Smart Business

Interview Gunther Kegel and Frank Hansen: Connectivity is the Big Game-Changer

IoT is growing up and the needs and demands of customers are changing fast. For companies, simply connecting a couple of components is no longer enough – customers want solutions that can do whatever the job requires. System vendors and distributors are under pressure to provide fully integrated IoT environments out of the box instead of a box full of gadgets and devices.

by Tim Cole

Smart Industry asked two industry heavyweights to share their out-look on where the IoT industry is heading: Dr. Gunther Kegel, president and CEO of Pepperl+Fuchs, a worldwide specialist in both factory and process automation; and Frank Hansen, vice president for technical resources and marketing at Avnet Silica, the European semiconductor specialist division of Avnet, a leading global technology distributor.

Why does it make sense to bring you both here together instead of interviewing you separately?
Kegel: I think it’s time to under-stand that we’re all working for solutions for our customers – and that requires not only an in-depth understanding of the application but also the ability of the suppliers to find and implement the right solution. More and more, we have to work together, not only in pairs but in even larger teams to give our customers the perfect application. We have our partners that provide us with parts for these solutions and I truly believe they should become a more integral part of the solution design rather than just supplier. The application-specific knowledge from our company, as well as from our sup-pliers, is essential to create new and novel solutions that help our customers be more successful.

What I hear you saying is that, instead of components, customers want solutions tailored to specific needs and tasks. How does this trend affect the market for IoT in general and sensors in particular?
Kegel: Nobody implements Internet technologies just for the sake of the technology, especially not on the shop floor. The real value of technology to the customer lies in the data-driven business models it enables. The shop floor feeds these business models with the appropriate data and that’s why these new IoT concepts are very much more solution-oriented than they are component-oriented. IoT today is about business and business solutions. Components are still important to our legacy business, of course. In fact, we are experiencing tremendous growth both in our solutions business and at the same time in our components business. Many customers don’t want to do the integration themselves so, eventually, they turn to integrated products – but the solution itself consists of more than just components. Electronics, mechatronics, and software all play a part in building the final solution. It would be wrong to say that we are moving away from components into solutions; it’s additive. We create more business by offering solutions.

Many customers don’t want to do the integration themselves so, eventually, they turn to integrated products.
Dr. Gunther Kegel, President and CEO of Pepperl+Fuchs
{bqalt}

 

Frank, Avnet has a history in components. Isn’t this solution stuff a whole new ball game for you?
Hansen: Definitely! But the old buzzword IoT is moving into real applications and use cases. More than 80 percent of our current engagement with customers is in areas like predictive maintenance. That’s the entry point for things like data warehousing, data analytics, and machine learning. With the current economic situation, time to market and an optimized cost of ownership strategy are key. We, at Avnet, can support our customers with our own resources, or can help them establish external ecosystems to bring solutions to their customers.

A few years ago, everybody believed software was eating the world. Today, the buzz is mainly around cloud solutions, data analytics, and artificial intelligence. How should vendors react?
Kegel: Whereas we see lots of software integration on the office floor, we don’t have as much on the shop floor, yet. Okay, we see some automation-specific digital communication capabilities emerging, but it’s not what you could really call the Internet of Things. So, for us as a leader in providing industrial sensors and sensor solutions, connectivity is the big game-changer. In manufacturing, sensors are nothing new. Today, while we still have the same sensing capabilities, it’s about data and measuring values in order to turn them into data, and then connecting them to the Internet of Things. This requires a lot more electronics and software, which needs to be integrated with the components. We have to convince our customers that we not only understand their applications but can also connect existing components in ways that are simple and competitive.

Achim Berg, the president of Bitkom, the German IT manufacturing association, recently warned that a lack of experienced personnel in areas such as data analysis and sensors is throwing a monkey wrench into the development of the so-called “Industry 4.0” in Germany. Do you agree?
Kegel: Yes and no. I think we are well-prepared to bring the Inter-net of Things to the shop floor. That’s because Germany as the world leader in machine engineering and automation technology basically owns the shop floor. What we need to do now is to turn these elements on the shop floor into the “things” of the Internet of Things. I think we are in a very good position; we may even lead the race at this point. The question remains, however: what do we do with all this data? Do we also lead the world in terms of machine learning, artificial intelligence, and big data? Unfortunately, the answer is clearly no. There is no company in Germany that is really relevant in social media and all these new data-driven business models. That’s the missing link: can we turn our dominant position on the shop floor, where all the data is generated, into a leading position in the use of that data and turning them into a customer advantage?

Despite all the hype, Industry 4.0 is still very much in its early stages, in Germany at least. Mechanical engineering still seems to be struggling to digitize; does that worry you?
Hansen: No, not really. I think that what we are seeing right now is a certain degree of internal digitalization but these machines are not completely connected yet. They don’t share this internal data with others as they would in an entirely data-driven business model. In their laboratories, in their R&D centers, these machine builders are all completely digitized. Everybody understands that the machines of tomorrow need to reach across the boundaries of factories by connecting suppliers and users; they completely understand the necessary balance between machines and the supply chain. Honestly speaking, I don’t know of a single manufacturer who is not eagerly working on getting their machines fully digital and fully connected. At the moment, end users – customers – are frankly not yet that interested. They talk more about partial digitalization that gives them direct access to the machine and to the sensors, but they don’t yet think about connecting these ma-chines to all the other machines or machine centers, even across the boundaries of their factory. So, this is more an evolutionary step-by-step approach. I don’t think it’s the machine builders that are holding us back. These guys are ready and waiting to roll out fully digital machines as soon as the customers start asking for them.

The machine builders are ready and waiting to roll out fully digital machines as soon as the customers start asking for them.
Frank Hansen, Vice president for technical resources and marketing at Avnet Silica
{bqalt}

 

There are lots of really old ma-chines standing around in the shop floors in Germany that were built long before anyone dreamt of connecting them to anything. What do we do about the legacy systems?
Kegel: Legacy machines are a perfect example of how you can add new connected sensors that feed the relevant data into a data-driven business model. You may still be using the mechanical part of the old machines and simply install new control mechanisms with all the digital connections you need. Most of the control-system manufacturers offer this kind of exchangeability even for some very old machines and, if you don’t want to do this, you can still add some sensors that allow you to plug in to cloud applications and connect even the oldest machines to a server architecture which feeds into a cloud application. Okay, if you’re, say, a super-large process plant, like BASF, that operates steam crackers that were built 20 years ago, then it might be difficult to imagine how we can digitalize all this without exchanging all the sensors and the different actuators. However, digitalizing all this might be very costly and hard to do, especially while the steam cracker is still running. The truth is that we have different kinds of industries that need different kinds of migration strategies toward the digital technology. But I don’t think that the machine builders are really the limiting factor.

Frank, is that your experience, too?
Hansen: Absolutely. I totally agree. At this year’s Hannover Fair Digital Days, all the companies were offering their own cloud solutions, and everybody also had their own connectivity solutions available. They’re all already showcasing their solutions which are currently available on the market.

What will the greatest impact of artificial intelligence be to your sector?
Kegel: First of all, we need to understand that artificial intelligence is not really new. I was doing my PhD 30 years ago and guess what? It was about integrating multi-sensor signals into robot control by means of artificial intelligence. Artificial intelligence is not any kind of magic. It’s really a set of defined algorithms that have been developed over the last couple of years simply by adding tremendous computational power, tremendous data storage, and, most importantly, labeled data that can be used to train these algorithms much more efficiently than we could 30 years ago. Users on the shop floor need to return to what I like to call tool realism. We need to understand, in realistic terms, what value these tools can create. It’s not that they’re standing the entire industry on its head, it’s more about an evolutionary approach to things like predictive maintenance. Understanding machine behavior and predicting machine behavior by analyzing standard data sets is crucial. When you talk about an image processing system, this is, today, dominated by deterministic algorithms to which you can apply machine learning algorithms. The advantages these new algorithms are bringing with them will give us the next push – but I would like to state that I’m not overly optimistic that this is really going to change the industrial world overnight.

Frank, sensors themselves seem to be getting smarter. Is artificial intelligence moving to the edge?
Hansen: Definitely, yes. As Gunther already said, most of our time is spent working with customers to give them an understanding of how artificial intelligence or ma-chine learning can improve their business case today. Customers must be brought to understand what they can get out of their data by developing new uses and new business cases. This is our main focus when we talk with customers about artificial intelligence and machine learning.

IoT is growing up and the needs and demands of customers are changing fast. For companies, simply connecting a couple of components is no longer enough – customers want solutions that can do whatever the job requires. System vendors and distributors are under pressure to provide fully integrated IoT environments out of the box instead of a box full of gadgets and devices.

Tags:

Leave a Reply

Your email address will not be published. Required fields are marked *

*