Khris Kammer, Information Partner and Competency Manager, and Keith McPherson, Director, Market Development, Rockwell Automation

Only 14 percent of executives say that their plant-floor data is integrated with enterprise systems.[1] This low number means the majority of industrial operations are not taking advantage of the wealth of valuable data from their automation systems.

The best way to tap into the valuable data generated in plants is by associating it with other data in order to get a holistic view of plant operations. This is called contextualization.

Tapping into contextualized data and sharing it across the enterprise allows manufacturers to address the age-old challenge of how to get more productivity out of existing assets.

The Traditional State of Manufacturing Information Systems

Traditionally, industrial operations have used a combination of centralized databases and local data storage for information systems. Special purpose and high speed data was, and is commonly kept locally, on-board, or near to the systems and equipment that need to access it frequently. Certain data necessary for contextualized decision-making is deemed worthy enough to be managed centrally, or at least “rolled up” to a central data warehouse. The thought behind this approach is that the information would be much easier to access and report on if it was all centralized.

Inevitably, this approach requires moving the data from its originating point to the central location and system, whether immediately when collected, or some period of time later. The data in the centralized warehouse takes on a life of its own, and is decoupled from the originating system. Which is the appropriate “system of record”? Will the originating system and the central data warehouse be synchronized for the entire lifecycle of the data? This can be very difficult or time consuming.   

Today, more and more industrial operations are moving away from that method toward a federated approach, in which the data exists where it makes sense to exist. This means the data can often stay where it originated. A federated approach allows the systems of record to manage the lifecycle of their own data autonomously, while an Enterprise Manufacturing intelligence (EMI) layer consolidates and exposes the data for consumption and analysis.

For example, information about an asset – a pump or a valve for example – can reside in multiple systems. Service records are kept in the Computerized Maintenance Management System (CMMS). Manufacturer specifications and recommended service intervals are buried in a spreadsheet available online. Actual pump performance and state data is in the control system and a plant-level Historian. Line and equipment schedules are in the ERP or MES system. All of this data is valuable for maintenance planners determining optimal service schedules. Does this mean that the data should be centralized into a planning database? Not necessarily. Using EMI software as an abstraction layer for the various autonomous data sources, decision makers of all types can make use of data without care of the underlying structures, formats, and data access methods, and without the cost of maintaining a central copy of the data.

Robust Network Fabric

Having data federated at the time of reporting and analysis requires a stable and robust network fabric. The system must have the capability to allow information to reside in different places, but also to bring it together in real time for the necessary analysis and visualization. One of the biggest reasons that industrial operations are able to take advantage of a federated data model is that there have been advancements in the speed, reliability, and security of network infrastructure.

Ethernet and the Internet Protocol (IP) has for decades enabled Information Technology applications, but it has now developed to a point where it can enable Operational Technology applications like equipment configuration, alarming, and process data collection – making it a viable infrastructure to handle all data movement in the plant. In addition, manufacturers who utilize standard EtherNet/IP networks can take advantage of the many innovative technologies that were developed for the commercial world, but are quickly migrating to the plant floor, such as video cameras, RFID readers, digital tablets, and security swipe-cards. These and other open-standard, IP-enabled devices are helping manufacturing and process operations reach new heights of production quality, efficiency, security and safety.

Beyond having a reliable network fabric, a federated approach requires strong security at all levels. This includes VPNs and firewalls on the network to regulate who can access the network, as well as application-level controls to regulate who can access who data on the network.

Now more than ever, manufacturers are facing global economic pressures and trying to get more productivity out of existing assets. Historically, data has not been shared easily across the enterprise because of the difficulties involved with locating, moving, and translating data throughout different systems.

Today, enabling technologies like secure, robust network infrastructures and EMI software that can provide meaningful analyses by bringing once disparate data together for effective sharing of data across the enterprise.

For more information, visit www.industrial-ip/org.

About the Authors:

Keith McPherson has been with Rockwell Automation since 1991, currently serving as Director of Market Development. In addition to his work at Rockwell Automation, Keith serves as a board member of MESA International. He holds a Bachelor of Science in computer engineering from the University of Pittsburgh.

Khris Kammer has been with Rockwell Automation since 2006, currently serving as Information Partner and Competency Manager. Khris is also active in the larger industrial community, holding the chair of the MESA Technical Committee since 2003, and working collaboratively with other industry professionals to make the Connected Enterprise a reality He has a degree in mathematics and computer science from Rochester Institute of Technology.

Advertisement