Skip to content
LumenaryAcademy
Systems 80 min

Vehicles and propulsion

What it takes to move — tractive effort, rolling resistance and drag computed for a car and an e-bike; combustion and electric powertrains compared over a drive cycle; jet thrust, propulsive efficiency and range; the rocket equation and staging; and ships and trains in brief.

Code tested · 6AI-draftedAwaiting review by a mechanical engineer

Your guides: Chloe · The Designer, Margaret · The Test Engineer, Gideon · The Checker

MotivationH

What it takes to move

A bicycle, a car, a freight train, a container ship, an airliner and a rocket all solve the same problem: keep a mass moving against whatever resists it, carrying the energy to do so. What resists them differs enormously. The car fights the air and the deformation of its tyres; the train, almost only the air; the ship, the waves it makes; the airliner must also hold itself up; the rocket must carry everything it will ever push against. How they carry their energy differs too: food, petrol, batteries, kerosene, heavy fuel oil, liquid oxygen and hydrogen.

This lesson follows the mechanical engineer's method across all of them: draw the forces, compute the power and energy, and follow the energy back to where it came from. It uses the dynamics of particles and rigid bodies, the engine cycles of Power and refrigeration cycles, the drag of Viscous flow and pipes and the motors of Mechatronics and actuators. The numbers it produces explain a great deal about the world: why motorway speed matters so much, why electric cars do best in towns, why airliners have huge fans, and why rockets come in stages.

DefinitionsH

Road vehicles: tractive effort

θF_tF_rF_dN_rN_fGmg
A car climbing a gradient at angle θ. The driven wheels push the road backwards, and the road pushes the car forwards with the tractive force F_t. Resisting it: rolling resistance F_r at the tyres, aerodynamic drag F_d, and the component of the weight mg along the slope. The normal forces N_f and N_r carry the rest of the weight.

Along the road, Newton's second law for the car (with its spinning wheels and drivetrain counted as a few per cent of extra mass) says that the force the road exerts on the driven tyres must supply four things.

Definition— road load and tractive force

The tractive force needed to drive a vehicle of mass mm at speed vv and acceleration aa up a gradient of angle θ\theta is

Ft=Crr mgcos⁡θ⏟rolling+12ρ CdA v2⏟aerodynamic drag+mgsin⁡θ⏟gradient+meff a⏟acceleration,F_t = \underbrace{C_{rr}\,mg\cos\theta}_{\text{rolling}} + \underbrace{\tfrac12\rho\,C_dA\,v^2}_{\text{aerodynamic drag}} + \underbrace{mg\sin\theta}_{\text{gradient}} + \underbrace{m_\text{eff}\,a}_{\text{acceleration}},

where CrrC_{rr} is the rolling-resistance coefficient (about 0.007–0.012 for car tyres on asphalt), ρ\rho the air density, CdC_d the drag coefficient and AA the frontal area (their product CdAC_dA is what matters), and meffm_\text{eff} the mass plus the equivalent mass of rotating parts. The first three terms are the road load. The power at the wheels is P=FtvP = F_tv.

Rolling resistance comes from the tyre flexing as it rolls, and losing some of the energy as heat in the rubber, not from friction with the road; Rigid-body dynamics showed that static friction at a rolling contact does no work. Aerodynamic drag grows with the square of speed, so the power to overcome it grows with the cube. Gradients are expressed as a percentage, 100tan⁡θ100\tan\theta: a 6 per cent motorway climb is θ=3.4°\theta = 3.4°.

ComputationC

A car and an e-bike, computed

The program computes the road load of a mid-sized car (1500 kg with its driver, Cd=0.28C_d = 0.28 on 2.3 m² of frontal area, Crr=0.010C_{rr} = 0.010) at several speeds, and of an electrically assisted bicycle with its rider (100 kg, an upright rider with CdA≈0.55C_dA \approx 0.55 m², Crr=0.006C_{rr} = 0.006). In the European Union and many other countries such pedelecs may assist only while the rider pedals, up to 25 km/h, with a continuous motor rating of 250 W.

python
import numpy as np

g, rho = 9.81, 1.20                          # m/s^2; air density near sea level, kg/m^3

def road_load(m, Crr, CdA, v, grade=0.0):
    """Tractive force (N) to hold speed v (m/s) up a gradient: rolling + aerodynamic + grade."""
    th = np.arctan(grade)
    return Crr * m * g * np.cos(th) + 0.5 * rho * CdA * v**2 + m * g * np.sin(th)

# A mid-sized car with its driver, and an electric-assist bicycle with its rider.
car = dict(m=1500.0, Crr=0.010, CdA=0.28 * 2.3)     # drag coefficient 0.28 on 2.3 m^2 of frontal area
bike = dict(m=100.0, Crr=0.006, CdA=0.55)           # rider sitting upright

print("car on a level road")
print(" speed    rolling   drag    force   power at   energy at the wheels")
print("(km/h)      (N)     (N)     (N)    wheels(kW)   (kWh per 100 km)")
for kmh in (30, 50, 90, 120, 130):
    v = kmh / 3.6
    roll = road_load(car["m"], car["Crr"], 0.0, v)
    F = road_load(**car, v=v)
    print(f"{kmh:5d}   {roll:7.0f} {F - roll:7.0f} {F:7.0f}   {F * v / 1e3:7.1f}       {F * 1e5 / 3.6e6:6.1f}")

v = 25 / 3.6
F = road_load(**bike, v=v)
print(f"\ne-bike at 25 km/h on the level: {F:.1f} N, {F * v:.0f} W, {F * 1e5 / 3.6e6:.2f} kWh per 100 km")
# Up a 6 % hill, rider 100 W plus the motor's 250 W: what speed can they hold?
lo, hi = 0.1, 15.0
for _ in range(60):
    mid = (lo + hi) / 2
    lo, hi = (mid, hi) if road_load(**bike, v=mid, grade=0.06) * mid < 350.0 else (lo, mid)
print(f"e-bike up 6 %, 350 W in total: {3.6 * lo:.1f} km/h")
battery = 500.0                              # Wh
for share in (1.0, 0.5):
    print(f"range on the level at 25 km/h with a 500 Wh battery, motor supplying {100 * share:.0f} % "
          f"(85 % efficient): {battery * 0.85 / (share * F * v) * 25:.0f} km")

# Rolling resistance per tonne-kilometre: rubber tyres on asphalt against steel wheels on rails.
for name, Crr in (("car tyres", 0.010), ("e-bike tyres", 0.006), ("rail wheels", 0.0015)):
    print(f"{name:13s}: rolling resistance costs {Crr * 1000 * g * 1000 / 3.6e6 * 1000:.1f} Wh per tonne-km")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

At town speeds the car's rolling resistance dominates; by 90 km/h drag has overtaken it, and at 130 km/h drag is more than three-quarters of the load. The power needed at the wheels rises from 3.1 kW at 50 km/h to 23.5 kW at 130 km/h, and the energy per kilometre nearly triples. The e-bike needs about 150 W at 25 km/h, which a fit rider can supply alone for a while; it uses 0.61 kWh per 100 km, about an eighteenth of what the car needs at 90 km/h, carrying the same person. On a 6 per cent climb, rider and motor together (350 W) hold 17 km/h, and a 500 Wh battery that supplies half the effort lasts about 140 km on the level.

ComputationC

Combustion and electric powertrains compared

A powertrain turns stored energy into tractive force. In a car with an internal-combustion engine, the engine follows (roughly) the Otto or Diesel cycle of Power and refrigeration cycles; the best car petrol engines reach a peak brake efficiency of about 40 per cent at their best operating point, and the largest marine diesels exceed 50 per cent. But a car engine rarely runs at its best point. At light load its friction and pumping losses are a large fraction of its output; it idles at traffic lights; and every stop throws the car's kinetic energy away as heat in the brakes. Gearboxes with many ratios, start–stop systems and hybrids exist to fight exactly these losses.

An electric powertrain has a battery, an inverter and a motor, usually with a single fixed gear, because an electric motor delivers high torque from rest and runs efficiently over a wide speed range. From battery to wheels it is typically 85 to 90 per cent efficient across most of its operating range; charging loses perhaps another 10 per cent. And it can run backwards: regenerative braking turns the motor into a generator and returns part of the braking energy to the battery.

The program compares the two over a home-made test cycle (not an official one): stop-start town driving, a country road at 80 km/h with one slowdown, and a motorway run at 120 km/h. The petrol car is modelled with a Willans line, a classic engine model in which fuel power is a fixed loss whenever the engine turns plus the output divided by a marginal efficiency. The electric car is 250 kg heavier, for its battery, with the same body.

python
import numpy as np

g, rho, dt = 9.81, 1.20, 0.1

def cycle(legs, accel=1.0, decel=1.0):
    """Speed trace (m/s) from (speed in km/h, seconds held) legs, ramping at the given rates."""
    v, out = 0.0, []
    for kmh, hold in legs:
        target = kmh / 3.6
        while abs(v - target) > 1e-9:
            v = min(target, v + accel * dt) if target > v else max(target, v - decel * dt)
            out.append(v)
        out += [v] * int(round(hold / dt))
    return np.array(out)

# A home-made test cycle (not an official one): stop-start town driving, a country road, a motorway.
parts = {
    "town":     cycle([(50, 30), (0, 30)] * 6 + [(30, 20), (0, 30)] * 4),
    "country":  cycle([(80, 300), (50, 40), (80, 300), (0, 10)], accel=0.8),
    "motorway": cycle([(120, 600), (0, 10)], accel=0.6),
}

def wheel_power(v, m, Crr, CdA):
    a = np.gradient(v, dt)
    F = Crr * m * g * (v > 0) + 0.5 * rho * CdA * v**2 + 1.05 * m * a   # 5 % extra for spinning parts
    return F * v                                                         # W; negative when braking

PETROL_KWH_PER_L = 8.9                       # lower heating value of petrol, about 32 MJ per litre

def petrol_car(v):
    """Fuel power (W): a Willans-line engine (6 kW of losses whenever it runs, 35 % marginal
    efficiency), 92 % efficient gearbox, 4 kW of fuel while idling, fuel cut off when slowing."""
    P = wheel_power(v, 1500.0, 0.010, 0.644)
    fuel = np.where(P > 0, 6e3 + P / (0.92 * 0.35), 0.0)
    fuel = np.where(v == 0, 4e3, fuel)
    return P, fuel

def electric_car(v):
    """Battery power (W): 88 % from battery to wheels and back, regeneration up to 50 kW,
    0.3 kW for electronics; a heavier car (battery) with the same body."""
    P = wheel_power(v, 1750.0, 0.010, 0.644)
    batt = np.where(P > 0, P / 0.88, -0.88 * np.minimum(-P, 50e3)) + 300.0
    return P, batt

print("           km   at the wheels    petrol car                       electric car")
print("                 used   braked    L/100km  kWh/100km  wheels/fuel    kWh/100km at plug")
for name, v in list(parts.items()) + [("all", np.concatenate(list(parts.values())))]:
    per100 = 100 / (v.sum() * dt / 1000)                 # converts kWh to kWh per 100 km
    kwh = lambda p: p.sum() * dt / 3.6e6 * per100
    P, fuel = petrol_car(v)
    _, batt = electric_car(v)
    print(f"{name:9s}{100 / per100:5.1f}  {kwh(P.clip(min=0)):5.1f}  {kwh(-P.clip(max=0)):5.1f}"
          f"      {kwh(fuel) / PETROL_KWH_PER_L:5.1f}    {kwh(fuel):5.1f}      {100 * kwh(P.clip(min=0)) / kwh(fuel):3.0f} %"
          f"          {kwh(batt) / 0.90:5.1f}")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

In town the petrol car turns only 20 per cent of its fuel into work at the wheels, and half of that work, 6.0 of 11.6 kWh per 100 km, is then thrown away in the brakes. On the country road it does best, 26 per cent, and 29 per cent on the motorway, where the engine works hardest and closest to its efficient region. Over the whole cycle it uses 5.6 L/100 km, 50 kWh of fuel energy. The electric car uses 17.2 kWh per 100 km from the plug, about a third as much, and its advantage is greatest in town (11.5 against 57 kWh), where regeneration recovers braking energy and there is no idling, and smallest on the motorway (21 against 56 kWh), where drag dominates and both powertrains are working near their best.

ComputationC

Aircraft: thrust and propulsive efficiency

An aircraft engine produces thrust by throwing air backwards. If air enters at the flight speed v0v_0 and leaves at the jet speed vjv_j, with mass flow m˙\dot m (and the exit pressure equal to the ambient), momentum gives the thrust

T=m˙ (vj−v0).T = \dot m\,(v_j - v_0).

The useful power is Tv0Tv_0, but the engine also leaves kinetic energy in the wake, 12m˙(vj−v0)2\tfrac12\dot m(v_j - v_0)^2 per second, so the propulsive efficiency is

ηp=Tv0Tv0+12m˙(vj−v0)2=21+vj/v0.\eta_p = \frac{Tv_0}{Tv_0 + \tfrac12\dot m(v_j - v_0)^2} = \frac{2}{1 + v_j/v_0} .

The same thrust can come from a little air accelerated a lot or a lot of air accelerated a little, and the second is far more efficient. That single fact shaped the modern airliner.

compressorburnerturbinefanair inbypass air: lots of it,a little fastercore jet,hot and faster
A high-bypass turbofan, cut in half along its axis (schematic). A large fan pushes most of the incoming air through the bypass duct, speeding it up only modestly; a small part enters the core, a gas turbine on the Brayton cycle (compressor, burner or combustor, turbine) whose turbine also drives the fan. Modern engines pass ten or more kilograms of bypass air for each kilogram through the core.

How efficiently an engine uses fuel is measured by its thrust-specific fuel consumption (TSFC), the fuel mass flow per unit thrust; a modern large turbofan in cruise burns roughly 15 g of kerosene per second for each kilonewton of thrust. An aircraft's range then follows from the Breguet range equation, which integrates the fuel burnt as the aircraft gets lighter:

R=v0g TSFC LD ln⁡mstartmend,R = \frac{v_0}{g\,\text{TSFC}}\,\frac{L}{D}\,\ln\frac{m_\text{start}}{m_\text{end}},

with L/DL/D the aircraft's lift-to-drag ratio in cruise. Three disciplines appear as three factors: propulsion (speed over TSFC), aerodynamics (L/DL/D) and structures (how much of the take-off mass can be fuel).

python
import numpy as np

g = 9.81
v0 = 240.0                                   # cruise speed, m/s (about 860 km/h, Mach 0.8 up high)
T = 60e3                                     # thrust needed from one engine at cruise, N

print("exhaust   air through   propulsive   kinetic energy left")
print("speed     the engine    efficiency   in the wake")
for vj in (600.0, 450.0, 330.0, 290.0):
    mdot = T / (vj - v0)                     # thrust = mass flow x speed increase (exit pressure = ambient)
    eta_p = 2 / (1 + vj / v0)                # useful power T v0, divided by T v0 + wake kinetic power
    wake = 0.5 * mdot * (vj - v0) ** 2
    print(f"{vj:5.0f} m/s  {mdot:6.0f} kg/s     {eta_p:5.2f}        {wake / 1e6:5.1f} MW")

# Range: the Breguet equation for a jet airliner, and the same aircraft carrying batteries instead.
TSFC = 15e-6                                 # thrust-specific fuel consumption: 15 g per kN per second
LD = 18.0                                    # lift-to-drag ratio in cruise
LHV = 43.2e6                                 # kerosene, J/kg
eta_0 = v0 / (TSFC * LHV)                    # overall efficiency: thrust power / fuel power
fuel_fraction = 0.30                         # share of take-off mass that is fuel burnt in cruise
R_jet = v0 / (g * TSFC) * LD * np.log(1 / (1 - fuel_fraction))
print(f"\noverall efficiency of the turbofan, fuel to thrust power: {eta_0:.2f}")
print(f"jet range burning {100 * fuel_fraction:.0f} % of take-off mass: {R_jet / 1e3:,.0f} km")
# Batteries do not get lighter as they discharge, so range = eta * e * (L/D) * (battery fraction) / g.
e_batt = 0.25 * 3.6e6                        # 0.25 kWh per kg of battery pack, in J/kg
eta_e = 0.75                                 # battery to thrust power: motor, inverter, propeller
R_batt = eta_e * e_batt * LD * fuel_fraction / g
print(f"battery-electric range with {100 * fuel_fraction:.0f} % of take-off mass as batteries: {R_batt / 1e3:,.0f} km")
print(f"energy per kg: kerosene {LHV / 3.6e6:.1f} kWh, battery pack {e_batt / 3.6e6:.2f} kWh "
      f"(ratio {LHV / e_batt:.0f})")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

A pure turbojet with a 600 m/s jet is 57 per cent propulsively efficient and leaves 10.8 MW of kinetic energy in its wake to make 60 kN; a turbofan whose mixed exhaust leaves at 290 m/s makes the same thrust at 91 per cent, by moving seven times as much air. That is why airliner engines have grown from slim turbojets in the 1950s to fans over three metres across, why bypass ratios have risen past ten, and why designers study open rotors and ever larger fans; the limits are the weight and drag of the nacelle and fitting it under the wing. The turbofan's overall efficiency, from fuel energy to thrust power, is 0.37, and the Breguet equation gives a range of about 10 500 km when 30 per cent of the take-off mass is fuel.

The same aircraft carrying batteries instead, of 0.25 kWh/kg (an optimistic pack-level figure as of 2026), would fly about 370 km, even though its electric propulsion is twice as efficient, because kerosene stores 48 times as much energy per kilogram and batteries do not get lighter as they discharge. Battery-electric flight therefore makes sense first for small aircraft on short routes, while long-haul aviation looks to sustainable aviation fuels, made from wastes, crops or captured carbon and hydrogen, and, further off, to hydrogen itself.

DefinitionsH

Rockets and the rocket equation

A rocket carries its propellant and throws it backwards; nothing outside is pushed against, which is why rockets work in a vacuum. As the propellant burns the rocket gets lighter, and that changes everything.

Theorem— The Tsiolkovsky rocket equation

A rocket ejecting propellant at a constant speed vev_e relative to itself, with no external forces, changes its speed by

Δv=veln⁡m0mf,\Delta v = v_e\ln\frac{m_0}{m_f},

where m0m_0 is its mass before the burn and mfm_f after. Engine makers quote vev_e as the specific impulse Isp=ve/g0I_{sp} = v_e/g_0 in seconds, with g0=9.806 65g_0 = 9.806\,65 m/s².

Proof

In a short time dt\mathrm dt the rocket, of mass mm and speed vv, ejects a mass dmp\mathrm dm_p at speed v−vev - v_e and keeps mass m−dmpm - \mathrm dm_p moving at v+dvv + \mathrm dv. With no external force, momentum is conserved: mv=(m−dmp)(v+dv)+dmp(v−ve)mv = (m - \mathrm dm_p)(v + \mathrm dv) + \mathrm dm_p(v - v_e). Expanding and dropping the product of two small quantities leaves m dv=ve dmpm\,\mathrm dv = v_e\,\mathrm dm_p. The rocket's mass falls by dmp\mathrm dm_p, so dm=−dmp\mathrm dm = -\mathrm dm_p and dv=−ve dm/m\mathrm dv = -v_e\,\mathrm dm/m. Integrating from m0m_0 to mfm_f gives Δv=veln⁡(m0/mf)\Delta v = v_e\ln(m_0/m_f).

∎

The logarithm is the whole story of rocketry. Every extra unit of Δv\Delta v multiplies the required mass ratio m0/mfm_0/m_f by the same factor, so a rocket that must reach low Earth orbit (about 7.8 km/s of orbital speed, plus roughly 1.5 to 2 km/s lost to gravity and drag on the way up, minus a little help from the Earth's rotation, about 9.4 km/s in all) must be mostly propellant. The escape is staging: drop the empty tanks and engines of the first stage once they have done their work, so that the rest of the flight does not have to accelerate them.

python
import numpy as np

g0 = 9.80665                                 # standard gravity, used to define specific impulse
dv_orbit = 9400.0                            # m/s to reach low Earth orbit, including gravity and drag losses

def payload_fraction(dvs, isps, eps):
    """Payload / lift-off mass for stages flown in order. dvs: delta-v of each stage (m/s);
    isps: specific impulses (s); eps: structural fraction of each stage, structure/(structure+propellant)."""
    frac = 1.0
    for dv, isp, e in zip(dvs, isps, eps):
        R = np.exp(dv / (isp * g0))          # Tsiolkovsky: mass before / mass after the burn
        f = (1 / R - e) / (1 - e)            # what this stage hands to the next, as a share of its own start
        if f <= 0:
            return 0.0
        frac *= f
    return frac

print(f"mass ratio needed for {dv_orbit / 1e3} km/s in one stage:")
for isp in (300, 350, 450):
    print(f"   Isp {isp} s (exhaust {isp * g0 / 1e3:.2f} km/s): {np.exp(dv_orbit / (isp * g0)):5.1f}, "
          f"so {100 * (1 - np.exp(-dv_orbit / (isp * g0))):.1f} % of lift-off mass must be propellant")

print("\npayload as a share of lift-off mass (stage structure 6 % of stage mass, upper 10 %)")
for isp in (300, 350):
    print(f"   one stage, Isp {isp} s: {100 * payload_fraction([dv_orbit], [isp], [0.06]):.2f} %")
for name, isp2 in (("kerosene upper stage, Isp 345 s", 345), ("hydrogen upper stage, Isp 450 s", 450)):
    splits = np.linspace(0.2, 0.8, 601)
    fr = [payload_fraction([s * dv_orbit, (1 - s) * dv_orbit], [300, isp2], [0.06, 0.10]) for s in splits]
    i = int(np.argmax(fr))
    print(f"   two stages, {name}: {100 * fr[i]:.2f} %, with the first stage giving {100 * splits[i]:.0f} % of the delta-v")

# Where the lift-off mass goes in the best hydrogen-upper-stage design (per cent).
s = splits[i]
R1, R2 = np.exp(s * dv_orbit / (300 * g0)), np.exp((1 - s) * dv_orbit / (450 * g0))
pi1 = (1 / R1 - 0.06) / 0.94                 # upper stage + payload, as a share of lift-off mass
pi2 = (1 / R2 - 0.10) / 0.90                 # payload, as a share of the upper stage's start mass
parts = {"stage 1 propellant": 0.94 * (1 - pi1), "stage 1 structure": 0.06 * (1 - pi1),
         "stage 2 propellant": pi1 * 0.90 * (1 - pi2), "stage 2 structure": pi1 * 0.10 * (1 - pi2),
         "payload": pi1 * pi2}
print("\n" + ", ".join(f"{k} {100 * v:.1f} %" for k, v in parts.items()))
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

With kerosene and liquid oxygen (Isp≈300I_{sp} \approx 300 s), a single stage would have to be 95.9 per cent propellant. If its tanks, engines and structure weigh 6 per cent of the stage, which is already excellent, nothing is left for payload. With a better engine (IspI_{sp} = 350 s) a single stage delivers 0.5 per cent, too thin a margin to survive the first kilogram of extra structure. Two stages deliver 2.5 per cent with a kerosene upper stage and 4.8 per cent with a hydrogen one, whose higher exhaust speed makes it worth giving the upper stage more of the work: the optimum first stage then supplies only 31 per cent of the Δv\Delta v.

one stageIsp 350 stwo stagesH₂ upperpropellant 93.5 %stage 1 propellant 62.3 %stage 2: 26.1 %structure 6.0 %structure 4.0 % and 2.9 %payload 0.5 %payload 4.8 %
Where the lift-off mass goes, computed above. Top: a single stage with Isp 350 s is 93.5 per cent propellant and 6.0 per cent structure, leaving 0.5 per cent for payload. Bottom: the best two-stage design with a hydrogen upper stage, 62.3 per cent first-stage propellant, 4.0 per cent first-stage structure, 26.1 per cent second-stage propellant, 2.9 per cent second-stage structure and 4.8 per cent payload.

Since the 2010s, reusable first stages that fly back and land, first flown in routine service by SpaceX's Falcon 9 and now pursued by companies and agencies in the USA, China, Europe and elsewhere, have changed the economics without changing the equation: the propellant needed to return costs payload, which the reuse repays in hardware.

ExplanationH

Ships and trains, in brief

Trains move freight and people with remarkably little energy, for two mechanical reasons. A steel wheel on a steel rail barely deforms, so its rolling-resistance coefficient is about 0.001–0.002, five to ten times less than a tyre's: the program above found 4 Wh per tonne-kilometre for rolling resistance on rails against 27 Wh for car tyres. And a long train hides most of its wagons in the wake of the ones in front, so its drag per tonne is small. Electric trains add regenerative braking, feeding energy back into the overhead line when they slow down, and high-speed lines, from Japan's Shinkansen (1964) to the networks of Europe and China, compete with flying on journeys of a few hundred kilometres.

Ships carry most of the world's traded goods. A hull pushing through water loses energy to friction on its wetted surface and, increasingly as speed rises, to the waves it makes; at typical speeds the power needed rises roughly with the cube of speed. So slow steaming pays: 10 per cent slower needs about 0.93=0.730.9^3 = 0.73 of the power, and since the voyage takes 1/0.9=1.111/0.9 = 1.11 times as long, about 0.92=0.810.9^2 = 0.81 of the fuel, a saving of nearly a fifth at the cost of time. The largest ships are driven by slow-speed two-stroke diesel engines turning the propeller directly, which are among the most efficient heat engines ever built. The International Maritime Organization's 2023 strategy aims at net-zero emissions from international shipping by or around 2050, and ships running on methanol, liquefied natural gas and, experimentally, ammonia, together with wind-assist sails and rotors, are being built and tested as of 2026.

ExplanationH

Vehicles and propulsion in professional practice

Testing to a cycle. Cars are certified by driving them on rollers through a standard drive cycle, such as the WLTP adopted at UNECE and used in the European Union, Japan and a number of other countries, or the cycles of the US Environmental Protection Agency; the lesson's home-made cycle imitates the idea, not any official cycle. Engineers then build longitudinal vehicle models like the one above, with measured engine maps, gearbox losses and battery models, to choose gear ratios, size batteries, tune hybrid strategies and predict real-world range, often in tools such as those described in Modelling dynamic systems.

Engines and aircraft. Aircraft engines are certified by national authorities against standards for safety, noise and emissions coordinated through ICAO, after years of rig, ground and flight testing. Engine makers in the USA, the UK, France, Germany, Canada, Japan and China and their partners compete on TSFC, weight, durability and noise; the airframe makers on L/DL/D and structural weight.

Space. National agencies, among them NASA, ESA, Roscosmos, ISRO, JAXA and CNSA, and a growing number of companies design launch vehicles with exactly the trade-offs computed here, stage by stage, with far more detailed models of gravity and drag losses and trajectories. Lumenary ME returns to them in Space and fusion engineering, and to the plants that generate the electricity for vehicles in Power plants and energy conversion.

History & global context

A short history

Richard Trevithick ran a steam locomotive on rails in Wales in 1804. Karl Benz patented his petrol-engined Motorwagen in 1886, and in 1888 Bertha Benz drove it over a hundred kilometres to prove it practical; Rudolf Diesel's engine followed in the 1890s, and electric cars, popular around 1900, lost out to petrol for most of the twentieth century before returning with lithium-ion batteries. Konstantin Tsiolkovsky published the rocket equation in 1903, and Robert Goddard flew the first liquid-fuelled rocket in 1926. Frank Whittle in Britain and Hans von Ohain in Germany independently invented the jet engine in the 1930s, and von Ohain's engine first flew in 1939. A Soviet R-7 rocket put Sputnik into orbit in 1957, and the first high-bypass turbofans entered airline service around 1970, with the wide-body jets they made possible.

Common misconceptionsH

Common misconceptions

Misconception

"Jet engines and rockets work by pushing against the air behind them." They push on the gas they throw backwards, and the gas pushes back on them (Newton's third law). Air resistance outside is a nuisance, not a help: a rocket works better in a vacuum, and a jet engine needs air only as the stuff it accelerates and burns its fuel in.

Misconception

"Driving a bit faster costs only a bit more fuel." Once drag dominates, energy per kilometre grows with the square of speed and power with the cube. From 120 to 130 km/h, 8 per cent faster, the car above needed 13 per cent more energy per kilometre and 22 per cent more power.

Misconception

"Twice the propellant gives a rocket twice the speed." Because of the logarithm in the rocket equation, each extra kilogram of propellant must also accelerate all the propellant still unburnt. Doubling the mass ratio adds only veln⁡2v_e\ln 2, about 0.7 exhaust speeds, to the Δv\Delta v, however large the rocket.

Misconception

"Electric vehicles use no energy and cause no emissions." They use about a third of the energy of a comparable petrol car at the plug, and none of their emissions come from the exhaust pipe, but the electricity must be generated and the battery made. How much cleaner they are depends on the electricity mix and the vehicle's life, which is why careful studies state both.

Exercises

Exercises

Use g=9.81g = 9.81 m/s² and ρ=1.20\rho = 1.20 kg/m³ for air unless an exercise says otherwise. Two exercises ask for Python functions; run them in the editor before submitting.

1. Vehicles and propulsion

Warm-upHuman

A car has a drag area CdA=0.60C_dA = 0.60 m². What aerodynamic drag force does it meet at 100 km/h in still air (ρ=1.20\rho = 1.20 kg/m³)? Give the answer in newtons, to the nearest whole number.

2. Vehicles and propulsion

Warm-upHuman

At high speed, almost all of a lorry's road load is aerodynamic drag. If it speeds up from 60 km/h to 120 km/h, by roughly what factor does the power needed to overcome drag increase?

3. Vehicles and propulsion

CoreHuman

An e-bike and rider (100 kg in all, Crr=0.006C_{rr} = 0.006, CdA=0.55C_dA = 0.55 m²) climb a steady 8 per cent gradient at 15 km/h. How much power, in watts, must rider and motor deliver together at the wheel? Round to the nearest watt.

4. Vehicles and propulsion

CoreHuman

Why do modern airliners use high-bypass turbofans rather than the slim turbojets of the 1950s?

Choose one

5. Vehicles and propulsion

CoreHuman

A rocket stage has a specific impulse of 311 s and a mass ratio m0/mf=4.0m_0/m_f = 4.0. What Δv\Delta v does it give in free space, in m/s, to the nearest whole number? (Use g0=9.806 65g_0 = 9.806\,65 m/s².)

6. Vehicles and propulsion

CoreHuman

A container ship normally crosses an ocean at 20 knots. Its owners slow it to 16 knots. If the propulsion power varies as the cube of speed, by what percentage does the fuel burnt on the crossing fall? Round to the nearest whole number.

7. Vehicles and propulsion

CoreComputational

Write breguet_range(v0, tsfc, LD, m_start, m_end, g=9.81) returning a jet aircraft's cruise range in metres from its speed, thrust-specific fuel consumption (kg/(N·s)), lift-to-drag ratio and its masses at the start and end of cruise.

The starter uses the fuel fraction in place of the logarithm. Run the tests to see why that underestimates the range, then fix it.

Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

8. Vehicles and propulsion

StretchComputational

Write stack_delta_v(stages, payload) returning the total Δv\Delta v (m/s) of a multistage rocket. stages is a list of (propellant mass, structure mass, specific impulse in s), bottom stage first; each stage burns all its propellant and is then dropped; payload is carried by the top stage. Use g0=9.806 65g_0 = 9.806\,65 m/s².

The starter gives far too optimistic an answer for two or more stages. Run the tests, then fix it.

Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

Further reading and resources