Connectivity and Context to Build Smarter Factories. Atos provides IoT and Digital Transformation services to transform modern manufacturing.
Click here to read the complete illustrated article as originally published or scroll down to read the text article.
Atos provides IoT and Digital Transformation services that redefine how people, machines and devices, factory processes, supply chains, and end-users can all interact to transform modern manufacturing. The Internet of Things (IoT) is a real thing. Gartner predicts more than 20 billion things will be connected by 2020. The ability to connect to individual machines as well as networked devices is generating unprecedented quantities of precise data that can help manufacturers work more efficiently and develop smarter factories.But what does IoT really mean in terms of practical application? Most importantly, how exactly does it make manufacturers smarter? And smarter about what?
Atos is a global digital transformation partner for manufacturers to design and implement Industry 4.0, creating value from insights derived from connected assets throughout the product lifecycle, manufacturing process and supply chain. The interaction among machines, products, people, processes, and resources triggers positive actions that lead to higher quality, improved productivity, competitiveness, and growth. But as a number of Atos associates directly engaged in the digital enablement of smart manufacturing point out, there’s a big difference between just connecting things and deriving real value from these connections.
“IoT and Industry 4.0 are buzzwords everyone sort of knows what is meant by, but may not be quite sure what it actually entails,” notes Terrance Bey, Chief Technology Officer, Manufacturing, Retail and Transportation for North America. “What you read is not often the reality. Which is that despite the hype that everyone is implementing IoT, really only about five percent of manufacturers are on the way there.”
So what, then, exactly is IoT? Or, perhaps a better way to put it is, what exactly should IoT mean?
“In the past, you’d collect data, and then make queries about it,” Bey explains. “In IoT, you can make very complex queries against data in real time.”
That’s one key aspect, but IoT is even more than that. Arnaud Langer, Head of IoT and Big Data Platforms for North America, explains, “It’s not just connecting machines, but processes and people as well. If you’re just connecting a machine to server and not anything else, you won’t be successful,” You have to go beyond just collecting data to learning from the data, otherwise, what’s the point? The objective of IoT is better quality, better communications, better everything. Our role at Atos is to help make everything manufacturers do mesh together and realize that objective.”
He adds, “There is no one basic IoT solution. Atos works with partners, as well as our in-house components to provide a customized solution to the customer that also integrates existing legacy systems and machines. While many new machines today come with IoT connectivity built-in, it’s not unusual for us to work with machines that are 20 or 30 years old.”
Nor is it the case that customers necessarily come to Atos and ask for an IoT implementation. “Customers come to us with a problem, and we solve that problems with IoT,” Langer says.
The Manufacturing Data Journey: Realizing IoT
According to Josh Fiorenza, Vice President of North American Operations IoT Practice, the digital transformation of manufacturing is a data journey that focuses the data requirements of the individual line of business (LOB) defined through market use cases. “IoT is about defining business outcomes and insights. IoT cannot be defined by a sensor, a gateway, a data lake, or an analytic platform, but is focused on the customer’s digital dividend. Digital dividend can be described as activating the data across the line of business and enabling smart business services and decisions,” he says. “You need to go past the IT department and develop a roadmap that thinks about other pieces of the business operations beyond the applications on the manufacturing floor.” This data journey involves a five-phase process:
1. Workshop (1-2 days)
This is essentially a one- to two-week brainstorming session to define desired outcomes of an IoT implementation. Success depends on involving all stakeholders. Fiorenza points out, “You want all the subject matter experts (SMEs) and other process experts involved not only because of their knowledge and experience, but for them to become change agents to promote what actions need to be taken as the data journey continues. This, in particular, is when you want to involve seasoned members of the workforce with decades of experience and ask them questions. The purpose is not just to find out information about how they do their jobs, as important as that is, but also to get them invested in how this can help them do their job better.”
2. Incubator (4-6 weeks)
Use case development and prioritization based on the desired outcomes and blueprinted data across the mechanical LOB. You need to create a framework, and then plug in what needs to be accomplished,” Bey says.
“In collecting data, it’s important to look at history,” Langer says. “It’s not just how often does a red light go off on a machine to indicate a problem, but, based on historical trends, can we predict at what intervals the red light is likely to go on, and then what proactive measures can we take to anticipate the problem and keep the red light off.”
Fiorenza adds, “If the data shows we’ve got 192 hours of failures, we know that’s costing the business lost revenue in terms of lost productivity and downtime. What we need to do next is look at what data we can find that tells us how we can predict failures and proactively reduce downtime.”
“Then if you predict a new failure rate, you must look at how this affects other aspects beyond the manufacturing floor itself,” Langer emphasizes. “For example, if our desired outcome is a lower failure rate, what then are the implications for capacity planning?”
3. Proof of Value (3 – 6 months)
Deploy the solution in the LOB to test the value of the desired outcomes defined in the use cases. “This is more than just a pilot proof of concept,” Fiorenza notes. “It’s the actual implementation. Everything is connected, all the dashboarding is completed, but it’s within the LOB. Here’s where we test the decisions made by the SMEs and process experts in the previous phases to see if we can achieve the desired outcomes and the expected results. And probably learn even more based on the data that is generated.” The Proof of Value phase is essential to confirm the business case of the use case and enable the scale out.
4. Scale-Out (6 – 12 months)
Implement the IoT strategy across LOBs, integrate the data into larger corporate ERP and other systems. In 2019, the IoT initiatives will be all about scaling out. This does not necessarily mean a final solution, but one that may lead to it. “We’ve defined benefits derived within an LOB,” Fiorenza explains. “Now we need to redefine these benefits as they affect other LOBs. For example, now that we have new maintenance intervals, we can develop new dashboards that display how far in advance to order parts in anticipation of those new maintenance intervals.”
5. Operations and Evolution Management (1 year +)
This is the last phase, which takes a year or more to complete. “Now we have a mature platform on which to improve decision making,” Fiorenza says. “The IoT strategy is in place, data is being collected and ingested into the analytic platform, and the algorithms are trigging decisions that create business outcomes. So we’ve defined what you might call the ‘new normal’ and now we need to take that training set of data, apply machine learning and AI (artificial intelligence) to make it even more robust.”
Bey adds, “You also need to evaluate your old technologies, which in manufacturing, in particular, can be pretty old. At this stage, you determine what do we actually need to keep and integrate and what you don’t. Then build future testing for additional systems that can accommodate your needs and incorporate new applications.”
“We don’t like to rip and replace an existing customer’s legacy environment, if only because that’s a big one-time cost expenditure,” Fiorenza points out. “Ideally, we want to use 60 to 80 percent of what the customer already implemented, identify integration points and develop the remaining 20 percent of the IoT use case. An additional integration benefit of an end-to-end IoT solution is that our customers are beginning to integrate with other IoT platforms or ERP systems such as OEM suppliers creating a data ecosystem and closing the data loops.”
He adds, “Ultimately, an IoT strategy must be across the end-to-end business process value chain involving the OEM and suppliers from the point of receiving parts, making a finished product, maintaining those products in inventory, sales and distribution, as well as ongoing service and maintenance of those products with end-users.”
Here’s an example where IoT ecosystem proves real value. “If we’re connected to a supplier’s ERP system, we’re now able to capture data points that we never had insights into before. For example, we’re having a failure in the production line of a high tensile yarn manufacturer due to product quality issues,” Fiorenza explains. “As a result, we can now look into our supply chain data and determine that an OEM vendor is delivering solvents that are impacting our finished product. The data within the blueprinted LOB generated proactive alerts allowed us to investigate the SLA (Service Level Agreements) of the OEM vendor. Perhaps you could have stumbled upon the problem without an IoT platform, but it would have taken you a lot longer or never discovered at all.“ And the longer it takes, the more it costs the manufacturer in terms of scrap product due to poor quality and lost productivity.
All these connected processes and people, machines, and departments create enormous quantities of data. Further complicating matters is the need to process mounds of this data in real-time. The solution is the Atos Digital Enablement platform.
“It’s a cloud-based infrastructure that provides flexibility and immense processing power,” Bey explains. Perhaps the best testament to this approach is that Atos uses it as the International Olympic Committee’s (IOC) Worldwide Information Technology Partner to help manage the Olympic and Paralympic Games. As you can imagine, there are a multiplicity of channels to manage, from mobile to online to broadcast involving an array of suppliers and stakeholders.
“It’s a seven-year cycle to plan and implement for each Olympics major event,” Fiorenza explains. “We decided to jettison the ‘build each time’ approach to ‘build once and use many times’. As an example, for the PyeongChang games, Atos moved to a cloud architecture and IP-enabled data center to deliver critical IT services and provide a consistent multichannel customer experience accessible anytime, anyplace, anywhere from our IT backbone operational center in Barcelona, Spain.”
As just one example of how the cloud delivery model optimizes productivity, Atos can now deliver a virtual competition site in four hours, as compared with three weeks to move and test equipment at a physical location. Plus, because it is standardized, it can be applied to other situations with the same ease.
When data goes from residing on devices to residing on the cloud, there is a legitimate fear of risking public exposure. “If it’s really a key concern of our customers, we can still put services on premises and provide end-to-end encryption on every device,” Langer says.
That, however, loses the flexibility and cost effectiveness, not to mention the processing power, of a cloud-based solution. And the reality is that nothing is ever perfectly secure. But as the IoT Security Foundation points out, the vast majority of IoT vulnerabilities can be prevented simply and cost-effectively.
There are four key touchpoints that underpin IoT security:
- Securing sensors and their operations
- Ensuring the confidentiality and integrity of data in transit
- Securing stored data
- Securing access to information.
Atos IoT security services are certified to comply with the most stringent standards, tapping into a powerful network of 13 global data centers and 80 regional centers to ensure continuity and reliability of all critical IoT data sources. In addition, Bey points out, “Atos also partners with the leading experts in cybersecurity to ensure data integrity.”
Proper security goes beyond the technology itself, but also managing the human interface. “It’s not so much about keeping intruders out than it is who you are letting in,” Bey says. “You need to ensure the right users have the right levels of access. Understand the gaps where security breaches could occur and vigorously monitor them.”
Just as data analytics provides predictive capabilities to optimize IoT operations, the same tools provide enhanced security. “We can monitor and look for abnormalities at the device level,” Fiorenza says, “to not only anticipate potential attacks, but take actions to prevent and quarantine attacks before they occur.”
Perhaps one of the biggest challenges to an IoT strategy has nothing to do with technology. It’s connecting people to the value of IoT. “The knee-jerk reaction is that IoT is going to eliminate my job or change it in ways that I won’t like or add tasks to make my job more difficult,” Fiorenza says. “You want to get them out of that kind of valley of despair as soon as possible.” Bey adds that, “It’s tough especially for veterans who been in the manufacturing for 25 or 30 years who don’t necessarily have a software background and are used to focusing on quality processes as opposed to technology, to figure out what they need to do to fit in to the paradigm.”
Consequently, Langer emphasizes, “Change management is a very big part of any IoT implementation. People need to understand that technology is there to help them, not replace them.”
Bey adds that, “IoT isn’t reducing jobs, it’s reusing jobs.” However, new jobs require new skill sets. Providing training to obtain these skills is, consequently, an important part of any change management program.
Without the understanding that technology is intended to benefit workers as well as the business as a whole, employees may not only not only use it poorly, they may outright resist it. An extreme example of that, Fiorenza points out, occurred with truck drivers in Europe. Devices installed in vehicles to track travel were frequently disconnected or otherwise sabotaged. This was because drivers felt the purpose was to monitor and control them and find reasons to fire them as opposed to provide data to improve efficiency and lessen driver fatigue.
To that end, Langer points to the importance of transparency and involvement. “People need to know in advance what is coming with an IoT implementation and how it is going to directly affect them and their jobs. That’s just being fair to your employees. But it’s also important to the success of your IoT strategy. The people who do the work are your best data source. We need to rely on the work process knowledge and experience of employees to design an effective IoT platform and provide them with tools to react smartly in the IoT environment.”
Fiorenza adds, “It’s important to involve stakeholders and SMEs to become change agents. There’s no better advocate for an IoT strategy than the people who benefit from it.”
Perhaps one of the most effective ways to accomplish this is through dashboards that display data in easily understandable graphic formats. “There’s no point in collecting data if people can’t see it,” Langer notes. “Dashboards are therefore key. You can customize them to display information relevant to a specific individual as well as throughout an organization. And it’s up to the individual how deep they need to go into the data. Some might want to look at the underlying algorithm, others just want to see the overall problem.”
“When you can show seasoned worker that a machine they’ve been running anywhere between 350 to 500 Centigrade has a lower threshold of failures if we run the device at 400 to 42.30 Centigrade, that’s when they begin to see the value in IoT,” Fiorenza says. Another example Fiorenza cited above in the yarn manufacturer whose products are used to make bulletproof vests. “Quality and breakage are big concerns,” he says. “You’ll always get resistance to making changes unless you can show the value. When data demonstrates how to take a 60 percent breakage rate and reduce it to 15 percent, you don’t have much resistance anymore.”
Connecting to Customers
Perhaps the ultimate value of an IoT strategy goes beyond the supply chain and the manufacturing floor to the direct connection to customers. “The ultimate aim is to understand consumer patterns. Machine learning helps us understand how best to use the tools to make products. Artificial Intelligence and Digital Twin technologies provide actionable insights into end-consumer behaviors,” Bey says.
“Everything is fine if you’re making the same widget over and over, but it’s not so fine when you need to tinker with the design,” Bey says. “Tools like a blockchain allow for micromanagement of the change process. So, let’s say you need a special insignia for one customer run. You enter the request in blockchain ledger instead of an email chain. So now it’s easier to find any breaks or flaws in the change process, you’ve got more accountability, and more accessible information about the change process to better satisfy the next end-user request for customization.”
Indeed, a Digital Enablement Platform empowers even individual consumers to make change requests directly. “You’re a car maker. A customer places an order online for one of your cars but wants a specialized matte-finish. That change can be communicated from a web page and implemented in real-time on the production floor to fill the order,” Bey notes.
There’s also the potential to move production from the manufacturer directly to the consumer. “That’s where we’re headed with 3D printing technology,” Bey says. “For example, a shoemaker introduces a special style of shoe that’s only going to be available for a short time and only so many are going to be made available. Instead of doing a limited production run and maintaining inventory in expectation of eventually selling out, you provide instructions to a third-party retailer. A customer gets a barcode online, brings it to the store, and the product is made right there.”
Which is why, according to Bey, “The idea of IoT is not just to reduce production costs and be more efficient—though it certainly is that—but also better determine what individual consumers want and then rejigger processes to cost-efficiently provide it and be able to charge a premium for a specialized product.”
Moving forward into 2019 and beyond, Atos is focusing on helping clients scale out their IoT strategies. “A lot of enterprises are stuck in pilot purgatory. We want to help move them along the roadmap and get more out of their digital journey to realize the broader benefits of IoT”
To that end, Atos is looking to develop more turnkey solutions. “There are a lot of IoT startups that provide small pieces of the puzzle. We’re seeking to partner and integrate them into the Atos framework. As a large systems integrator of IoT technologies, Atos can take what we already know and build a platform that can be implemented quickly.