How Nestle is taking advantage of autonomous technology

Nestlé SA, the Swiss multinational food and drink giant, may be more than 150 years old, but that doesn't mean the household brand is stuck in the past when it comes to manufacturing processes.

In fact, unlike many major manufacturers still clinging to traditional automation, Nestlé is quietly doubling down on the digital transformation of its manufacturing and supply chains. Its goal is to create competitive gaps through data, artificial intelligence (AI), and predictive analytics. In many factories, the company is beginning to deploy Internet of Things (IoT) connected devices with remote-sensing technologies, along with autonomous vehicles and collaborative robots.

The company believes digitalization will help it generate efficiencies, create agility, and provide new platforms for growth. Achieving all of that, however, will take more than the systems themselves. Having true factories of the future will require underlying intelligence and localized or edge computing so everything can run autonomously as more connected endpoints come online.

Please read: Moving from automated to autonomous manufacturing 

We recently met with Ralf Hagen, engineering manager for Nestlé Deutschland, whose team is five years into transitioning several German factories from automated to autonomous systems. Here's what he had to say about how the program is progressing and where he sees it going.

Why is Nestlé on this autonomous technology journey?

Our vision is to use autonomous technology to handle the challenges we are facing in a more localized and reliable way than we have in the past.

Automation has been central to our operations for decades, as it has for many manufacturers. We've had classic client-server structures driving the business. But as digital technology becomes more prevalent, there are more limitations in that approach because of the dependency on server farms located away from factory floors. This central computational approach is no longer benefiting us.

Please read the report: Edge computing yields deeper insights, faster

For example, one of our server farms recently experienced a failure of a hard disk array and an un-activated monitoring [system] that led to an avalanche of failures for all mounted devices. We had to rely on our computing and software vendors to help get everything back on track. And we experienced 30 hours of downtime in the process. Fortunately, this happened on a Friday afternoon, and by Sunday, we were back on track. But it was a problem that this occurred at all. Had it hit earlier in the week, it could have taken down an entire factory. This was an awakening for us—recognizing these system dependencies and realizing that autonomous technologies could help tackle the problem.

How does autonomous technology solve these continuity challenges?

It frees us from relying entirely on automation and centralized systems. It helps ensure contextual data about systems and operations is always available, even with the failure of a disk array. It helps spread contextual data around instead of placing it [in a server farm] and then sending it forward or upwards in a linear direction. Autonomous systems give us the opportunity, even in a hacking event, to continue operations. At the same time, we avoid the kind of waste that occurs anytime you need to shut down to correct, say, a PO [production order].

Please read: How enterprises need to rethink business continuity planning

By keeping things going, we also avoid the waste and mechanical wear and tear that can occur whenever you hit stop. To accomplish this level of autonomy, you must move your technology and compute power close to where plant floor operations data is generated. You're still talking about the same power, but it's situated differently. Hardware isn't just centralized in a room somewhere. It's in cabinets closer to the machines. This is really a step forward that we are trying to take. We are putting the brains back down with the lines where they should be.

Have you already shifted computation power from server farms to factory floors?

At the moment, it's only logically there, not physically. But we look forward to the types of advances the independence of autonomous systems will provide.

For example, with OT [operational technology] security, you need a good foundation in IT security policy. The problem is IT security can be performance killing. Performance requires energy. It requires time. And you have to balance that with your need to secure systems. We are considering how autonomous technology might provide that equilibrium in the background without us having to do much along the way.

It sounds as if that might be one of the main attractions of autonomous systems for you.

Correct. We are surely not going to have a sufficient number of skilled systems administrators in our company to do this because we are a production company. We are not an IT company. We are not an infrastructure company. We are just a simple producer of food. We want systems to be like the iPhone. Everyone has one in their pocket but doesn't have to think about how it works. It just does. On the industrial side, we can learn from this consumer experience. We want to remove the complexity. It is not beneficial. And this is the kind of autonomous state we would like to achieve.

Please read the report: Accelerating the future

That's why we are on a journey to restructure [our operations]. We are trying to re-layer our network layers, for example, according to functionalities. This means a lot of data generation and sharing will happen on the factory floor instead of somewhere else.

I remember not long ago people were talking about "sensor to cloud." This makes absolutely no sense to me. You want to pull reports directly from sensors in real time without having to send data to the cloud only to retrieve it again five minutes later, when it might not be relevant. With autonomous systems, the sensor [information] stays put, and you only send information to the cloud that really needs to be in the cloud for some reason.

We've heard autonomous systems can help better control production quality to, say, adjust the mix of ingredients in candy bars or the printing on labels. True?

Yes, the idea is that your MES [manufacturing execution systems] for tracking and documenting the transformation of raw ingredients into finished products has access to the information needed to make such improvements. As opposed to basing necessary information in monolithic central servers, you arm your local systems to access all the parameters—locally or from SAP, ERP, and other databases—they need to autonomously complete a PO. This is what we're striving to achieve.

Has COVID-19 affected the drive to implement autonomous systems at Nestlé?

It has played a big role in our drive to digitalize. With the pandemic, we have been told that delivering this [autonomous technology] is not just urgent, it's over-urgent and overdue. A lot of things have been made possible the last six or seven months that would have taken two or more years to accomplish previously. Hopefully, we can maintain that rate of change.