Dimensional Analysis & Modeling
Dimensional analysis converts a many-variable fluid problem into a smaller relation among dimensionless groups. It is both an error detector and a disciplined basis for experiments, scale models, and uncertainty-aware extrapolation.
Learning objectives
- Distinguish physical dimensions from the units chosen to report them.
- Test governing balances for dimensional homogeneity and consistent reference quantities.
- Construct independent Buckingham Pi groups with defensible repeating variables.
- Select Reynolds, Froude, or Mach similarity from the dominant physics.
- Scale model measurements and quantify consequences of incomplete similarity.
- Verify dimensions, regimes, trends, and experimental uncertainty.
Dimensions & Units
Dimension, unit, and basis
A dimension identifies a physical kind; a unit assigns its numerical scale. Fluid mechanics usually uses mass \(M\), length \(L\), time \(T\), and temperature \(\Theta\). Record whether pressure is gauge or absolute and temperature is absolute before forming gas or Mach groups. A change of units cannot change a dimensionless result.
| Dimension | Symbol | SI Unit |
|---|---|---|
| Mass | m | kg |
| Length | L | m |
| Time | t | s |
| Temperature | T | K |
| Electric current | I | A |
| Amount of light | C | cd |
| Amount of matter | N | mol |
Physical meaning
Dimensions reveal the type of balance represented by each term. They cannot supply a missing numerical coefficient, but they can prove that an proposed additive balance is impossible.
Dimensional Homogeneity
Every additive term in an equation must have the same dimensions. You cannot add apples and oranges — or pressure and velocity. A term with the wrong dimensions is a sure sign of an error made earlier in the analysis.
Nondimensionalization
Dividing every term of a homogeneous equation by a group of variables and constants that carries those same dimensions renders the equation nondimensional. If the resulting terms are also of order one, the equation is normalized. This process makes dimensionless groups — Reynolds number, Froude number, and the like — appear naturally.
Problem-solving strategy
- Define the system or control volume and the response to be predicted.
- List every independent variable, including properties and boundary scales.
- Write a dimension matrix and determine its rank rather than merely counting symbols.
- Build independent groups, then verify every exponent and dimension.
- Choose similarity requirements from the governing force or wave balance.
- Audit regimes, uncertainty, and any unmatched group before extrapolating.
Dimensional Analysis & Similarity
Purpose of Dimensional Analysis
It serves three ends: to collapse many variables into a few dimensionless groups (so experiments need fewer runs), to obtain scaling laws that predict prototype behavior from a model, and to predict trends in the relationships between parameters.
Geometric similarity — the model is the same shape as the prototype, scaled by a constant factor. Kinematic similarity — the velocity at any point in the model flow is proportional, and parallel, to the velocity at the corresponding point in the prototype. Dynamic similarity — all forces in the model flow scale by a constant factor to the corresponding forces in the prototype. All three together give complete similarity.
The Froude Number
A key dimensionless group in free-surface flows, interpreted as the ratio of inertial force to gravitational force. Matching it between model and prototype governs ship-hull and open-channel testing.
Recognition strategy
A free surface and gravity waves identify Froude similarity; Reynolds matching is checked afterward rather than imposed simultaneously.
Given information
- \(L_m/L_p=1/25\), \(V_p=10.0\ \text{m/s}\), and equal \(g\); water properties are at the same stated laboratory temperature.
- Both pressures are atmospheric gauge zero; \(x\) is positive downstream and speed is positive in \(+x\).
Model and interpretation
The control volumes span corresponding spillway reaches. Geometric and Froude similarity are exact stated constraints; neglect of viscous and surface-tension scale effects is an approximation requiring later regime checks.
Verification
Units: The square-root length ratio is dimensionless, leaving the required velocity unit \(\text{m/s}\).
Signs and direction: Positive \(2.00\ \text{m/s}\) means downstream, consistent with the declared \(+x\) direction.
Plausibility: A model five times smaller in linear velocity scale has equal Froude number; Reynolds number will be lower by \(125\) for equal viscosity and must be reported as an unmatched effect.
Engineering interpretation
This setting reproduces gravity-wave kinematics, not complete dynamic similarity. The test report must quantify Reynolds and Weber scale effects before prototype decisions are made.
The Method of Repeating Variables & the Buckingham Pi Theorem
A dimensionally homogeneous relation among \(n\) dimensional parameters can be rewritten as a relation among \(k = n - j\) independent dimensionless groups (Π’s), where \(j\) is the number of primary dimensions represented (the reduction).
The Six-Step Procedure
1. List all \(n\) parameters of the problem. 2. Write each one’s primary dimensions. 3. Guess the reduction \(j\) (usually the number of primary dimensions present) and set \(k = n - j\). 4. Choose \(j\) repeating parameters that together contain all the primary dimensions (never choose the dependent variable). 5. Combine the repeating parameters with each remaining one to form \(k\) dimensionless Π’s. 6. Check that each Π is dimensionless and write the final functional relationship.
Recognition strategy
The task supplies a complete variable set and requests reduction, so a dimension matrix and repeating-variable method apply.
Given information
- Six variables use \(M,L,T\); pressure is a positive gauge difference \(p_{\rm in}-p_{\rm out}\).
- Properties are uniform at the stated reference temperature; \(x\) and \(V\) are positive downstream.
Model and interpretation
A steady duct control volume is assumed. The variable dependence is a supplied approximation: compressibility, roughness, entry effects, and fittings are excluded.
Verification
Units: Both \(\Delta p\) and \(\rho V^2\) have units \(\text{kg}/(\text{m}\!\cdot\!\text{s}^2)\); every reported group is dimensionless.
Signs and direction: The positive pressure-loss definition agrees with positive downstream velocity and avoids hiding a minus sign.
Plausibility: Three groups equal \(n-r\), and increasing \(L/D\) should increase loss; dimensional analysis deliberately does not determine the function or coefficient.
Engineering interpretation
The reduction organizes experiments but does not prove fully developed flow or provide a friction correlation. Those require separate data and regime validation.
Experimental Testing & Incomplete Similarity
Reducing the Experimental Burden
Plotting the dependent Π against the independent Π’s collapses what might be a multi-dimensional study into a single curve, slashing the number of experiments and producing results that apply across scales.
Incomplete Similarity
Often it is impossible to match every Π at once — matching the Reynolds number on a small model, for instance, may demand impractical speeds or introduce compressibility or cavitation. When not all groups can be matched, the test has incomplete similarity, and engineers use careful extrapolation, correlations, and judgment to bridge the gap to the prototype.
For the Pi method, the make-or-break step is choosing repeating variables: include the parameters you most expect to matter, make sure they collectively cover every primary dimension, and never pick the variable you are solving for. Get that right and the algebra falls into place. This chapter ties the whole course together — every dimensionless group you’ve met (Re, Fr, Ma) is an instance of it.
Scaling a measured response
A dependent response is commonly normalized by a characteristic inertial scale. For rotating propulsors, a supplied thrust coefficient may take the form \(C_T=T/(\rho n^2D^4)\). If geometry and all governing independent groups match, equal \(C_T\) transfers measured thrust; if Reynolds or cavitation groups differ, the transfer is only an approximate extrapolation.
Recognition strategy
Equal advance conditions and geometry make the thrust coefficient \(C_T=T/(\rho n^2D^4)\) the appropriate supplied similarity form.
Given information
- \(D_m=0.300\ \text{m}\), \(D_p=1.20\ \text{m}\), \(n_m=15.0\ \text{s}^{-1}\), \(n_p=3.75\ \text{s}^{-1}\), \(T_m=18.0\ \text{N}\), and equal supplied advance coefficient.
- Water density at the common reference temperature is supplied as \(998\ \text{kg/m}^3\); ambient absolute pressure is common, and thrust is positive forward.
Model and interpretation
The rotating-fluid control volumes have geometrically similar propellers at the same supplied advance coefficient, hence the same stated dimensionless operating point. Equality of \(C_T\) is approximate because Reynolds number and cavitation number are not demonstrated equal.
Verification
Units: The multiplier is dimensionless, so model newtons scale directly to prototype newtons.
Signs and direction: Positive thrust remains forward under geometrically corresponding rotation and the declared convention.
Plausibility: Although rotational speed drops by four, diameter rises by four and the \(D^4\) effect yields a net factor of sixteen; verify Reynolds and cavitation margins before use.
Engineering interpretation
This is an extrapolated test estimate, not certified performance. Prototype testing must address Reynolds mismatch, free-surface effects, pressure uncertainty, and cavitation.
Chapter recap
- Homogeneity detects impossible balances but not missing dimensionless constants.
- Buckingham Pi reduces variables according to dimension-matrix rank.
- Similarity follows dominant physics; unmatched groups define model risk.
- Before moving on, be able to construct, interpret, match, and audit Pi groups with uncertainty.