Stone Shi is executive editor-in-chief, Control Engineering China. Courtesy: Control Engineering China
- Industrial edge computing and artificial intelligence (AI) contribute to automation innovation and effectiveness.
- Automation vendors include edge computing and AI among recent offerings.
Edge computing and artificial intelligence (AI) are helping to advance industrial computing and machine learning to increase factory floor applications. Since the rapid growth of industrial automation technologies in the 1970s and the development of computer technology, new technologies continue to advance the capabilities of the programmable logic controller (PLC), programmable automation controller (PAC), industrial PC (IPC), distributed control system (DCS), and process control system (PCS).
In the 21st century, the pace of technological development seems to be slowing down. It is more difficult to see the birth of major technological innovations and disruptive products, which may lead some to wonder if automation technology has reached a development “ceiling?”
With the rise of edge computing and AI, however, automation technology seems to have found its future direction for development. More automation manufacturers have joined the product development and the practical application of edge computing and AI technologies. What changes can edge computing and AI bring to automation?
Edge computing, cloud advantages
Industrial edge computing brings the advantages of the cloud to the field level. In the past, before there was no edge computing, the cost of developing and maintaining related applications at the factory floor level was very high. Edge computing platforms allows engineers to develop lower-cost, newer applications such as shop-floor data analysis and quality prediction. Edge computing also bridges the gap between traditional operational technology (OT) and information technology (IT) and integrate the advantages for both.
AI advances machine optimization
What benefits does AI provide for industrial automation? At the factory level, operations efficiency and accuracy can be improved through AI applications, making execution simpler and more efficient. In the past, people wrote control logic programs based on their understanding of mechanical and physical rules. Limits increase with increased complexities. If machine learning (ML) and deep learning methods are introduced, the hidden rules can be deduced and analyzed and optimal responses can be calculated automatically.
Edge computing and AI applications can produce the outstanding value of “two swords.” AI models need to be iterated and upgraded continuously, and edge computing can support distribution and iteration. Industrial edge computing has formed very good data collection capability; collected data also can be provided to AI. AI inside edge computing can ensure a lot of data is deduced locally, reducing sensitive data leaks. Through the back-end development mechanism, AI applications can be distributed with the help of high-level development frameworks. With AI’s and edge computing’s support for each other, 1+1>2.
Automation industry AI, ML examples
In the past two years, many automation products support edge computing.
Advantech is currently pushing the edge computing controller, which performs traditional logic control and motion control and independently runs a desktop operating system (OS). High-level language programming, processing audio and video and run-time application programs can be implemented on the IPC and be connected to the cloud.
Siemens released the WinCC Unified HMI in April; the human-machine interface (HMI) can run applications and do edge computing.
In AI automation applications, more products are emerging.
In 2019, Rockwell Automation released the AI module LogixAI of ControlLogix controller, to realize AI on PLC.
This year, Siemens released a neural network computing unit TM NPU for the S7-1500 controller, which can analyze and deduce relevant data at field level based on neural network.
On the latest TwinCAT3 platform from Beckhoff Automation, the TF380x machine learning inference engine and TF381x neural network inference engine are integrated.
Various machine learning models, such as Matlab from MathWorks and TensorFlow open-source software from Google Brain Team can be imported into the controller for AI analysis.
Industrial edge computing, industrial AI
Integrating AI with the help of autonomous machines, cognitive engineering and edge technologies increases automation effectiveness to improve the future of industry. At present, it’s increasingly obvious industrial edge computing and industrial AI will increase automation innovation.
Stone Shi is executive editor-in-chief, Control Engineering China. Edited by Mark T. Hoske, content manager, Control Engineering, CFE Media, [email protected]
KEYWORDS: Industrial edge computing, artificial intelligence
Industrial edge computing and artificial intelligence (AI) contribute to automation innovation and effectiveness.
Automation vendors include edge computing and AI among recent offerings.
How can AI and edge computing help your industrial applications?