids give data at fixed points. The material derivative is the bridge between the two, and it is the source of the convective acceleration that appears in Euler's, Bernoulli's and the Navier–Stokes equations.
Key ideas
Lagrangian description. Label each particle by its starting position r₀ = (x₀, y₀, z₀) at t = 0 and describe its position as r(r₀, t). Velocity and acceleration of that particle are ordinary time derivatives: V = ∂r/∂t and a = ∂²r/∂t² with r₀ held fixed. This is natural for solid mechanics, for tracking droplets, ice crystals, pollutants or a balloon, and for particle-based numerical methods. It becomes awkward for a continuum of billions of particles.
Eulerian description. Describe the flow as fields: V(x, y, z, t), p(x, y, z, t), ρ(x, y, z, t), T(x, y, z, t). You ask "what is the velocity at this point at this time?", not "where is particle 37?". Pitot probes, hot-wire anemometers and grid-based CFD are all Eulerian. A flow is steady if no field quantity at a fixed point changes with time (∂/∂t = 0); particles may still accelerate as they move through a steady field.
Material (substantial) derivative. For any property f carried by a particle, the rate of change following the particle is D f/Dt = ∂f/∂t + (V·∇) f.
- ∂f/∂t is the local rate of change at a fixed point (zero in steady flow).
- (V·∇) f = u ∂f/∂x + v ∂f/∂y + w ∂f/∂z is the convective rate of change, due to the particle moving to a place where f is different. Applied to velocity it gives the particle acceleration. Even in steady flow, a particle passing through a nozzle accelerates because the convective term is non-zero.
Converting between descriptions.
- Eulerian to Lagrangian: solve dx/dt = u(x, y, z, t), dy/dt = v, dz/dt = w with the initial position. The solution r(r₀, t) is the particle path (pathline).
- Lagrangian to Eulerian: compute V = ∂r/∂t, then eliminate r₀ in favour of the current position (x, y, z). Both routes must give the same acceleration, which is a good check.
Where each is used. Continuity, momentum and energy equations are usually written for a fixed control volume (Eulerian) using the Reynolds transport theorem, which converts a system (Lagrangian) law into a control-volume form. Pathlines, streaklines and streamlines, covered next, are defined using these ideas.
Formulas
r = r(r₀, t), V = ∂r/∂t |r₀, a = ∂²r/∂t² |r₀
- r: particle position (m); r₀: label (initial position, m); V: velocity (m/s); a: acceleration (m/s²). Lagrangian.
V = u(x,y,z,t)·i + v(x,y,z,t)·j + w(x,y,z,t)·k
- Eulerian velocity field; u, v, w in m/s.
Df/Dt = ∂f/∂t + u·∂f/∂x + v·∂f/∂y + w·∂f/∂z
- Material derivative of any field f (e.g. T in K, ρ in kg/m³); units of f per second.
a_x = ∂u/∂t + u·∂u/∂x + v·∂u/∂y + w·∂u/∂z (similarly a_y with v, a_z with w)
- Acceleration components (m/s²): first term local, remaining terms convective.
dx/dt = u, dy/dt = v, dz/dt = w
- Particle-path equations; integrate with r = r₀ at t = 0.
Worked examples
Example 1 (standard): acceleration in a stagnation-point flow. Given: V = 2x·i − 2y·j (m/s, x and y in m), steady. Find the acceleration at (1, 2).
a_x = u·∂u/∂x + v·∂u/∂y= (2x)(2) + (−2y)(0) = 4x = 4 m/s².a_y = u·∂v/∂x + v·∂v/∂y= (2x)(0) + (−2y)(−2) = 4y = 8 m/s².- |a| = √(4² + 8²) = 8.94 m/s². Answer: a = 4i + 8j m/s², |a| ≈ 8.94 m/s², even though the flow is steady.
Lagrangian check: dx/dt = 2x gives x = x₀·e^(2t), and dy/dt = −2y gives y = y₀·e^(−2t). Then d²x/dt² = 4x₀·e^(2t) = 4x, matching step 1. A particle starting at (1, 1) is at (e, 1/e) = (2.718, 0.368) m after 0.5 s, and moves along the hyperbola xy = 1.
Example 2 (GATE level): unsteady one-dimensional flow. Given: u = x²·t (m/s), v = w = 0. Find the local, convective and total acceleration at x = 2 m, t = 1 s.
- Local:
∂u/∂t= x² = 4 m/s². - Convective:
u·∂u/∂x= (x²t)(2xt) = 2x³t² = 2 × 8 × 1 = 16 m/s². - Total: a_x = 4 + 16 = 20 m/s². Answer: a_x = 20 m/s² (local 4, convective 16).
Example 3: temperature following a balloon. A weather balloon rises at 5 m/s through a steady ISA troposphere (dT/dz = −0.0065 K/m). DT/Dt = ∂T/∂t + w·∂T/∂z = 0 + 5 × (−0.0065) = −0.0325 K/s: the balloon sees cooling although the field is steady.
Common mistakes
- Writing the acceleration as only ∂V/∂t. That misses the convective terms and gives zero acceleration in every steady flow, which is wrong.
- Thinking "steady" means "particles do not accelerate". Steady means nothing changes at a fixed point.
- Forgetting the product of velocity and gradient in the convective term (writing ∂u/∂x instead of u·∂u/∂x).
- Mixing the two descriptions: differentiating an Eulerian field with respect to t and calling it particle acceleration.
- Dropping the cross terms v·∂u/∂y in 2-D flows.
For GATE AE
Typical questions give a 2-D velocity field and ask for the acceleration at a point (local, convective or total), whether the flow is steady, or the material derivative of temperature or density. Some ask you to find a particle path by integrating dx/dt = u. Practise partial differentiation quickly and always evaluate all four terms of each component.
Quick check
- In steady flow, is the local acceleration zero, the convective acceleration zero, or both?
- For V = 3x·i − 3y·j, what is the acceleration magnitude at (1, 1)?
- Which description does a fixed Pitot tube in a wind tunnel use?
- What equation gives the particle path from an Eulerian field?
Answers: 1. Only the local acceleration is zero 2. a = 9i + 9j, |a| ≈ 12.7 m/s² 3. Eulerian 4. dr/dt = V(r, t) with the initial position given
Interview questions
All Fluid Mechanics interview questionsTry answering each one aloud before you open it.
1.What is the Lagrangian description of fluid motion?Concept
The Lagrangian description of fluid motion focuses on tracking individual fluid particles as they move through space and time. It describes the path and properties of each particle, such as velocity and acceleration, as a function of time. This approach is similar to tracking the trajectory of a single object in classical mechanics.
2.What is the Eulerian description of fluid motion?Concept
The Eulerian description of fluid motion examines the changes in fluid properties at fixed points in space over time. Instead of following individual particles, it looks at how properties like velocity, pressure, and density change at specific locations. This approach is often used in fluid dynamics because it simplifies the analysis of flow fields.
3.Explain the main differences between Lagrangian and Eulerian descriptions.Concept
The main difference between Lagrangian and Eulerian descriptions is the perspective on fluid motion. The Lagrangian approach tracks individual fluid particles, focusing on their trajectories and properties over time. In contrast, the Eulerian approach examines the flow properties at fixed spatial points, observing how these properties change over time. The Lagrangian method is more intuitive for tracking specific particles, while the Eulerian method is more practical for analyzing flow fields.
4.Why is the Eulerian description often preferred in fluid dynamics analysis?Application
The Eulerian description is often preferred in fluid dynamics because it simplifies the analysis of complex flow fields. By focusing on fixed points in space, it allows for easier mathematical modeling and computational simulations. This approach is particularly useful for studying steady-state flows and large-scale fluid systems, where tracking individual particles would be impractical.
5.What happens if you apply the Lagrangian description to a turbulent flow?Application
It is valid but hard: particle paths in turbulence are chaotic, so neighbouring particles separate rapidly and you would need to track enormous numbers of them with very fine time steps. In practice turbulence is computed in Eulerian form (RANS, LES, DNS on a grid), while Lagrangian tracking is used for specific questions such as turbulent dispersion of droplets, soot or pollutants, where statistics of many particle paths are wanted.
6.How can the Lagrangian and Eulerian descriptions be used together in fluid dynamics?Application
The Lagrangian and Eulerian descriptions can be used together to provide a more comprehensive understanding of fluid dynamics. For example, Lagrangian particle tracking can be used to study the dispersion of pollutants in a flow field, while the Eulerian approach can analyze the overall flow characteristics. Combining both methods allows for detailed insights into both individual particle behavior and the broader flow patterns.
7.In what scenarios would the Lagrangian description be more advantageous than the Eulerian description?Application
The Lagrangian description is more advantageous in scenarios where the focus is on individual particle trajectories, such as in the study of sediment transport, pollutant dispersion, or the motion of bubbles in a fluid. It is also useful in cases where the history of a particle's path is important for understanding its current state or behavior.
8.Consider a fluid particle moving in a velocity field described by v(x, y, z, t) = (2t, 3t², 4t³). What is the position of the particle at time t = 2 if it starts from the origin?Numerical
To find the position of the particle at time t = 2, integrate the velocity components with respect to time. The position components are: x(t) = ∫2t dt = t² + C₁, y(t) = ∫3t² dt = t³ + C₂, z(t) = ∫4t³ dt = t⁴ + C₃. Given the initial condition that the particle starts from the origin (0, 0, 0), we find C₁ = 0, C₂ = 0, C₃ = 0. Thus, at t = 2, the position is (2², 2³, 2⁴) = (4, 8, 16).
9.A steady velocity field is V = (x², y², z²) in m/s with coordinates in m. Using the Eulerian description, what is the velocity at the point (1, 1, 1), and is a particle there accelerating?Numerical
In the Eulerian description you simply evaluate the field at the point: V(1, 1, 1) = (1, 1, 1) m/s, magnitude √3 ≈ 1.73 m/s. The field is steady, but the particle there still accelerates through convection: a_x = u·∂u/∂x = x²·2x = 2x³ = 2 m/s², and likewise a_y = a_z = 2 m/s².
10.How does the choice between Lagrangian and Eulerian descriptions affect computational fluid dynamics (CFD) simulations?Application
In CFD simulations, the choice between Lagrangian and Eulerian descriptions affects the modeling approach and computational requirements. The Eulerian description is typically used for grid-based methods, where the flow field is discretized into a mesh. This allows for efficient computation of flow properties over large domains. The Lagrangian description, on the other hand, is used in particle-based methods, which can be more computationally intensive but provide detailed particle tracking. The choice depends on the specific requirements of the simulation, such as the need for detailed particle paths or large-scale flow analysis.
11.What is the material derivative and why does it have two parts?Concept
The material derivative Df/Dt = ∂f/∂t + (V·∇)f is the rate of change of a property following a fluid particle, written in Eulerian field variables. The first term is the local change at a fixed point (non-zero only in unsteady flow); the second is the convective change because the particle moves to a place where f is different. Applied to velocity it gives the particle acceleration that Newton's second law needs, which is why convective terms appear in Euler's and the Navier–Stokes equations.
Finished this topic? Mark it so your progress, study plan and readiness keep up.
Stuck on something here?