Dimensional Analysis & Similitude - Theory & Concepts

Dimensional Analysis & Similitude

Learning Objectives

Express hydraulic variables in MLT and FLT dimensional systems and test equations for dimensional homogeneity.
Apply the Buckingham Pi theorem and select valid repeating variables.
Interpret Reynolds, Froude, Euler, Weber, Mach, Strouhal, and cavitation parameters.
Distinguish geometric, kinematic, and dynamic similarity.
Derive model–prototype velocity, discharge, time, force, and pressure scale relations.
Include fluid-property ratios in Reynolds and other similarity laws rather than assuming identical fluids automatically.
Recognize scale effects, incompatible similarity requirements, roughness limitations, and distorted-model consequences.

Dimensional analysis reduces a physical problem to relationships among dimensionless groups. Similitude uses those groups to transfer observations from a model to a prototype. Hydraulic models are reliable only when their scale convention, dominant forces, fluid-property ratios, boundary roughness, and unavoidable scale effects are stated explicitly.

Dimensions and Units

A dimension describes the physical nature of a quantity, while a unit provides a numerical standard. Velocity has dimension $LT^{-1}$ whether it is reported in m/s, ft/s, or km/h. Dimensional analysis is independent of the unit system as long as units are internally consistent.

Common MLT Dimensions

Quantity	Symbol	Dimensions
Length	$L$	$L$
Time	$t$	$T$
Velocity	$V$	$LT^{-1}$
Acceleration	$a$	$LT^{-2}$
Mass density	$\rho$	$ML^{-3}$
Dynamic viscosity	$\mu$	$ML^{-1}T^{-1}$
Kinematic viscosity	$\nu$	$L^2T^{-1}$
Pressure or stress	$p$	$ML^{-1}T^{-2}$
Force	$F$	$MLT^{-2}$
Surface tension	$\sigma$	$MT^{-2}$
Power	$P$	$ML^2T^{-3}$

MLT and FLT Systems

The MLT system uses mass, length, and time as fundamentals. The FLT system uses force, length, and time, with mass derived from $M=FT^2/L$ . Do not mix an MLT dimension for one variable with an FLT dimension for another in the same exponent calculation.

Dimensional Homogeneity

Every additive term in a physically valid equation must have the same dimensions. Homogeneity can reveal a missing length, gravity term, density, or conversion factor, but it cannot prove that a numerical coefficient or physical model is correct.

Dimensional Consistency Check

STEP-BY-STEP

Replace every variable by its fundamental dimensions.
Reduce exponents of $M$ , $L$ , and $T$ on each side.
Confirm that all additive terms share identical dimensions.
Check that arguments of logarithmic, exponential, trigonometric, and similar functions are dimensionless.
Separately verify unit-system constants in empirical equations.

Dimensionally Homogeneous Can Still Be Wrong

Both $V=\sqrt{gH}$ and $V=100\sqrt{gH}$ are dimensionally homogeneous, but only the correct coefficient and model assumptions can be established from theory or experiment. Dimensional analysis determines form, not complete physics.

Buckingham Pi Theorem

If a phenomenon is described by $n$ dimensional variables involving $r$ independent fundamental dimensions, the relation can be rewritten using $n-r$ independent dimensionless groups:

F(q_1,q_2,\ldots,q_n)=0 \quad\Longrightarrow\quad \Phi(\Pi_1,\Pi_2,\ldots,\Pi_{n-r})=0

Buckingham Pi Method

STEP-BY-STEP

List all variables required by the physics, including dependent variables and relevant fluid, geometric, and forcing quantities.
Express every variable in fundamental dimensions.
Determine the dimensional rank $r$ , not merely the number of dimension symbols written.
Select $r$ repeating variables that collectively contain all fundamental dimensions and cannot form a dimensionless product by themselves.
Combine each nonrepeating variable with the repeating variables raised to unknown powers.
Set the net exponents of every fundamental dimension to zero and solve.
Check independence and physical meaning of the resulting Pi groups.
Compare with known limiting cases and available experimental data.

Variable Omission Cannot Be Repaired Algebraically

Buckingham Pi analysis cannot recover a variable omitted from the original list. If roughness, gravity, surface tension, compressibility, geometry ratio, or unsteadiness affects the phenomenon, it must be included before forming groups.

Example: Pressure Loss in a Rough Pipe

Suppose pressure drop depends on $\Delta p$ , density $\rho$ , mean velocity $V$ , diameter $D$ , length $L$ , dynamic viscosity $\mu$ , and roughness $\varepsilon$ . Dimensional analysis can produce

\frac{\Delta p}{\rho V^2} = \Phi\left( \frac{\rho VD}{\mu}, \frac{\varepsilon}{D}, \frac{L}{D} \right)

The left group is an Euler-type pressure coefficient; the right groups are Reynolds number, relative roughness, and relative length. Darcy–Weisbach and the Moody chart provide the empirical/theoretical closure for this functional relation.

Geometric Similarity

All corresponding lengths have constant scale ratio and corresponding angles are equal. Shape ratios, relative roughness, gate openings, curvature, and boundary details must also be represented when they influence the flow.

Kinematic Similarity

Corresponding velocity vectors have constant scale ratios and similar directions, producing geometrically similar streamline and pathline patterns. Time and acceleration scales follow from length and velocity scales.

Dynamic Similarity

Ratios of the dominant forces are equal between model and prototype. This is achieved by matching the appropriate dimensionless numbers, not merely by matching geometry.

Scale Convention Used in This Lesson

The simulator and formulas use

\lambda=\frac{L_p}{L_m}

so a 1:20 model has $\lambda=20$ . Every scale written below is prototype divided by model. Other references may use the reciprocal; never copy exponents before confirming the convention.

Reynolds Number

The ratio of inertia to viscous effects:

Re=\frac{\rho VL}{\mu}=\frac{VL}{\nu}

Reynolds similarity is central to internal flows, boundary layers, wakes, drag, and viscous-dominated model behavior.

Froude Number

The ratio of inertia to gravity effects:

Fr=\frac{V}{\sqrt{gL}}

Froude similarity governs free-surface waves, spillways, rivers, hydraulic jumps, ship waves, and gravity-driven open-channel flows.

Euler Number

The ratio of pressure to inertia effects:

Eu=\frac{\Delta p}{\rho V^2}

It appears in pressure coefficients, meter behavior, pumps, turbines, nozzles, and pressure-driven flows.

Weber Number

The ratio of inertia to surface-tension effects:

We=\frac{\rho V^2L}{\sigma}

Weber similarity matters for jets, droplets, bubbles, air entrainment, small waves, capillary structures, and very small free-surface models.

Mach Number

The ratio of flow velocity to acoustic speed:

Ma=\frac{V}{c}

Mach similarity becomes important when density changes caused by compressibility are significant, including gas flow and high-speed liquid transients.

Strouhal Number

A dimensionless unsteadiness or frequency parameter:

St=\frac{fL}{V}

It is used for periodic vortex shedding, oscillating gates, waves, pulsating flows, and time-dependent model tests.

Cavitation Number

A pressure-margin parameter comparing local absolute pressure above vapor pressure with dynamic pressure:

\sigma_c=\frac{p_{abs}-p_v}{\tfrac12\rho V^2}

Cavitation similarity may be required for spillways, gates, valves, pumps, turbines, and high-velocity conduits.

Selecting the Governing Similarity Law

Use Froude similarity when gravity and inertia dominate a free surface.
Use Reynolds similarity when viscosity controls separation, friction, or internal-flow structure.
Include Weber similarity when surface tension influences breakup, aeration, or small-scale waves.
Include cavitation number when vapor formation is a design risk.
Include Mach or Cauchy-type compressibility parameters for pressure-wave and high-speed gas problems.
Include Strouhal similarity for periodic or transient behavior.

The governing law follows from the force and time scales of the actual phenomenon, not from the type of laboratory facility available.

Froude Model Scale Laws

Prototype-to-model ratios for equal gravity and Froude similarity.

For $Fr_p=Fr_m$ and $g_p=g_m$ ,

\frac{V_p}{V_m}=\lambda^{1/2}

\frac{t_p}{t_m}=\lambda^{1/2}

\frac{Q_p}{Q_m}=\lambda^{5/2}

For equal fluid density,

\frac{F_p}{F_m}=\lambda^3, \qquad \frac{P_p}{P_m}=\lambda^{7/2}

Variables

Symbol	Description	Unit
$P$	Power	W
$F$	Force	N

Pressure Scale under Froude Similarity

Hydrostatic or dynamic pressure differences scale as $\rho V^2$ . With the same fluid, $\Delta p_p/\Delta p_m=\lambda$ . If densities differ, multiply by $\rho_p/\rho_m$ .

General Reynolds Model Scale Laws

Includes prototype-to-model kinematic-viscosity ratio instead of assuming equal fluids.

For $Re_p=Re_m$ ,

\frac{V_p}{V_m} = \frac{\nu_p/\nu_m}{\lambda}

\frac{Q_p}{Q_m} = \lambda\left(\frac{\nu_p}{\nu_m}\right)

\frac{t_p}{t_m} = \frac{\lambda^2}{\nu_p/\nu_m}

Equal-Fluid Reynolds Scaling Can Produce Very Low Prototype Velocity Ratios

When $\nu_p=\nu_m$ , Reynolds similarity requires $V_p/V_m=1/\lambda$ . A small water model would therefore need a model velocity much larger than the prototype velocity, often creating impractical pressure, power, cavitation, or free-surface conditions.

Interactive Scale Calculator

The simulator uses the same $\lambda=L_p/L_m$ convention and includes the kinematic-viscosity ratio in Reynolds scaling.

Hydraulic Model Scaling Simulator

Model : prototype = 1 : $\lambda$ , where $\lambda=L_p/L_m$ . All displayed ratios are prototype divided by model.

Governing similarity law

Geometric scale1 : 10

Model velocity1.00 m/s

Model discharge0.10 m³/s

Prototype prediction

3.162 m/s

velocity

31.623 m³/s

discharge

Governing equality: $Fr_m=Fr_p$
$V_p/V_m$: 3.16228
$Q_p/Q_m$: 316.22777
$t_p/t_m$: 3.16228

V_p/V_m=\lambda^{1/2}

Q_p/Q_m=\lambda^{5/2}

Froude similarity matches inertia and gravity. Reynolds, Weber, and cavitation effects may remain distorted at small model scales.

Incompatibility of Froude and Reynolds Similarity

With the same fluid and gravity, a reduced-scale model generally cannot satisfy both $Fr_p=Fr_m$ and $Re_p=Re_m$ . Froude similarity requires $V_p/V_m=\lambda^{1/2}$ , while Reynolds similarity requires $V_p/V_m=1/\lambda$ . These can both hold only for $\lambda=1$ unless fluid properties or gravity are changed.

Scale Effects

A scale effect is the systematic model–prototype difference caused by dimensionless groups that are not matched. Common examples include:

Excessive viscous influence in small Froude-scaled models.
Surface-tension distortion of shallow waves, jets, and air entrainment.
Roughness that cannot be reduced geometrically below material grain size.
Different boundary-layer thickness and flow separation.
Premature or suppressed cavitation.
Air entrainment and bubble size that do not scale geometrically.
Instrument resolution and leakage becoming large relative to model discharge.

Managing Scale Effects

STEP-BY-STEP

Identify the dominant dimensionless group and match it exactly where feasible.
Calculate the unmatched groups in both model and prototype.
Keep the model sufficiently large to move secondary groups beyond known threshold ranges.
Adjust fluid properties, pressure level, temperature, or roughness where safe and practical.
Perform tests at multiple scales or operating points to quantify trends.
Calibrate against prototype or field data when available.
Report uncertainty and the range over which extrapolation is valid.

Distorted Model

A model that intentionally uses different horizontal and vertical length scales or modifies another geometric ratio. River and estuary models often exaggerate vertical scale so that shallow depths, slopes, and velocity fields can be measured.

Consequences of Geometric Distortion

Distortion changes slope, curvature, hydraulic radius, wave propagation, sediment motion, secondary circulation, and sometimes the governing dimensionless parameters. Scale relations must be re-derived from the chosen horizontal and vertical scales; undistorted Froude exponents cannot be applied blindly.

Roughness Does Not Scale Only by Length

Even if geometric roughness height is scaled, the hydraulic effect depends on relative roughness and Reynolds number. Model surfaces may need artificial roughness or calibration to reproduce prototype resistance, especially in rivers, floodplains, and turbulent conduits.

Similarity in Numerical Models

Computational models do not eliminate similitude concerns. Grid spacing, time step, numerical diffusion, turbulence closure, wall treatment, multiphase interface resolution, and boundary conditions create numerical scale effects analogous to laboratory limitations. Validation and mesh/time-step independence remain essential.

Hydraulic Model Planning Workflow

STEP-BY-STEP

Define the prototype question and response quantities to be predicted.
List all physically relevant variables and derive or identify dimensionless groups.
Rank gravity, viscosity, pressure, surface tension, compressibility, and unsteadiness effects.
Choose and document the scale convention.
Select the dominant similarity law and derive every required scale ratio from it.
Calculate unmatched dimensionless numbers and assess scale effects.
Design instrumentation, uncertainty, boundary roughness, and operating range.
Validate model behavior against theory, limiting cases, and available prototype data before extrapolation.

Key Takeaways

Dimensional homogeneity is necessary but does not prove physical correctness.
Buckingham Pi reduces variables only after a complete and physically justified variable list is chosen.
Always declare whether the length scale is model/prototype or prototype/model before using scale exponents.
Froude similarity governs gravity-driven free-surface flow; Reynolds similarity governs inertia–viscosity balance.
General Reynolds scaling includes $\nu_p/\nu_m$ and should not silently assume identical fluids.
Weber, Strouhal, Mach, Euler, and cavitation parameters can control hydraulic phenomena ignored by a single-law model.
Reduced-scale models usually contain unavoidable scale effects; these must be quantified, managed, and reported rather than hidden.

Previous TopicHydraulic Machinery - Examples & Applications

Quiz Me

Next TopicDimensional Analysis & Similitude - Examples & Applications

Prev Next

Quiz Me