Skip to content
LumenaryAcademy
Frontier 80 min

Sustainable engineering

Counting a product's impacts over its whole life — a worked life-cycle comparison, design for disassembly, repair and recycling, decarbonising steel, cement and other heavy industry, efficiency standards, and the engineer's responsibility.

Code tested · 5AI-draftedAwaiting review by a mechanical engineer

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

MotivationH

The impact is decided at the drawing board

A pump's purchase price is a few per cent of what it costs over its life; most of the rest is electricity. A car body's material choice fixes emissions at a smelter or blast furnace years before the car is driven, and fixes what can be recovered from it fifteen years later. A concrete floor slab poured a little thicker than needed carries that extra cement's CO₂ for the life of the building. The choices that decide a product's environmental impact are, overwhelmingly, the choices engineers make on paper.

They add up. Producing materials, above all steel, cement, aluminium, chemicals and plastics, accounts for roughly a quarter of the world's CO₂ emissions; steel and cement are each responsible for something like 7–8% on International Energy Agency estimates. Mechanical systems (motors, pumps, fans, compressors, engines, heating and cooling) use most of the world's energy. So mechanical engineers sit on both sides of the problem: what things are made of, and how much energy they use.

This lesson is about doing that accounting honestly and acting on it: a worked life-cycle comparison, designing for the loops of repair and recycling, the chemistry and engineering of decarbonising heavy industry, efficiency standards, and what it means to be responsible for it all. It builds on the life-cycle assessment introduced in Lumenary Chemistry's sustainable chemistry, on vehicles and propulsion and on power plants and energy conversion.

DefinitionsH

Life-cycle thinking for engineers

Life-cycle assessment (LCA), standardised in ISO 14040 and 14044, has four phases: goal and scope, inventory, impact assessment and interpretation. For engineered products a few decisions dominate the outcome:

Definition— Functional unit and system boundary

The functional unit is the service being compared, stated so that alternatives deliver the same thing: not "a door panel" but "a door panel meeting the same crash and stiffness requirements, carried for 200 000 km in a given car". The system boundary says which stages are counted: cradle to gate (materials and manufacture), cradle to grave (adding use and disposal), or cradle to cradle (adding recycling into new products).

Definition— Recycling allocation

When material is recycled, its burdens and benefits must be split between the first life and the next. The cut-off approach gives recycled input free of its earlier production burden and gives no credit for recycling at the end of life; the avoided-burden (substitution) approach credits the product at end of life for the virgin material its scrap will replace. Each is defensible; mixing them double-counts the benefit. Construction products' environmental product declarations under EN 15804 report the end-of-life credit separately (as "module D") for exactly this reason.

The data come from databases such as ecoinvent, from industry associations, and increasingly from suppliers' own environmental product declarations (EPDs), verified documents that follow product-category rules. Early in design, a simpler eco-audit, popularised by Mike Ashby at Cambridge, estimates energy and CO₂ phase by phase from a bill of materials and a usage profile, enough to see which phase dominates and where design effort should go.

ComputationC

A worked comparison: is lighter greener?

Car makers save mass by replacing steel with aluminium or carbon-fibre composite, because a lighter car needs less energy to accelerate and roll. But aluminium and carbon fibre take far more energy to make than steel. Which wins? The comparison below is for one body part with the same function made three ways (a 30 kg steel part, an 18 kg aluminium part, a 12 kg composite part), over 200 000 km, with recycling by the cut-off approach. The emission factors are rounded values within published ranges; a real study would take them from databases and suppliers' declarations and would test their uncertainty.

python
# Life-cycle greenhouse-gas comparison of one car body part made three ways (cradle to grave, cut-off
# approach for recycling). Every number is a stated, rounded assumption within published ranges.
LIFE_KM = 200_000
options = {  # name: (part mass in kg, cradle-to-gate kg CO2e per kg of finished part)
    "steel (blast furnace route)":        (30.0, 2.5),
    "aluminium, primary, average grid":   (18.0, 16.0),
    "aluminium, 70 % recycled content":   (18.0, 0.3 * 16.0 + 0.7 * 0.8),
    "carbon-fibre composite":             (12.0, 27.0),
}
# Use phase: extra energy needed to carry each kilogram for one kilometre
PETROL_L = 0.3 / 100 / 100            # litres per km per kg (0.3 L/100 km per 100 kg, engine resized)
PETROL_CO2 = 2.8                      # kg CO2e per litre, well to wheel
EV_KWH = 0.5 / 100 / 100              # kWh per km per kg (0.5 kWh/100 km per 100 kg)
cars = {"petrol car": PETROL_L * PETROL_CO2,
        "EV, grid 0.40 kg/kWh": EV_KWH * 0.40,
        "EV, grid 0.05 kg/kWh": EV_KWH * 0.05}

base_mass, base_ef = options["steel (blast furnace route)"]
for car, per_kg_km in cars.items():
    print(f"\n{car}: mass-induced use emissions {per_kg_km * 1e6:.0f} g CO2e per tonne-km")
    print(f"  {'option':34s} {'production':>10s} {'use':>7s} {'total':>7s}   break-even vs steel")
    for name, (mass, ef) in options.items():
        prod, use = mass * ef, mass * per_kg_km * LIFE_KM
        if name.startswith("steel"):
            note = "-"
        else:
            extra = prod - base_mass * base_ef                     # extra production emissions
            saving = (base_mass - mass) * per_kg_km                 # saved per km
            km = extra / saving
            note = f"{round(km, -3):,.0f} km" if km <= 1e6 else "never (over 1,000,000 km)"
        print(f"  {name:34s} {prod:8.0f} kg {use:5.0f} kg {prod + use:5.0f} kg   {note}")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.
steelprimary Alrecycled Alcompositebreak-even 21 000 km211 000 km165 000 kmvehicle life020040060080050100150200250distance driven (thousand km)cumulative kg CO2e
Cumulative greenhouse-gas emissions attributable to the part in a petrol car, from the program's assumptions. Each line starts at the part's production emissions and rises with the fuel needed to carry its mass. Recycled aluminium overtakes steel after about 21 000 km; primary aluminium only after about 211 000 km, beyond this car's assumed 200 000 km life.

Three lessons come out of the table. First, in a petrol car, lightweighting with primary aluminium breaks even only after 211 000 km, just beyond the car's assumed life; carbon fibre does somewhat better (165 000 km), and aluminium with high recycled content pays back within two years of driving (21 000 km). The production route matters as much as the material. Second, in an electric car on an average grid, carrying mass costs less, so the break-even distances stretch to 89 000 km for recycled aluminium and beyond any car's life for the others. Third, on a clean grid, mass hardly matters at all, and the lowest-carbon part is simply the one with the lowest production emissions: here, steel. As grids decarbonise, the centre of gravity of vehicle LCAs moves from the exhaust pipe to the factory, and the same shift is happening for buildings, appliances and machines.

Worked examplesH

Design for disassembly, repair and recycling

The circular economy idea is to keep products and materials in use at their highest value for as long as possible. Its loops are ordered: the inner ones keep more of the value and the energy that went into the product.

materialspartsassemblyuselandfill or burningrecycle: recovers the material onlyremanufacture, refurbishrepairreuseinner loops keep more value and embodied energy
The loops of a circular economy for an engineered product. Inner loops (maintenance and repair, reuse) keep the most value; remanufacturing restores products to as-new condition; recycling recovers only the material. Landfill and incineration are losses from the system.

Each loop depends on design decisions made long before:

  • Repair needs access: fasteners rather than adhesives, standard screw heads, modules that can be swapped, spare parts and documentation. Fairphone in the Netherlands and Framework in the United States sell phones and laptops designed around user repair, and the EU's 2024 right-to-repair directive and its new ecodesign regulation are pushing the same principles into law, alongside right-to-repair laws in several US states and France's repairability index.
  • Remanufacturing restores used products (engines, gearboxes, starter motors, construction-equipment components, office printers) to as-new specification with a warranty, typically for a fraction of the energy and material of new manufacture. It needs cores that come back (often through a deposit), parts that survive disassembly, and wear surfaces that can be rebuilt, for instance by the directed energy deposition of advanced manufacturing.
  • Recycling works best with few materials that separate cleanly. Mark plastic parts with their polymer (ISO 11469 sets the marking), avoid bonding incompatible materials together, avoid coatings that contaminate the melt, and keep alloys that recycle together in the same product. A subtle problem is tramp elements: copper, from wiring and motors that end up in shredded scrap, cannot be removed from molten steel, and above a few tenths of a per cent it makes steel crack during hot rolling.
python
# Copper in recycled steel: copper cannot be removed in the furnace, so it accumulates.
# Mass balance: product Cu = share of scrap * Cu in scrap + (1 - share) * Cu in virgin iron.
CU_VIRGIN = 0.01                       # wt % copper in iron made from ore
limits = {"deep-drawn sheet": 0.10, "general structural": 0.25, "reinforcing bar": 0.40}  # wt %, illustrative
CU_SHREDDED = 0.35                     # wt % copper in shredded end-of-life car scrap, illustrative

def max_scrap_share(cu_scrap, limit):
    if cu_scrap <= limit:
        return 1.0
    return (limit - CU_VIRGIN) / (cu_scrap - CU_VIRGIN)

print(" copper removed     Cu in scrap     largest scrap share in the product")
print(" before shredding     (wt %)     " + "".join(f"{name:>20s}" for name in limits))
for removed in (0.0, 0.5, 0.8, 0.95):
    cu = CU_VIRGIN + (CU_SHREDDED - CU_VIRGIN) * (1 - removed)
    shares = "".join(f"{max_scrap_share(cu, lim):19.0%} " for lim in limits.values())
    print(f"      {removed:4.0%}          {cu:5.3f}      {shares}")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

With the illustrative numbers, shredded car scrap can make up only about a quarter of the charge for high-quality sheet, though all of it can go into reinforcing bar. Remove half the copper before shredding (pulling the wiring harnesses and motors) and the share for sheet doubles; remove 80% and the scrap can make sheet on its own. Studies from Cambridge have argued that unless copper is separated better, the growing supply of scrap steel will increasingly be confined to low-grade uses. Design that lets the copper come out (harnesses that are easy to pull, motors that unbolt) is design for recycling.

ComputationC

Decarbonising steel, cement and heavy industry

Steel and cement are hard to decarbonise because much of their CO₂ comes from chemistry, not from burning fuel. Making iron means removing oxygen from iron ore, and the blast furnace does it with carbon from coke. Making cement clinker means heating limestone, CaCO3→CaO+CO2\mathrm{CaCO_3 \to CaO + CO_2}, so the CO₂ comes out of the rock itself. The program computes both from the chemistry, with rounded energy figures.

python
# Heavy industry from first principles: where the CO2 of cement and steel comes from (rounded assumptions)
M = {"Ca": 40.078, "O": 15.999, "C": 12.011, "Fe": 55.845, "H": 1.008}
M_CaO, M_CO2 = M["Ca"] + M["O"], M["C"] + 2 * M["O"]

# --- Cement: CaCO3 -> CaO + CO2 in the kiln, plus fuel to reach about 1450 C, plus electricity
cao_in_clinker = 0.65                               # mass fraction of CaO in clinker
process = cao_in_clinker * M_CO2 / M_CaO            # t CO2 per t clinker, from the limestone itself
fuel = 3.4 * 0.095                                  # 3.4 GJ/t clinker at 0.095 t CO2/GJ (coal and petcoke)
clay = 2.5 * 0.095                                  # calcined clay: ~2.5 GJ/t at 750-850 C, no carbonate
elec = 0.10 * 0.45                                  # 100 kWh per t cement at 0.45 t CO2/MWh
print(f"clinker: {process:.2f} t CO2/t from calcination + {fuel:.2f} from fuel = {process + fuel:.2f}")
cements = {  # (clinker share, calcined clay share)
    "Portland cement (95 % clinker)": (0.95, 0.0),
    "blended, 75 % clinker + limestone/slag": (0.75, 0.0),
    "LC3: 50 % clinker, 30 % calcined clay": (0.50, 0.30),
}
for name, (k, c) in cements.items():
    t = k * (process + fuel) + c * clay + elec
    print(f"  {name:42s} {t:.2f} t CO2/t cement; in concrete at 320 kg/m^3: {320 * t:3.0f} kg CO2/m^3")
ccs = 0.95 * (process + fuel) * (1 - 0.90) + elec  # capture 90 % of the kiln's CO2, store it underground
print(f"  {'Portland cement, 90 % captured at the kiln':42s} {ccs:.2f} t CO2/t cement (plus the capture energy)")

# --- Steel: reducing iron ore. Hydrogen route Fe2O3 + 3 H2 -> 2 Fe + 3 H2O
h2 = 1000 / M["Fe"] * 1.5 * 2 * M["H"]              # kg H2 per tonne of iron, stoichiometric
kwh_h2 = h2 * 52                                    # electrolysis at about 52 kWh/kg H2
print(f"\nhydrogen to reduce a tonne of iron: {h2:.1f} kg, about {kwh_h2:,.0f} kWh of electrolysis")
routes = {  # (direct t CO2/t steel, electricity MWh/t steel)
    "blast furnace + basic oxygen furnace": (2.1, 0.1),
    "scrap in an electric arc furnace": (0.10, 0.5),
    "hydrogen DRI + electric arc furnace": (0.10, kwh_h2 / 1000 + 0.8),
}
for grid in (0.45, 0.05):
    print(f"grid at {grid:.2f} t CO2/MWh:")
    for name, (direct, mwh) in routes.items():
        print(f"  {name:38s} {direct + mwh * grid:4.2f} t CO2/t steel  ({mwh:.1f} MWh/t)")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

Cement. Calcination alone releases about 0.51 t of CO₂ per tonne of clinker, more than the fuel (0.32), so even a kiln run entirely on clean energy would emit most of its CO₂. The levers are: less clinker per tonne of cement, by blending in limestone, slag, fly ash or calcined clay (LC3, developed at EPFL in Switzerland with partners in India and Cuba, cuts the cement's CO₂ by about a third in this model); less cement per cubic metre of concrete, through better mix design; less concrete per building, through efficient structural design; and carbon capture and storage at the kiln. In 2025 Heidelberg Materials opened the first industrial-scale capture plant at a cement works, at Brevik in Norway, designed to capture about 400 000 t of CO₂ a year for storage under the North Sea.

Steel. The blast-furnace route emits about 2 t of CO₂ per tonne of steel. Melting scrap in an electric arc furnace emits a fraction of that, but scrap is limited by how much steel reaches the end of its life, and by its quality, as the copper calculation showed. For new iron, the leading route is hydrogen direct reduction: hydrogen strips the oxygen from iron ore pellets, leaving water instead of CO₂, and the resulting sponge iron is melted in an electric arc furnace. The chemistry needs 54 kg of hydrogen per tonne of iron, about 2.8 MWh of electrolysis, so the route uses roughly 3.6 MWh per tonne in all, and it is only low-carbon on a clean grid: 0.28 t CO₂/t at 0.05 t/MWh, but 1.73 at 0.45, hardly better than the blast furnace.

Status as of 2026, factually: Sweden's HYBRIT project (SSAB, LKAB and Vattenfall) made fossil-free sponge iron in a pilot plant from 2021 and delivered the first steel from it that year; Stegra (formerly H2 Green Steel) has been building a large hydrogen-based plant at Boden in northern Sweden; China's Baowu and others operate hydrogen-rich shaft furnaces; and several European projects have been delayed or shelved on cost, including ArcelorMittal's plans for two German plants in 2025. Molten oxide electrolysis, which reduces iron ore directly with electricity, is at pilot scale. In aluminium, ELYSIS, a joint venture of Alcoa and Rio Tinto in Canada, is developing inert anodes that release oxygen instead of CO₂ during smelting. Industrial heat below about 200 °C (for food, paper, chemicals) can increasingly come from industrial heat pumps, and higher temperatures from electric heating and heat storage, which the cost of electricity relative to gas largely decides.

ExplanationH

Efficiency standards and the rebound effect

For products that use energy, efficiency regulation has been one of the most effective climate policies of the last thirty years, because it changes millions of purchases at once. Electric motors are the clearest case: they use a large share of the world's electricity, and a motor running most of the year can use more than its purchase price in electricity within weeks. The International Electrotechnical Commission's standard IEC 60034-30-1 defines efficiency classes for motors, from IE1 (standard) up to IE4 (super premium), with IE5 defined for some types. Many economies set minimum energy performance standards on them. The EU's ecodesign rules, for example, have required IE3 for most motors from 0.75 to 1000 kW since July 2021 and IE4 for those from 75 to 200 kW since July 2023, and the United States, China, India, Japan and many other countries have their own minimums and labels. Pumps, fans, compressors, transformers, boilers, refrigerators and air conditioners have similar rules, and the EU's 2024 ecodesign regulation extends the approach from energy to durability, repairability, recycled content and digital product passports.

Efficiency gains are not always fully realised. The rebound effect: when using something becomes cheaper, people use more of it. More efficient cars are driven somewhat further; cheaper lighting is used more. Studies generally find direct rebound effects that take back some of the saving but not all of it, with larger effects in poorer households and economies where demand is unmet. William Stanley Jevons observed in 1865 that more efficient steam engines had increased Britain's coal consumption, by making coal-powered industry more profitable. Rebound is a reason to pair efficiency with clean supply, not a reason to waste energy.

ExplanationH

The engineer's responsibility

Professional codes of ethics now say plainly that engineers answer for environmental consequences. The UK Engineering Council and Royal Academy of Engineering's Statement of Ethical Principles lists respect for life, law, the environment and public good among its four principles; the American Society of Civil Engineers' code asks engineers to protect the environment and to practise sustainably; and engineering bodies from Canada to Australia, and the World Federation of Engineering Organizations, say similar things. In practice that means:

  • Count before claiming. An environmental claim about a product ("carbon-neutral", "green", "eco") needs a stated functional unit, boundary, method and data, and is tested by the same scrutiny as a strength calculation. Regulators are tightening this: an EU directive applying from September 2026 bans generic environmental claims that cannot be substantiated.
  • Don't shift the burden. A design that cuts CO₂ by moving pollution to a mine, a smelter or a disposal site in a poorer region has not solved the problem; LCA's multiple impact categories exist to reveal this.
  • Say what you don't know. Uncertain inputs, contested methods and missing data belong in the report, not in a footnote.
  • Design for the whole life. Specify repairable, durable, recyclable designs when the brief allows, and raise the question when it does not.
  • Respect the people affected. The communities near a plant, a mine or a dam bear its local impacts, and their knowledge and consent matter.
Optional Christian reflection · Counting the cost

A reflection from the Christian tradition that inspires Lumenary Academy. It is not part of the mathematics and you are welcome to skip it.

"For which of you, intending to build a tower, sitteth not down first, and counteth the cost, whether he have sufficient to finish it?" (Luke 14:28, King James Version). Jesus used an engineer's habit as a picture of honest commitment: before beginning, reckon what the whole undertaking will take. Life-cycle thinking is that habit extended to costs that fall on others, and on the future. This reflection is optional, and nothing in the lesson depends on it.

Common misconceptionsH

Common misconceptions

Misconception

"Lighter is always greener." Only when the use-phase saving outweighs any extra production burden within the product's life. With carbon-intensive light materials, or on clean electricity, the heavier part can have the lower life-cycle footprint.

Misconception

"Recyclable means recycled." A material is recycled only if the product is collected, the material separated cleanly and the recyclate used in place of virgin material. Mixed materials, adhesives, coatings and contaminants like copper in steel decide whether "recyclable" happens.

Misconception

"Hydrogen steel is zero-carbon." Only if the hydrogen comes from low-carbon electricity, and the route uses about 3.6 MWh per tonne. On an average grid its emissions are close to the blast furnace's.

Misconception

"An LCA gives one true number." It gives an answer conditional on its goal, boundary, allocation method and data. Two honest studies can disagree because they answer different questions; reading the scope is part of reading the result.

Exercises

Exercises

1. Sustainable engineering

Warm-upHuman

In a cement kiln, limestone decomposes: CaCO3→CaO+CO2\mathrm{CaCO_3 \to CaO + CO_2}. How many kilograms of CO₂ are released by calcining 1000 kg of pure calcium carbonate? (Molar masses: CaCO₃ 100.09 g/mol, CO₂ 44.01 g/mol.) Give the nearest whole number.

2. Sustainable engineering

CoreHuman

In hydrogen direct reduction, Fe2O3+3 H2→2 Fe+3 H2O\mathrm{Fe_2O_3 + 3\,H_2 \to 2\,Fe + 3\,H_2O}. How many kilograms of hydrogen does the reaction need, at the very least, to produce 1000 kg of iron? (Fe 55.845 g/mol, H₂ 2.016 g/mol.) Give the nearest whole number.

3. Sustainable engineering

CoreHuman

A steel product may contain at most 0.10 wt% copper. Iron made from ore contains 0.01% copper; the available scrap contains 0.19%. What is the largest percentage of scrap in the furnace charge that keeps the product within its limit?

4. Sustainable engineering

CoreComputational

Write break_even_km(m_base, ef_base, m_new, ef_new, per_kg_km). The baseline part has mass m_base (kg) and production emissions ef_base (kg CO₂e per kg); the alternative has m_new and ef_new; carrying 1 kg for 1 km causes per_kg_km kg CO₂e. Return the distance in km after which the alternative's cumulative emissions are lower. Return 0.0 if the alternative is already better at production, and None if it never catches up.

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

5. Sustainable engineering

CoreComputational

A motor delivers p_out_kw kW of shaft power for hours hours a year. Write motor_savings(p_out_kw, hours, eff_old, eff_new, price, extra_cost) returning a tuple (kwh_saved_per_year, payback_years) for replacing a motor of efficiency eff_old with one of efficiency eff_new, where electricity costs price per kWh and the better motor costs extra_cost more.

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

6. Sustainable engineering

StretchHuman

A supplier's declaration for an aluminium extrusion uses 80% recycled content, with the recycled input counted as nearly burden-free (cut-off approach). The product designer then also claims a credit for the aluminium being recycled at the end of the product's life (avoided burden). What is wrong?

Choose one

7. Sustainable engineering

CoreHuman

You are redesigning a domestic appliance for repair and recycling. Which changes help? Select all that apply.

Choose all that apply

8. Sustainable engineering

Warm-upHuman

A city replaces its street lights with LEDs that use 60% less energy per lumen. Two years later its street-lighting electricity has fallen by only 40%, because it has lit more streets and kept lights brighter for longer. What is this called?

Choose one

Further reading and resources

  • ISO 14040 and ISO 14044, Life cycle assessment · International Organization for StandardizationThe principles, framework, requirements and guidelines of LCA.
  • Materials and the Environment: Eco-informed Material Choice · Michael AshbyButterworth-Heinemann (3rd edition, 2021). The eco-audit method for engineers, with material data.
  • Sustainable Materials: With Both Eyes Open · Julian Allwood and Jonathan Cullen · 2012UIT Cambridge. How much we could cut emissions from steel and aluminium by using less and keeping it longer.
  • Utilization of structural steel in buildings · Muiris Moynihan and Julian Allwood · 2014Proceedings of the Royal Society A 470, 20140170. Finds most beams in real buildings carry well under half the load they could.
  • How will copper contamination constrain future global steel recycling? · Katrin Daehn, André Cabrera Serrenho and Julian Allwood · 2017Environmental Science & Technology 51, 6599–6606.
  • IEA, Iron and Steel and Cement tracking · International Energy AgencyEmissions, technologies and progress in heavy industry.
  • HYBRITSSAB, LKAB and Vattenfall's hydrogen-based ironmaking development in Sweden.
  • LC3 ProjectLimestone calcined clay cement, developed by EPFL with partners in India, Cuba and elsewhere.
  • IEC 60034-30-1, efficiency classes of line-operated AC motors · International Electrotechnical Commission
  • Ecodesign for Sustainable Products Regulation, Regulation (EU) 2024/1781 · European Union
  • Ellen MacArthur FoundationExplanations and case studies of circular-economy design.
  • Statement of Ethical Principles · Engineering Council and Royal Academy of EngineeringThe UK engineering profession's ethical principles, including respect for life, law, the environment and public good.
  • openLCAFree, open-source LCA software.