The IoT Meets Process Control: Strategies for the Future
Process control, AI, IoT, latency, resilience, connectivity, and autonomy are discrete concepts, but the lines between them are becoming blurred as distributed intelligence and computing power continue to grow.
Industrial process control has been around since the first century, automating processes to minimize or eliminate human intervention. Now, however, with AI, ML and other advancements in the IoT, much is evolving. Bob Banerjee, VP Products, EPIC IO, discusses the blurring lines between IoT and process control and what it entails for organizations going forward.
Since the 1920s, process control has been used to automate production lines for everything from automobiles or baked goods to precision optics and semiconductors. In contrast, IoT technology has only been around since the internet and is used to collect data from various environments so it can be analyzed and acted upon. But with Industry 4.0 revolutionizing the way companies manufacture, improve and distribute their products by integrating IoT, cloud computing and analytics, and AI and machine learning into their production facilities, the lines between process control and IoT are blurring.
See more: What Does Enterprise Metaverse Mean for IT/OT Convergence?
IoT Versus Process Control
We typically think of process control as being used in continuous production processes or production lines. In the 1760s, process control inventions were aimed to replace human operators with mechanized processes. For example, Oliver Evans created a water-powered flour mill that operated using buckets and screw conveyors in 1784. Henry Ford applied the same theory in 1910 when the assembly line was created to decrease human intervention in the automobile production process.
Today, process control systems help us achieve production consistency, economy and safety levels that cannot be matched by manual control. In most cases, process control networks are hard-wired to deliver very high levels of reliability and resiliency with very low latency. After all, if a production line is cranking out thousands of widgets an hour and its process control system measures some defective widgets, it needs to move those off the line quickly.
In contrast, IoT networks have historically been focused on collecting data rather than controlling devices. A true IoT network sees the bigger picture more than just remote sensing. It “knows” not only when a door is open or a light has been left on but also that there are people still in the hallway, for example. In smart buildings, IoT networks can be used to monitor ambient light (and then inform an industrial control system to change lighting intensity according to preset thresholds) or to monitor and report on the average number of people in a room to facilitate space planning.
Another static example of IoT would be monitoring the levels inside two wastewater tanks sitting side by side in a construction yard. Even as one becomes full, the system would only issue a gentle warning since it knows that based on the rate of recent deposits in the tanks, there’s plenty of time before they are both full and need recycling. A truly smart IoT network would even know the tank emptying schedule and calculate the probability of overflow. True IoT, coupled with AI, feels lightyears beyond merely remote sensing.
But there are many real-time applications, including those where wireless connectivity is essential, where IoT is now making inroads into traditional process control territory. Some of these are time-sensitive and cannot afford the lag times associated with slow or disrupted connectivity. Others are mission-critical and cannot afford a single point of failure. Layers of reliability create shells of autonomous functionality, protecting the overall performance against established threat vectors. If the IoT sensors’ cellular connection dies, the local LAN keeps running everything until cell service is restored. If the LAN fails, each node on the LAN knows what it needs to do and just keeps doing it until everything reconnects and they can report their status. If the node itself dies or loses power, it must resume seamlessly upon resurrection.
Integrating IoT and Process Control: An Example
One recent example of IoT technology to control a process is a food safety system protecting produce in transit. The food industry has traditionally relied solely on refrigeration to control the levels of pathogens like E.coli or Listeria, but refrigeration doesn’t kill pathogens – it simply slows down the rate at which they multiply. In long-haul shipping scenarios (bringing fruit from Chile to the eastern United States, for example), even slow pathogen growth is undesirable. Scientists have shown that by adding ozone to the atmosphere inside a container, it’s possible to actually kill these pathogens.
Recently, EPIC iO debuted a system that injects small and precise ozone concentration into a food transport container to control pathogen levels without affecting the product it’s trying to protect. (The company’s research has shown that ozone injection can significantly extend the life of produce by limiting common pathogens.) This briefcase-sized system injects ozone into a refrigerated container, monitors those levels every few seconds, and adjusts its output accordingly. The system is self-contained, so the decisions about ozone concentrations are made instantly, and the IoT is used to collect data about ozone levels over time and upload it to a data lake in the cloud.
In this instance, the lines between IoT and process control have become blurred. This briefcase device brings IoT and AI into the world of process control – it finally meets the demands of reliability, precision, responsiveness, and reliability.
Bringing AI into the Picture
The system described below can be augmented with video, which is a good example of adding AI. Imagine monitoring the container with a video camera and then automatically shutting down the ozone flow if someone enters the container. (Ozone is a powerful oxidant with many industrial and consumer applications, but you don’t want to inhale more than you have to). Sure, you could use more IoT devices, such as motion sensors, but what if you’re covering 10,000 square feet? A single camera would find that easy.
To accomplish this automated surveillance, AI analyzes video images and takes appropriate action if it detects a human presence. Video analysis is a common application for AI engines for applications like facial recognition.
The Future Ahead
As cities and organizations seek to streamline their IT networks and converge IT and operational technology networks, AI and IoT will play an increasingly important role in the automation of all processes. For example, a remote sensor reports a truck’s cargo temperature and relative humidity. A more sophisticated IoT network merges that data with GPS information and the rate at which disinfecting ozone is being consumed by the cargo, not just of that truck but of the hundreds of trucks in the fleet. That IoT data ends up in a data lake – a priceless data repository waiting for AI and data scientists to mine it for patterns and anomalies.
Process control, AI, IoT, latency, resilience, connectivity, and autonomy are discrete concepts, but the lines between them are becoming blurred as distributed intelligence, and computing power continues to grow.
How are you leveraging AI and IoT advancements for process control? Tell us on Facebook, Twitter, and LinkedIn.
MORE ON IOT
