One of the challenges we as an industry now face is how to manage the data available from today's integrated systems, and convert it to knowledge for action. I see the process of getting from data to action as requiring four levels of transformation: data (raw data collected from field sensors, operators, purchase orders, inventory levels, etc.), information (putting this information into context of place, time, relative amounts), knowledge (how the change in information affects the stability of the operation), and action (doing something to maintain equilibrium in the system).
For the asset management system to even start the four steps, the data has to arrive in a format that can be understood by both the sender and receiver. But data flow is one half of the equation. The other side is the physical layer and associated protocols that are all defined by standards, which define the format in which the information is to be converted to bits, then the bits into a signal, so it can be transmitted and received as packets of information. It's likely that a single packet of information moving from a field sensor to the control room and then to the manufacturing execution system (MES) or enterprise resource planning (ERP) system will use one or more of these physical media, as well as multiple protocols.
A review of the protocol standards confirms that they too reflect the fact that a single physical layer will not meet all the requirements for getting data from the field to the controller and back to the field, which is why, for example, we have three flavors of HART. And in most cases, once the data enters the controller and control system, other protocols typically using some form of Ethernet packet come into play.
Though each of these protocols presents information in a different way, they're all designed to supply information important to the industry they serve, so the information fits into three broad categories:
- Measurement—The process measurement and often support for secondary and typically up to tertiary variables. For outputs, these can include the primary output as well as feedback information or other process variables.
- Status—One or more bits providing information on the present health of the device. In the process industries, this information is typically compliant with NAMUR NE-107, Self-Monitoring and Diagnosis of Field Devices.
- Diagnostics—Information on the health of the device, typically including, as a minimum, body temperature, program self-checks, etc., by which the device confirms that it's operating within its design constraints and assigned/configured operating range.
One group trying to help find and make use of the nuggets of information available from today's smart sensors is the ISA108, Intelligent Device Management, committee that's defining standard templates of best practices and work processes for implementation and use of diagnostic and other information provided by intelligent field devices in the process industries. Part of that work will include models for the flow of information from devices through the various systems that help users take the correct action at the right time to keep devices operating within their limits, prevent control action from using unreliable/bad data, or schedule the maintenance required without impacting operations.
Unfortunately, at least until the ISA108 committee completes its work, I see a disconnect between automation asset management systems, which are good at taking the rich amount of data from smart field devices (more about them next month) and converting it to actions for an instrumentation technician or controls engineer, but not so good at linking to other maintenance systems to schedule support labor, such as scaffolders, pipefitters, millwrights, etc. Will the work underway at IEC on the digital factory and similar initiatives help integrate all the disparate standards into a unified whole? Only time will tell.