New Year’s resolution: You want to improve your process, cut downtime, increase efficiency, better your product quality and make your food even safer. Who doesn’t? And today, there are all kinds of technology you can throw at these “wants,” but if you’re just flinging technology at processes and equipment without a clear goal in sight, real solutions could be elusive.

Problem is, unless your plant is new and has a single vendor providing a complete turnkey process/packaging/palletizing solution with a consistent communications infrastructure, you probably have a smattering of hardware, equipment and software scattered throughout your facility. So you’d like to get some of the existing islands of automation and equipment “talking” to each other and providing you with some useful information.

To achieve your goals, you’ve probably heard that putting together a sensor-based IIoT network may cure all your communication ills and lack-of-actionable information. Not many years ago—instead of a fancy catch phrase—we would have simply called this a “controls upgrade,” which would have used some form of proprietary industrial networking. 

In addition, wireless sensors (often associated with IIoT) are anything but new, but fortunately many wireless sensors today can be had for a fraction of the cost of a dedicated industrial wireless networked sensor—because the former often borrow on existing commercial wireless Ethernet or cellular technologies. However, they may not offer some of the technologies that their more expensive industrial counterparts offer—such as self diagnostics, mesh networking, security and industrial protocols like WirelessHART.

So, how do you implement a sensor-based IIoT system that can fulfill your needs now and into the future? Do you install it alongside your present automation systems? Should it be cloud based? Do you integrate it into your existing system(s)? Before answering those questions you need to ask: What do you want to monitor and accomplish—and why? 

I asked system integrators, sensor and controller manufacturers, parts makers and system software providers to help us answer some basic questions, keeping in mind that key to any IIoT solution—same as a controls upgrade—is to create a system that can be easily maintained into the future without becoming obsolete.

Challenges of existing systems

If operators are continuously finagling with equipment to control a process, the reasons could be several, but two general categories crop up: the equipment and software weren’t designed to handle the control needed today—and/or there’s a maintenance issue with one or more components. Either way, more data is needed, which usually means more sensors, up-to-date controllers that can process the data, historians/HMIs and analytics that can deliver useful or actionable information. Is a sensor-based IIoT system up to the task?

“We have seen users with older systems that simply cannot provide the information they now want, as well as adequate systems that were not configured or set up to provide the needed information,” says Steve Pflantz, CRB associate/senior application engineer. The latter occurs when not enough focus is placed on the vendor’s design, and the client receives a solution that is not quite aligned with expectations. 

It comes down to whether the system is designed with the right instrumentation to measure or sense something, and then if the system is configured correctly to give the user the data in a usable format. The issue is that once a piece of equipment is built and installed, it becomes more costly and challenging to make modifications, says Pflantz. “It is best to design in what you need. If that does not happen, users are forced to deal with added cost and downtime of the equipment to make the changes.” 

Aging equipment can pose other problems, either unintentionally or intentionally. What do I mean by unintentional? Michael McEnery, PE, president of McEnery Automation, a Control System Integrators Association (CSIA) certified systems integrator (SI), explains: “We see legacy systems or control hardware that do not match the plant standards.” These are difficult or impossible to modify by plant personnel and as a result, accessing valid and useful production data is limited. If the control system has networking capability, a software driver can be found to communicate with it, but PLC or HMI program reviews or modifications are typically required to format the data into usable information. This can be prohibitively expensive, especially on older controllers or systems with poor or outdated documentation.

But, intentional lock-outs can be costly as well. Kevin McClusky, Inductive Automation co-director of sales engineering, explains: “Many companies struggle to access data from existing OEM equipment for a variety of reasons.” The information could be locked down from external access, there could be networking issues or lack of infrastructure, or there may be restricted access until a fee is paid. For these systems, often the only information accessible is on the screen, and operators may end up using clipboards to collect the data. 

“Many of the challenges we’ve seen start with the ability to connect existing control systems and legacy devices,” says Pete Atkin, GM & VP for Samsara Industrial, a CSIA member. There are a lot of powerful solutions for processing your data today, but all the computing power in the world doesn’t help if you can’t get your operational data off your assets first. Simple, cost-effective ways of easily connecting to these devices are critical for obtaining actionable data.

“Aside from this, one of the biggest concerns customers have with IIoT is ending up with even more data silos,” adds Atkin. Being able to deploy open platforms with modern APIs is critical to ensure data can be used across platforms. If food and beverage processors can do this, they can get very useful information fed directly into their top line business systems and start using that data to streamline and drive efficiency across their operations.

Better data helps both process and maintenance 

On one hand, operational data has been used to improve plant productivity, says Takayuki Sugizaki, manager, Yokogawa Electric Corporation. This is already achieved by collecting data from process sensors using both wired and wireless factory networks. On the other hand, equipment data for maintenance has been used to improve the efficiency of equipment maintenance. In the past, most equipment data were not online, quantified or visualized, but measured manually by operator rounds.

“Whether you manufacture beer, wine or whiskey, the No. 1 concern as a beverage processor is to ensure that your fermentation is made at an optimum quality level and is also easily repeated,” says Anne-Sophie Kedad, Emerson/ASCO director of marketing—process industries. Generally speaking, the more automation a processor uses, the better the repeatability to ensure each brand maintains its quality. Automated processes coupled with sensors and electronic diagnostics make the operator’s life easier in dealing with maintenance issues. “In the IIoT world, you can analyze that the time needed to chill your recipe is taking longer, so having this data can help the beverage processor determine which components need maintenance before they fail,” says Kedad.

Some users with “legacy” instruments such as HART protocols may be closer to IIoT than they think. Ola Wesstrom, Endress+Hauser food and beverage industry manager, suggests there are both technical and mindset reasons why processors may be lacking the operational data they want. “Technical can be as simple as not realizing that the I/O in many PLCs is HART-enabled, allowing the PLC to recognize digital HART signals from instruments. Older systems may require upgrades of older PLCs to add this capability.”

Wesstrom says the mindset reason is that often additional information from the instruments is not even considered when looking at how IIoT can support more efficient manufacturing. “The most common approach is data aggregation of common process variables, thus leaving out the available, and valuable, information.”

Ask basic questions first

This is where operations and senior management need to sit down and whiteboard what it is they want to accomplish together. They need to establish some priorities. Do they want to improve maintenance and cut downtime? Do they want to improve a process that produces inconsistent product? Are there problems on a particular line, e.g., bottlenecks or jams? How ready are they for implementing such a system? Should they contact suppliers or system integrators?

“My recommendation would be that if whomever you work with to assess an IIoT implementation does not have a readiness appraisal procedure that they can and will share with you, find another resource,” says Dan Harder, director of business development at Malisko Engineering, a CSIA Certified Member. “There are some companies trying to capitalize on the excitement of the IIoT wave and are selling and installing systems that will not live up to the expectations of the customer.”

The very first thing is to decide what a processor hopes to accomplish with IIoT, says Robert Miller, senior product marketing manager for Mitsubishi Electric Automation. Identify a problem that needs to be solved, say downtime on a specific machine or line. What machine, what line, etc? Then identify the deliverables; what is the target? Should the system deliver 10% more uptime; what does that relate to in terms of cost benefit? The single most important thing is identifying the specific need. Work with suppliers and integrators to define the project specifics.

While coming up with a list of questions that can benefit every processor identically is next to impossible, McEnery provides a good punch list that can be found in the sidebar, “Key questions to implementing a sensor-based IIoT.” Whatever system is chosen, keep in mind that someone will need to be trained to support it and any software, says McEnery.

Whether a processor wants to accomplish a digital transformation of the business or just gather data to monitor a specific metric, generate a report or analyze a machine or line, processors will need to check if data can be easily gathered from existing PLCs, devices or connected control systems, says Inductive Automation’s McClusky. If it’s not immediately available, installing an IIoT-based sensor network may be the best option.

If a processor should decide to look into an IIoT-based sensor network, McClusky suggests doing the following first:

  • Make sure that adding new wireless technologies doesn’t negatively impact available bandwidth for existing wireless communications.
  • Ensure the system has store and forward, so you can weather network blips without losing data.
  • Make sure the chosen hardware and software uses standards-based communication protocols to future-proof the system.
  • Discuss with the appropriate parties to decide who will maintain the network.

While a processor may be satisfied with product quality and food safety, maintenance and uptime are key considerations, and can be addressed by a bolt-on sensor-based IIoT system. “Companies should begin by evaluating how connected devices (portable and permanently installed) could help with condition monitoring, which is the most cost-effective way to manage assets,” says James Henry, SKF USA strategic marketing, rotating equipment performance digitalization. Unplanned downtime is seven to 10 times more expensive than planned maintenance. Manufacturers can begin to reduce unplanned downtime by using these connected devices that are selected based on the needs of each specific application, depending on the criticality of the equipment. Site assessments also help establish a baseline for equipment performance. Having this historical data allows companies to create KPIs for future improvements and to document ROI, says Henry.

Improve existing systems or use an IIoT overlay?

Some motors today may have built-in vibration sensors to track bearing wear, but the idea of a bolt-on sensor for motors goes back a few years. For example, SKF has a complete retrofit sensor system (SKF Pulse) for motors, which can monitor bearings and relay the data to SKF intelligence via the cloud, which then in turn, lets users know how much life is left on a bearing. While motor bearing monitoring is a predictive maintenance (PdM) necessity today, this is one of the earlier applications of sensor-based IIoT as an overlay or bolt-on system—or whatever you’d like to call it.

Does it make sense to use a bolt-on or overlay system? “From a perspective of being as minimally disruptive as possible, this is a sound strategy,” says Malisko’s Harder. The recommendation could be to address the sensors on the processes or devices that the client is most concerned about and start there. You can then build the rest of the network out as time and resources are available.

Adding IIoT sensors throughout the process can be an economical way to gather additional information from the plant floor, says Carrie Lee, Omron Automation Americas product manager - controllers. However, it’s often easier to leverage existing plant networks when possible. Using an IIoT system parallel to existing controls is advantageous because it allows food and beverage processors to gather more data without disrupting working systems. This can also be beneficial in adding a layer of redundancy. Potential disadvantages would be the need to add another channel to configure, process, and share data, adds Lee.

Should you use wired or wireless sensors with IIoT? “That would depend on the architecture,” says Endress+Hauser’s Wesstrom. If digital-enabled I/O is available, then a wired approach is proper. If not, wireless edge devices become a viable option for core data points. IIoT networks can be on premises or cloud based; it really comes down to company policy, says Wesstrom. “We think cloud-based is the way to go for this type of information. Keep in mind that it is data aggregation, and the cloud based systems do not allow anyone to hack into and change things when set up correctly, and the data itself would be useless to someone from outside, not understanding the context.”

“We have seen new sensors for vibration and temperature monitoring of motors be implemented with wireless transmitters to new software applications specifically designed for monitoring these devices and passing this data to the process historian, whether it is on premises or in the cloud,” says McEnry. These sensors are not duplicating existing sensors in the control system. The location of the data repository does not really affect the decision whether to integrate into existing control systems or to bypass them with new IIoT systems.

Another example is with pneumatics. An IIoT edge computing gateway breaks down end-of-stroke sensor signals and evaluates pneumatic cylinder deceleration, says Emerson’s Kedad. It then analyzes the data internally and sends actionable information to specific people, the cloud, or to the parent MES or ERP systems via the OPC UA interface.

“Many new smart sensors and actuators have IIoT capabilities built in, but that doesn’t mean replacing your ‘old-school’ devices or adding new and more costly ones are the only two options,” says Bill Dehner, AutomationDirect technical marketer. IIoT bridges like the AutomationDirect STRIDE Pocket Portal can gather data from existing devices via discrete or analog signals, or a Modbus connection, and send it directly to the cloud, allowing users to create their own DIY cloud-based data solution.

IIoT sensors could be a cheaper option, but may provide less data than connecting to the existing equipment, says Inductive Automation’s McClusky. Take a close look at your goals, as newer analytics technologies like machine learning/artificial intelligence often perform much better with more available data. 

“One approach we’ve seen a large number of companies taking is connecting IIoT data collection devices directly to PLCs rather than installing new sensors, allowing for the advantages that IIoT communication protocols provide (low bandwidth, massive data collection, easy communication over a local network or WiFi/cell/satellite, centralized data access), while reading the sensors that are already installed on the system,” says McClusky. “IIoT devices can either be connected to the cloud or on premises, depending on user preferences.”

“It very well may be more practical and economical to add on sensors and control components as a separate system and leave the existing system as is,” says CRB’s Pflantz. That is not always the case, and an experienced automation professional should be able to help decide which way to go: bolt on all new stuff or incorporate it into the existing controls. “The older your existing system, or less standardized, the more likely a separate system is your best choice,” says Pflantz.

“Whatever you add could be connected to a SCADA or MES system if planned and designed accordingly,” Pflantz adds. “I would envision some instruments, I/O, maybe a controller that goes to some PC-based SCADA application serving as a historian or reporting tool, and then that data being passed to the IT world and up to a MES if desired. “ 

The caveat here is to proceed with caution, says Pflantz. Adding on a separate solution parallel to the existing controls can be done, but you are creating parallel systems that do things for the same equipment or process, and further down the road, can become a control system management nightmare. Careful what you build as you will have to maintain it. Understand your control system and data system roadmap and make sure what you do now will not come back and cause you more problems down the road. Also, Pflantz suggests that when processors use the cloud, they should implement best practices for the connection and have a sound cybersecurity system in place.

Plan carefully

As we’ve seen, installing a sensor-based IIoT system can be a good way to gain intelligence about a process or maintenance issue, and the cost can be less than a full-fledged control system. You can start out with a small system that you can move around from equipment to equipment or line to line to monitor a problem without making the installation permanent. The best advice may be in knowing what you want to accomplish before you commit to a project. Suppliers, engineering houses and system integrators stand ready to help.

Key questions to implementing sensor-based IIoT

If you’re planning a sensor-based IIoT system, it might be a good idea to sit down with a unbiased third-party like a system integrator who can get you on the right track in evaluating such an investment. When planning a new project, be sure to consider the following:

  • What are the most important pieces of information that are needed by the plant and why?
  • Are these data points available from existing control platforms, and if so, what is the budgetary estimated cost per data point to integrate them into a process historian directly through the control system architecture?
  • What is the estimated cost per data point for installing a duplicate instrument or sensor, including maintenance costs?
  • For data points not available, are there any plans to upgrade the existing control platform? What are other costs of maintaining an outdated control system vs. the benefits of a modern control system?
  • What is the desired time frame for implementation? Does it allow for a multi-year replacement of existing control systems?
  • Is the data coming from new instruments? Could these be installed wirelessly? Does a wireless network exist in all the areas of the plant where the new wireless devices would be installed?
  • What will be the ongoing plan to support any new software installed for an IIoT system?
  • Who will be trained to support any new software? Will it be a plant or corporate resource? Will an outside resource be required?

—Michael McEnery, PE

For more information:
Automation Direct,
Inductive Automation,
Malisko Engineering,
McEnery Automation,
Mitsubishi Electric Automation, Inc.,
Omron Automation Americas,
Samsara Industrial,
Yokogawa Electric Corporation,