Dimensional Analysis & Similitude

Every physical quantity in fluid mechanics can be reduced to a combination of fundamental dimensions. There are two primary systems used in engineering:

• MLT System: Uses Mass ($M$), Length ($L$), and Time ($T$).
• FLT System: Uses Force ($F$), Length ($L$), and Time ($T$). Force and Mass are related by Newton's Second Law ($F = MLT^{-2}$).

Note: Temperature ($\theta$) is also a fundamental dimension but is rarely needed in standard incompressible hydraulics.

To apply dimensional analysis, engineers must be able to convert variables into their fundamental $MLT$ components. Common examples include:

• Velocity ($V$): $LT^{-1}$ (distance per unit time)
• Acceleration ($a$): $LT^{-2}$ (velocity change per unit time)
• Force ($F$): $MLT^{-2}$ (mass times acceleration)
• Density ($\rho$): $ML^{-3}$ (mass per unit volume)
• Pressure ($P$): $ML^{-1}T^{-2}$ (force per unit area)
• Dynamic Viscosity ($\mu$): $ML^{-1}T^{-1}$ (shear stress divided by velocity gradient)

Instead of conducting experiments varying ten different parameters individually (which requires massive effort), the Pi theorem reduces the problem to a relationship between just a few dimensionless groups. This drastically reduces the number of experimental tests required to fully characterize a physical system.

Engineers build scale models of dams, spillways, aircraft, and ships because full-scale testing is too dangerous or expensive. To reliably extrapolate model data to the full-scale prototype, complete similarity (similitude) must be achieved.

Complete similitude is a strict requirement governed by three sequential levels:

1. Geometric Similarity (Scale): The model must be an exact 3D geometric replica of the prototype. Every linear dimension must scale by the exact same ratio ($L_r = L_m / L_p$). Angles must be preserved exactly.
2. Kinematic Similarity (Motion): The fluid flow patterns (streamlines) must be geometrically similar. The ratio of velocities and accelerations at corresponding points between the model and prototype must be constant ($V_r = V_m / V_p$). This requires geometric similarity to be satisfied first.
3. Dynamic Similarity (Force): The ratios of all corresponding forces (e.g., inertial vs. viscous, or inertial vs. gravitational) acting on the fluid particles must be identical in both the model and the prototype ($F_r = F_m / F_p$). This is achieved by matching specific dimensionless numbers.

It is physically impossible to match the ratios of all forces simultaneously unless the model and prototype are the exact same size using the exact same fluid. Therefore, engineers identify the dominant forces in a specific problem and match the dimensionless number that represents the ratio of those dominant forces.

Reynolds similarity ($Re_{model} = Re_{prototype}$) must be achieved when viscous forces dominate the flow behavior. This is the primary criterion for modeling fully enclosed flows, such as flow inside pressure pipes, wind tunnel testing of airplanes, or flow around deeply submerged submarines. Gravity does not affect these flows because there is no free surface.

Froude similarity ($Fr_{model} = Fr_{prototype}$) must be achieved when gravity is the dominant force dictating flow behavior. This is strictly required for any flow featuring a free liquid surface, such as open channels, rivers, spillways over dams, weirs, and surface waves on ships. In these models, gravity drives the flow and surface waves, while viscosity is considered a secondary effect.

Euler similarity is considered when analyzing systems driven entirely by pressure differences where viscous and gravitational forces are negligible. It is commonly used in testing flow through orifices, valves, nozzles, and analyzing cavitation onset in pumps.

Mach similarity is crucial when fluid compressibility cannot be ignored. In civil engineering, water is generally considered incompressible, so Mach similarity is rarely needed. However, it becomes critical when analyzing extreme pressure transient events, such as "water hammer" shockwaves traveling through a closed pipeline.

When modeling massive, shallow bodies of water like wide rivers, estuaries, or harbors, maintaining a strict 1:1 scale ratio for both length and depth is problematic.

If a river is $1\text{ km}$ wide and $5\text{ m}$ deep, a $1:1000$ geometrically similar model would be $1\text{ m}$ wide but only $5\text{ mm}$ deep. At a $5\text{ mm}$ depth, surface tension and laminar friction completely dominate the flow, ruining the Froude similarity required to model the river accurately.

To fix the shallow depth problem, engineers exaggerate the vertical scale ratio ($L_{rv} = y_m / y_p$) compared to the horizontal scale ratio ($L_{rh} = L_m / L_p$).

• Distortion Factor ($D$): Defined as $D = \frac{L_{rv}}{L_{rh}}$.
• Trade-offs: Distorted models successfully maintain turbulent flow and measurable depths, but they skew velocities and slopes. The engineer must apply complex scaling adjustments to model velocity and discharge to compensate for the geometric distortion.