You've probably been wondering about the talk on IoT, Industrial Internet of Things, or data-driven manufacturing. How do you actually get the project done and at the same time be able to realize the value of remotely monitoring your machines, getting data from the plant, finding and eliminating inefficiencies, managing machine health, and simply getting visibility into the factory floor. Very few companies actually talk about the details of how to do it and also talk about the lessons learned. We're here to actually pull back the curtains a little bit and give you more insight into how it happens. This might be a good guide for you to do it yourself, if you like...you're more than welcome to do so. This is also a good guide for you to work with an IIoT systems integration partner to get the job done as well.
We'll talk first about what IoT is and the value very briefly to get everybody on the same page. Then, we'll talk about the process for planning a project, and then implementing that project. Lastly, we’ll review a lot of lessons learned we've gathered over time doing this kind of work.
IoT is taking data from the real world, converting it into valuable information, and using it to gain some advantage or value.
The idea is that you want to use this data to enable more efficient operations, eliminate inefficiencies, find and eliminate reasons for downtime, eliminate scrap to increase production, and be able to do more with what you have.
Data-driven manufacturing or IIoT is the process of pulling data from machines, operators, and other systems. Then, converting that data to viable information to create one version of the truth. Then provide that to people across the organization to enable them to make data-driven decisions to improve the operations of the company. This is necessary because manufacturing is becoming much more competitive. Costs of raw materials are going up, especially during Covid-19. Good machine operators are much harder to find. Equipment is getting more expensive to purchase. Your competitors are becoming more aggressive, and many other factors. With these changes in manufacturing, it makes it more difficult to provide higher quality products to your customers at the lower prices they are demanding. You need to be able to dig deeper into your organization to know what’s going on, to improve operations, and become more efficient than you already are. The goal is to do more with what you have. That is one of the big keys - to be able to do more with what you have.
So, what a company should be doing is connect to the machines, operators, and people to make an integrated IIoT solution that can share data. Keep in mind this isn’t something that you do as a single project. You don’t simply connect to the machines, watch them for a little while, finish the project, and then you move on.
This is more of a long-term strategy, and I don’t say this as a scare tactic, not at all. I’m simply saying this because the idea of becoming a data-driven manufacturing company is going to become a new way of life. Think about what’s happened in the sports of baseball and basketball. There’s data all over the place. Cameras and sensors around each baseball park to track the pitch, speed, and position of the baseball as it’s pitched by the pitcher, and then where and how the ball is hit by the hitter. Then the exit velocity and exit angle of the ball, the exact path of the ball, where it lands, the position of every player, and how they pass the ball around to execute plays. Add to that data about the nutrition and health of each player, and all kinds of other aspects about baseball. Think about what happened with the Oaklands A’s back in the 1990s and how they revolutionized baseball using data to their advantage. That data enabled them to find advantages, find gaps in how baseball was typically played and managed. Those gaps were opportunities they exploited. They exploited those gaps by improving how they played and who they brought in as players to do more with what they had in resources from the owner. They had a lot of success but eventually, all of the other baseball teams caught on and data analytics became the new way to operate in baseball.
This is exactly what’s happening in manufacturing - manufacturers are now realizing to become more competitive in their respective industries they need to dig deeper into their organizations to get data, use that data to their advantage, and try to do more with what they have. Therefore, this is going to be the new way to operate. The idea is if you want to compete you are going to need to find the gaps, find the opportunities, and exploit those opportunities with better
data-driven decisions. So with that in mind, let’s talk about the planning process of an IIoT project.
One of the first things you need to do in planning your project is to assess the business. The idea is to figure out where you are at, where you are going, and identify that gap. Define what challenges you will encounter to get from here to there. Then score those challenges based on complexity, cost, and the value of solving those challenges. Once you’ve scored them, list those challenges in order of priority. Which ones have the highest value to solve and the lowest cost or complexity? Then create your roadmap for how to solve various challenges in your plant potentially using IIoT solutions and data to solve those challenges.
Once you have your roadmap established then you can take a look at one of the top items in that list that is low in cost, complexity, and high in value. Select one of those as your proof of concept or pilot project as an IIoT solution and then start planning that project.
For planning the project, the next step is then to look at your plant and figure out how to execute the project. Typically when we start a process with a customer we will discuss a few different things to establish a plan. Some of it is going to be the information to assess the business, where it’s at, where it’s going, the gap, and the challenges. Once the roadmap is where you
want it, everything we try to do in the proof of concept or the pilot project then should align with those objectives and goals at a high level. At that point, we will start to look at those first challenges that were selected for the proof of concept and go through a step-by-step process to determine the type of solution to create. From the beginning, we don’t want to make any assumptions about what technology will be used.
We start first at the top by creating and defining the business challenge. From there we then determine what the hypothesis is to try to solve that business challenge.
Then we attempt to determine the information that’s needed to solve the hypothesis and determine how to solve that challenge with metrics, KPI, etc. Once we’ve clearly defined the information that’s necessary, we define the data that’s needed to solve that challenge. We try to limit the data as well so that we don’t have to capture too much. The keep-it-simple principle is very important in these kinds of situations.
Once we have defined that, then we attempt to define the technology we need to get that data. It’s at this last step that we finally understand what the solution is potentially going to look like.
In the last step, we're finally able to define the technology to get the data. Here we’re talking about looking at the PLCs we have to connect to, looking at the registers on the PLCs that contain the tags, our data points. Or we are looking at the controllers on the CNC machines. This also includes defining the communication and data protocols, the IoT gateway devices, and the sensors we need. We are also choosing the software platforms to use in the cloud or
on-prem to develop the screens and whatnot which all data consumers will use, from operators to CEOs.
Once the scope of work for the project is defined in that process, we plan how to execute that project. The pilot project typically has a few steps in it.
For the onsite survey, there are a lot of details to gather and this is where mistakes can be made and details can be missed which can have a big impact on the project later on. This is where you can effectively end up painting yourself into a corner, so to speak. Even though you may know your operations internally very well, if you are implementing a project yourself we still recommend taking an outsider’s view of your facility. There can be a lot of benefits in looking at your business from a new, outside perspective.
It’s sort of like a house or a business in your neighborhood you walk by or drive every day but you never really notice it. Then a friend or relative comes in to visit that hasn’t been around for a while and they say “Hey, look at that!” You say “Oh wow I didn’t realize that, I’ve never noticed it before.” The same thing here can prove to be really important.
So when going through the site survey there are four areas that we look at, assess, and document. They are:
For the machines, we are typically looking at what type of machines are in the plant. Do they have PLCs or controllers? When was the machine manufactured? What is the model, or serial number? Firmware versions? These details will be very important once you start to put together the technology solution.
Look for how you can connect to the controllers and the CNC machines, the PLCs, or the other brains of the machines whatever it might be. Do the FANUC controllers in the CNC machines have “embedded ethernet” or are there ethernet ports on the CNC machines, or in the PLCs? Is MTConnect or FANUC FOCAS or FOCAS2 within the controller itself? Is that an option that can be turned on? Or is there RS-232 connectivity, 9- or 25- pin connectivity?
Or, are there I/O ports that you need to work with if nothing else is available. If there is a panel with relays etc. or similar, then do you have the electrical drawings so a controls engineer can look at those drawings and understand where you can grab certain signals from the machines?
For the processes in the manufacturing plant- what does the company manufacturer? How is it manufactured? What is the process? What are the routes? What documents are used, like travelers which move with the product as it goes through the plant? What information is recorded on the travelers and other documents? Typical cycle times at each station? Wait times? How many machines does an operator typically run? How is the data gathered?
Manually on a piece of paper? Excel sheet? Then what happens? Or is it entered digitally and then what happens to that data then? Is it consumed, by who? What form is it in? All of these questions are important for understanding existing operations, how the new IIoT solution can improve operations, and how it should fit into existing or, better yet, updated operations.
For the operators, there are questions about how trustworthy the data is which they enter on paper. How trustworthy is the accuracy of the data when they have to enter downtime or scrap reasons? How trustworthy is production data if they need to enter that? Is it possible to capture that data from machines automatically? What is the level of comfort that the operators have with technology? Are they open to using an iPad, or an Android tablet with a modern browser to be able to open a web application there? Or is there an opportunity to using an HMI that can be developed with simple screens to show them the data, and help drive some of their decisions for how they operate their machines? What their next tasks are? Are operators open to change?
Level of interest to do better? To work at a better company? To help the company get better by doing the right things? Again, all important questions to ensure the IIoT solution will deliver significant value and fit within how the company operates now and in the near future as operations improve with the solution.
Keep in mind you might be surprised by operators. We've run across any number of manufacturers that don’t have good culture or good relationships with their operators. On the other hand, when their operators are trusted, when the company works with the operators to help them do the right thing they're often pleasantly surprised by how their operators work.
What is the IT network connectivity like on the plant floor? Is there an operational technology Network that potentially is segregated from the IT Network? Is there an existing network, or does a new network within the plant need to be set up? What's the layout of the building such as walls or corners that are a challenge for Wi-Fi? Is there a lot of radio frequency noise from VFD (Variable Frequency Drives)? How reliable will the Wi-Fi connectivity be such that if the Wi-Fi?
Will it be necessary to run ethernet cables to each machine?
Are there cybersecurity requirements for manufacturers and their own customers? The customers might be Tier 1 or Tier 2 suppliers in defense, aerospace, or automotive industries which have some regulations to abide by around cybersecurity.
And then what is the plan to set up the required IT infrastructure which doesn’t exist yet? Does this need to be planned out, down to the point of planning subnets, IP addresses, etc.?
Finally, we can look at the lessons learned. With all of the work that we've discussed above for a pilot project, it looks like there are a lot of details to cover and there are. However if you have a strong team, strong management, and some experience on your side whether it's internal or working with a trusted partner, the projects can generally go smoothly and have success in the end.
There is a lot of press about failed IIoT projects. Having said that, there are reasons for those failures. Those reasons typically around lack of clarity around what the project is going to accomplish, lack of focus on the project itself (prioritizing it relative to other projects), and keeping relevant stakeholders involved. Success comes from everyone having a clear focus of the objectives to attain, a solid plan to achieve those objectives, maintaining focus (along with priorities relative to other projects) on achieving the objectives, and solid leadership.
With those thoughts in mind, I'm going to go through some quick Lessons Learned on our side, and on the side of our customers over time. Looking at these lessons, you might find something that can help you operate your projects more efficiently and have a higher chance of success.
Gathering data from machines is typically a fairly straightforward process, but it does require some work, some thought, and some experience. One of the first things that one should do when looking to pull data from a machine is to look at what possibilities there are for pulling data directly from the machine itself. Don't start with trying to add external sensors when there's even a remote possibility that you might be able to get data directly from the machine’s control system. The reason is that data might be better, more reliable, actually easier to get, and easier to maintain over the long-term versus adding external sensors to the system.
When you look at a machine and it doesn't look like it's possible to connect to that machine or you’re just not really sure how, call the vendor or the manufacturer of the machine. Honestly, you might be charged by that manufacturer or the vendor to open a port or get software tools (software drivers or adapters) that enables the data to be sent to an IIoT platform. Those drivers may cost anywhere from a couple hundred dollars to a couple thousand dollars per machine.
However, even in the short and medium-term, the return on IIoT projects can still be very significant and very fast because the data (when transformed into valuable information) can be incredibly valuable.
Once you've expired that option of the manufacturer not being able to provide any connectivity into the machine then you look at what kind of I/O is available? This takes us back to the situation defined earlier in this article looking for Ethernet, RS-232, etc. If you get to the point where you need to open panels and look for relays, etc. you generally should have the electrical drawing of the machines. A good controls engineer can work with those documents or trace signals to find the right signals to pull from the machine.
Once you have identified the signals to pull then you need to get those signals into your IIoT platform. There are multiple ways to do that. The best approach there is to use Industry 4.0 devices, which provide a variety of open technologies to connect with including REST APIs,
MQTT, etc. When you find proprietary, Industry 3.0 protocols like Modbus, etc. it’s best to move that data to more open, modern protocols on its path to the IIoT platform.
If all of the above doesn't work then finally you look at external sensors. One of the easiest and least complicated to add is a CT (Current Transform) sensor. Add this to the third leg of the power for the machine, and/or the spindle motor, or some similar power source. The idea with a CT sensor is to pull information to indicate whether the machine is off, or if the machine is on and not running, not doing something valuable. Or if the machine is running and doing something valuable- adding value in the lean manufacturing sense of the word. You can also add many other types of sensors that detect temperature, vibration, humidity, etc. With these sensors, the same challenge exists to get the data into the IIoT platform using recommended Industry 4.0 devices, like PLCs, remote I/O, or gateway devices.
Another lesson learned over the years is to make sure that you are focusing on the business challenge first, and not the technology. We have seen a number of situations where companies realize they need to do something to improve operations. Many of those people are engineers and, as a result, they often focus on finding the technologies for the solution first. We’ve even seen some billion-dollar companies get themselves wrapped around the axle by telling the engineers to go forth and “IoTize.” Those engineers then focus so much on the PLCs, sensors, and trying to gather as much data as possible that they never actually get anything done. The business challenge must be clearly and succinctly defined first. Once that’s defined, then you can work down the set of steps outlined above to get to the point of selecting sensors and devices to get the data.
Another idea is that this IIoT pilot and this IIoT effort is not a project but rather a long-term strategy. Becoming a data-driven organization is a new idea. And that data for the data-driven decisions are going to come from where the actual business happens in a manufacturing company- on the plant floor. So connecting to 1 or 2 machines and then potentially connecting to a few other machines to get basic data is great. But it doesn’t stop there. There are a lot of other opportunities to use the data from your operations to improve on-time delivery, production schedules, supply chain, product quality and so much more.
Don’t forget to focus on the idea that you are driving your company toward becoming a
data-centered company. You’ll be focusing on consuming and using data to make decisions. This is something that companies often forget. They will pull data from a machine, monitor it for a little while, put that data up on a big tv screen on a plant wall, then everyone eventually begins to ignore that data. There should be some internal processes to work with that data and use that data to your advantage. This includes team meetings to find issues and solve those challenges by watching the metrics, initiating lean kaizen events, and using the data from your efforts to confirm if those lean efforts have actually improved your operations.
Another lesson learned that we’ve seen with a lot of companies is to be careful of limiting yourself by thinking your ERP system can also be your IIoT system. It can’t. Some ERP systems may in fact allow you to capture some production and scrap data. But there’s a lot more happening on the factory floor than counting parts and counting scrap. ERP systems are
higher-level transactional systems to manage customer orders, inventory, supply chain, jobs, work orders, people, and accounting. It isn’t a product for understanding what’s going on on the factory floor. Understanding what goes on on the factory floor is typically an MES (Manufacturing Execution System), a MOM (Manufacturing Operations Management), or a SCADA (Supervisory Control and Data Acquisition) system. These kinds of systems are a little bit lower down in the ISA-95 levels of manufacturing. What we’re really talking about for manufacturers and what they need to implement more of are MES or MOM systems that are monitoring machines, resources, people and schedules, work orders, recipes, and current activities.
Start with a small project, 1 or 2 machines. Even with 3 or 4 machines is probably too much. Keep it simple. Always keep in mind to be flexible, nimble, agile. You’re going to be learning as a company about what’s going to work in your organizations and what won’t work. Therefore, you’ll need to adjust along the way.
Of course, you don’t want to let projects go sideways and last 5x longer than intended to last. But be ready to be flexible, Within reason of course. The way you measure that level of reason for flexibility is based on the business challenge you are trying to achieve, the approximate cost it’s going to take to solve that business challenge, and the value of that business challenge.
Lastly, within the lessons learned, you must have a strong project manager on your side to manage and drive the project. That project manager should know the situation, the business challenges, and the scope of the project. They should have a certain level of authority within the organization to make sure the work gets done. They should know and understand the plant floor, the people who work and manage the plant floor, have good relationships with them, and be very good communicators. Lastly, they should be available to help run the project. If you’re missing one or more of these attributes in your project manager, the likelihood of project failure increases substantially.
There are many options and ways you can go about starting your first IIoT project. We have a lot of experience in the industry. Hopefully sharing our thoughts and lessons that we learned along the way will help you on your own project whether you do it yourself or work with a trusted partner like us. We have established an immense amount of value in capturing and really utilizing data for multiple manufacturers. Now it’s time for you to take action!
Written by the Ectobox team.
Check out our upcoming webinar with exctobox: How to Start your First IIoT Project!
You may also be interested in reading: