An aluminum producer asked us to evaluate energy-generation choices for a smelting operation. The assignment was not simply to find the cheapest electricity. A smelter is a continuous electrochemical process, so the value of a power source depends on its reliability, operating profile, capital cost, and interaction with the plant.
Why Aluminum Is an Electricity Business
Primary aluminum is made through the Hall-Héroult process. Refined alumina is dissolved in molten cryolite inside a reduction cell. A powerful direct current passes from carbon anodes through the molten bath to a carbon-lined cathode. Aluminum ions gain electrons and collect as molten metal; oxygen from the alumina reacts with the carbon anodes, consuming them and producing carbon dioxide. Keeping hundreds of cells hot and energized requires an enormous, steady supply of electricity.
An interruption is more serious than lost production for an hour. If the molten bath cools and freezes, restarting can be slow, expensive, and damaging. The plant therefore valued dependable baseload power differently from intermittent or interruptible power, even when the latter appeared cheaper per megawatt-hour.
Electricity Can Also Make Oxygen
Electricity can be converted into chemical products as well as aluminum. In water electrolysis, an electrolyzer uses electrical energy to split water into hydrogen and oxygen: hydrogen forms at the cathode and oxygen at the anode. If a site has a use for both gases, the oxygen stream can become a useful coproduct of hydrogen production.
Nor does electrolysis automatically make oxygen economical. The equipment, electricity, water treatment, compression, storage, and value of the hydrogen all matter. The useful planning question is whether an integrated site can obtain more value from its energy system than it would by evaluating electricity, hydrogen, and oxygen separately.
Modeling Power as Part of Production
We compared generation and supply portfolios across fuel prices, capital costs, availability, maintenance outages, emissions, ramping limits, and the plant's continuous load. Candidate configurations could combine firm grid supply, dedicated generation, backup capacity, and interruptible sources, with operating rules for normal and adverse conditions.
The model also preserved the operational asymmetry: a surplus megawatt-hour might have modest value, while an unexpected shortage could threaten an entire potline. That meant the least expensive nominal portfolio was not necessarily the least expensive risk-adjusted portfolio.
The Result
The engagement reframed energy as part of the production system rather than a commodity purchase made beside it. The prototype let management compare configurations on total economics, resilience, and the value of useful coproducts. For an aluminum smelter, power strategy and manufacturing strategy are inseparable.
