Dennis Kolberg, head of research at the department of innovative factory systems at the German Research Centre for Artificial Intelligence (DFKI), discusses developing and creating the Smart Factory and how Industry 4.0 will result in a paradigm shift in production operations
How will I4.0 change the automotive manufacturing industry?Dennis Kolberg (DK): I4.0 is a revolution in terms of changing how we manufacture things, rather than just the introduction of new technologies. There is a technology ‘push’ as we have access to new, cheaper technologies, for example if you compare a current smart phone to say a personal computer from 10-15 years ago, the phone is much cheaper and more sophisticated. On the other side there is a market ‘pull’ as the demand for customised products increases. Now you can visit a website and create your own personalised items such as mobile phones, etc.
This puts greater demands on the production process and the process has to adapt to meet new market requirements. So we see these two things coming together in I4.0; we need to change the production environment with intelligent production systems that are able to reconfigure themselves. Also with affordable, high-performance IT systems and software now available it enables these new requirements to be fulfilled. The automotive industry is already quite advanced in this area as they are facing this problem now; they have shorter lead times for products, shorter lifecycle times and more variation of products.
Can you give us an overview of the Smart Factory and its aims?The Technologie-Initiative SmartFactory KL e.V. is a non-profit organisation, base at the German Research Center for Artificial Intelligence (DFKI) in Kaiserslautern. It consists of 48 members, including enterprises and institutes from all over the world, ranging from small SMEs to big, multinational companies such as IBM, SAP, Siemens, Bosch, etc. These have come together to work on I4.0 to develop the shift from research to application. We are always looking for new partners and a key thing is we are not just discussing how we can develop manufacturing operations, we are also showing how I4.0 works and can be applied.
For example, at the 2014 Hannover Fair, we presented the first vendor-independent I4.0 demonstrator. We got together with our partners and decided we would show what was possible in an I4.0 production operation. We defined our vision and created a large scale (7 metre) production line realising the I4.0 paradigms and enabling visitors to the fair to touch and see the ideas in operation. With this we demonstrated that I4.0 is not just about technology but a paradigm shift in the way we should design production lines in the future.
What will be the hardware/software requirements needed to realise this paradigm shift?Before I go into this I should describe what the future I4.0 production environment looks like to offer some context for the hardware/software requirements. Our vision for this is built around four pillars: modular production, the need to be very flexible and able to change production operations or processes quickly; the need for digitalisation, having digital communication from the field device level, up to IT systems and the cloud, so having a common language used by all systems; user orientation, people often regard I4.0 solely as a technology driven paradigm but you need to address the needs of the companies looking to implement this concept; lean operations, this is from two points of view. The first is to look at existing lean manufacturing paradigms (Toyota is a good example for automotive).The second is that we offer solutions that are lean in their application. We have learned from previous mistakes that simply applying more technology to a problem won’t provide a solution on its own. We realised that the solutions we were setting up didn’t need to be too complex. For the user they should be simple but effective.
From these four pillars you can derive the requirements in IT and hardware. With the first of these you need clear interfaces to allow all parts of the process to communicate, using an agreed common language. In addition the technical terms and expressions used in this language must be explained and understood by all. There are some protocols in use, such as OPC UA which is used a lot in Europe, that are capable of this.
From a hardware point of view we need the machines to be able to collect more data and communicate this through an open interface. This is straightforward with new machines but many companies have older generations of machines in use so they need to retrofit existing machines. This approach relates to edge computing and requires the machine to be fitted with additional sensors, which are now more affordable and widely available. These will allow the required data to be gathered, processed and communicated.
Could you explain how the ‘Plug and Produce’ concept fits into the Smart Factory process?As we mentioned earlier one of the pillars of the Smart Factory is modularisation. When you consider producing customised products you need to have a very flexible manufacturing environment, which is realised through modularisation. If you think about a USB connection for computers, it allows almost any device to be instantly connected and to function directly. This is the idea behind Plug and Produce. We took this approach and defined a standard for a single common connector that can connect all of the modules in our production operation. This single connector provides [compressed] air pressure, high voltage and secure connectivity, and will connect all modules in the production line. This also helps when planning any future changes to the line, knowing that all parts can easily be connected.
How are human workers being integrated into the Smart Factory?It's an important element to include human workers in the I4.0 concept. Back in the 1990s, industry made the mistake of thinking one could run production environments without any human involvement. But this wasn’t possible as there is always a situation where even the smartest machine needs for some level of human intervention. As we have developed these I4.0 technologies we have always included the human element in the system. So how can these new technologies support the human worker? One example is the challenge of increasing customisation. This can require frequent training updates for the production operatives dealing with rapidly changing product types and this isn’t always feasible. So the worker needs support systems to guide them through the production process rather than time consuming training; something like a virtual simulation system. It’s also important for the worker to get real-time information from the machines so they can be fully aware of the status of the process and identify any problems, and be guided to rectify the issue.
This will mean we need more people who are able to use these technologies and this will see a shift in training and skills. A good example was the shift in the skills needed to repair cars; where as once technicians where trained as ‘mechanics’, now they are trained in mechatronics to be able to work on complex vehicle systems. But I think this is happening now and it doesn’t matter how old or young you are, people today have a good understanding of tablets and mobile devices.
What level of artificial intelligence (AI) do you envisage being used in I4.0 systems?I think with the growing access to production data and the use of algorithms to process this information, I4.0 has set a baseline for the use of AI. Predictive maintenance is a good example of how we are starting to use AI in processing machine data. While AI is essential to manage and process the huge amounts of data now being gathered, we still need the human element to be the last part of the process to make the final decision.
Do you see these smart production networks growing beyond individual company boundaries, with the sharing of production data throughout the value chain?Certainly the development of networks should grow beyond just machine-to-machine. Companies are realising that they can’t just rely on their own proprietary solutions; they must embrace the use of open source protocols. Automotive customers are realising the potential of these protocols and are encouraging their suppliers to use them. To develop these open interfaces companies need to collaborate, even if they are competitors. This is something we have encouraged and demonstrated within our Smart Factory project.
Security of these big open networks is definitely a concern. There are fears regarding the loss of intellectual property but I think in some cases it is less important who owns the data than who is using the data.