It’s hard to read an industry blog or website without seeing bold statements about the industrial internet of things (IIoT). Much of what you read is loaded with eye-popping statistics, industry buzzwords, and complex diagrams. The end result for the reader can be an overwhelming sense of confusion and lack of direction as to how to begin.
Chances are your company already has IIoT in use: Manufacturing industries have been implementing IoT concepts for more than 20 years. For example, the HART communication protocol was developed in the early 1990s to provide digital data from smart sensors previously providing one data point. The estimated installed base of HART devices is more than 40 million. Other open standards such as EtherNet/IP and Foundation Fieldbus developed in the late 1990s enabled placing sensors directly on networks.
So don’t be overwhelmed worrying about the right way or wrong way. Let’s cut through the fog of IIoT and focus on implementation advice for those beginning their journey.
1. Find a user first. “If you build it, they will come” might work for Iowa farmers chasing baseball ghosts, but don’t count on it for success with your IIoT project. Before you start, find end users who are committed to using the data and who will champion your project by assisting with funding, implementing procedures to use the data, and informing management of the final results.
2. Common examples of IIoT in use. Applications that can benefit from the use of IoT tools and strategies include:
- Preventive maintenance – provide motor run time or equipment cycles.
- Predictive maintenance – provide warnings for conditions that indicate impending device failure.
- Energy consumption – provide power meter data to assist in decisions to take equipment on- or offline to avoid peak use energy penalties.
- In-line quality sensors – provide real time data to the QA department to reduce labor for regulatory sampling and allow for quick reaction to quality issues with raw materials or finished products.
- Batch optimization – provide batch data and trends to process engineers for analysis to optimize mixing, heating, and cleaning times or correct overadditions of expensive minor additives.
- PID loop tuning – time series process data can be analyzed for PID loop tuning improvements, mechanical performance issues of control valves or for finding root causes of process variations.
- Inventory planning – provide real-time inventory information about raw materials, finished goods, and work in progress for improved scheduling, shipping, and ordering.
3. Network infrastructure. As the amount of data and number of users continues to increase, the bandwidth, stability, and security of your network must also improve. Focus first on two basic concepts: network segmentation and boundary protection. An excellent resource to increase your understanding is NIST Special Publication 800-82, “Guide to Industrial Control Systems (ICS) Security” (https://plnt.sv/1711-AZ). Because network design is critical, this is an area where using a system integrator with specialized expertise is recommended.
4. Don’t get lost in the fog, or the cloud, or the edge. IIoT is an overwhelming concept with many buzzwords, options and opinions, many of which are not critical and can be interpreted in different ways. Most IIoT applications are not currently using the cloud, and many control engineers don’t have the foggiest idea about fog computing. Don’t let your fear of not using the latest technology keep you from moving ahead.
5. Let’s stay together. Options for implementation methods and technologies can be overwhelming, but the benefits of working with a vendor you already know and trust will probably outweigh benefits and features offered by other software vendors. If you already have a good relationship with an automation vendor and an installed base of its products, strongly consider that company’s offerings before moving on.
6. Conscious uncoupling. An important concept of modern IIoT implementations will be the decoupling of devices and applications through use of emerging IoT protocols. A software solution that accesses the field device directly and bypasses your automation control system might be the best fit if large amounts of data are required, the data can be processed outside of your control system, and the control system has no other need for that data.