IEN: Which R & D hotspots (the Semantic Web, industrial wireless networks, SOA, autonomous agents, etc.) are close to commercialization?
Ken Hall: The Semantic Web provides useful functionality for manufacturing; the ability to describe the data and its context will enable much easier integration between IT and manufacturing and even between manufacturing groups. The concepts are in development and need to be proven on the web before it can be applied to manufacturing, which is a few years away.
Industrial wireless networks are now starting to show up in many applications. In many cases, wireless is replacing slip rings and flexible cables in current applications. Use of wireless for operator interfaces is also increasing. The growth is slow with very little risk taking -- which protocol wins on the factory floor will be interesting to watch. Rockwell Automation is using 802.11abg for wireless HMIs and limited control and is investigating Bluetooth, Zigbee, and 802.14.4 for use in wireless data collection.
SOA is rapidly growing in the IT world and we expect it to move quickly into manufacturing. SOA is simple, open, and uses standards. Concerns are mainly with performance and interoperability with multiple vendors, such as IBM, Microsoft, BEA, and open source.
Rockwell Automation has been developing autonomous agents, which are implemented in the company''s ControlLogix controllers, for a number of years. We have delivered test systems for the Navy and are currently working with customers in a number of other applications to build evaluation and pilot systems. Rockwell Automation expects implementations of autonomous agents deployed in factory systems in the next 5 years.
IEN: What advances do you see in Ethernet-based COTS solutions? Open standards and data sharing?
Ken Hall: As the cost of Ethernet continues to drop, we expect to see more and more devices connected to Ethernet. The Rockwell Automation Ethernet implementation uses standard Ethernet, TCP/IP, and UDP with no proprietary extensions. Our systems are based on determinist sampling using time synchronization and not only deterministic delivery. Rockwell Automation has been a big contributor to the IEEE 1588 time synch over Ethernet standard and is a big supporter of OPC and its newer OPC UA and XML standards. Rockwell Automation is also interested in leveraging commercial IT technology to interface manufacturing to IT where its appropriate.
IEN: How about production management, collaborative manufacturing, plant intelligence and visibility, advanced process control?
Ralph Kappelhoff: The space between the plant floor automation and the business level is currently a very fragmented market. The complexity comes from a variety of issues, with a key one being the need to mirror both the specific production processes of the plant and the specific business processes of the company. Both are inherently owned by other systems, in a third system sometimes called an MES system. More customers are making standardization decisions in this space and they are coming to the conclusion that the best value alternative is to have the ERP vendors provide the transactional, business critical aspects of the plant floor systems and have the automation vendors provide the real-time, process critical aspects of their plant floor systems. Thus, we are seeing a drive to have automation vendors and ERP vendors work much more closely together to provide standard solutions. Customers expect to realize functionality such as production manager, collaborative manufacturing, and plant intelligence over the existing infrastructure they have been putting in over the past 10 years vs introducing additional levels of complexity to their Manufacturing IT infrastructure.
Advanced process control continues to be an important issue but many customers are getting bigger results from tighter integration of the automation with the human processes via integrated workflow. Getting the human interaction more tightly linked to the manufacturing automation system has huge potential to reduce errors, improve quality, reduce the cost of poor quality, and increase throughput. In many cases, the sensor for advanced process control is a manual sample or a human-directed analysis of a complex data set, thus a tightly integrated workflow mechanism is a fundamental part of the equation.