Starting today, we will be publishing a series of articles to help you better understand each of the 7 transformations of the Factory of the Future journey developed by Agoria and Sirris. We will be starting with the Digital Factory transformation, taking a closer look at 'Machine Learning', a technology at the crossroads of data science, artificial intelligence and process engineering.


Machine Learning to support the development of the Digital Factory

Agoria supports companies in their transformation into a factory of the future. Within this framework, we will be publishing a series of articles to help you better understand each of the 7 transformations of the Factory of the Future journey, and we will be taking a closer look at a few of its components.

Today, within the framework of the Digital Factory transformation, you will discover Machine Learning, a promising technology that will soon be a must-have in industry, and see an example of its application at Duracell.

 

Manufacturing companies and Machine Learning: a 'match made in heaven'. And so it would seem at first glance, but… too few manufacturing companies today capitalize on the gigantic advantage that Machine Learning holds in store for them.

What is Machine Learning?

In a nutshell, Machine Learning can be described as the juncture of Data Science, Artificial Intelligence and Process Engineering.

  • Data Science

Thanks to the enormous computing power of today’s computer systems and the affordability of computer memory (which did not use to be the case) enormous volumes of data can now be mined and processed. Data Science is an interdisciplinary field for mining, processing and analysing large volumes of data. The aim is to distil insights from the data, which can be useful for business success in 1001 ways, e.g. predict buying behaviour, make stock management more efficient, optimise energy and traffic flows and… optimize the operation of machinery.

  • Artificial Intelligence

Artificial Intelligence uses the aforementioned data to find solutions to existing problems. And the great thing about AI is that problems that used to be difficult if not impossible to solve, can now be tackled successfully.

AI software do not just execute tasks passively – they learn from the data. AI systems learn from mistakes, try out better alternatives, and look for solutions in creative ways – in a word, all the things that used to be possible only with human intelligence.

  • Process Engineering

Process Engineering can in turn make rewarding use of Data Science and Artificial Intelligence – especially when it comes to designing, optimizing and innovating industrial processes, production machines and machine tools, so as to make them smarter, more effective, more cost-efficient, and more flexible.

Data Science, Artificial Intelligence and Process Engineering come together in Machine Learning. And it should be clear by now: Machine learning is no longer remote and of no concern of mine, but holds enormous promise for our manufacturing industry. What is more, anyone who still doubts the need to get aboard the Machine Learning train today, will probably be sweeping the platform tomorrow. Put another way, not getting on board is not option.

Strategy for manufacturing companies

And yet we see that today only 58% of Belgian SMEs in the manufacturing industry are actively capturing data. Even fewer SMEs, 39%, actually do something with their data. And just 14% are already testing, implementing or using AI.

Congratulations are in order for all those who have already taken steps. But there is still a lot of work to be done and time is running out. Faster players are lurking on the market. And small, agile start-ups can nowadays carve a gigantic market share for themselves and wipe out traditional values in no time at all – there are precedents.

So, manufacturing companies: you must make sure that you have a Machine Learning playbook. CEOs: you must work together with data scientists to innovate your business model and processes by integrating Machine Learning. Become data-driven – this is not a hype… it is a must! And don’t get stuck in the planning and modelling phase. Model, test and implement. And then, assess, model, test and implement again. And so forth, and so on.

Also ask the right questions:

  • How do I structure the mass of data that will roll in?
  • What infrastructure will I need for the implementation of Machine Learning?
  • What are the limitations, what can Machine Learning do and what can’t it do?
  • How am I going to prepare my staff for the switchover to Machine Learning – to use, maintain and make the system smarter?

A lot of Agoria members can help you with this.

Machine Learning , Business As Usual?

Machine Learning may very well be a technological story, but human beings and the process still take centre stage. Machine Learning is just a ripple in the water if no one asks the ‘machine’ the right questions. Even more important: the success of integrating Machine Learning in your manufacturing company is inextricably linked to a good relationship between man and machine. So you must endeavour for a support base and a broad consensus about the benefits of Machine learning -- on the commercial, operational and human front -- in all parts of your company. Because those benefits are there. And they are colossal.

And however revolutionary the technology may be, applying and integrating it is still a matter of going through a process. A machine learning project is not that different in and of itself from another digital project. And you will notice that only when you start working with it!

Winston Churchill also had a good piece of advice for us, which is ever so relevant in the context of Machine Learning and the manufacturing industry: "If you don't take change by the hand, it will take you by the throat." Best of luck and success with the transformation.

Agoria stands behind you, in particular by organizing exchanges between technology providers, innovative manufacturing companies and progressive start-ups, and by sharing the insights afterwards. We are looking forward to your initiatives and experiences also!

Le machine learning est une des composantes de la transformation Digital Factory évaluée dans le FoF Scan. Vous voulez savoir où en est votre entreprise en matière de transformation en factory of the future ? 
Remplissez le FoF Scan ADMA(European Advanced Manufacturing Support Centre dont Agoria et Sirris sont partenaires.)

Machine Learning & Digital Twin at Duracell

Responding rapidly to realtime information has been common practice at Duracell in Aarschot for some time. However, the step to Machine Learning is being taken now. The aim is to anticipate machines correctly on the basis of predictive trends in the data.

Paul Nuyts, Industry 4.0 project leader at Duracell : "We have taken the first step on the way to machine learning with a worldwide pilot project on a product mixer. It is crucial for us to know which conditions contribute to a good or to a less good product. Up to now we investigated this subject through laboratory measurements. From now on we want to do that with a self-learning algorithm that takes into account all possible parameters such as product, process, materials, environment, etc. With this pilot project, we want to show the Duracell organization worldwide that machine learning is the way to go!”

Because the Duracell pilot project records all product parameters, this Machine Learning project is also an important step towards Digital Twin, which is a virtual representation of the product. Digital twins are particularly useful for product design, simulation, control, optimization and maintenance.

"Our Smart Digital Assistance project goes even more emphatically on that direction,” Mr Nuyts says. "We are working on this with a subsidy from the Vlaamse Organisatie voor Innovatie en Ontwikkeling (VLAIO) [Flemish Organization for Innovation and Development]. There is always some drift in production: gradual deviation from the values under the influence of changing environmental conditions and slight differences in the composition of ingredients. Monitoring and adjusting this requires a lot of data. All line sensors and all SAP, process, environmental and storage data come together via an interface on the Omega platform, which essentially forms a Digital Twin of the product and the related process."

 

If you have an idea of what you could do with AI within your company, perhaps because you have already followed our “AI in Business” online course, then take part in our Agoria AI Discovery Sessions and be challenged by Artificial Intelligence (AI experts).