Most of a nickel refinery’s energy bill vanishes into its rotary kilns and electric arc furnaces. Studies show heat recovery, smart slag chemistry, and cleaner power can slash that demand — with Indonesia now nudging smelters toward renewables.
Industry: Nickel_Mining | Process: Refining
Nickel refining is a game of heat — and lots of it. For laterite ores processed in Indonesia’s rotary kiln–electric furnace flowsheets (RKEF; a rotary kiln for calcining/prereduction and an electric arc furnace for smelting), thermal steps dominate energy use, while crushing and grinding sip only a few kilowatt‑hours per tonne (mdpi.com) (researchgate.net).
A typical laterite feed (1.8–3.0% Ni, 17–34% moisture) is dried to roughly 10% before calcining and smelting (researchgate.net). Drying 1 tonne of ore from ~30% to 10% moisture removes ~200 kg of water and takes ~0.5–1 gigajoule (GJ; one billion joules; ≈150–300 kWh) per tonne, typically via coal or gas combustion (researchgate.net).
By contrast, high‑pressure acid leaching (HPAL; dissolving limonitic ore in hot sulfuric acid at ~250 °C) spends more than 90% of its energy heating slurry water (nickelinstitute.org). In such acid circuits, precise reagent addition hardware such as dosing pumps is standard industrial practice.
Process routes and unit operations
In RKEF (rotary kiln–electric furnace), ore moves from crushing/grinding to a rotary dryer (or the kiln itself) for moisture removal, then through kiln calcination/prereduction and electric arc furnace (EAF) smelting, followed by ladle furnace refining (mdpi.com) (researchgate.net). HPAL (high‑pressure acid leaching) instead dissolves the nickel at pressure/temperature in sulfuric acid; its energy mostly heats water in the slurry (nickelinstitute.org).
Where the kilowatt‑hours go
Thermal stages dominate. Wei et al. (2020) modeled ferronickel (FeNi, 35% Ni) at about 110 GJ of primary energy per tonne of alloy (≈30 GJ per tonne‑Ni) (mdpi.com). In per‑alloy terms, the same work reports roughly 174 GJ/t for pure nickel, 369 GJ/t for nickel oxide, 110 GJ/t for Fe–Ni, and 60 GJ/t for nickel pig iron (NPI) (mdpi.com). Per unit Ni, FeNi is ~309 GJ/t‑Ni vs 174 GJ/t for Ni metal (mdpi.com).
Wei et al. also attribute, by case: Ni metal (refining used ~44% of 174 GJ), NiO (smelting ~45% of 369 GJ), FeNi (calcining ~44% of 110 GJ), and NPI (smelting ~67% of 60 GJ) (mdpi.com). Put plainly, the hot‑stage furnaces — the kiln and EAF — are the largest energy consumers; auxiliary loads like hoisting, conveyors, pumps, and grinding/crushing remain a few kWh per tonne and often below 5% of total (researchgate.net) (mdpi.com).
State‑of‑the‑art RKEF units run around 75–100 MW and still consume on the order of 300–400 kWh per tonne of dry ore; even advanced furnaces (>75 MW, water‑cooled walls) have only pushed specific energy “below ~400 kWh/t,” according to a 2022 review (mdpi.com). A large RKEF complex feeding ~20 kt ore/day (~5 kt FeNi/day) can draw on the order of 10–20 GWh per day (20 MW) per furnace — mostly electricity for the EAF and fuel (coal/natural gas) for the kiln (researchgate.net).
Advanced furnaces often run water‑cooled walls, a reminder that cooling loops are critical infrastructure; operators typically plan for chemistry control programs, including cooling‑tower chemicals, in parallel with energy upgrades.
Major energy consumers
Academic and industry analyses concur that “the rotary kiln and electric arc furnace are the primary energy consumers” in ferronickel smelting (researchgate.net) (mdpi.com). For FeNi, Wei et al. show ~44% of total energy in kiln calcining and a comparable share in EAF smelting (mdpi.com), while exergy studies report >30% exergy destruction in the kiln alone — a thermodynamic tally of irreversibility and loss (researchgate.net).
Efficiency‑improvement strategies
Slag and charge chemistry. Adjusting flux and slag composition can lower melting points and improve heat transfer. Zhang et al. (2023) report that increasing slag basicity and FeO (and reducing Cr₂O₃) in a CaO–MgO–SiO₂–Al₂O₃–FeO–Cr₂O₃ system lowers smelting temperature and time, “helpful for reducing the temperature and time of smelting, and reducing the energy consumption of the RKEF process” (rd.springer.com).
Furnace design and scale. Upgrading to high‑voltage AC or DC EAFs and scaling to 75–100 MW has cut specific consumption; some designs achieve under 400 kWh/t dry ore, whereas older, smaller units exceed 500 kWh/t (mdpi.com). Still, “only marginal improvements are possible with [conventional] process design,” reviews caution (mdpi.com).
Fuel switching and by‑product reuse. Modeling shows dividends from using rotary‑kiln off‑gas to preheat/dry ore and recycling EAF vent gas into the kiln as supplemental fuel. One analysis predicted saving ~1.64 t/h of coal (≈46% reduction) via the first measure and ~3.21 t/h (≈38% reduction) via the second (researchgate.net) (researchgate.net). For HPAL plants, note that while no coal is used, very large electric power is consumed for pumps/heaters and mostly to heat slurry water (nickelinstitute.org).
Carbon source and energy mix. The electricity mix matters. In Wei et al., powering a ferronickel plant with ~70% hydropower lowered energy use to ~110 GJ/t (6 t CO₂‑eq) versus ~174 GJ/t (14 t CO₂) with coal‑based power — about a 37% cut in specific energy and ~64% in emissions (mdpi.com). Indonesian policy signals are aligned: regulators are pushing smelters toward renewable/grid alternatives (ebtke.esdm.go.id) (finance.detik.com).
Heat recovery and integration
Biggest prize: waste heat. In an RKEF kiln, roughly 25% of input energy leaves in flue gases, ~15% in vaporizing moisture, and ≈8% via shell radiation; overall energy efficiency is ~66%, meaning about one‑third of fuel energy is lost as hot gas/water (researchgate.net). Exergy analysis similarly finds 33.1% exergy destroyed in the kiln (researchgate.net).
Flue‑gas heat recuperation. Capturing heat from kiln exhaust (often several hundred °C) to preheat and dry incoming ore can displace much of dedicated dryer fuel, per analyses advocating heat exchangers on the kiln and integrated drying (researchgate.net) (researchgate.net). Where flue‑gas/boiler trains are added, plants typically address deposition risks with programs that include scale inhibitors.
Shell/radiation capture. Case studies on similar rotary kilns (e.g., a 30 m cement kiln) show ~4,980 kW of waste heat recovered via annular absorbers — enough to generate ~0.86 MW by an organic Rankine cycle (ORC) turbine (researchgate.net). Rough scaling suggests a longer Ni kiln (e.g., 50–80 m) could yield tens of MW of recoverable heat, usable for multi‑MW power or feed heating (researchgate.net).
EAF off‑gas. The arc furnace off‑gas — a hot admixture of CO, CO₂, and steam at ~1000 °C — can be routed to the kiln or a waste‑heat boiler. Recycling this gas into the kiln alone could cut ~38% of kiln coal usage in one model (researchgate.net). Multi‑line parks can cross‑feed waste heat so one furnace’s off‑gas preheats another line’s feed.
Power integration and policy. Indonesia is pressing smelters onto renewable/grid alternatives, with policy statements targeting 100% renewable (or gas‑based) power for new smelters, especially NPI (ebtke.esdm.go.id) (finance.detik.com). The new Weda Bay nickel refinery (8–10 GW capacity) aims for solar PV to supply ~50% of power by 2025 and ~60–70% by 2030 (ebtke.esdm.go.id) (finance.detik.com).
What the numbers imply
Across case studies, the kiln/EAF pair is the energy center of gravity. Wei et al. put FeNi at 110 GJ/t (35% Ni) with refining accounting for ~44% in Ni metal, smelting ~45% in NiO, calcining ~44% in FeNi, and smelting ~67% in NPI (mdpi.com). Dryers and mills are small contributors; thermal steps dominate (researchgate.net). Even at 75–100 MW, specific consumption hovers around 300–400 kWh/t of dry ore (mdpi.com).
Heat recovery looks decisive: dust‑laden flue accounts for ~25% of kiln input heat, moisture vaporization another ~15%, and radiation ~8%, implying large recoverable streams (researchgate.net). Integrations such as kiln‑off‑gas drying and EAF‑off‑gas recycling show modeled coal savings of ~40–50% (researchgate.net). Implementing heat integration — preheating, flue boilers, ORC turbines — could plausibly halve net fuel use, while cleaner electricity (e.g., ~70% hydro) cut specific energy ~37% and emissions ~64% in modeling (mdpi.com).
As Indonesia leans on smelters to shift power sources (ebtke.esdm.go.id) (finance.detik.com), investment priorities emerge: high‑efficiency furnace designs, waste‑heat boilers/ORCs, and renewable back‑ups. On the utility side, supporting systems — from cooling to fluids handling — tend to move in lockstep with energy retrofits; integrators typically fold in items like water‑treatment ancillaries as upgrades roll out.
Each numeric claim above is supported by process models or case studies: Wei et al. (energy 110 GJ/t; GHG 6 tCO₂/t for FeNi with hydropower) (mdpi.com), Zhang et al. (slag tweaks reducing temperature/time) (rd.springer.com), Liu et al. (coal savings 1.64–3.21 t/h via heat reuse) (researchgate.net), Quintero‑Coronel/Rong et al. (kiln heat losses and 33.1% exergy destruction) (researchgate.net), and Indonesian sources (renewables targets, Weda Bay plans: ~50% solar by 2025; ~60–70% by 2030; 8–10 GW capacity) (ebtke.esdm.go.id) (finance.detik.com). The synthesis points toward practical levers to capture multi‑MJ and multi‑tonne‑CO₂ gains per tonne of nickel.
Sources: Authoritative technical and industry studies — mdpi.com; researchgate.net; researchgate.net; rd.springer.com; ebtke.esdm.go.id; finance.detik.com.
