Moving the Dial to Decarbonise

David Wong and colleagues discuss how the world’s aluminium smelters could enable the rapid uptake of renewables in global power systems

DECARBONISING the energy sector by 2050 is a key aim of the International Energy Agency, as one of the outcomes of the Paris Agreement. While renewables uptake must continue and intensify, the unpredictable nature of new sources of variable renewable energy, particularly solar and wind, poses a major technical challenge for power systems worldwide. As variable renewable energy increases, generation becomes less dispatchable and more variable, so the question remains: “what grid-scale solutions should be used to balance electricity demand with generation?”

At Energia Potior, we see primary aluminium smelters becoming significant pieces in the global grid-balancing puzzle. The informed chemical engineer might ask however: “Hold on! Isn’t this an industry renowned for highly-intensive and inflexible energy demand? How could smelters be useful in balancing power systems?” This is certainly true in the way smelters have been designed and operated in the past – shutting power to a smelter for even a few hours can lead to catastrophic freezing of the cells’ molten electrolyte and metal contents, requiring multimillion-dollar investments to restart operations.

This inflexibility in energy requirements is the primary reason aluminium smelting is recognised as one of the big three hard-to-abate sectors for CO₂ emissions (along with cement and steel production). Electricity generation from fossil fuels accounts for around 80% of greenhouse gas (GHG) emissions from aluminium smelting globally. It is also the key driver behind the EnPot technology, which Energia Potior has spent the last decade developing.

EnPot is an enabling technology that converts a smelter from a huge inflexible energy consumer, to one that can modulate energy use and support renewable grids. At the heart of the technology is a mechanical system of heat exchangers, engineered to not alter the electrochemical process, but allow smelters to modulate energy use up and down without disrupting the internal heat balance of the pots.

This moves smelters beyond the constant-baseload “straight-jacket” to one that enables flexible operations and energy use, to the order of ±20–30% of nominal baseload demand.

As introduced in issue 891 (September 2015, p52-54), this equips smelters to act as “virtual power plants”, providing critical demand-side response services to the grid, enabling greater uptake of renewables. During times of insufficient generation in the grid, the smelter can reduce power demand, minimising the need for dispatchable backup generators. Conversely, smelters can soak up excess power during times of surplus generation, rather than curtailing generation (particularly from variable renewable energy).

How EnPot works

To understand the technology we must first do a quick revision of the Hall-Héroult smelting process (see issue 864, June 2013, p45-47). Aluminium metal doesn’t exist naturally, so we first refine bauxite to create alumina (Al₂O₃), then smelt alumina into aluminium, by dissolving alumina in a bath of molten cryolite (Na₃AlF₆, plus additives) so that electrolysis can occur, typically at temperatures between 940–980°C. Direct current (anywhere from 100–600 kA) is passed through the cells, with liquid aluminium metal deposited at the cathode, while anode carbon is oxidised to produce carbon dioxide (CO₂).

Molten cryolite, however, dissolves not only alumina, but everything else including carbonaceous and refractory pot linings and steel shells. The only thing that can contain molten cryolite on the sides of the cell is a “ledge” of frozen cryolite. A temperature gradient is required to maintain a freezing point isotherm, or else catastrophic failure will occur. With reduced heat generation it won’t be long before the molten electrolyte solidifies. Conversely, reducing the heat losses will melt the protective layer.

Heat generation within the cell is via electrical resistance, while heat loss is via fume extracted from the cell to gas treatment centres (50%), and natural convection through the sidewalls (35%) and bottom (15%) of the cell.

The electric current therefore performs two tasks, each accounting for about 50% of the total energy used:

• to provide the energy for the electrochemical reaction; and
• to maintain the critical heat balance inside the pot.

This is why almost all smelters still operate under a constraint of constant, 24/7 electricity supply, to maintain the balance between the energy supplied and the natural heat loss to ensure the protective frozen ledge is stable. 

This is where EnPot technology comes in, enabling artificial manipulation of the heat loss, to keep things in balance when the energy input is varied. This is achieved by fitting patented shell heat exchangers to the sidewalls of every cell. The system draws ambient air over the sidewall through the heat exchangers, using the air as a transfer fluid to absorb energy. Air is drawn through ducting to an external suction fan (see Figure 1).

At the nominal smelter set-point, the fans will be working at a slow rate to recreate the natural heat loss through the sidewalls of the cell (see Figure 2 – Neutral mode).

Increasing fan speed will create a greater cooling effect (Figure 2 – Cooling mode) and allow for increased energy use (with corresponding increased metal production, but also greater heat generation). Decreasing fan speed will reduce the heat transfer (Figure 2 – Insulating mode) and allow for energy use to be reduced (with corresponding decrease in production and heat generation, without freezing cells).

This controlled heat transfer enables a smelter to vary its energy consumption within a ±20% range almost instantaneously, for an indefinite time period (the range can be extended to ±30% if transitioned over a longer period), while maintaining the heat balance for stable, efficient cell operation.

The fans can even be turned off completely in the case of a full power outage to the smelter, creating a “full insulation mode”, to both reduce the harmful effects on operations, and extend the duration that cells can go without power and still safely restart.

The system can be retrofitted to any modern cell technology, and can be installed on cells whilst in operation, or during cell reconstruction.