Metso Insights Blog Mining and metals blog Digitalization of primary aluminium production: solutions and challenges
Mining Metals refining
May 31, 2019

Digitalization of primary aluminium production: solutions and challenges

Last month, I talked about some of the challenges of digitalization of primary aluminium production, which is, by the way, a good example of many process industries and why progress is slower than we perhaps think. In late April, I had the privilege to present on the state of digitalization in the aluminium industry at the CRU World Aluminium Conference in London. In this article, I wanted to relay some of the findings and concentrate on what already is out there and what are some of the critical choices aluminium producers face to get started.
Aluminum smelter

There are many buzzwords out there when it comes to digitalization – Industry 4.0, Advanced process control, Big data etc. The reality is that the definitions of these terms are not always clear-cut, and in complex process industries like alumina refining and aluminium smelting, automation and digitalization can sometimes get muddled. Therefore, it’s important we have a clear definition of what the practical applications really are.

At Outotec, we have developed a simple five-stage framework for applications in complex process industries that serves as a testament of our maturity in different stages of the value chain:

1) Measuring and managing data: Systems including both hardware and software for analyzing, sensing and monitoring processes, as well as equipment visualizing measurement results without analyzing them at least in the advanced fashion.

2) Sensing and diagnosing: Software solutions and systems that draw data from assets and processes, which provide analytics and visualization of asset availability, performance and condition for asset maintenance and performance management.

3) Simulating and creating a “twin”: Stand-alone systems that support decision making and learning by exploring process and equipment behavior and testing different scenarios.

4) Optimization: Applications that stabilize, advise and optimize by connecting to process control and equipment, having an algorithm and actionable suggestions to adjust process control parameters manually or in closed loop in order to improve efficiency, performance and availability while reducing costs.

5) Robotics and autonomous operation: Systems and equipment that can autonomously operate based on sensing and gathering information dynamically.

It’s important to note that while the scale is not linear, 1-5 do represent a scale of complexity and can each exist either at the plant level, or at the individual equipment level.

Aluminum production digital application categories

With the definitions clear, we interviewed our main customers, performed a wide search in literature from recent years to understand where we are in different parts of the value chain. Next, we’ll go through just three examples of digitalization in different value chains in the primary aluminium industry: alumina refining, carbon area and potlines.


Alumina refining  automation and process control

Alumina refining

In Alumina refining, there is a lot already out there. It’s important to remember, however, that the difference between automation and digitalization, especially in a complex process plant like an alumina refinery, is not always clear-cut. As individual examples, we’re seeing interesting applications in, for example, Charge Control – i.e. how much alumina is in the liquor, giving a refinery a signal on how much bauxite to charge, online XRF to understand composition of bauxite and therefore changing the recipe of the bauxite, conductivity measurements to understand condensate quality and online charge control as well as infrared Soda measurement. Looking at the whole plant, we’ve already seen plant level APC systems using feed forward / cascade loops - but these are not based on fundamentals, i.e. the chemistry. Multivariate parameter optimizers, on the other hand, where we are seeing first references, use fundamental process knowledge to deliver savings. We can also see quite a bit of predictive equipment maintenance applications coming, be it in grinding mills or refinery parts.

Looking at the big picture, there are three striking observations. Firstly, the latest state-of-the-art refineries are not state-of-the-art in adopting these digital solutions – the digital applications are rather being developed in pockets of existing capacity.  Secondly, there is a lot out there, but still it is all quite piecemeal, i.e. different refineries are developing different applications so the whole benefits are not being felt anywhere. Finally, we are not yet seeing real big data analytics come into play at the whole refinery level – it requires a more integrated approach between process areas.

Rod tracking

Carbon Area

In the carbon area, there are far fewer existing applications than in alumina refining. The first digital products in this area were Anode and Rod tracking systems and there are many out there. However, these typically have been missing most of the analysis, and only now is that side of things being developed, as the tools have also improved.  What we’re also seeing are developments in predictive maintenance. The real beef here is the loop from raw materials to the anode impact on potlines, and these are still under development.

Having said that, we’ve come quite far on automation – and nowadays in a state-of-the-art rodshop, the only must be manned station is really the casting – and even that is due to operators’ preferences, as opposed to a technological constraint.



In smelters, the main thing to think about is energy consumption. What we can see is a lot under development – for example, Hydro has a very extensive and ambitious vision of the smelter of the future – and the first steps, i.e. data management systems, are being put in place. We’ve also seen now a couple of references in plant simulation and twins, for example in Greece. Several partnerships have been signed on co-developing artificial intelligence applications. Integration of areas is lacking, however, and we often hear that potline operators and carbon area have conflicting targets and do not talk to each other enough. We’re also seeing very little development between alumina and aluminium.


Outotec has already been involved in many of the above areas. In Calcination, our new Pretium Calciner Optimizer, which works fully alongside any DCS or control system in place, is showing promising results with up to 10% energy savings in a brownfield installation – in just one calciner, we are saving up to 55 kt of CO2 per annum! (and there are hundreds of alumina calciners out there!) At the same time, in aluminium smelting, our current focus is on building the chain of analytics required from the raw materials of anodes all the way to the potline performance, as well as the recycling loop in the rodding shop. We have a solid basis for this – we delivered the first anode rod tracking system over 10 years ago, and with the hardware already there, now we are using our common Outotec proprietary tools to develop the analytics. First references in the rodshop are being delivered now.

Digitalization is a hot topic nowadays – not just because it’s what everyone is supposed to talk about, but because it brings real benefits. The key, however, is to have very clear goals in mind at all times – in order not to invest in digitalization just for digitalization’s sake.

Mining Metals refining