Industrial Ethernet is moving from theory to reality.
|The G2-2 fieldbus manifolds from Numatics combine the flexibility of a scalable I/O system with a modular pneumatic valve manifold. An onboard Web server makes the node accessible from any standard browser for configuration, testing, and retrieving technical documentation. The manifold can also send e-mails triggered by specific events, such as diagnostic information or automatic preventive-maintenance scheduling. As a slave on an Ethernet/IP network, the unit controls up to 164 outputs and 96 inputs, and features plug-in valves, subbases, and I/O modules.|
Director of Technology Development
Knowledge is power. That's why manufacturers worldwide are spending billions of dollars annually on networked industrial control systems. Companies such as IBM, Microsoft, Oracle, and their many partners sell software for everything from enterprise-resource planning (ERP) to customer-relationship management (CRM). On the hardware side, computer controlled (CNC) machine tools, and other production equipment are tailor-made to be linked to enterprise-level networks such as Ethernet.
But where does the knowledge actually reside? Despite advances in manufacturing technology, successful and transparent industrial networking and control from the enterprise level to the device level are about as rare as the paperless office.
The goal of such systems -- making money, maintaining good customer relationships, and improving productivity up and down the supply chain -- requires more than efficient scheduling and inventory control, although it's vital to pay close attention to both. Forging truly stellar customer relationships involves knowing your process intimately and identifying every opportunity to increase quality, save money, and deliver on time, every time. This requires process knowledge right down to the device level.
The problem is that most enterprise-wide network solutions take a top-down approach, indiscriminately capturing large amounts of data and revealing little information, such as diagnostics, which would be immediately useful to the shop floor. Part of the problem is that information from field devices -- motors, valves, or actuators, for example -- must cross several layers before reaching the enterprise network. This information usually encounters a number of control networks with proprietary or open communications protocols acting as barriers along the way.
Fieldbus protocols such as ControlNet, Profibus-DP, DeviceNet, and A-B 1771 RIO bundle data for transmission over the network. These protocols are designed for either the control or device layers of individual industrial networks. While effective, they make it difficult to transparently transfer diagnostic and performance data to the enterprise level for analysis.
Ethernet is the ultimate destination for this information. First developed in 1973 by DEC, Intel, and Xerox, engineers now have a certain comfort level with Ethernet. They know the terminology and how to configure Ethernet networks, and can build relatively low-cost, dependable networks. It's been said that not many electrical engineers come out of college knowing typical fieldbus networks, but most everybody knows Ethernet. And while technically Ethernet can be simply defined as a physical and data-link layer technology, it is one with many benefits and far-reaching possibilities.
Industrial control networks, though, are demonstrably different from most information networks. Extending a network to the plant floor, where diagnostic information resides and decision-making needs to happen, causes concern for a variety of reasons.
First of all, there are not a lot of proven Ethernet-capable industrial devices. Drives, motors, actuators, valves, and other devices that feed digital data directly to information networks like Ethernet represent a significant departure from earlier statistical-process-control systems. Using Ethernet at the device or I/O level in manufacturing is currently in the early-adopter phase. Primary criticisms revolve around whether commercial off-the-shelf Ethernet-capable devices can meet industry-specific requirements of ruggedness, durability, noise immunity, and intrinsically dependable operation. The traditional lack of industrially hardened components and connectors exacerbates this perception. However, suppliers are starting to announce Ethernet-compatible products.
A recent report from ARC Advisory Group, Dedham, Mass., forecasts shipments of Ethernet-capable industrial devices such as drives, motors, actuators, and valves will reach nearly 5 million annually by 2005. (For more information, visit www.arcweb.com.)
Widespread penetration at the device level will be a key determinant of whether industrial Ethernet emerges as a common network architecture for manufacturing enterprises. Wide adoption of Ethernet-capable devices is a logical progression along the learning curve to easier integration, improved performance, and lower total costs. After all, the goal of using one common network platform from the enterprise level to the device level throughout a manufacturing operation is certainly worth pursuing.
While we can delve into a discussion of the IEEE 802 standard based on the Open Systems Interconnection (OSI) model that specifies the physical and data-link layers, at the end of the day, Ethernet is a network designed to move large amounts of data very fast, but one with various added benefits. Many in industry understand it conceptually, but many more need to realize where Ethernet is headed and the options it brings to an industrial organization.
For example, what does it mean to a company to have an embedded Web server on an industrial device? For one thing, it means device manufacturers can embed technical manuals right at the unit, as opposed to having to search and obtain technical information from the Internet or other sources. It also lets users easily access configuration information for that device. Traditional fieldbus networks require separate software packages to configure a valve or actuator. Now this information can reside at the device and is accessible via the Ethernet connection and a standard PC running Web-browser software.
Users can also access diagnostic information for devices with any standard browser. This includes the ability to see inputs, set outputs, and view common errors. Notification capabilities via e-mail are another option. Where it was once possible to determine device status through software polling routines, now that device can be configured to notify appropriate personnel proactively without software polling, essentially making such devices self-monitoring.
Hardware, always assumed to be unchanging, is now becoming transparent to industrial organizations. Ethernet-capable devices let companies develop products and processes around standard protocols, making such processes accessible through existing IT software and infrastructure.
The possibilities are endless. Operating data is delivered to enterprise software directly from the plant floor. Configuration and diagnostic information is stored at the device, accessible through a standard web browser via a local connection, LAN, or even over the Internet. E-mail capabilities on the device itself offer regularly scheduled notification of performance and scheduled maintenance to internal or external personnel. Efficiency utilization factors can be tied directly to accounting software.
Many more possibilities exist, including some not yet imagined. Integrating production information in real time throughout an industrial organization is closer to reality than ever before. It has been said that if knowledge is power, then building the right network is the path to power because it lets a company know its process right down to the device level. The way to get there is Ethernet.