Continuum Mechanics & Fluid Dynamics

CF01.01 begins from an observable question, not a formula list. The chapter asks: distinguish parcel-scale averages from molecular descriptions and declare the resolution window. First declare the material or fluid parcel, spatial region, time interval, fields, units, reference frame, initial data, boundary data, and whether the description is material or spatial. A symbol is not meaningful until its type and domain are named.

Read the basic types aloud. X labels a material point, x is its current position, χ(t,X) is the motion, u(t,x) is velocity, ρ is mass per current volume, p is scalar pressure only in an isotropic stress split, σ is Cauchy stress, n is an oriented unit normal, and D/Dt=∂t+u·∇ follows a moving parcel. Never add quantities with incompatible units or tensor types.

Separate kinematics, balance, and constitutive response. Kinematics describes motion and deformation without choosing a material. Balance laws encode mass, momentum, angular momentum, and energy. A constitutive law closes unresolved stress, heat flux, pressure, viscosity, or surface force under stated symmetry and thermodynamic restrictions. Conservation alone does not choose Newtonian, inviscid, compressible, or incompressible behavior.

A local differential equation and an integral control-volume law are two representations whose equivalence needs regularity, orientation, and transport theorems. Write every boundary flux with the outward normal and control-surface velocity. When solutions have shocks, corners, interfaces, or only weak derivatives, return to distributions or integral balances before using pointwise chain rules.

The frame ledger distinguishes a change of observer from a physical deformation. Translation and rotation can alter component formulas while objective scalar or tensor relations transform in a prescribed way. Before accepting a constitutive equation, test its units, isotropy or material symmetry, frame indifference, dissipation sign, parameter range, and behavior under rigid motion.

The exact finite calibration is: CF01.01: a declared reference length L=4 m and speed U=3 m/s give the advective time L/U=4/3 s. A second clock T=4 s gives Strouhal number St=L/(UT)=1/3. These labels organize one scaling ledger; they do not determine a governing equation or a flow regime without the omitted material and force scales. Keep fractions, π, radicals, tensor entries, signs, and units until the final line. This row is evidence for the declared finite data only. It does not prove existence, uniqueness, stability, convergence, regularity, or a limit unless a separate theorem connects those claims.

Nondimensionalization is a typed substitution, not deletion of units by inspection. Choose reference length, time, velocity, density, pressure, temperature, and force scales; substitute them into every term; divide by one common scale; and retain each resulting dimensionless group. A Reynolds, Mach, Froude, Rossby, Weber, or Strouhal number is meaningful only with its declared scale choices.

Mass balance tracks density and volume together. In a material volume, no mass crosses the moving boundary; in a fixed or arbitrary control volume, relative velocity creates boundary flux. The differential continuity equation follows only after applying the transport and divergence theorems with sufficient regularity. Incompressibility can mean volume-preserving motion, zero velocity divergence, or constant density under additional hypotheses; state which meaning is used.

Momentum balance converts boundary traction and body force into acceleration of mass. Cauchy's stress theorem represents traction by σn under locality and balance assumptions; angular-momentum balance gives stress symmetry in the classical nonpolar setting. Pressure is not always the entire stress. For a Newtonian fluid, viscous stress depends on the symmetric velocity gradient and bulk response, with coefficients constrained by dissipation.

Energy and entropy answer different questions. The first law balances kinetic and internal energy with work and heat. The second law restricts constitutive choices through nonnegative entropy production after temperature and entropy flux are declared. A positive viscosity coefficient supports viscous dissipation in the standard model, but thermodynamic consistency still depends on density, temperature, pressure, heat conduction, and boundary fluxes.

Pressure in incompressible flow acts as a constraint multiplier that enforces divergence-free velocity together with boundary conditions. Taking divergence can yield a pressure Poisson equation, but its boundary data and normalization must be derived. Applying a Helmholtz projection removes gradients only in the declared domain and function space; topology and boundaries affect the decomposition.

Vorticity is curl u and records local rotation of velocity, not the angular velocity of every finite parcel. Curling Euler or Navier–Stokes removes pressure under suitable smoothness and reveals transport, stretching, baroclinic production, and diffusion. Two dimensions lack the three-dimensional stretching term, which changes the regularity and cascade story. Circulation statements require an oriented material loop and exact force hypotheses.

Weak solutions retain integral or distributional balances after classical derivatives fail. State the test-function class, divergence constraint, initial trace, energy inequality, pressure interpretation, and boundary behavior. Weak existence is not weak uniqueness. A strong solution may control a weak one through weak–strong uniqueness only when a relative-energy or stability argument and every integrability hypothesis are available.

For three-dimensional incompressible Navier–Stokes, finite-energy weak solutions exist in standard settings, while global smoothness and uniqueness for arbitrary smooth data remain unresolved. The course may derive energy bounds, local strong theory, continuation criteria, and partial regularity under exact hypotheses. It must stop before claiming the open global regularity conclusion.

Compressible flow couples density, momentum, and energy with an equation of state. Characteristic speeds depend on sound speed and state. Smooth compression can steepen into shocks, so weak conservation and entropy admissibility replace multivalued classical flow. Vacuum, contact discontinuities, heat conduction, viscosity, and boundary conditions each alter the theorem; no scalar shock rule covers every system.

Boundary layers arise because small viscosity multiplies the highest spatial derivative while viscous and inviscid boundary conditions can disagree. Derive the stretched coordinate and leading balance from the scaled equations. A formal Prandtl profile is not automatically well posed or attached. Separation, adverse pressure gradients, instability, and Kato-type dissipation criteria require separate hypotheses and evidence.

Linear stability studies the spectrum or evolution operator of a linearized equation around a named base flow. Spectral stability does not preclude transient amplification for a nonnormal operator, and linear growth does not alone prove nonlinear transition. State the perturbation norm, boundary conditions, wavenumbers, Reynolds number, time interval, and whether the conclusion is modal, energy-based, or nonlinear.

Rotation, stratification, gravity, and free surfaces add time and length scales. Coriolis terms do no work in the basic kinetic-energy identity but redirect momentum. Buoyancy exchanges kinetic and potential energy. Free-boundary curvature produces surface-tension forces and requires kinematic plus dynamic conditions on an unknown interface. Every reduced model needs a declared asymptotic limit and error boundary.

Turbulence statistics separate a field into resolved mean and fluctuation only after an averaging operator is declared. Reynolds stresses introduce an unclosed correlation, not an extra known force. Energy spectra and structure functions depend on normalization, domain, stationarity, homogeneity, isotropy, and ensemble or time averaging. Scaling ranges observed over finite data are evidence, not universal exact laws.

A fluid computation needs a scheme ledger: mesh geometry, variable placement, numerical flux, pressure coupling, time integrator, boundary closure, linear and nonlinear tolerances, conservation residuals, divergence defect, stability restriction, refinement path, and exact output rows. A smooth image cannot replace these records. Verification addresses equations and code; validation against a physical domain needs separate observations and uncertainty.

The bounded theorem bridge for this unit is: consistent reference scales convert a dimensionally valid balance into a dimensionless equation whose coefficients expose competing mechanisms, while the model remains valid only between its declared microscopic, geometric, and observational scales. Rewrite it as named hypotheses, mechanism, conclusion, and excluded stronger claim. The failure mutation is: mix units, leave fields untyped, confuse observer change with material change, infer a regime from one number, or extend the continuum model below its resolution scale. Identify the first line of the argument that uses the removed property, then state the strongest conclusion that still survives.

Finish by teaching What a continuum field represents to a learner who has not seen continuum mechanics. Name every field and unit, draw the control object, derive the exact finite row twice, identify the solution class, audit the theorem bridge, activate one failed hypothesis, and explain why the surviving finite observation is weaker than a global fluid theorem.