Image in modal.

The Industrial Internet of Things (IIoT) is a popular buzzword these days, and with good reason: The inclusion of IIoT technologies on processing equipment can bring increased uptime by predicting impending equipment failures or maintenance needs. Integrating IIoT technologies adds the capability to diagnose issues in real time with access to cloud data. This allows problems to be remedied during scheduled downtime.

As IIoT proliferates, original equipment manufacturers (OEMs) would like to successfully launch IIoT in their business and offer a level of connectivity in their equipment. Being aware of underlying maintenance or control issues that could cause an equipment shutdown can provide the foresight to allow them to:

  • Order spare parts.
  • Schedule maintenance.
  • Adjust the control settings and algorithms to prevent serious downtime.

Adding IIoT technology just for the sake of getting on the bandwagon, however, may not deliver on the promises that IoT could bring to industrial equipment.

To understand how OEMs are able to deliver predictive failure and remote diagnostics, it is important to first understand the barriers that previously prevented this type of capability in equipment.

In the past, many hurdles limited the ability to record and retrieve the information needed to diagnose equipment in the field. The first hurdle was retrieving data already stored on the machine — on a PLC or within datalogging hardware or software. This required either sending personnel to the equipment, having the end user download the data or using a virtual private network (VPN) to remotely access the data. Getting the necessary information via these methods can be time-consuming and frustrating. In addition, while a VPN can provide access to a few individuals, it can create vulnerabilities in the end-user or OEM network. For this reason, many companies prohibit an outside vendor or computer from gaining access via a traditional VPN.

The next hurdle was determining the type and amount of data that would be relevant and useful to monitor the equipment health in the field. Historically, the type of sensors, data formats, storage capacity and connectivity of the controller or PLC have limited the amount of remote diagnosis that can be accomplished. It required the engineers and programmers to predict ahead of time all of the necessary information and algorithms that would be used to diagnose issues automatically or remotely. Sensors were expensive and could be unreliable, which made it difficult to add sensors and determine what information they could provide. This, in turn, limited the number of sensors being added to the equipment. Combined with the limited storage capacity of many industrial controls, these conditions limited remote diagnostic capabilities.


PH 1121 Wisconsin Oven IIoT Dashboard Notext

Interactive dashboards display current operational data for the operators as well as OEM technical and engineering staff and facilitate real-time collaboration. The dashboards allow the OEM engineers and the end-user operators to review historical information. (The data shown was selected at a specific time in the past.) Noted at the top of the industrial furnace is N1, a motor sensor that can display temperature and vibration data for real-time analysis. Photos credit: Wisconsin Oven (Click on image to enlarge.)


With the introduction of supervisory control and data acquisition (SCADA) systems, the centralization of control and information inside a processing plant was possible. This gave operators the ability to monitor and control the process internally. It was a great leap forward for engineers and technicians within the plant: They learned about their processes and were able to coordinate the control of all of the equipment used in the processes. Using centralized process control, they could collect data that allowed them to improve those same processes.

While SCADA improved how a plant and process were operated, it had its own limitations, however. SCADA recorded and displayed information that was pertinent to the process yet only lent access to the individuals that used the equipment. Recorded data only provided insight into the process. Additionally, it was rare to have an engineer or technician working at the plant that could identify which information could be used to diagnose and predict future failures of the many types of equipment in the plant. Equipment manufacturers typically could not gain access to the data — or more relevant data on the machine itself — so they had no incentive to add sensors to the equipment unless it was needed for alarms or control.

With the advent of widespread access to IoT gateways and secure cloud service providers (CSP) in the consumer and commercial markets, a new opportunity arose that enabled access to extremely valuable data. There were challenges, however, in the different industrial communication protocols and the lack of connectivity on the plant floor.

A few IoT providers recognized that distinction, and a new branch, called the Industrial Internet of Things, created a place for companies that could specialize in this emerging technology. To fully take advantage of the IIoT technologies, the first step was to move the plant SCADA system to the cloud. Doing so proved to be extremely successful in improving the plant process. One drawback, however, was that this typically only provided access (and benefits) for the end user of the equipment, and to those companies that had internal experts on the equipment. The knowledge and innovation rarely propagated back to the engineers and technicians who manufactured and supported the equipment.

Some OEMs of industrial components and equipment began to offer more sensors and cloud access for end users. This was the next logical step in the path to remote diagnostics and the development of predictive failure algorithms. It seemed a quick and easy way to provide access to the information collected and stored on the machine by moving that access to the cloud. No longer would end users be required to hardwire and navigate the myriad of communication protocols and the esoteric register lists for the data they wanted on the machine. Instead, OEMs could provide access to the cloud database and create dashboards with the KPIs that were important to the end users. But again, there still existed barriers to real remote diagnostics and failure prediction.


PH 1121 Wisconsin Oven IIoT Predictive Failure

Dashboards allow OEM service technicians to review equipment data before going to the process plant. With this system, the heatup rate profile had changed significantly, and the heat output for the temperature had increased greatly. This is a classic example of a progressive failure of a component. While the equipment had not failed and was maintaining its process setpoint without warnings, the data predicted failure. While performing routine maintenance, the technician was able to correct the issue and avoid future downtime. Photos credit: Wisconsin Oven (Click on image to enlarge.)


Today, equipment manufacturers are adding this level of data acquisition and cloud access to their products and providing access to the data — not only to the end user but also to their own engineers and service technicians.

By allowing their engineers to add sensors on the equipment in-house and in the field, these OEMs have created a new wave of innovation. Inexpensive sensors and data-rich components are being added to equipment so engineering teams can analyze the data. This creates a cycle:

  • Additional sensors gather and provide more data.
  • Additional information allows the engineering teams to detect problems and impending failures.
  • Additional information also allows the teams to analyze data in the field to offer cost-effective remote diagnostics.
  • Having a great understanding, brought about from the additional data, allows manufacturers to embed predictive failure algorithms into the equipment.

By combining data from more intelligent components, engineers and technicians have access to a plethora of data and are no longer limited to adding sensors and I/O modules for every new piece of data. For example, a new, relatively inexpensive vibration sensor can bring in hundreds of data points and built-in algorithms. Industrial PID controllers, PLCs and industrial control components can pass on thousands of different data points that may be useful for diagnostic purposes.

The key to success is to continue to efficiently determine which sensors and what data can be used to create algorithms that are capable of predicting failure. This requires intimate knowledge of what a propagating failure looks like and requires combining that knowledge with the ability to create and add algorithms to the equipment. Equipment manufacturers successful in IIoT and predictive-failure analysis encourage and enable coordination of their suppliers, engineers and service technicians.

One last crucial component is for these OEMs to collaborate with the end users of the equipment. At first glance, it is easy for a processor to fear giving access to equipment data to the OEM of that equipment. It may seem like a level of privacy is being lost; however, the OEM already knows how the equipment is supposed to operate in the process. In order for plant operators to be successful, they would have had to purchase equipment that is already designed specifically for that process, or they would have had to convey that important information in the development of customized equipment.

What actually happens by allowing IIoT and cloud data access to the OEM is that more privacy is achieved. The OEM can now only see its equipment operating. Previously, without this feature, a processor would have to allow OEM engineers or service technicians into the plant, where they would see proprietary processes as well as all of the manufacturing equipment, the chemistry and recipes, and the products being processed. This creates a privacy situation that may allow information to be communicated to competitors inadvertently. By contrast, allowing OEMs to access IIoT data for their equipment ensures that processors have more control over the privacy of the process and products while increasing productivity and reducing downtime.