Ship Resistance Components: Full Breakdown

Total hull resistance is the force opposing a ship’s forward motion through water at steady speed. It determines the effective power required at the hull surface, links directly to propulsion system sizing, and drives fuel consumption and carbon intensity under MARPOL Annex VI regulation 23 (EEDI) and resolution MEPC.333(76) (EEXI). This article traces the full decomposition from William Froude’s 1874 two-component model through the modern ITTC-78 power-prediction method, covering the dimensionless coefficients, the Froude-Reynolds similarity conflict, model-to-ship extrapolation, added resistance in waves, and the hull-form and coating strategies that reduce each component. The companion Holtrop-Mennen resistance calculator implements the dominant empirical method in practice; the ITTC 1957 friction coefficient calculator computes $C_F$ alone; and the form factor (Prohaska) calculator derives $(1+k)$ from low-speed model data.

The Froude decomposition and its modern extension

William Froude, working at the Torquay tank in 1874 on HMS Greyhound trials, proposed dividing total resistance into two parts. Frictional resistance was estimated from a flat plank of the same wetted area and length, and residual resistance was the remainder obtained by subtracting that estimate from measured total resistance. The decomposition was operational, not physically rigorous, but it was experimentally tractable and it drove ship-model testing methodology for the next century.

The modern form retains Froude’s structure but replaces his flat-plank friction with the ITTC 1957 correlation line and adds a form factor to capture three-dimensional viscous effects:

where $R_V = (1+k)\,R_F$ is the total viscous resistance (frictional plus viscous-pressure), $R_W$ is wave-making, $R_{app}$ is appendage, and $R_A$ is air resistance. In dimensionless form, dividing each component by $\tfrac{1}{2}\rho V^2 S$ (where $S$ is wetted surface area):

The correlation allowance $C_A$ (sometimes written $\Delta C_F$) is added to $(1+k)\,C_F$ during model-to-ship extrapolation to account for hull roughness and any remaining systematic difference between model and ship behaviour. It is distinct from $C_{A\!A}$, which is the air-resistance coefficient. In the ITTC-78 method the full resistance equation at ship scale is:

The total resistance coefficient $C_T$ equals $R_T / (\tfrac{1}{2}\rho V^2 S)$ and is dimensionless. For a fully laden Panamax bulk carrier at 13.5 knots, $C_T$ is roughly $3.8 \times 10^{-3}$; for a post-Panamax container ship at 22 knots it is closer to $5.5 \times 10^{-3}$, with wave-making contributing a much higher share.

Frictional resistance and the ITTC 1957 line

Physical origin

When a ship moves through water, a boundary layer forms over the hull surface. In the laminar sub-layer immediately adjacent to the hull, momentum transfer occurs through viscosity; beyond that the boundary layer is turbulent for virtually all ship-scale flows. The tangential (shear) stress integrated over the wetted surface gives the frictional resistance. For a flat plate in fully turbulent flow the frictional coefficient depends only on the Reynolds number $R_n = VL/\nu$, where $V$ is ship speed, $L$ is waterline length, and $\nu$ is kinematic viscosity (about $1.19 \times 10^{-6}$ m$^2$/s for seawater at 15 °C, or $1.13 \times 10^{-6}$ m$^2$/s at 20 °C).

ITTC 1957 correlation line

The ITTC 1957 line, adopted at the 8th International Towing Tank Conference, is:

This is the standard for all model-to-ship extrapolation in practice. At a ship Reynolds number of $3 \times 10^8$ (a 200 m vessel at 14 knots in 15 °C seawater), $C_F = 1.46 \times 10^{-3}$. At $R_n = 10^9$ (a 300 m VLCC at 15 knots), $C_F = 1.35 \times 10^{-3}$. The line is deliberately not an attempt to represent the physical flat-plate friction accurately; rather, it is a correlation tool calibrated so that $C_A$ averages to zero across the tank-to-ship data sets that underpinned the 1957 decision.

The Schoenherr (ATTC) line predates ITTC 1957 and is solved implicitly from:

It gives lower $C_F$ values at high Reynolds numbers and was superseded by ITTC 1957 for merchant-ship extrapolation, though it remains in use in some US naval design contexts. The Schoenherr friction calculator computes both lines side by side for comparison.

Frictional resistance and effective power

The frictional resistance force is:

For a 50,000 DWT bulk carrier with $S = 8{,}500$ m$^2$ at 13.5 knots (6.94 m/s) in seawater ($\rho = 1{,}025$ kg/m$^3$):

The frictional contribution to effective power at that speed is $P_{E,F} = R_F \times V \approx 304 \times 6.94 \approx 2{,}110$ kW. At typical overall propulsive efficiency of about 0.65, this translates to roughly 3,250 kW of delivered engine power just to overcome hull friction. That figure illustrates why coating condition management and biofouling prevention are the cheapest interventions available to a ship operator.

Speed and Reynolds scaling

The frictional resistance varies with $V^{1.825}$ to $V^{1.95}$ across the practical speed range (the exact exponent falling as $R_n$ rises and the ITTC line flattens). It does not follow a pure $V^2$ law because $C_F$ itself decreases with increasing Reynolds number. For most merchant ship types at service speed, frictional resistance accounts for:

• 70-80% of calm-water $R_T$ for tankers and bulk carriers at $F_n < 0.18$.
• 55-65% for container ships and LNG carriers at $F_n = 0.22$-$0.28$.
• 35-50% for fast ferries and naval vessels at $F_n > 0.35$.

The dominance of friction at low speed is why slow-steaming bulk carriers benefit strongly from hull coating investment, whereas fast container ships must also address wave-making.

Form factor and viscous pressure resistance

Three-dimensional viscous effects

A hull is not a flat plate. The flow accelerates around the bow, decelerates toward the stern, and the resulting adverse pressure gradient can cause boundary layer thickening and partial separation at the stern. The additional pressure drag from this effect is the viscous pressure resistance $R_{VP}$. It is physically distinct from wave-making (which is an inviscid free-surface phenomenon) but cannot easily be separated from friction in a single measurement.

The ITTC-78 method handles this by the form factor $(1+k)$, introduced by Hughes (1954). The viscous resistance is written as:

so that the form factor captures the three-dimensional amplification of the flat-plate friction. This is equivalent to assuming that viscous pressure resistance is proportional to, and scales with, the frictional resistance as a function of Reynolds number.

Measurement: the Prohaska method

The form factor is determined experimentally by running model resistance tests at very low Froude numbers (typically $F_n < 0.12$) where wave-making is negligible. At those speeds:

Plotting $C_T/C_F$ against $F_n^4/C_F$ and extrapolating the linear fit to the $y$-axis gives $(1+k)$. The form factor (Prohaska) calculator automates this regression. Typical ranges:

A tanker with block coefficient $C_B = 0.83$ and a full, full-bodied stern has a markedly higher form factor than a container ship with $C_B = 0.67$ and a fine, hollow run. The form factor also increases with transom area and stern flare.

ITTC-78 vs ITTC-57 extrapolation

The older ITTC-57 method used $(1+k) = 1.0$ (flat-plate friction only) and absorbed all three-dimensional viscous effects into $C_A$. The ITTC-78 performance-prediction method introduced the explicit $(1+k)$ term, reducing the required $C_A$ and improving consistency between different model basins. For vessels with extreme hull forms the difference between the two methods in predicted full-scale $C_T$ can reach 4-6%.

Wave-making resistance and the Kelvin wave system

Physics of the Kelvin wake

A ship moving at constant speed on a calm surface disturbs the pressure field in the water and at the surface, generating a pattern of transverse and divergent gravity waves. Lord Kelvin (1887) showed that, independent of ship speed, the wave pattern is confined within a half-angle of approximately 19.47 degrees from the ship’s track. Inside that angle, the wave energy radiates to infinity; the work done in maintaining it is the wave-making resistance.

The wave system has two primary sources: the bow pressure field (which generates a crest at the stem) and the stern pressure field (which generates a trough approximately at the quarter-length). The transverse bow and stern wave trains interfere. At certain Froude numbers, the interference is constructive (crests add), producing a resistance hump; at other Froude numbers it is destructive (crest meets trough), producing a hollow. The principal hump for a ship with $L_{WL}/B = 6$ to 7 occurs near $F_n = 0.48$-$0.52$. A secondary hump is present near $F_n = 0.27$-$0.32$. Most loaded cargo vessels operate well below $F_n = 0.30$, but container ships at service speed often sit on the shoulder of the secondary hump.

Froude number and speed dependence

The wave-making resistance coefficient $C_W$ increases steeply with $F_n$. In the range $F_n = 0.20$-$0.35$, $C_W$ roughly doubles for each 15% increase in speed. Wave-making resistance therefore scales faster than $V^2$: depending on hull form and Froude number, $R_W \propto V^{4}$ to $V^{6}$ in the rising part of the resistance curve. This is why the “speed-cubed law” for power (derived from the admiralty coefficient method, which includes wave-making) understates the steepness for ships operating above their hull speed threshold.

The wave resistance coefficient can be written:

where the functional dependence is determined empirically (Holtrop-Mennen, Hollenbach), by slender-body theory (Michell 1898 integral for fine hulls), or by 3D potential-flow panel methods (Dawson, NEWDRIFT). For the ITTC extrapolation method, $C_W$ is obtained from the total model resistance minus the ITTC-1957 viscous component:

This is then transferred directly to ship scale without correction, since wave-making obeys Froude similarity and does not depend on Reynolds number when the model speed is set to match $F_n$.

Resistance hump and hollow in practice

The following table summarises the speed-dependence pattern and the fraction of total resistance at typical operating points:

Note: ranges are indicative from published systematic-series and full-scale trial data; actual values depend on hull form coefficients, loading condition, and appendage configuration.

Bulbous bow: wave cancellation mechanism

A bulbous bow reduces wave-making resistance through phase cancellation. The bulb creates a pressure wave roughly 180 degrees out of phase with the wave generated at the main stem waterplane. The two waves partially cancel, reducing the net energy radiated in the Kelvin wake. A well-designed bulb reduces $R_W$ by 5-15% at the design speed & draft. The gains are speed-specific: a bulb calibrated for laden condition at 14 knots loses effectiveness at 12 knots and can increase resistance in ballast at shallow draft. The bulbous bow retrofits article covers the retrofit design methodology and the post-2020 reassessment driven by widespread slow-steaming.

Transom sterns

Most post-1980 commercial ships use a flat or nearly-flat transom stern. At low speed, the transom drags a ventilated dead-water region behind it, adding extra pressure drag. Above a critical Froude number (roughly $F_n^{LT} = 0.30$-$0.35$, referenced to the transom half-beam), the flow detaches cleanly from the transom edge and the dead-water region collapses. Holtrop and Mennen’s 1984 refinement includes a transom pressure resistance term $R_{TR}$ that accounts for this transition. The Holtrop transom resistance calculator evaluates $R_{TR}$ for given transom geometry and Froude number.

Model-to-ship extrapolation: the Froude-Reynolds conflict

Why similarity cannot be simultaneously satisfied

The central practical problem in model testing is that two dimensionless parameters govern the resistance: Froude number $F_n = V/\sqrt{gL}$ governs gravity (wave) effects, and Reynolds number $R_n = VL/\nu$ governs viscous (friction) effects. For a geometrically similar model at scale ratio $\lambda$ (model length $L_m = L_s/\lambda$):

• Froude similarity requires $V_m = V_s/\sqrt{\lambda}$.
• Reynolds similarity requires $V_m = V_s \cdot \lambda$.

The two conditions are simultaneously satisfiable only for $\lambda = 1$ (full scale) or for a fluid with kinematic viscosity $\nu_m = \nu_s/\lambda^{1.5}$ (which does not exist at model scales of $\lambda = 20$-$80$). In practice, towing tanks operate at Froude similarity. A model at scale 1:40 runs at $V_m = V_s/\sqrt{40} = 0.158\,V_s$, giving $R_{n,m} = R_{n,s}/40^{1.5} = R_{n,s}/253$. A 200 m ship at $R_n = 4 \times 10^8$ corresponds to a model at $R_n = 1.6 \times 10^6$, a factor of 250 lower. Since $C_F$ changes by about 15% per decade of Reynolds number in the turbulent regime, this discrepancy matters and is why the ITTC 1957 line and form factor exist: to compute and subtract the correct friction at both model and ship Reynolds numbers rather than assume the friction correction is negligible.

The ITTC-78 extrapolation procedure

The procedure applied by every towing tank for a powering prediction is:

1. Run resistance model tests at a range of Froude numbers matching the service speed range. The model is fitted with turbulence stimulators (studs or sand strips) at about 5% of $L_{WL}$ from the bow to trip the boundary layer to turbulent and avoid premature laminar separation.
2. Measure $C_{T,m}$ at each speed.
3. Subtract the ITTC-1957 viscous component: $C_{W} = C_{T,m} - (1+k)\,C_{F,m}$.
4. Add back the ship-scale viscous component: $C_{T,s} = (1+k)\,C_{F,s} + C_W + C_A + C_{app} + C_{A\!A}$.

The correlation allowance $C_A$ is applied at step 4. The ITTC recommends a default $C_A = 0.5 \times 10^{-3}$ for a newly painted hull with average roughness height $k_s \approx 150\,\mu$m. A worn or fouled hull may require $C_A = 1.0$-$2.0 \times 10^{-3}$; an exceptionally smooth coated hull (silicone foul-release) may justify $C_A = 0.2 \times 10^{-3}$. The correlation allowance calculator implements the ITTC-78 formula for $C_A$ as a function of measured mean hull roughness.

Effective power and its components

Definition

Effective power is:

It is the power that would be required to tow the bare hull at speed $V$ with no propulsive losses. For a laden Aframax tanker ($R_T \approx 650$ kN at 14.5 knots / 7.46 m/s):

Delivered power $P_D = P_E / \eta_D$, where $\eta_D$ is the quasi-propulsive coefficient, typically 0.60-0.72 for single-screw vessels. For that tanker at $\eta_D = 0.67$, $P_D \approx 7{,}240$ kW, and shaft power $P_S \approx 7{,}500$ kW after shaft-bearing and stern-tube losses.

Components of PE

Breaking $P_E$ into its resistance components:

where $P_{E,V} = R_V \cdot V = (1+k)\,C_F\,\tfrac{1}{2}\rho V^3 S$ is the dominant term for slow ships. The strong cubic dependence on speed of each term (because $R \propto V^2$ and $P = R \cdot V$) is why a 10% speed reduction cuts $P_E$ by about 27% and fuel consumption per unit time by close to the same fraction, all other things being equal.

EEDI and EEXI connection

The Energy Efficiency Design Index (EEDI, MARPOL Annex VI Reg.20) and the Energy Efficiency eXisting ship Index (EEXI, Reg.23) both link hull resistance directly to the regulatory carbon intensity rating of a vessel. Both indices are proportional to engine power, and engine power is sized to provide the required effective power at the design (or, for EEXI, the reference) speed. A 1% reduction in $C_T$ at the design waterline propagates to a 1% reduction in required $P_E$, roughly 1% reduction in required $P_B$ (after constant propulsive efficiency), and therefore a 1% improvement in the attained EEDI. The EEXI, EPL and ShaPoLi article and the what is EEDI article cover the regulatory framework; hull resistance reduction through form optimization, coating selection, and anti-fouling management is one of the primary compliance pathways.

Appendage resistance

Components and typical magnitudes

Every structure below the waterline that is not part of the main hull contributes appendage resistance. The principal items and their typical resistance fraction for a single-screw vessel:

For a twin-screw vessel (container ships, RoPax, ferries), shaft bossings and struts can push total appendage resistance to 6-10% of $R_T$. The Holtrop-Mennen method estimates appendage resistance via wetted-area form factors assigned to each appendage type; the Holtrop appendage resistance calculator implements those coefficients.

Reduction strategies

Rudder bulbs reduce the rudder-induced drag by recovering some of the propeller’s rotational kinetic energy (they also improve propulsive efficiency). Streamlined bow thruster tunnel covers reduce the 0.3-1.2% tunnel drag to near zero when thrusters are inactive. Twin-skeg designs replace open bossings with a more fairing hull form and typically reduce appendage resistance by 30-50% relative to an equivalent twin-shaft arrangement.

Air resistance

Calculation

Air resistance acts on the above-water superstructure, deck cargo (container stacks), and the hull above the waterline. The standard formulation is:

where $\rho_{air} = 1.225$ kg/m$^3$ at 15 °C & 1 atm, $V_{ar}$ is the apparent wind speed (vector sum of ship speed and true wind velocity), $A_T$ is the projected transverse (frontal) area of the vessel above the waterline, and $C_X$ is the longitudinal aerodynamic force coefficient, which depends on wind heading angle and superstructure geometry.

For calm-water headway with no true wind, $V_{ar} = V_{ship}$ and $C_X$ is the head-on drag coefficient, typically 0.5-0.8 for most merchant ships.

Wind coefficients

The two standard empirical datasets are:

• Isherwood (1972): regression on wind tunnel data for typical merchant superstructure types, published in RINA Transactions. Coefficients cover bow, quarter, beam, and stern wind headings. The wind resistance calculator implements Isherwood coefficients.
• Blendermann (1994): an extended dataset covering a wider range of vessel types including container ships with high stack loads, car carriers, and cruise ships. Recommends type-specific coefficients for $C_X$, $C_Y$ (lateral), and the moment coefficient.

Fraction of total resistance by vessel type

Air resistance is 2-5% of calm-water $R_T$ for most cargo ships with low freeboard superstructures. It rises substantially for vessels with large frontal areas:

• Container ships (full container stack, 23,000 TEU): $A_T$ of 2,000-2,500 m$^2$ and $R_{A\!A}/R_T$ of 7-12% in head-on conditions.
• Cruise ships and large passenger vessels: superstructure $A_T$ of 3,000-4,500 m$^2$; air resistance 10-18% of $R_T$.
• Car carriers (PCTC): very high and boxy topsides; $R_{A\!A}/R_T$ of 10-16%.
• Loaded VLCCs: low freeboard, small accommodation block; $R_{A\!A}/R_T$ of 1-3%.

For wind-assisted propulsion systems (Flettner rotors, wing sails, kite sails), the aerodynamic framework is the same: $C_X$ becomes a thrust coefficient rather than a drag coefficient when the apparent wind angle is favorable. The net benefit depends on the ratio of thrust force along the ship’s course to the rotor/sail’s own drag plus any increase in hull resistance from leeway.

Hull roughness and the correlation allowance

Effect of roughness on friction

A newly built hull with modern ablative antifouling typically has mean hull roughness $k_s$ in the range 100-200 $\mu$m (measured by the BM-measure or BSRA roughness gauge). At ship Reynolds numbers, this roughness is hydraulically transitional and increases $C_F$ above the smooth-surface ITTC value. The increase is captured in the correlation allowance $C_A$.

Schultz (2007) tested flat plates coated with representative fouling conditions in a recirculating water channel and found that:

• A typical newly painted hull (roughness $\sim$150 $\mu$m, Schultz “AF-SPC”) adds $C_A \approx 0.4 \times 10^{-3}$ over the smooth-plate value.
• A lightly fouled hull (slime plus scattered barnacles, $\sim$500 $\mu$m equivalent) adds $C_A \approx 1.0$-$1.5 \times 10^{-3}$, a power increase of 10-15%.
• Heavy barnacle or calcareous fouling ($\sim$3{,}000 $\mu$m) can increase resistance by 50-60% for a vessel operating in the friction-dominated regime.

That last figure means a VLCC with heavy hull fouling needs 50-60% more effective power to maintain speed, which typically translates to an equivalent increase in fuel consumption, since the propulsive efficiency changes only modestly. The service allowance calculator quantifies the combined effect of roughness, fouling, and weather margin on the delivered power budget.

Coating strategies

Foul-release coatings (typically polydimethylsiloxane, PDMS, or fluoropolymer-based) maintain lower surface energy than conventional antifouling and achieve $k_s$ in the 80-120 $\mu$m range after application. Schultz (2007) showed that a foul-release surface at $k_s \approx 90\,\mu$m reduced $C_F$ by about 2% relative to a conventionally painted hull, corresponding to roughly 0.6-1.0% fuel saving at the operating point of a tanker. Hull polishing at drydock (to $k_s < 100\,\mu$m) and groove textures aligned with the flow direction have been tested by several yards as additional friction-reduction measures.

Added resistance in waves

Sources of added resistance

In irregular seas, a vessel experiences mean added resistance $R_{AW}$ in addition to the calm-water value. The sources are:

1. Wave reflection at the bow: incident waves are partly reflected; the resulting momentum change adds a force opposing ship motion. This term dominates at short wave periods (wavelengths below about $0.4 L_{WL}$).
2. Radiation damping: the ship’s heaving, pitching, and rolling motions in waves generate radiated wave energy. The energy required to maintain those motions is the radiation-damping added resistance; it dominates near the natural heave/pitch resonance, typically at $\lambda/L \approx 0.6$-$1.0$.
3. Mean second-order forces: oblique waves generate a lateral drift force and yaw moment; the rudder counteracts yaw, adding rudder drag.

Calculation methods

The main methods for practical calculations are:

• Salvesen, Tuck & Faltinsen (1970, STF method): strip theory, frequency-domain solution for all six degrees of freedom; the industry workhorse for seakeeping and added resistance prediction. See seakeeping for the full framework.
• Boese (1970): simplified strip-theory formula applicable during early design.
• Maruo (1957): far-field momentum formulation; theoretically rigorous for head waves.
• STAWAVE-2 (ITTC 2014): empirical formula derived from systematic tank tests; the method recommended in ITTC Procedure 7.5-02-07-03.9 for CII performance calculations and ISO 15016:2015 (speed-power-performance trials). It uses hull geometry, Froude number, and wave parameters directly. The STAWAVE-2 calculator implements this formula.

For CII compliance projection under MARPOL Annex VI Reg.28, IMO’s interim guidelines (MEPC.1/Circ.896) allow the use of STAWAVE-2 or an equivalent method to correct log-speed-derived fuel data for sea-state conditions.

Weather margin

In ship design practice, a “sea margin” or “service allowance” of 15-25% is added to calm-water trial power to define the continuous service rating. The allowance covers average year-round sea conditions on the trade route, wind resistance, and minor fouling between drydockings. For vessels operating in high-latitude North Atlantic or North Pacific trades, the allowance may reach 30%. The service allowance calculator provides the standard IMO method per ISO 15016 for route-specific correction.

Shallow-water resistance effects

When depth-to-draft ratio $h/T < 4$, the channel confinement modifies both the wave system and the local flow velocity around the hull. Three mechanisms increase resistance:

1. Froude depth number effect: the wave phase speed in shallow water is $\sqrt{gh}$ rather than $\sqrt{gL}$. As ship speed approaches $\sqrt{gh}$, wave-making resistance rises sharply (the “ship-critical” condition). Most bulk-carrier port approaches operate at $F_{nh} = V/\sqrt{gh}$ of 0.5-0.7, well below criticality, but the resistance increase is still material at 3-15%.
2. Blockage: in narrow channels, the hull’s blockage of the cross-section (blockage coefficient $m = A_{MS}/A_{channel}$) increases local flow velocities, which raises frictional resistance and modifies the pressure distribution.
3. Squat: the decreased pressure under the hull causes sinkage and trim change (squat effect), increasing the wetted surface area.

The Schlichting (1934) correction and Lackenby (1963) refinement convert deep-water resistance to shallow water by adjusting effective ship speed at the same resistance value. The shallow-water resistance calculator applies both methods.

Hull-form coefficients and their resistance implications

The naval architecture coefficients article covers the full set of form coefficients. In the resistance context, the key ones are:

• Block coefficient $C_B = \nabla / (L_{WL} \cdot B \cdot T)$: the primary indicator of hull fineness. Increasing $C_B$ from 0.65 to 0.80 increases the form factor $(1+k)$ by roughly 0.10-0.15 and moves the wave-resistance hump to lower Froude numbers, reducing the design speed for minimum wave-making.
• Prismatic coefficient $C_P = \nabla / (A_{MS} \cdot L_{WL})$: controls the longitudinal distribution of displacement. Froude’s original observation was that resistance is minimized for a given speed when $C_P$ matches a specific value; the “optimum $C_P$” as a function of $F_n$ is a standard design chart (SNAME PNA Figure 8, originally derived from Series 60 data). Roughly, the optimum $C_P \approx 0.59 + 0.54\,F_n$ for $F_n = 0.15$-$0.35$.
• Midship coefficient $C_M = A_{MS}/(B \cdot T)$: affects stern fullness and bilge radius; a full midship section with no deadrise tends to increase the form factor.
• Waterplane coefficient $C_{WP} = A_{WP}/(L_{WL} \cdot B)$: governs the longitudinal distribution of restoring waterplane area and influences wave system characteristics.

The block coefficient calculator provides a cross-reference of these coefficients; the Holtrop-Mennen resistance calculator accepts all four as direct inputs.

Empirical resistance prediction methods

Holtrop-Mennen (1982, 1984)

Holtrop and Mennen developed their regression method at MARIN (Netherlands Maritime Research Institute) from a database of 334 ship model tests run between 1973 and 1982, augmented in the 1984 paper with full-scale trial data. The method predicts total calm-water resistance by regression equations for:

• Frictional resistance via ITTC 1957 and Holtrop form factor $k_1$.
• Wave resistance $R_W$ as a function of $C_P$, $L_{WL}/B$, $B/T$, $F_n$, and the half-angle of entrance $i_E$.
• Transom pressure resistance $R_{TR}$.
• Model-ship correlation allowance $R_A$.
• Appendage resistance via wetted-area $\times$ appendage form factors.

Accuracy is typically $\pm$5% for hull forms within the training data range ($C_P = 0.55$-$0.85$, $F_n = 0.10$-$0.45$, $L_{WL}/B = 3.9$-$9.5$). The Holtrop-Mennen calculator implements the full 1984 regression. Subsidiary calculators address specific terms: form factor $k_1$, wave resistance, transom resistance, appendage resistance.

Hollenbach (1998)

Hollenbach (HSVA, Hamburg) regressed a database of 434 model tests (1980-1995), with separate regression branches for single-screw and twin-screw vessels, and specific attention to the range $F_n = 0.14$-$0.32$ where most slow-steaming container ships and bulk carriers operate. His method gives better accuracy than Holtrop-Mennen for full-form single-screw hulls and for vessels with $L_{WL}/B < 5.0$. The Hollenbach calculator covers both single and twin-screw branches.

Taylor-Gertler series and Series 60

The Taylor Standard Series (1910, extended by Gertler 1954) covered a systematic variation of prismatic coefficient and displacement-length ratio for naval hull forms. Series 60 (Todd, 1963) was a SNAME-sponsored follow-on covering merchant ships with $C_B = 0.60$-$0.80$ and $L_{WL}/B = 6.0$-$8.5$. Both series remain in use for early conceptual design and as validation benchmarks; the Taylor-Gertler resistance calculator provides rapid estimates from that dataset.

CFD approaches

Reynolds-averaged Navier-Stokes (RANS) solvers with free-surface volume-of-fluid methods now provide full resistance prediction at typical towing-tank accuracy ($\pm$2-5%) for well-resolved grids. The barrier to adoption is grid generation time and computing cost, not accuracy: a single resistance run for a new hull form at five speeds requires 8-24 hours on a 64-core workstation with an $\sim$5 million cell mesh. CFD is used routinely in newbuild optimization (bow form, bulbous bow shape, stern-tube area) and is becoming standard for retrofit performance prediction (bulbous bow removal for slow-steaming, stern duct design). For the purposes of ITTC-recommended uncertainty analysis, CFD predictions carry an additional validation uncertainty term.

Froude scaling and the model-to-full-scale gap

Froude scaling law

Froude’s law states that resistance coefficients ($C_T$, $C_W$) are equal for geometrically similar hulls at the same Froude number, regardless of absolute size. This is the basis for all model testing: the wave-making component transfers directly. The quantitative statement is:

where $\lambda^3$ is the scale-volume ratio and $\rho_s/\rho_m$ is the density ratio (fresh water in most tanks, salt water at sea). Since $C_W$ is dimensionless, $C_{W,model} = C_{W,ship}$ at the same $F_n$.

The full-scale extrapolation gap arises entirely from the friction discrepancy. At the same $F_n$, the model’s $C_F$ (per ITTC 1957) is higher than the ship’s $C_F$ because $R_{n,model} < R_{n,ship}$. If the naive assumption $C_{T,ship} = C_{T,model}$ were used without the ITTC 1957 correction, the ship power would be overpredicted by 8-15% for typical merchant vessels. The ITTC-78 procedure corrects this by computing $C_F$ at both Reynolds numbers and transferring only $C_W$.

Tank correlation factors

Even after applying ITTC-78, systematic biases remain between towing-tank predictions and full-scale trial results. These biases are tank-specific (MARIN, HSVA, SSPA, NMRI all have slightly different historical databases and turbulence stimulator practices) and hull-form-dependent. Major model basins maintain proprietary correlation databases that adjust predictions to match their historical full-scale accuracy.

The ISO 19030 standard (published 2016, covering monitoring of changes in hull and propeller performance) uses the same $C_T$ framework to track in-service degradation from trials through the vessel’s operational life.

Resistance reduction: hull-form and operational strategies

Hull-form measures

The resistance components each have specific hull-form levers:

• Frictional: minimize wetted surface area $S$ for the given displacement. The optimum hull form for minimum $S$ is a semicircular cross-section (minimum perimeter), but structural and stability requirements prevent this. Fine L/B ratio and moderate $C_M$ reduce $S/\nabla^{2/3}$. See wetted surface area for the Mumford, Denny, and Henschke formulas.
• Viscous pressure: minimize form factor by ensuring smooth stern lines with no abrupt fullness increase aft of the widest section. Stern bulbs and asymmetric stern fins reduce bilge vortex shedding and lower $(1+k)$ by 0.02-0.05.
• Wave-making: optimize $C_P$ and the longitudinal distribution of buoyancy for the design $F_n$. Bulbous bow for wave cancellation at design speed. Narrow entrance half-angle $i_E$ (below 25 degrees) reduces bow-wave amplitude.
• Appendage: minimize wetted appendage area. Use twin-skeg instead of open bossings on twin-screw vessels. Fit bow thruster covers.
• Air: reduce frontal area $A_T$ of superstructure and minimize $C_X$ through streamlined superstructure design. For container ships, optimized container-stack arrangement (echeloned, sloped top tier) reduces $C_X$ by 5-15%.

Operational measures

The trim optimisation article covers how even-keel versus design-trim adjustments typically reduce $R_T$ by 1-4% at given displacement by shifting the center of the wave system. For bulk carriers and tankers, optimal trim at each loading condition is now tracked by onboard performance-monitoring systems. Air lubrication systems inject microbubbles under the flat-bottom hull to reduce frictional resistance, with claimed savings of 5-8% in full-scale trials for VLCC and bulk-carrier hull forms. Energy-saving devices (wake-equalizing ducts, pre-swirl stators, propeller boss cap fins) operate primarily on propulsive efficiency rather than hull resistance, but the two interact: the wake field behind the hull governs the inflow to those devices.

For the powering calculation pipeline from resistance through to fuel consumption and CII projection, the ship resistance and powering article covers propulsive coefficients, wake fraction, and thrust deduction factor. For the upstream naval-architecture context (displacement, waterplane, form coefficients), see naval architecture coefficients. Propeller design and interaction with the hull wake are covered in propeller theory.

Limitations

Empirical method validity ranges. Holtrop-Mennen and Hollenbach are regressions against specific training databases. Both methods degrade for hull forms outside those databases: very high-speed planing vessels ($F_n > 0.50$), semi-displacement forms, multihulls, unconventional stern arrangements, and vessels with $C_B > 0.86$. Extrapolating beyond the stated validity range can produce errors of 15-30%.

Form factor uncertainty. The Prohaska regression to determine $(1+k)$ is sensitive to the low-speed resistance data quality. Turbulence stimulators must fully trip the boundary layer at the test speeds; if they do not, the extrapolated $(1+k)$ is too low and the full-scale power is underpredicted. ITTC Procedure 7.5-02-02-01 specifies the stimulator geometry and placement for this reason.

Scale effects on wave-making. Froude similarity fails in detail for transom sterns because the free-surface intersection at the transom is a function of $F_n$ referenced to transom geometry, not hull length. For hulls with large transoms ($A_T/A_{MS} > 0.25$), the Holtrop transom correction introduces additional scatter of up to $\pm$8%.

Added resistance in waves: method-specific scatter. STAWAVE-2 was calibrated against head-sea conditions at unit wave steepness. For oblique seas, beam seas, or vessels with unusual bow geometry (flare, spoon bow, X-bow), the method may underpredict added resistance by 20-40%. RANS-based seakeeping calculations in irregular waves are more reliable but are still not routine at design stage.

Roughness and fouling uncertainty. Schultz (2007) measured equivalent sand roughness under controlled laboratory conditions. In-service hulls have heterogeneous roughness distributions (better at the waterline, worse at the flat bottom) that are not captured by a single $k_s$ value. The standard ITTC $C_A = 0.5 \times 10^{-3}$ is a statistical average; individual ships deviate by $\pm 0.3 \times 10^{-3}$, corresponding to $\pm$5-8% in effective power at the friction-dominated operating point.

Shallow water limits. Schlichting-Lackenby corrections are one-dimensional; they do not account for channel shape, bank proximity, or unsteady squat at low under-keel clearance. For port approach at $h/T < 1.5$, dedicated computational hydrodynamics or scale-model tests in a shallow-water channel are required.

CII and operational corrections. Using calm-water resistance to project CII performance ignores wave-induced speed loss, which varies with weather pattern and trade route. On North Atlantic or North Pacific routes, the annual mean speed loss from waves may be 3-8%; for operators projecting CII compliance three years ahead, this uncertainty is larger than the EEDI/EEXI noise floor.

See also

• Ship resistance and powering
• Wetted surface area
• Naval architecture coefficients
• Propeller theory
• Bulbous bow retrofits
• Trim optimisation
• Seakeeping
• Squat effect
• Wind-assisted propulsion
• Air lubrication systems
• Energy-saving devices
• EEXI, EPL and ShaPoLi
• What is EEDI
• What is CII
• Hull form design
• Block coefficient
• Holtrop-Mennen resistance calculator
• ITTC 1957 friction coefficient calculator
• Form factor (Prohaska) calculator
• Correlation allowance calculator
• Hollenbach resistance calculator
• Holtrop wave resistance
• Holtrop transom resistance
• Holtrop appendage resistance
• Holtrop form factor $k_1$
• Air resistance calculator
• Wind resistance calculator
• STAWAVE-2 added resistance calculator
• Shallow-water resistance calculator
• Schoenherr friction line calculator
• Service allowance calculator
• Froude speed scaling calculator
• Reynolds number calculator