Select region:
English (APAC)
aumina refinery in the distance over bay

Is a floating alumina refinery possible?

Jock Armstrong Sub-sector Leader

Jock Armstrong | Global Light Metals Sub-sector Lead | 18 August 2017

The Shell ‘Prelude’ project is embarking on a revolutionary step-change in liquefied natural gas (LNG) processing; a 488m long floating facility that will process, liquefy, store and transfer LNG at sea, directly above the natural gas field. Could the same floating concept be applied to an alumina refinery moored at a shipping berth close to a bauxite supply?

Imagine a portable alumina refinery asset that can be moved to another location and continue to operate…  

Traditionally, greenfield alumina refineries have been placed in the most economic location, whether that is at or near the bauxite mine, an alumina smelter, or a reliable, cheap energy source. However, as bauxite reserves are depleted, or energy prices increase, the operating costs and viability of the refinery can be significantly impacted, decreasing its value. In addition, politically unstable or socially fragile regions can also impact the value of an alumina refinery, leading to an asset that can no longer be operated as intended. 

A portable refinery can eliminate these risks. It can be moved to another location, and with some infrastructure capital, can continue to operate effectively as a refinery. What’s more, a portable refinery can potentially hold a higher value to its owner than a land-based refinery in the same region.

Following similar principles to recent floating liquefied natural gas projects, the following article explores design considerations for the portable floating alumina refinery, called ‘Going FAR’.

The Floating Alumina Refinery (FAR)

During operations, Going FAR would be located at a purpose-built shallow water mooring facility. The hull of the refinery will be used to store liquid raw materials such as fuel, caustic, desalinated water and other liquid chemicals.  In-process liquors would be stored in separate compartments to provide liquor volume surge for the refinery.

Alumina would be directly loaded from the refinery onto ships for transport to the smelter or customer, while inventories of bauxite, lime, spares and parts would be stored in onshore facilities. As an alternative, it would also be possible to support the storage of materials and services with barges or ships, as opposed to land-based infrastructure. 

Employees would live in nearby onshore facilities. Major maintenance will also be carried out in onshore facilities, however moving equipment easily between ship and shore is seen to be major challenge that would need to be resolved during the design process. Bauxite residue will be neutralized and back loaded to the mine, returning the residue to its original location. 

When the bauxite reserve is finally depleted, or other factors make the operation no longer viable, tug boats would move Going FAR from its moored location to a new bauxite supply.

Refinery capacity

Typical land-based refineries can produce between 100 and 150 ktpa per hectare of the refinery process areas, excluding stockpiles, utilities and services, and residue storage. Going FAR will require new processes and technologies so that the facility is more compact, with a layout design that maximizes the use of space. 

Vertical integration of facilities in the layout design, like utilizing multiple floor levels, will help reduce its footprint. Using Prelude as a guide, the area for Going FAR would be around 36,000 m2 (3.6 hectares), which gives an estimated refinery capacity in the range of 1.1 to 1.6 Mtpa.


A key criterion in the technology selection for Going FAR is compact, high-efficiency units that consume a small footprint volume (i.e. length x breadth x height). Some key areas of the refinery that require significant changes in technology from a land-based asset, are examined below.

Water balance: Seawater cooling could be a suitable technology.  It has no evaporation losses and therefore requires no additional make-up water. However, the environmental impacts of elevated temperature need to be managed and the risk of caustic contamination removed. 

To minimize the total evaporation requirement, water input to the process needs to be minimized. The water inputs from washing product hydrate and mud are discussed further in following sections however, management of other parasitic water inputs are just as important. 

Going FAR will have limited online liquor volume surge capacity and evaporation. As such, the design of the refinery will need to minimize potential sources of parasitic water  including purge water and hose water.  

Grinding: Wet grinding of the bauxite using caustic is assumed.  This required all Bayer unit operations using caustic liquors to be contained on the ship, significantly reducing the transfer of caustic between land and the ship. This strategy minimizes the potential for loss of caustic containment to the environment by eliminating sources. 

Bauxite transport from the mine to the ship will be by whatever is the most viable method for the given location – whether it be truck, conveyor, rail or pipeline capsule. The latter offers an interesting fit with the Going FAR concept, as it is relatively adaptable to a range of logistic and topological conditions. 

Digestion: In a typical low temperature digestion refinery, significant evaporation occurs in digestion. With limited space on Going FAR, a separate evaporation facility only adds to its total footprint. The design temperature of the digestion facility will need to be set to cater for all evaporation. 

To meet the minimum water balance requirements of the Bayer circuit, a digestion temperature of 160°C will be required. This may require an additional flash heating stage in digestion, which will increase the footprint by 15-20 percent. However, providing the same evaporation rate in a separate facility would use a significantly larger footprint. 

Mud separation: Separation of the residue from the digestion slurry has conventionally used gravity separation thickeners. Due to the space constraints on Going FAR, an alternative technology will be required. Direct filtration of the undiluted blow off slurry would significantly reduce the footprint. Recent advances in vertical pressure filter technology, with fully-automated cleaning cycles, enables the filters to have a very short cycle time (minutes). This provides a much higher filtration rate compared to older style filters, with cycle times of several hours.

Residue washing and disposal: For residue washing on land-based facilities, four to six gravity separation vessels, each with a diameter of 16–20m were traditionally required. For it to fit on Going FAR, a step or technology change will be required. With residue washing using a two-stage pressure filtration process, the viscosity will be higher for the first filtration stage, requiring a larger number of filters, compared to the second stage,  due to the relationship of viscosity and caustic concentration.

For the same caustic losses the wash water rate can be reduced in a filter washing circuit. This reduces the evaporation capacity required. However, the trade-off is an increase in pressure filters for residue washing due to the increase in viscosity. For a reduction in wash water ratio from two to one, the number of operating filters increases by 20 percent.

The cleaning requirements of the filters, processing high-scaling slurries, will influence the number of spare filters. Residue pressure filtration has been proven as a viable technology for the last washing stage. However, using pressure filters for residue with high caustic and liquor phase alumina concentrations is not proven and would require pilot testing and trials to prove the operability of the concept.

Automated chemical wash facilities may be required to remove hydrate scale to maintain filtration rate across different cycles. The discharge from the final pressure filter stage is expected to be a cake of 22-30 percent moisture, suitable for transportation. The residue cake can then be back loaded via the bauxite transport system to the mine area.  Suitable neutralization and/or encapsulation procedures would be employed, based on site requirements.

Precipitation: Novel ideas will be required to reduce the precipitation footprint of Going FAR. Large growth precipitation holding times are required to give high yields, which might not be an optimum solution.  

There is a trade-off of yield versus flow in precipitation to give an optimum holding time. A holding time of 12-15 hours would minimize the size of the precipitation area, however, after considering the impact on other plant areas and product quality issues, it is likely that the optimum yield would be closer to 20 hours. 

The tank volume required for a 20 hour holding time is the biggest challenge to achieving a sufficiently compact footprint to make the concept viable. A move away from conventional cylindrical tanks may allow all or part of the precipitation volume to be stored in the hull, providing improvements in agitation technology are found to drive this.  

Classification and filtration: Product and seed will be classified using hydrocyclones. The cyclones can be located on top of the precipitators, allowing the use of gravity flow to seed filtration, or direct to precipitation. Alternatively, the cyclones can be located on the ship floor level utilizing the head in the precipitators to drive the flow at the required cyclone pressure.

For Going FAR, all hydrate seed will be filtered to minimize spent liquor recycle, thereby reducing the required precipitation volume. Locating seed disc filters on top of the precipitation will provide for a compact footprint, as additional filtration capacity does not consume valuable ship floor space.

Calcination: Calcination of hydrate to alumina is a very energy-intensive activity, resulting in smelters that are typically located close to sources of low-cost energy. While calcination will be possible on Going FAR, do the economics support it? Should Going FAR ship hydrate or ship alumina? Shipping hydrate would allow for the calcination energy to be sourced from a cheaper energy supply, but the free and bound moisture are now transported along with the hydrate, which would increase transport costs. 

Preliminary analysis indicates a break-even price of about US$60-US$80 per barrel of oil, depending on freight distance and gas prices at the calcination location.  The economic decision on the location of calcination needs to be based on the predicted long-term energy price. Despite a recent drop in oil price it is assumed that the long-term oil price will continue to rise. Thus it has been assumed Going FAR will calcine hydrate to alumina prior to transportation.

Steam power station and utilities: The power demands for Going FAR will be slightly higher than those of a land-based refinery. This is because flow will be high as a result of the reduced holding time, which increases power consumption. 

Also, the multiple level layout increases the height of some facilities, which will require increased pumping power. While clever layout and design can utilize gravitational energy to an advantage, there will be inefficiencies that increase the pump power costs above conventional refineries.

A supply of clean water will be essential to replace water losses in the refinery. Reverse osmosis desalination can provide Going FAR with a compact, secure clean water source, using only a third of the footprint of multi-flash desalination, and at much lower cost. Desalinated water would be stored in the hull for use in the power station and throughout the refinery.

Other considerations

Operations and maintenance: Automation of several maintenance and operations activities will be required. Investment in specialized cranage, lifting devices and tools will be necessary to remove equipment for maintenance through tight access envelopes. The layout design will therefore have two competing forces: 

  1. Minimize the footprint and operations and maintenance envelopes to eliminate wasted space
  2. Provide sufficient maintenance and operations envelopes to complete the required tasks ensuring acceptable health and safety risks.

Clear definition and description of all operating and maintenance tasks, logistics and scheduling will be required early in the design process. Risk assessments of all operations and maintenance activities will be required during each step of the design to ensure that health and safety requirements are satisfied.

Environmental considerations: Going FAR will require several measures to protect the neighboring seawater environment. Its design basis limits the transfer of caustic liquors to and from the ship (reducing the risk of environmental contamination) and all caustic liquors will need to be contained on the ship to prevent environmental harm. In addition, the design will need to ensure that liquors from unplanned failures and events are fully contained within the ship.

It is important to maintain high environmental standards, and not become complacent about environment issues, although the impact of a small caustic release into seawater poses lower environmental consequence compared to the release of crude oil from the hydrocarbon industry.

Constructability: Going FAR does not require any earthworks, foundations or civil infrastructure for the refinery areas, which will reduce both cost and schedule. Modularization, and the lack of civil works, will provide a significant schedule advantage of several months, enabling production much earlier than a conventional refinery while also improving the economics.

An estimated cost for Going FAR is in the range of US$2000-US$2500 per tonne of alumina, using low construction labor costs in Asia. Due to the multiple layer compact concepts required for Going FAR, materials cost is assumed to be two times the quantity of steel per tonne of production when compared to typical land-based refineries. 


While the capital cost for Going FAR is higher than a land-based greenfield refinery, this new concept brings other advantages, such as reduced political and sovereign risk in certain regions, and the ability to unlock small isolated bauxite deposits. In addition, Going FAR would have a shorter construction schedule, producing first alumina ahead of conventional land-based refineries. 

Factoring in these advantages, it seems that a favorable business case for this concept is much closer to reality than we originally thought. However, the technical risk associated with the construction of the first Going FAR facility is a significant barrier to overcome. The risks associated with operating complex and compact plants on large floating facilities will soon be better understood thanks to the Shell Prelude project. The lessons learnt from the design and operation of Prelude, and future floating facilities, will be invaluable to the Going FAR concept.

The Bayer technical risk for Going FAR can be reduced with elements of the technology installed in existing refineries as part of efficiency upgrades. Many of the technical improvements needed for Going FAR provide significant opportunity to reduce operating costs in existing refineries. 

With continuing improvements and innovation in technology, it might only be years  not decades  before the first ‘floating’ alumina is produced from Going FAR.

This article is based on the Going FAR (Floating Alumina Refinery) paper - recipient of the 2016 TMS Light Metals Award. This award recognizes the individual excellence of a paper published in the preceding year’s volume of Light Metals. This paper also won the Light Metals Subject Awards – Alumina/Bauxite.