Edge Computing Trends in Industrial Sectors: Speed, Connectivity, and Real-time DataMarch 24, 2021 by Jeff Kerns
This article will cover two trends and two drivers that are advancing edge technology and what that means for the future of manufacturing.
Edge computing is reducing limitations in manufacturing while creating new challenges for engineers. While edge technology incorporates many devices and systems, there are a few common trends in the industrial sector.
Overall trends in 2021 are similar to previous years; speed, control, and additional capabilities are moving IT and OT closer than ever.
Connectivity and Speed on the Edge
It is no secret that manufacturing is becoming more connected. IT and networks are becoming bigger and faster in a mad dash to gain data from all the inventoried material to each vibration or temperature shift of a process. Early adopters of connected devices gained value and ROI by using data for predictive or preventive maintenance.
As connectivity scales through a manufacturing plant, it also expands into departments. With enough data, executives can make more informed business decisions. The demand for more data increases the need for infrastructure, networks, bandwidth, and speed.
Industrial Ethernet, wireless networking, and standards have progressed to gigabytes per second speeds. Despite increases, industry continues to demand faster speeds, more data, and broader bandwidths.
Decentralizing the PLC
High-speed production is driving manufacturing and centralized PLCs might be too slow for some applications. Decentralizing the PLC was one solution to reduce latency in high-speed applications.
PLCs are expensive. A PLC was typically centralized due to the cost associated with protecting the sensitive electronic equipment inside it from the potentially dangerous environment outside the PLC cabinet.
However, technology advancements, cost reductions, and the demand for a latency solution led to a decentralized trend. Breaking up the PLC to process data closer to the source reduces latency and is also a simple definition for edge computing.
Operating Dynamically in Real-time on the Edge
Feedback from connected devices improves process capabilities. Moving computing closer to the source of the data, edge computing increases the speed of these capabilities. But why does industry need so much data and high-speeds?
Modern commerce may be the largest driver for speed. Online ordering and short delivery times are standard. The ability to move products and packages faster is necessary to satisfy this demand and capture market share.
Nano pulses in action, with a rate of up to 50 parts per second. Screenshot used courtesy of Smart Vision Lights [PDF]
Automated pick and place operations, machine vision inspection, even label reading are pushing the limits of line speeds. The need for dynamic and real-time operation adds to the challenge of faster speed. A simple label or code-reading operation may require a camera to capture a thousand images per second. For a system to read a label or code, proper lighting may be required.
Examples of Real-time on the Edge
Now the LED lighting and camera must not only operate in synchronization, but do it repeatedly within 1/1000th of a second. For sub-microsecond timing, an external LED driver may introduce latency based on how far the driver is from the light source. According to Smart Vision Lights, even a 2-foot cable would introduce impedance and latency challenges.
High-speed real-time challenges become increasingly difficult with dynamic lines. Early motion control limited automation to one thing. For example, a label reading process works well if a packaging line only has one side of the box, and the label is in one location and orientation. It is expensive and inefficient to have a line for different-sized boxes.
Adding various package sizes to a line means the label heights will change. In a label or code reading process, multiple label distances would require a camera to change focal lengths at extremely short periods of time to capture accurate data. Fortunately, companies such as Cognex have developed a liquid lens that can change the focal length in tens of milliseconds.
Cognex liquid lens technology. Image used courtesy of Cognex
To increase the ability to capture accurate repeated images for machine vision applications, smart lighting companies are introducing real-time dynamic solutions that adjust lighting intensity based on need.
Continuing with the label or code reading example, in a flexible packaging line of various heights, changing the intensity of the light can greatly improve data accuracy. These dynamic systems working in synchronization with each other give options and solutions that didn’t exist just ten years ago. While many technologies advanced to make this happen, dynamic real-time data processing would not be at the speed it is today without computing power at the edge.
Real-time Edge Solutions
One of the developments keeping edge computing on time is the OPC UA standard. Edge computing could mean that multiple processors are working in synchronization with each other. It is imperative that each device works by a universal clock, and each piece of data is associated with some type of timeline.
A model of the OPC UA architecture. Image used courtesy of the OPC Foundation
An accurate timeline allows many pieces of data from multiple systems and devices to be combined into one cohesive historical process. These real-time and dynamic advancements are why industry requires speed, consumes bandwidth, and is adopting advanced software such as AI.
To make accurate decisions fast, software is also becoming more dynamic. Machine learning and other AI programs are becoming more common in manufacturing.
Edge and the Future of Manufacturing
Edge computing will continue to push for more speed and greater bandwidths while delivering more real-time dynamic solutions.
What is the Future of Edge Computing?
While edge computing will take automation and manufacturing into the future, the demand for people with the skills to understand, develop, and implement these systems will increase. Additionally, finding people with experience will become more difficult as more baby boomers are retiring than there are currently students in some of these fields.
Adopting competitive edge computing technologies will hinge on the ability of companies to work with educators to produce the skills necessary for advanced technology applications. Overall, with or without a skills gap, technology will continue to become smarter, faster, and operate more dynamically in real-time.