Global Director Industrial Automation
Mitsubishi Electric Europe
Malte currently applies his 20+ years of experience in the automation industry as a Global Director within Mitsubishi Electric Europe. He combines his knowledge from Lenze, Bosch to lead management across F&B, CPG, FMCG & Life Sciences for Mitsubishi. He is an expert in Industry 4.0, edge computing, AI & robotics.
Section 1: Current Automation Strategy
1.1. What is your current automation strategy? What are the intended goals?
We have a thesis – human work must improve the product quality, everything else can be done by automation and robots.
The current automation strategy is to have simple human manned processes replaced with automation and robotics.
And then don’t dismiss people, that’s not the strategy. Instead, educate people, bring people to a higher skill level, so that their work has an impact on product quality. The big challenge is to implement this into the industry to find the right solution.
There are two fields; greenfield sites & brownfield sites. On greenfield sites we are installing a new factory, installing a new line for production. For greenfield sites we have everything ready to plan, and we are changing from classical mass production manufacturing lines to automating everything sequentially. In the greenfield sphere we are implementing cells. We are implementing scalable smaller cells where we have a high level of modularity. You can have more cells; the cells are easy to adapt to new products or slight changes to a product without reworking tools because it’s all software side based. These cells are the future of our production.
It is important to understand that in our industry, mass production still happens but you always have variation, in size, flavor and packaging style. The packaging style can change dramatically for the different seasons, different colors. And therefore, you need automation technologies that are flexible, that are adaptable, and with low downtime. And that is precisely what we are installing in greenfield sites.
The interesting part is the brownfield site. Take your big customers like your Nestlé, they have more than 400 fully equipped factories out there. In these factories it is expected that they have a higher throughput, higher product quality, and to do things simpler with better processes. And in these cases, we are implementing collaborative robots into these factories to simplify processes to co-work with humans. This frees up humans to do more inspection work.
1.2. What are the challenges to implementing automation in your industry?
First is micro stoppages, in existing lines, they are all designed with 98% reliability but the big problem right now is that some of these lines have a productivity of 50-55% because of micro stoppages. We are analyzing these lines, the production data and the reliability to find out what we can do against these micro stops. If you can eliminate the micro stops, you are not changing anything in the hardware or the software, most of the time it’s flow, having the right buffers, deliveries made just in time, and so you can easily bring the productivity from 50% up to 70%. And that is a big bargain for all these big factories.
Productivity on these lines is dependent on humans, we have typically 24/7 production, so we have three shifts of operators. These people are setting the operation of the line, and what we observe is the quality of the product changes over production time. This is one of the big challenges on how to get reliable output.
How can we automate decisions made by humans? How can we get this into a process that is better controlled, perhaps even having predictive elements. For life sciences it is quite further along but you still have these long boring processes, and sometimes it’s a fail and you have to throw everything away.
The next challenge is adding automation to older machines, and then we also come to different types of automation, by adding sensors, having a better view onto your process, becoming less independent on human interaction. And that is, I think, the really important thing that our industry needs. The classical automation like PLCs, this is pure automation, and that is very, very good. But now we have to do something about the bigger outline to get more reliability in production.
1.3. What are the different types of automation you have invested in?
The biggest impact right now is digitalization of production. In greenfield sites, that’s where the latest and best is being invested where digitalization is a must, and it’s standard. But we are talking about brownfields. How to get a 25 year old machine that has strong value, it is productive, it has 20 more years of operational lifetime and it is mechanically wonderful, but how can you know in advance when things are going to go wrong? To solve this, we have to add sensory technology which measures abrasion, temperature of gearboxes which all makes your system more reliable.
The other part is monitoring. Many factories are still working with paper written recipes in food & beverages. We are investing into SCADA, MES and ERP solutions that are connecting all the standalone silos of production and interconnecting them within one factory to communicate to each other and harmonize the processes. That helps prevent micro stoppages and you have a seamless interconnection based on data so you can make decisions.
We also invest into very flexible actuators, for example, robots. With robots, it’s very easy as they can do a lot of different tasks and it only requires a change of software. You can choose a different software and the robot will do a different task. And then for brownfield sites, robotics must be collaborative because your operator is still relevant there and you don’t have space for big fences or laser curtains. We need robots that are safe but can do high speed when there is no operator around, that’s called hybrid robots.
1.3.1. What proportion of your budget are these technologies taking up of your budget?
In automating manual processes with robots or even advanced cyber techniques it is at least 30-35%. The process control, this MES, SCADA, where you interconnect everything, they have seen another 25-28% of investment. And then you have the classical automation, we had robots and edge computing solutions that are in place to collect the data and digital improve the process is around 30-35%.
In the next three, four years, the percentage for classical automation will go down a little bit and we will see the software side grow; the control of the software, the analytics of the software, the handling of the data, there will be further investment into these areas. And then you have the robots, and the robots are on the way. In our industries, the sale of robotics is doubling every year. This is a really fast growing market. Still, we are waiting for new kinds of robots, robots that are better for the process industry, like H1 grease, IP69
1.3.2. Have the investments met your ROI expectations?
When measuring ROI you have two values; what more do you produce; and how much waste you do not create with these new technologies.
For example, a big chocolate producing machine, it’s 25 years old and is operated manually. If you put sensors in, we can predict one breakdown. Mechanical machines take at least a week to repair so we lose a week of production. And that money you can invest into automation technology, robots, sensors, diagnosis software. I think the return on investment is a no brainer. That’s the smallest part. It’s more the really clever way how to use this money you can invest to have the biggest impacts and to explain that to management, that’s really the challenge.
1.4. How has the pandemic changed your approach to automation?
The pandemic is an accelerator. We would have seen the same steps occur but at a much slower pace. The pandemic really boosted the industry.
Everybody is investing into digitalization technologies, so much so, that we have a crisis in the supply of integrated circuits and now we have problems getting the materials in place in time. It’s also a problem with the transportation failures in China in the harbors, they don’t get the rare earth materials out to the production factories.
There’s such a high demand in robotics. For example, in our company, we see that in the number of orders placed for robotics. Robotics is the actuator of digitalization, and it’s flexible, but the order numbers are so huge. To meet the demand, we have to be more independent of manual labor in our simple processes.
Section 2: Future Automation Strategy
2.1. What are your current unmet needs/pain-points?
The biggest pain point is resources. There is a lack of high-skilled people, even in factories, I’ve noticed that big end users in Poland, they have issues in having people run the factory over simple packaging processes. Of course, the solution here can be automation, but it takes time to build up.
Another pain point in general is getting the right resources such as polymers and metals. For the packaging industry, for example, aluminum is very important, and it is lacking. What is really key is to develop technologies to replace them, or if you have to use them, use them as minimally as possible, use them with materials you have in mass, combine them and use. For example, aluminum, only use very thin layers and have a multi-layer product.
2.2. What are your upcoming priorities in automation in the next 3-5 years? Why?
We have new technologies and we urgently need them in place. We have a lot of startups which is great in the R&D phase, but in production, you have no time for startups. So, the biggest challenge here is to have a networks of partners. Digitalization is not a job for one supplier, it is impossible to have one company do all elements of digitization. You have a network of open partners in different fields coming together talking about cloud computing, edge computing, analytics, AI, deep learning. So, these are all fields important for digitization. You can’t have one universal expert in the industry, but you have to build this network.
The priority is to build this network, have the right people coming together, but really keep startups to a minimum. You have to have startups, and I think every big company should have a startup within itself. This startup is allowed to fail. And I’ve seen that when they try to implement startup technology into production, and if that fails, then it’s a disaster. So, bring the technology to the right place, find the right partners, create platforms that are scalable.
2.2.1. Will we see an increase in acquisitions in the automation start up space?
Yes and no. You have some big players and the big problem that happens is if they’ve bought the start-up, they bring them into their structure of their big organization, and the character of the startup is quickly lost.
What I am observing is investment into startups, not really taking them over, but investing, holding shares so that they are not easily bought by another competitor, but keep them intact, keep them as a startup. That allows them to fail. Most big companies don’t understand that startups have to fail. They go down a path that is a dead end and then they have to turn around and try a new approach. So, there is a technique to failing. And at the end, you will have a wonderful, good solution.
Solutions are developed much faster because you go down this failing process and fail early enough. You can run in the wrong direction for years, and then you lose money. But you have to identify early enough that you are chasing the wrong path, and then you go down another avenue, and that’s much faster evolution than with R&D teams doing principle developments within a big companies.
2.3. What are the key emerging technologies that appeal to you in the future?
2.3.1. AI Tech
I think AI technology, deep learning, where machines recognize by themself what are the best methods. Systems analyze themselves for prescriptive maintenance, identifying problems and plan their own service interventions. So, techniques to improve or to keep the machine at a better level.
The real vision and focus on the future will be something else. If you use these new technologies, like AI analytics, deep learning, we will generate the factory for the future. You install this factory at a certain specification, you start to run it, and while it’s running, based on the data that is collected that are analyzed, deep analyzed, this factory gets better, and better, and better. So, productivity process will get better, even process quality, having robots controlled based on AI protocols, so the movements of the robots will get smoother and smoother, and at the same time, fast and faster. So, you have a higher productivity, there’s a better product quality, and that will be based on these new technologies.
2.3.2. Smaller customizable production cells
What we see in the food and beverage industry is that you produce using smaller and smaller cells. You are really getting away from the classical lines of production. Customers products can be customized in different production cells with AGVs, AMRs. There are also technologies for self-driving robotic devices for the different cells, so the future will be to get into parts of machines, production cells, packaging cells, and then there will be flexible devices in between where the products produced will move to the different production steps.
2.3.3. Digital Value Chain
In the life science industry, we are already at a very high level of automation. What is coming here is the interaction of the complete supply chain. From where the raw products are produced, how are they produced and how are they packed in these factories. This process generates more data on the product, and the next step will be to sell raw products as a smart raw product. You will deliver your product with sets of data that can be used by the next user to produce the final good for the consumer leading to further optimization. I call this a “digital value chain,” a chain of different companies that are automatically exchanging data, selling smart products to each other, resulting in a much better product for the end consumer.
Let’s take coffee capsules. In the production of the coffee capsules for one manufacturer, more than 200 million coffee capsules are destroyed for quality checks every year. This is a wasteful part of production. New technologies with deep learning vision systems check all the steps used in producing the end good, developing a digital picture of each single product produced so you do not need to take them out and destroy any products.
2.4. How important is integration across the automation ecosystem?
Every unit, everything you are using in automation ecosystems must be able to have the right interfaces. All are part of the complete solution. You have to have TSN, time-sensitive networks, that today, we have still our Profinet ethernet IP, CC-Link. These are all different standards, and it makes it so difficult to work together. So, get TSN, OPC UA for the IT world. These general networks, standards for communicating must be in place, and you have to have the ecosystem of experts bring this technology. And for this, you need one data highway, like TSN, to implement your entire automation ecosystem.