Industry 4.0 Implementation for Small and Medium-Sized Shops
Article From: May 2021 Manufacturing Engineering, ME Staff Report
The following is an edited and condensed version of an SME Smart Shop roundtable discussion on Job Shop Implementation of Industry 4.0, which was part of the IMTS Spark online learning series sponsored by AMT—The Association for Manufacturing Technology. The panel was moderated by Alan Rooks, editor in chief of Manufacturing Engineering magazine. The roundtable participants were:
- Chris Pollack, virtual technical application center manager, Siemens Industry Inc.
- Jeff Rizzie, director of digital machining, sales area Americas, Sandvik Coromant
- Tony Del Sesto, technical fellow, MxD
- Vivek Furtado, digitalization manager, machine tools, Siemens Industry Inc.
To access the entire roundtable on video, visit: https://tinyurl.com/SMEJobShop.
Rooks: How can you define the elements and goals of Industry 4.0 in a way that job shop workers can understand and embrace?
Rizzie: The complexity of Industry 4.0, or the appearance of complexity, is what keeps most people from implementing it. At Sandvik Coromant, instead of thinking of it as an Industry 4.0 problem, we call it the connected machine shop. The key is to think big, but start small, and move quickly. If you’re afraid the technology is going to change, don’t worry, it most certainly will! But there are some things that you can do today to get going along that path. It’s all about reducing waste out of the machine shop’s value stream.
Del Sesto: Start off with the fundamentals of manufacturing. The very foundation of manufacturing is continuous improvement and digital technologies do not displace that foundation—they build on it. One of the first steps is the need to measure, get information and translate that to digital. That’s what sensors are for. Sensors can help automate collection of digital information. After that, you need to communicate that information and analyze it. Then the question is what can you find out from that? Now you’re into the machine learning realm, or artificial intelligence. Now that you have everything digital, use that to drive that information back into your control systems. One of the big benefits of Industry 4.0 and digital manufacturing is that they allow you to automate your continuous improvement process.
Pollack: The fundamental purpose of Industry 4.0 is the automation of traditional manufacturing and industrial practices using modern technology. From the 100,000-foot view, we’re trying to take processes that reduce manufacturing efficiency and machine up time, and see which of them can be transitioned to digital. For smaller manufacturers, job setup time is a big issue. What part of that process can we remove to keep the machine efficient and running? Just by adopting something like a digital twin and third-party simulation verification, we can start to make this transition.
Furtado: People wanted more uptime and better performance 20 years back. But today, it’s easier to make decisions that get to those goals based on real data. The challenge is job shops are confronted with this huge IoT vision, and it can be overwhelming. But if you know your processes, you know where to look. Keep it simple. Try to identify three pain points. Get an understanding of your production, use your domain knowledge, and try to get at the low-hanging fruit.
Rooks: What are the key tools that job shops and other small to medium-sized enterprises need in order to start implementing Industry 4.0?
Rizzie: It starts with a culture of continuous improvement. Without that, these tools are really worthless. The first thing we do is send a team of people out onto the shop floor with clipboards and stopwatches and they collect all this information. That alone is not very lean-like, but that’s where digitalization starts to wring out the waste. When you move from manual to digital data collection, it helps gather this baseline information in more efficient and more accurate ways. We need to get our machine tools connected. We need to get information from those machine tools. They can’t be islands anymore. We have to be able to access that data and information so we have a true picture of our shop floors.
Del Sesto: MxD is the Digital Manufacturing Institute, and we certainly work on technology application. But another key area is workforce development. The number one issue comes down to people. If you’re on this journey, you have to think about who is going to install all this stuff, run it, maintain it, and fix it when it breaks. Ask yourself ‘what do I need to do from a training perspective?’ ‘How are my skill sets going to change?’ ‘How are my job descriptions going to change?’ ‘How are my people on the shop floor going to react to this?’ ‘What am I going to do from a change management perspective?’ Don’t dive head-first into the technology until you understand the people side of the equation.
Pollack: Building off of what Tony said, obviously there are all kinds of services and tools you can buy. And yes, it can be daunting to try to implement them and get personnel up to speed on utilizing this software and this equipment to start getting solid analytics. But step one is always understanding your process and knowing where you want to optimize. Today, machine tool dealers have realized that leaving it all up to the end-user, including the cost of the connection and maintaining these client connections and licenses, can be pretty pricey. As a result, more machine tool dealers are providing a service-based solution to their customers so they don’t have to have that knowledge. They don’t have to worry about training the staff on their own. They can rely on others that are servicing a broader market segment that already have the knowledge. Especially for the smaller shops, it’s a great way to get some base analytics without a huge investment.
Furtado: If you don’t know the area you’re trying to understand or the pain point you want to focus on, you definitely need some kind of a monitoring tool to get better transparency into your operations. Which machine is causing the problem? Is it a machine or is it the part program being generated? There are tools available that can tell you where your productivity is being influenced. However, most job shop owners already know where they can improve. You don’t need to focus on the whole job shop to move forward. You can focus on specific areas. Shops don’t want to be confronted with unpredictable machine downtime. There are tools that can tell you in advance when a machine is going to break down so you can do preventive maintenance. You can also have tools that show you when a machine is going out of accuracy and will influence part quality. Siemens offers a suite of digital tools that machines shops can implement to increase manufacturing performance.
Rooks: What should the goals be for the first year of an implementation, and for five years after that?
Rizzie: It’s about low-hanging fruit. Vivek, you mentioned that most job shops know they have problems and they know the glaring ones. And this is where technology can actually start to address some of that low-hanging fruit. When we talk about using digital technology to eliminate waste, it’s across the entire value stream. It’s not just in the machining application; it goes all the way back to design, planning, part processing, and CAM programming. The problem is we don’t communicate across those departments very well. There are lots of silos in machine shops. For example, in one shop, they thought they had measured overall equipment efficiency (OEE) between 72 and 76 percent consistently for the past five years. When we measured it with actual digital data, it was 42 percent. The answer to that riddle is that the shop was not using the right information before. The data was bad. With that information, the shop learned that maybe it didn’t need new machines to increase capacity; it needed to get more out of the ones it already had. For me, the long-term vision is about moving from a reliance on tribal knowledge to system knowledge. It’s gathering data over time. And Tony, you mentioned machine learning. That only comes with gathering data. You’re not going to do machine learning in year one. That will be in year two, year three, year four, and then you feed that knowledge back into the systems that are generating programming or processes.
Furtado: For the first year, it could be monitoring. It could be going from paperless to digital in whatever form. But eventually, something has to come out of it. For the first year, there should be a goal set, such as better quality or better productivity. It doesn’t have to be a great breakthrough. You can take a small step forward. And that could be just identifying the problem. But I think the next steps in the years ahead would be to resolve that problem and move on to other issues and topics to increase productivity, increase quality, and reduce downtime. And as Jeff said, you can start to break down silos in organizations with the transparency that data brings.
Del Sesto: As my colleagues have already mentioned, make sure you understand continuous improvement and have an appraisal of where your problems are. Understand what you’re trying to fix. I see so many companies just running off and investing in technology because their competitors are doing it, and they end up with a hammer in search of a nail. Understanding the problem is step number one. From there, start small. ROI is extremely important for a small business. Start small, but—and this is a big but—there is a danger with starting small, which is you end up with something that’s not scalable, or as you progress into your five-year plan you have a bunch of disparate systems that don’t work together very well, which defeats the whole purpose of Industry 4.0. One of the most important things you can do during the first year is put together a plan, a road map. As you go into your next five years, when you think about things like networks and communications and data structure, you’ve got to make sure that, as you make these investments, they will all work together long term. The best thing you can do over the next five years is continue to measure ROI. Determine if you are on the right track. And then, based on that, continue to refine that plan over time.
Pollack: I agree that it’s certainly going to differ from shop to shop, but in all cases I’m a big advocate of taking small bites. Start with a few machines or a few processes. Look at the machines doing most of your heavy lifting or a new technology or process you’re just bringing online. But as Jeff said, it’s really important to not lose focus on of the entire manufacturing process chain. So often, we all get stuck in the weeds of, ‘I’ve just got to look at analytics,’ or ‘I’ve got to pull data.’ And there’s a ton of stuff that can be optimized through digitalization, from the design segment to the programming piece. How do you manage your part programs? How do you manage your tool data? The answers to those questions will refine the process. And you don’t necessarily have to worry about connecting a machine tool to accomplish that. Look at your entire process holistically and figure out what’s best for you. But be careful that you don’t pigeonhole yourself. In some cases, a shop makes a big investment and gets software that works great on Brand A machine, but on the shop’s Brand B machine it won’t communicate. You need a solution that’s going to be as universal as possible.
Rooks: What are some real-life examples of job shops and other SMEs that are taking advantage of Industry 4.0 strategies today?
Pollack: We’ve worked successfully with institutes like Naugatuck Valley Community College and other schools to implement digital twin concepts that allow them to train more people. And that’s a great way to start implementing. Also, I had an opportunity to go to the Grob [machine tool] manufacturing facility in Mindelheim, Germany. They had a system where they had a centralized tool room that was setting up all the parts to be run. They had a pallet pool system that was basically like a small locomotive train. And this whole track system running this long cell of machine tools, and the system was completely lights out.
Rizzie: We have a solution called Machining Insights, a machine monitoring solution. It’s had different levels of success. Again, it goes back to the customer’s culture of continuous improvement. At one installation, it involved their weekly production information spreadsheet, where everybody on staff got an email and they had to fill out all the information for their area. Step one was rather than teach everybody how to get into all the different screens of the new solution, we just connected that solution to their spreadsheet. Now the spreadsheet is live, and it’s updated automatically from the machine. It saved them hours of time previously spent gathering information. That is a huge cost savings. And in our own factory we had a situation where maintenance was not prioritizing a repair because they didn’t see it as a big issue. Once we had the machines hooked up, the data told us we were losing 20 hours a week of machine time because of this problem. The problem was just a faulty $84 machine sensor. All of a sudden, that became a priority.
Furtado: Our Siemens digitalization team worked with a small job shop with ten machines, and they started on a similar path of machine monitoring and found bottlenecks. They located a machine that was running slowly and holding up the rest of production. At Siemens, we have different levels of software applications; some run on the machine, some on the controller. We also have factory- and cloud-based applications. In this case, the monitoring took place on a factory-based solution. After the problem was identified, we implemented an adaptive control and monitoring solution, which basically is a data optimization system within the control that checks the machining load and optimizes the feed rate. This let the machine run faster and we were able to reduce cycle times by 10-12 percent per part on average. That led to an ROI on the system within six months.
Del Sesto: At MxD, we maintain a 22,000 square-foot research factory here in Chicago. A lot of the work and projects we do are about retrofitting digital systems onto older legacy machines. No manufacturer is going to rip out all of their capital equipment just to go digital. As an example, we focused on one of our own Bridgeport-type knee mills. Some of them have been around since the 1930s, and they’re about as non-digital as you’re going to get. They are old, manual mills, but are still being used in a lot of job shops. Honestly, it’s probably one of the most used machines in our factory. We took an industrial current sensor that cost $100 and put that on the electrical supply lines to the spindle motor, and then we fed that into a Wi-Fi enabled microcontroller that we bought on Amazon for $18. We had one of our high school interns program it. So, for $150, we put our little home-brew system on this knee mill. Now we know when the spindle motor is turned on. And, because it’s an analog sensor, we also know if it’s cutting or not, if it’s under load or not. And because it’s Wi-Fi-enabled, it can now transmit that information. So, using this kind of simple system, if you’re a job shop and have ten or 20 machines, you can open a laptop and instantly see if those machines are cutting or not. And you can record that information historically over time. You can start to look at utilization. You can start to look at OEE. Even though you might have a bunch of old, non-digital machines, that opens up again a whole world of possibilities. A lot of shops say, ‘we can’t go digital because it’s expensive.’ But as our example shows, it doesn’t have to be expensive.