Ship resistance and powering: from hull to installed power
Every powering estimate answers the same question: how large an engine does this hull need to make this speed? The answer is built in three stages. First the resistance of the hull in calm water, then the losses between the engine flange and the water, then the margins that keep the speed available when the weather and the hull are no longer ideal.
This guide follows that chain for one ship, with the numbers produced by the Holtrop–Mennen calculator, the Wageningen B-series calculator and the propeller curve calculator. It is the estimate you make at the concept stage, before a model test or CFD, and the one a performance analyst makes when a ship has no usable reference curve.
The components of resistance
Holtrop and Mennen split the calm-water resistance into six parts:
R_T = R_F (1 + k1) + R_APP + R_W + R_B + R_TR + R_A
- Friction, R_F. The drag of a flat plate with the ship's wetted surface S, from the ITTC-1957 line: C_F = 0.075 / (log10 Rn − 2)², R_F = ½ ρ V² S C_F. The Reynolds number is Rn = V L / ν, with ν = 1.18831 × 10⁻⁶ m²/s for sea water at 15 °C.
- Form factor, 1 + k1. A hull is not a flat plate. Its curvature speeds the flow up along the sides and thickens the boundary layer aft, so the viscous resistance is the plate friction multiplied by a factor that depends on the fullness and the shape of the afterbody.
- Appendages, R_APP. Rudder, bilge keels, shaft brackets: their wetted area with the same C_F and their own factor 1 + k2 (1.3 to 1.5 for a rudder behind the stern, 1.4 for bilge keels, 3.0 for shaft brackets).
- Wave making, R_W, with separate pressure terms for a bulbous bow near the surface (R_B) and an immersed transom (R_TR).
- Correlation allowance, R_A = ½ ρ V² S C_A. Fitted to trial data, it covers a hull roughness of 150 µm, the still-air resistance and the model-to-ship scaling effects.
Wind, waves, shallow water and fouling are not in R_T. They are handled separately, with the wind resistance, added resistance in waves and shallow water calculators, or covered by the sea margin.
Froude number, Reynolds number and the cube law
Two dimensionless numbers govern the two halves of the resistance. The Reynolds number sets the friction coefficient, which falls slowly as the ship gets longer or faster. The Froude number, Fn = V / √(g L), sets the wave pattern: ships at the same Froude number make geometrically similar waves. The transverse wave the ship makes has a length λ = 2π V² / g, and it becomes as long as the hull at Fn = 0.40, the hull speed. The main resistance hump lies near Fn 0.5.
If the resistance coefficients were constant, resistance would rise with V² and power with V³. That is the cube law, and it is only a local approximation. C_F falls a little with speed, but the wave resistance grows much faster than V² as the Froude number rises, so the real exponent in P ∝ Vⁿ is above 3 for most ships: 3.2 to 3.5 is typical for container ships and ferries, and it climbs steeply towards the hull speed. The worked example below shows an exponent of nearly 3.9.
The Holtrop–Mennen method
The method is a statistical regression, published by Holtrop and Mennen in 1982 (International Shipbuilding Progress 29) and re-analysed by Holtrop in 1984 (ISP 31) on 334 models and 168 full-scale trials. It predicts each component above from the main particulars, so the inputs are the ones a designer has early:
- waterline length L, breadth B, draughts forward and aft, displacement volume ∇;
- midship and waterplane coefficients C_M and C_WP (the block coefficient calculator gives C_B, C_P and the others from the dimensions);
- the longitudinal centre of buoyancy lcb, in per cent of L forward of midships (negative aft);
- bulb section area and its height, immersed transom area, appendage area, stern shape;
- propeller diameter and blade area ratio, for the propulsion factors.
The 1984 variant has a new form factor and a wave resistance formula that reaches higher speeds. Use it for new work and say which variant you used, because codes that mix the two are common.
Range of validity
The regression describes hulls like those in its sample: roughly C_P 0.55 to 0.85, L/B 3.9 to 15, B/T 2.1 to 4.0, and Froude numbers up to about 0.45 for the 1982 formulas or 0.8 for the 1984 one. Inside that range expect ±5 to 10 % on power. Between Fn 0.40 and 0.55 the 1984 wave resistance is a linear interpolation that Holtrop himself calls "more or less arbitrary". Outside the range the formulas still return a number, and it is an extrapolation.
From effective power to brake power
The effective power is what it would take to tow the hull: P_E = R_T V. The engine has to deliver more, because the propeller works in the hull's wake and loses energy of its own.
T = R_T / (1 − t) thrust the propeller must give
V_A = V (1 − w) speed of advance into the propeller
η_H = (1 − t) / (1 − w) hull efficiency
η_D = η_H × η_R × η_O quasi-propulsive efficiency
P_D = P_E / η_D delivered power at the propeller
P_S = P_D / η_S shaft power
- Wake fraction w: the water at the propeller already moves forward with the hull, so the propeller advances more slowly than the ship.
- Thrust deduction t: the propeller lowers the pressure on the stern, so it must give more thrust than the towed resistance.
- Hull efficiency η_H combines the two. On a single-screw ship it is usually above 1, because the propeller recovers energy from the wake.
- Relative-rotative efficiency η_R corrects for the difference between open water and the uneven flow behind the hull. Holtrop gives η_R = 0.9922 − 0.05908 A_E/A_0 + 0.07424 (C_P − 0.0225 lcb) for a single screw.
- Open-water efficiency η_O = J K_T / (2π K_Q) belongs to the propeller alone, with the advance coefficient J = V_A / (n D). It comes from the B-series polynomials (Oosterveld and van Oossanen, 1975).
- Shaft efficiency η_S: bearing and stern tube losses, 0.99 by default. On a direct-coupled two-stroke engine the brake power is practically the shaft power.
Sea margin, engine margin and light running
P_S is the power on an ideal trial: calm deep water, clean hull. The engine is sized from this propeller design point (PD) with three margins, here in the form of the Everllence (formerly MAN Energy Solutions) project guide:
SMCR power = P_PD × (100 + SM) / (100 − EM)
SMCR rpm = n_PD × (SMCR power / P_PD)^(1/3) × (1 − LRM/100)
The sea margin SM is extra power to hold the service speed in average weather with a hull in average condition, traditionally about 15 %. The engine margin EM keeps the service point at 85 to 90 % of the specified maximum continuous rating (SMCR). The light running margin LRM is different in kind: it is an rpm margin. A fouled hull or a head sea makes a fixed pitch propeller "heavy", needing more torque at a given rpm. If the engine were laid out on the clean-hull propeller curve it would reach its torque limit before its rated power. So the layout curve is placed 4 to 7 % lower in rpm than the light curve, up to 10 % for ships that often meet adverse weather.
Worked example: a 205 m single-screw ship at 25 knots
The ship is the example hull of the 1982 paper: L 205 m, B 32 m, draught 10 m level, ∇ 37 500 m³, lcb −0.75 % of L, C_M 0.98, C_WP 0.75, bulb 20 m² centred 4 m above the keel, transom 16 m², appendages 50 m² with 1 + k2 = 1.5, U-shaped stern sections, an 8 m four-bladed propeller. Sea water at 15 °C, ρ 1025 kg/m³. The 1984 variant is used.
- Coefficients. C_B = 37 500 / (205 × 32 × 10) = 0.572, C_P = 0.572 / 0.98 = 0.583. V = 25 kn = 12.861 m/s, so Fn = 12.861 / √(9.807 × 205) = 0.287. All inside the range.
- Friction. Rn = 12.861 × 205 / 1.18831 × 10⁻⁶ = 2.22 × 10⁹, C_F = 0.001390. The wetted surface formula gives S = 7 381 m². With ½ρV² = 84.77 kPa, R_F = 84.77 kPa × 7 381 m² × 0.001390 ≈ 869.6 kN. The form factor is 1.185, so the viscous resistance is 1 031 kN.
- The other components. R_APP = 8.8 kN, R_W = 554.0 kN, R_B = 0.05 kN, R_TR = 0, and with C_A = 0.000352, R_A = 220.6 kN.
- Total. R_T = 1 031 + 8.8 + 554.0 + 0.05 + 0 + 220.6 = 1 814 kN. Wave making and its pressure terms are 30.5 % of it.
- Effective power. P_E = 1 814 kN × 12.861 m/s = 23 331 kW.
- Propulsion factors. w = 0.2749, t = 0.1984, so η_H = 0.8016 / 0.7251 = 1.105. Thrust T = 1 814 / 0.8016 = 2 263 kN; V_A = 12.861 × 0.7251 = 9.326 m/s.
- Propeller. In design mode the B-series calculator takes T, V_A and the 8 m diameter. With the shaft 5.8 m below the surface Keller's cavitation screen asks for a blade area ratio of at least 0.762, so take 0.77. The optimum is P/D = 1.006 at 101.8 rpm, J = 0.687, η_O = 0.641.
- Powers. With A_E/A_0 = 0.77, η_R = 0.991. η_D = 1.1055 × 0.9913 × 0.641 = 0.7024. P_D = 23 331 / 0.7024 = 33 215 kW and P_S = 33 215 / 0.99 = 33 550 kW.
- Engine. With 15 % sea margin, 10 % engine margin and 5 % light running margin: SMCR = 33 550 × 115 / 90 = 42 869 kW at 101.8 × (42 869 / 33 550)^(1/3) × 0.95 = 104.9 rpm. The service point is 90 % of that, 38 583 kW at 101.3 rpm. On sea trial the clean ship would reach 110.5 rpm at full power, 105.3 % of the SMCR speed and inside the 110 % limit of the standard load diagram.
Running the same hull over a range of speeds, with η_O held at 0.641 for simplicity, shows how the curve bends:
| Speed (kn) | Fn | R_T (kN) | Wave share (%) | P_E (kW) | P_S (kW) |
|---|---|---|---|---|---|
| 19 | 0.218 | 863 | 13.4 | 8 435 | 12 111 |
| 21 | 0.241 | 1 102 | 17.9 | 11 902 | 17 098 |
| 23 | 0.264 | 1 421 | 24.3 | 16 817 | 24 171 |
| 25 | 0.287 | 1 814 | 30.5 | 23 331 | 33 550 |
| 27 | 0.310 | 2 217 | 34.2 | 30 794 | 44 300 |
Between 21 and 25 kn the exponent is ln(33 550 / 17 098) / ln(25 / 21) = 3.87. The wave share more than doubles over the table, which is where the extra above the cube comes from. The Holtrop–Mennen calculator reproduces every line; send its thrust and speed of advance to the B-series calculator for step 7 and the shaft power to the propeller curve calculator for step 9.
A quick check with the Admiralty coefficient
The Admiralty coefficient A_c = Δ^(2/3) V³ / P, with displacement in tonnes, speed in knots and power in kW, condenses a power point into one number. Here Δ = 37 500 × 1.025 = 38 438 t, so A_c = 1 139 × 15 625 / 33 550 = 530. Merchant ships at service speed fall roughly between 350 and 600, so the estimate is plausible. A value far outside usually means a unit slip.
The coefficient is also the cube law in disguise, and the example shows its limit. Scaling the 25 kn point down to 21 kn with n = 3 gives 33 550 × (21/25)³ = 19 885 kW. The resistance calculation gives 17 098 kW. The cube law is 16 % high for a 16 % speed change, because the true exponent here is 3.87. Use the coefficient to check a number or to scale within a few per cent of speed, with a fitted exponent where you have one.
Common mistakes
- The sign and datum of lcb. It is measured from half the waterline length, positive forward. The 1982 paper prints −2.02 % for its example, relative to half L_PP; the value that reproduces its own intermediate results is −0.75 % of L_WL.
- Mixing L_PP and L_WL. The method uses the waterline length. Coefficients formed on different lengths are not comparable.
- Using design coefficients at ballast draught. C_B, C_M and C_WP all change with draught, and so does the bulb's immersion.
- Leaving η_O at a default. A guessed open-water efficiency moves the power directly: 0.62 instead of 0.641 raises P_S in the example by 3.4 %. Design the propeller, then go back.
- Feeding the propeller the ship's speed. The B-series polynomials need V_A = V (1 − w) and the thrust R_T / (1 − t), never V and R_T.
- Ignoring the cavitation screen. A blade area below Keller's minimum gives a slightly better efficiency on paper and an eroded propeller in service. Keller itself is only a first screen.
- Counting margins twice, or not at all. P_S is a trial figure. If your reference power already includes weather, do not add a full sea margin on top; if it does not, an engine sized on P_S alone cannot hold the speed in service.
- Confusing sea margin with light running margin. One is power, the other is rpm. A ship can have ample power in hand and still be unable to use it because the propeller is too heavy.
Limits of the method
Holtrop–Mennen is a statistical description of conventional displacement hulls. It does not know the details of the lines: two hulls with the same coefficients and different bow shapes get the same answer. Twin-skeg sterns, very full or very slender forms, fast craft near or beyond the main hump and modern optimised bulbs all sit at or beyond the edge of the sample. The wake and thrust deduction formulas are for conventional sterns, and the B-series represents its own blade geometry; skewed or tip-unloaded propellers differ by a few per cent.
The estimate is therefore a starting point with a ±5 to 10 % band, to be replaced by model tests or validated CFD before a contract speed is guaranteed. The engine layout in step 9 follows the propeller law P = c n³ for a fixed pitch propeller; the torque and speed limits, barred speed ranges and the rating field need the engine maker's own data.
Calculators used in this guide
- MaritimeHoltrop–Mennen resistance and powerHoltrop–Mennen (1982/1984) resistance components, effective and brake power, wake fraction and thrust deduction from main particulars.
- MaritimeWageningen B-series propellerWageningen B-series KT, KQ and efficiency from the 1975 polynomials at a given J, or the optimum pitch ratio and rpm for a thrust and diameter; Keller check.
- MaritimePropeller curve and light running marginEngine SMCR from the propeller design point with sea, engine and light running margins, the propeller curves, and the heavy running of a measured point.
- MaritimeAdmiralty coefficient and speed–power scalingAdmiralty and fuel coefficients from a known point; power or speed at a new displacement and speed (cube or chosen exponent), with fuel per day and per mile.
- MaritimeFroude number calculator for ships and boatsLength and depth Froude numbers of a ship or boat, the critical speed in shallow water, the hull-speed and hump speeds, and the wave-making regime.
- MaritimeBlock coefficient and hull form coefficientsBlock, midship, prismatic and waterplane coefficients of a hull from its dimensions, and the displacement Δ = C_B L B T ρ in salt or fresh water.
- MaritimeWind resistance (ISO 15016)Added wind resistance and power from the relative wind by ISO 15016:2025 Annex C with the Fujiwara coefficients or wind-tunnel data.
- MaritimeAdded resistance in waves (STAwave-1)ISO 15016 STAwave-1 added resistance in head waves from beam, bow length and significant wave height, with the power increase at speed.
Results are estimates. They do not replace the governing code, class or flag requirements, or a check by a competent person. Report an error in this guide
More guides
- Cargo securing calculation to CSS Code Annex 13
- How to do a draft survey calculation, step by step
- How EU ETS and FuelEU Maritime costs are calculated
- How to calculate a ship's CII rating, step by step
- Hull and propeller performance in service with ISO 19030
- How to calculate ship squat and under-keel clearance