Artificial intelligence is the real game-changer

Article published in ERJ's July/August 2018 issue 

AI technologies are already here and will gain widespread adoption in tire factories much sooner than the industry expects.

Just as smart tires can self-monitor and adapt to external driving conditions so too will tire factories ‘learn’ to self-adjust and autonomously operate at optimum levels.

This fascinating vision of the not-too-distant future was presented by Cimcorp technology director Jyrki Anttonen at the Future Tire Conference 2018, held 30-31 May in Cologne, Germany.

Like autonomous vehicles – of which smart tires are a key component – autonomous tire plants will employ software-based connectivity, ‘big data’ and Internet of Things technologies, Anttonen set out.

The tire industry, he noted, has already gone from individual device-level controls to integrated device controls and automated materials-handling systems. Production processes have also been integrated through MES software and connection of plants into ERP systems.

“This is more-or-less where we are today,” said the Cimcorp director, while also noting the growing influence of computing-power, which is increasing at an exponential rate, the flexibility of Cloud computing and the availability of very large data sets.

“These have all enabled the development of things like Big Data and finally artificial intelligence (AI), which enables us to make a big leap [forward] in the tire plant,” said Anttonen.

AI is not a new invention: the term having been first used in 1956 and is used today, for example in mobile-phone and online functions, without people realising.

Indeed, the first steps have already been taken towards establishing an AI factory as seen in areas such as predictive maintenance in which sensor data is collected, stored and analysed.

There are also machine-learning algorithms for analysing data and learning, over time, how to predict component failures. Many suppliers already have such capabilities, which Anttonen described as “more like basic Industry 4.0 or industrial IoT functionality… on a very low component or equipment level.”

Machine learning

Looking ahead, he foresaw the introduction of machine-learning algorithms to continuously capture data on the curing-time of the green tire or analyse KPIs such as robot utilisation.

Such algorithms, he added, would, likewise, learn how to position SKUs to get the optimum performance from the gantry-robot or any other aspect of storage within the tire factory.

And by collecting time-series data, an algorithm could analyse patterns so that it learns over time to predict the production of scrap tires in a plant.

“So we can react before we even produce the scrap tires,” said the Cimcorp presenter. “In this case it might be a human intervention or, taking it a bit further, it might be AI that responds to this situation [or] any other event that needs to be known about in advance.”

Anttonen also highlighted the potential role of AI in the complex task of managing the many variables around production planning and scheduling in a tire plant. This includes forecasting orders, seasonality and availability of different materials, equipment or operators – all of which tend to change unexpectedly.

Here, Anttonen said, “AI could simulate different scenarios and select the best way to react to a changed situation. And, as with machine-learning, the algorithm will get better and better and will make better and better decisions.

“Being able to process huge amounts of data, the algorithms could consider far more aspects than a human-being could, including also energy-usage or emissions.”

AI tire manager?

The Cimcorp executive went so far as to suggest that AI could become a member of the management team within tire manufacturing companies.

While the idea might seem far-fetched, Anttonen pointed out that a Finnish IT service company Tieto had already made such an appointment to the leadership team of one of its businesses units. Called AlicaT, the ‘director’ has since 2016 been given the duty of processing huge amounts of information.

“By reading hundreds of thousands of texts each day and from these texts, AliciaT [is] able to provide different company and market information to support the human members of the management team in their decision-making,” reported Anttonen.

“AliciaT can make recommendations but she is not making any decisions – at least not yet,” he added. “And there is nothing that prevents the use of AI in any other industry in this way, including of course the tire industry.”

Nothing mystical

Anttonen’s advice to tire makers was that they should not think of AI as “something mystical or something unreachable, it will be the new normal. Also, you don’t need to be able to develop AI, you only need to be able to utilise it.”

And as with any other investment, managers need to build a business case and define where AI can help the organisation.

Plan how to deploy and make the ROI calculations,” he said. “You should proceed business -led rather than technology-led.”