How will the Internet of Things (IoT) actually be implemented in an industrial enterprise? Many of the separate elements have been in place for a couple of decades, so the challenge is to understand what needs to be done to pull them together to function as a whole.
Part 1 of this four-part series defined IoT as a conceptual framework about communicating useful information between devices. Part 2 focused on specific practice and begian with a definition from Tom Moore, IHS Analyst II, Industrial Automation, who says, “IoT is literally collecting sensor data from IP addressable sensors anywhere in the world for anything and transmitting the data.”
There’s little doubt that the Industrial Internet of Things (IIoT) would be ideal for enabling the manufacturing and extraction industries to be more efficient and productive—producing more output with reduced input of time, effort and energy. Over the long term, once the costs of implementation have been recouped, it would reduce operating and production costs.
But like any ideal, the devil is in the details and with IIoT the scale of the details is daunting.
“The difficulty is the sheer amount of data that’s being generated,” says Moore. “You’re looking at terabytes of data from hundreds of sensors and you’ve got to store it and analyze it in order to do what you want with it.”
The challenges to adopting IIoT involve technology, economics and people.
Perhaps the most daunting challenge that must be faced even before addressing the issue of the amount of data, is the physical infrastructure requirement. The infrastructure must enable each automated system within a factory to communicate with every other, with the enterprise network and with the routes to the outside world including the Internet and cellular networks. The best way to implement that is with a system-wide IT network supporting IP-based control and communication. An IP-based network designed to be used in factories is Industrial Ethernet, which has the advantage of being compatible with the Ethernet that long has been used for office networking.
“The networking would ideally be Ethernet,” says Moore, “but industrial automation has a huge legacy of serial-based networking technologies.” The question, then, is how to convert from systems that use one or more serial-based networking technologies to a uniform IP-based system.
Various means exist to achieve that. Cost and difficulty aside, the most straightforward way would be to replace all controllers, sensors and cables that are not Ethernet compatible. Not many companies could or would do that.
Other networking solutions are available that are not as drastic, but all of them have in common the requirement that a physical network must be in place. Unfortunately, that is not always the case.
In the early days of industrial automation, there was no networking. Each device in a system was hard-wired back to a controller, usually a PLC. There would be separate PLC-based systems for different components of the overall process. As Moore points out, because it is not unusual for factory automation systems to have a 30–40-year life cycle, it would take a major investment of time and materials to install just the cabling backbone for an integrated network in such a plant.
A number of IP-based networking solutions could integrate the existing PLC control systems, but not only would they require additional hardware and software, but the network design would likely be complex. The layout and types of cable will have a big impact on system performance. If networks had never been installed, that would mean either controls engineers on staff would have to learn design practices for good networking layouts or the project would have to be left to specialists, perhaps from outside the organization. The physical design requires many decisions; for example, whether there would be a home run from every PLC to a controller, or would it be better to establish a zone strategy? Other issues include the number of nodes, the bandwidth requirements, noise sensitivity and the physical distance of runs.
If a physical network infrastructure already exists, it would most likely have to be converted from a serial fieldbus to an IP-based protocol such as PROFINET or EtherNet IP. This would incur additional costs in hardware and software and require additional expertise. Another issue is whether the installed network could handle the large amounts of data required for Industrial Ethernet. A system of routers, switches, HMIs and computers would have to be introduced, if none existed.
All of this means that a transition to IIoT is an expensive and time-consuming proposition. What’s more, there is also little data available for companies to use in calculating the return on investment (ROI). The question, therefore, is what would compel a company to undertake such a project? At this point, it likely would take visionary leadership willing to make a large capital investment with an uncertain ROI. The management of a medium size company, unburdened by layers of bureaucracy, might be more easily motivated.
Moore agrees, saying “it is often mid-size companies that jump first and it tends to be a visionary CEO or CTO that needs to make the decision.” The paradox, however, is that mid-sized companies are less able to make the capital investment.
“In the short run, this is disruptive technology. It will require changes and will threaten entities that do not have the resources or leadership to make the changes,” wrote Herman Storey, co-chair of the ISA Wireless Automation Standards Committee in the May 2013 issue of Consulting-Specifying Engineer. “The challenges of realizing the benefits of this technology will be more organizational in nature than technical.”
It might be more accurate to say that the challenges are equally financial and organizational. The technology exists—the problem is how to start using it.
Just as the automation architecture changes, so, too, will the way in which it is implemented by the people who must install, use and maintain it. This change in infrastructure must necessarily change the way operations are carried out. There are several potential cultural conflicts. IIoT requires the people in charge of particular processes to take a broad view of how their piece of the operation ties into the even broader, newly integrated system. Process people, controls engineers and IT specialists must learn to communicate with each other if they have not done so previously, and that can be difficult. After all, it is not easy to take someone else’s point of view.
“You’ve got enterprise IT and factory floor IT—there’s always a lot of conflict between them, particularly over the use of IP addresses,” says Moore. The problem is that adding addressable sensors pushes the limits of what’s available—and that limit is generally set by enterprise IT. “Then there’s the problem of cybersecurity,” he says. “Enterprise may have security like McAfee, but that might hinder what the factory floor needs to do and it’s not always well communicated. It’s been Enterprise IT who has the say at the end of the day, but ideally there needs to be a convergence between the two.”
In an article in the Industrial Ethernet Book issue 84/5, Dave Cronberger of Cisco Systems, wrote that “The typical factory involves specialists in areas ranging from robotics, to welding, to conveyor control. These workers are well versed about how to deliver solutions in their individual areas, but they’re typically not trained network experts.”
An article in the March 2010 issue of the Siemens Industry Journal also addressed some of the most pressing organizational issues. “In large organizations,” wrote Rolf Berth, a German psychoanalyst and management consultant, “every single person interested in change is opposed by five other people equally interested in maintaining the status quo.”
There is also a psychological aspect to this phenomenon: people tend to peg their identities to their work roles and change can provoke anxiety. Steve Elwart, PE, director of systems engineering for Ergon Refining, addressed the subject in the October 2014 issue of Control Engineering magazine where he used the analogy of the different stages of mourning after the death of a loved one. “When told about a new control scheme, some people in the control room may say ‘They’ll never do that. They’ll see it isn’t practical.’ At the extreme,” he writes, “there are those who will say, ‘I’m not learning anything new at this stage of life; I’m retiring.’”
Some practical causes exist for workers to resist changing to an integrated automation system. For one thing, the change can be difficult (and not just for older workers) to learn new technology. There are other fears. Any time major changes are made, there is a high probability that there will be initial problems that disrupt production. In this particular case there are many possible sources of difficulty.
Connecting systems that previously had been separate (for example, connecting formerly separate plant with enterprise-level networks) will almost certainly have unforeseen consequences that will require significant time and effort to address. What’s more, connecting to the outside world creates an opening for willful or accidental interference that could be devastating.
Addressing the Challenges
In order to initiate an IoT project, a good economic case must be made. An article in the Automation World Factory and Machine Automation Playbook, offers advice on how to go about that. Writing in the Playbook, Keith Campbell of Hershey’s engineering management team, says “Do it in their language, it’s easier to convince them. You must identify the ‘positive cash flows’ you create with a new system.” Peter Martin of Invensys says that although building the financial case is largely an accounting problem, engineers can contribute to solving it by using existing data from sensors to help establish financial numbers.
In planning for a transition to IoT, it’s also essential to involve people from all departments and levels, from hands-on personnel to management. The planning team should include workers from the factory floor, operations, controls, networking, IT, finance and top management. Fostering a sense of ownership by involving a range of stakeholders will improve the chances that the change will be successful.