How to Make Sense of Machine Data in Smart Factories

Although the majority of companies know about IoT, not many understand the value they can extract from machine data right now; it’s all about the communication ability between the shop floor and the business environment.

Thomas Berndorfer -
Factory buildings and wrenches
Illustration: © IoT For All

“Digital transformation” and “smart factory” are terms that have been around for a while, but only advanced manufacturers apply them to practice. The idea of machine-to-machine communication has raised initial excitement and high hopes, but now owners of mid-size and small factories are asking, “What’s next? How do I start adopting these technologies in my factory?”

While over 98 percent of business leaders understand the concept of the Internet of Things (IoT), many are not sure what IoT means for their business, according to research by Aruba Networks. This results in IoT not being implemented properly or being used only at a basic level instead of as a fully transforming process.

However, when we look at production plants that have grasped the essence of the Industrial Internet of Things (IIoT), it really boils down to the communication ability between the shop floor and business environment. In practice, it covers how machine data is transferred from equipment to standard business applications in the back office: ERP, MES, and CRM systems, and what conclusions are drawn from these data and how they are applied back to the production process. Overall, the success of digital transformation depends on the capability of an organization to make sense of numbers from the machines and to turn them into a competitive advantage.

Emerging Challenges for Business

Charles Kettering, the famed inventor and head of research for GM, said, “A problem well-stated is half-solved.” The winding road to a smart factory poses several emerging and existing challenges for companies, and manufacturers need to understand how to run a digital business model at their production facilities.

Why Do You Need a Smart Factory?

First and foremost, machines can generate a lot of data but throwing numbers to the cloud does not automatically mean you will obtain useful information from it. You need to have a context.

Factory equipment can provide the most accurate real-time information about production processes, but how can people start taking advantage of it and apply the knowledge to their day-to-day job?

Is it going to be about maintenance and about making sure that every part is up and running? Or is it about so-called predictive maintenance: resolving the breakdown before it occurs based on the aggregate data from many similar devices and early signs of the problem?

It can also be about the business side of the production: human operators making decisions based on the reports from the machines in the MES and ERP systems. Embracing Industry 4.0 dictates not only an equipment upgrade but an overall thinking paradigm shift.

Connecting Across Vertical Silos

For a factory to avoid being “islands of things” but to work along with the concept of the Internet of Things, production machines need to talk to each other and eventually to edge devices. Given the many systems involved, the question is, how do you set up such communication?

One of the challenges at a modern factory is that machines often communicate within individual vertical silos, transferring data from a device to HMI/SCADA, etc., where an operator keeps an eye at this particular part of the production. Managers have long known that valuable information resides in the systems that run their production facilities, but they don’t have direct access to it. The problem has often been that both worlds operate in isolation, with little or no connection between them. However, if we are talking about blending data and carrying it across various departments, the factory data exchange infrastructure should be built differently.

Data Travel to ERP

The next concern is data connectivity. How far down do you need to see in your production process? Is it only controllers, or do you actually need the data from each sensor, etc.? This will largely influence the amount of information you operate and the picture of your facility, including the whole plant in general, the individual production lines, and the particular devices being used.

Putting data from the physical world into the cyber world requires the planning of analytics. You need to understand if it’s an operator who will use the data for achieving better local control, or if the server automatically sends aggregate information up to the ERP or MES system. All in all, which numbers from the avalanche of data need to be sent forward for better analytics and planning at the factory level, and which data stays local for quick control and decision making.

Sharing data across divisions and third parties is another important issue. At advanced factories, an ERP system can automatically order raw materials or inform the maintenance service about the need for a check-up, based on notifications coming from the production facilities. This information exchange can also work for the benefit of machine builders: based on automatic reports arriving from various factories, they plan maintenance, substitute details in advance, and provide better customer service, saving clients from annoying and expensive downtime.

Machine Data Security

Finally, the ultimate issue for the “smart factory” is ensuring the security of open communication. Security occupies the first position among IoT developers’ concerns (31 percent), according to the 2019 report by Eclipse Foundation, with connectivity (21 percent) and data collection and analytics (19 percent) taking only the next two spots. Gartner reports that spending on IoT endpoint security solutions will increase to $631M in 2021. Opening up production facilities to the world requires corresponding protection measures.

From OPC to OPC Unified Architecture

Since the late 1990s, production automation has been built around the OPC server, enabling data travel from a device to an operator’s computer in the control room whether the hardware and software were from the same provider or not. The OPC protocol standardized the communication and preconditioned a breakthrough in manufacture automation. OPC applied the Microsoft Distributed Component Object Model (DCOM), which is part of the Windows operating system. In the early 2000s, the protocol launched industrial automation and enabled factories to transition to safer, more accurate and cost-effective production. In fact, many production plants around the world still rely on this technology.

However, soon after the software entered the market, the disadvantages of the OPC protocol became clear. Firstly, DCOM makes OPC limited only to the Windows operating system. Next, experts complain about the difficulties in finding and fixing software failures, and the impossibility of exchanging data over the web. And finally, OPC technology is not reliable in terms of information security.

This is why, in response to multiple requests from the industry, OPC Foundation created a new standard: OPC UA. It’s based on the previous versions but without drawbacks, and it suggests a totally new approach to machine-to-machine communication and industrial digital transformation.

Its new features include the following:

  • Platform independence. OPC UA doesn’t rely on OLE or DCOM technology from Microsoft anymore. You can launch it on any platform: Apple, Linux (JAVA) or Windows.
  • Easier configuration of data exchange between multiple computers. Instead of DCOM, use common data-exchange standards: TCP/IP, HTTP(S), and SOAP.
  • Easier data governance and maintenance. You can use structures or models, and group and give context to data tags or points. All this provides much more flexibility and makes it possible to connect your factory across the departments and not work only in vertical data silos.
  • Easy integration to SCADA. An OPC UA server can be a direct part of a PLC, which is built into a device. There is no need to buy and configure an OPC UA server.
  • Resource saving. With OPC UA, you don’t need many PCs that sit in the facility and translate between the systems.
  • Data encryption. OPC UA security is based on a multilayer concept covering application and transport layers. It contains authentication and authorization as well as encryption and data integrity by signing.
  • Flexibility of integration. UA APIs are available in several programming languages. Commercial SDKs are available for C, C++, Java, and .NET. Open-source stacks are available at least for C, C++, Java, Javascript (node) and Python.
  • Cost- and time-effectiveness. Availability of OPC UA connectors integrate your real-time machine data to CRM, ERP, BI, Exchange, and document management systems like SharePoint for improved planning, maintenance, and boosted effectiveness of your production.
  • Connection to legacy infrastructure. Availability of mechanisms for migration from OPC to OPC UA; OPC UA wrappers.

How to Start Using Machine Data for Digital Transformation

Manufacturers call OPC Unified Architecture the first communication protocol that removes borders between the enterprise and the factory floor. But, how do you actually connect your production to the systems that help manage your company?

Customer relationship management (CRM), enterprise resource planning (ERP), an manufacturing execution system (MES), a document management system (DMS), a business intelligence (BI) system, and a warehouse management system are the most common software that assists in managing a business. As many companies, regardless of size, use these applications in various combinations, it’s safe to generalize that the life of a factory is planned, analyzed, predicted, and reported in these systems. Being a fully automated smart factory means having accurate up-to-date information about the production process available anytime in these systems for analysis, in-time alerts, maintenance notifications, etc., depending on your priorities.

Integrating machine data in business systems can happen in two ways: 1) via custom development through OPC UA APIs, and 2) via a ready-made OPC UA connector that links machine data to ERP, CRM systems, etc.

In our opinion, the second scenario is better than custom programming. Firstly, you don’t waste time and money on an expensive programmer’s labor. Secondly, you don’t need to take care of further maintenance. A pre-built OPC UA connector is typically bug-free and ready to use from day one. It does require programming skills but very basic ones. Two-three hours are needed to understand the integration principles and to build the logic for connecting your enterprise’s machine data to business systems.

From the Device to the Enterprise

Why do you need to connect your machine data to ERP, CRM, and MES? An OPC-UA slogan is “from the device to the enterprise,” and in this case, “enterprise” means an intelligent enterprise resource planning system.

To make smart decisions, an intelligent ERP requires real-time IIoT data and analytics. They differentiate a company from the competition and constitute what is called a smart factory.

As ERP applications fulfill so many tasks—order management and processing, data collection, planning on an enterprise production level, business process management overall, and hooking up with other systems—the timely performance of intelligent ERPs provides a lot of advantages.

When you need to speed up processes, reduce manual labor and errors, and move faster between steps, grabbing real-time data as close as possible to the source, which can be from a particular production line or even a device, an integration of machine data via OPC UA to the ERP system is revolutionary.

Besides, such integration helps the culture shift we discussed above, as employees adopt the new technology more willingly if it’s presented via a familiar interface (and even if it’s mobile-friendly!). Ensuring user adoption remains a big challenge in IIoT, and presenting the data via the systems everyone knows and uses—ERP, CRM, MES. etc.—speeds up the transition to Industry 4.0.

Other components of an IIoT smart factory are predictive analytics, machine learning, blockchain report sealing, and semi-autonomous and autonomous decisions and actions. All these can be a part of your ERP system, which orders materials or takes shipment decisions based on multiple sources and huge amounts of contextual fast data, analyzed and leveraged in real-time.

Conclusion

IIoT has already turned into a demand for modern factories. To keep up with the competition, connecting your machine data to the management business systems is not optional anymore. OPC UA has been doing a good job in this area, providing many more opportunities and security measures to open communication in a production facility, so it’s time to reap the benefits of a connected smart factory.

Author
Thomas Berndorfer - CEO of Connecting Software, https://www.connecting-software.com/

Contributors
Guest Writer
Guest Writer
Guest writers are IoT experts and enthusiasts interested in sharing their insights with the IoT industry through IoT For All.
Guest writers are IoT experts and enthusiasts interested in sharing their insights with the IoT industry through IoT For All.